cs.CL [Total: 43]
cs.CV [Total: 157]
cs.AI [Total: 10]
cs.CY [Total: 2]
cs.MM [Total: 1]
cs.IR [Total: 6]
cs.SD [Total: 1]
cs.RO [Total: 6]
cs.LO [Total: 1]
cs.LG [Total: 9]
eess.IV [Total: 3]
cs.CR [Total: 1]

cs.CL [Back]

[1] SEAD: Competence-Aware On-Policy Distillation via Entropy-Guided Supervision cs.CLPDF

Chia-Hsuan Lee, Zelei Cheng, Yu Wang, Renkun Ni, Sambit Sahu

TL;DR: 本文提出SEAD方法，一种基于熵引导监督的能力感知在线蒸馏框架，用于解决在线蒸馏中教师监督质量依赖学生能力的问题。该方法通过熵作为统一探针，在三个尺度上调整监督策略：令牌级别的梯度分配、训练阶段的KL散度调度以及提示的难度课程安排。在OLMo-3模型上，SEAD在六个数学基准测试中相比普通在线蒸馏平均准确率提升4.8%。

Details

Motivation: 在线蒸馏中，教师监督质量受学生能力影响，导致不连贯的生成产生噪声梯度，已掌握的令牌产生冗余梯度，造成令牌、训练阶段和提示三个尺度的浪费，而现有方法采用均匀监督无法解决此问题。

Result: 在OLMo-3模型上，SEAD在六个数学基准测试中相比普通在线蒸馏平均准确率提升4.8%，消融实验证实了各组件间的超加性交互作用。

Insight: 创新点包括：使用联合师生熵对令牌进行分区并分配定制化梯度；基于余弦调度从正向KL散度退火到反向KL散度；以及基于能力门控的从易到难提示课程。这些组件相互依赖，共同解决了能力依赖的监督退化问题。

Abstract: On-policy distillation (OPD) has a property absent in offline distillation and RL: teacher supervision quality depends on student competence. Incoherent rollouts yield noisy gradients; already-mastered tokens yield redundant ones. This creates waste at three scales (tokens, training phases, and prompts) yet existing methods supervise uniformly. We introduce SEAD, which uses entropy as a unified probe of this competence-dependent degradation at three scales: (1) joint teacher-student entropy partitions tokens into zones receiving tailored divergences or zero gradient (approx. 50% skipped); (2) a cosine schedule anneals from forward to reverse KL as competence grows; (3) a competence-gated curriculum introduces prompts easy-to-hard. These components are symbiotically necessary: token selection requires coherent rollouts (curriculum), annealing requires monotonic improvement (also curriculum). On OLMo-3 (7B to 32B), SEAD achieves +4.8 avg accuracy over vanilla OPD across six math benchmarks, with ablations confirming super-additive interactions.

[2] Developmental Trajectories of Situation Modeling and Mentalizing in Transformer Language Models cs.CLPDF

Pamela D. Rivière, Cameron Jones, Sean Trott

TL;DR: 本研究从发展视角追踪了Olmo2和Pythia语言模型在训练过程中情境建模与心智化能力的发展轨迹。研究发现，错误信念任务的性能依赖于模型规模和训练量，在预训练后期出现，并通过后训练干预得到提升，但表现脆弱，易受非事实性动词影响。情境建模能力通常先于且优于心智化能力，但情境表征在某些方面存在不一致性。

Details

Motivation: 针对大型语言模型在错误信念任务中表现出的心智化能力，存在构念效度担忧，研究旨在通过发展视角追踪模型在训练各阶段心智推理行为及其前提条件的发展模式。

Result: 在Olmo2和Pythia模型套件上的实验表明，高于随机水平的错误信念任务性能取决于模型大小和足够的训练量，在预训练后期出现，且在最具诊断性的条件下，后训练干预（SFT, DPO）能最大程度提升性能。然而，性能脆弱，非事实性动词会增加错误信念归因。情境建模准确率通常先于且超过错误信念任务准确率。

Insight: 创新点在于采用发展视角和压力测试方法评估LLM能力，揭示了模型以发展上适当的顺序构建部分连贯的情境模型，但表现出令人惊讶的脆弱性，强调了追踪能力涌现过程对全面评估模型的重要性。

Abstract: Recent work suggests that Large Language Models (LLMs) are sensitive to the belief states of agents described by text, as measured by the false belief task (FBT), yet persistent concerns of construct validity remain. We adopt a developmental perspective, tracing the pattern of mental state reasoning behavior – and likely preconditions for this behavior – across multiple training stages in the Olmo2 and Pythia language model suites. We find that above-chance FBT performance depends both on model size and sufficient training volume, emerges relatively late in pretraining, and is most improved by post-training interventions (SFT, DPO) in the condition most diagnostic of mentalizing (False Belief, Implicit). However, FBT performance is fragile: consistent with past work, the use of non-factive verbs (e.g., thinks) increases false belief attributions even in the True Belief condition. To contextualize these findings, we track the emergence of situation modeling: the ability to report on basic factual properties of a described scene. Situation modeling accuracy generally precedes and exceeds FBT accuracy, yet situational representations also prove surprisingly incoherent in certain respects: when asked about the knowledge states of the Antagonist agent – who always knows the item’s true location – Olmo2 13b is consistently influenced both by the Target agent’s knowledge state and the presence of non-factive verbs. Together, these results suggest that larger, sufficiently trained models build partially coherent situation models in a developmentally appropriate sequence, yet display surprising fragility – highlighting the value of developmental and stress-testing approaches for evaluating LLM capabilities.

[3] Latent Bridges for Multi-Table Question Answering cs.CL | cs.AI | cs.DBPDF

Simone Varriale, Tamara Cucumides, Floris Geerts, Paolo Papotti

TL;DR: 本文提出了GRAB框架，一种用于表格问答的构造器-编码器-桥接流水线方法。该方法将关系数据转换为异构图，通过消息传递进行编码，并使用少量查询条件化的潜在标记将信号传递给大语言模型，为大语言模型提供紧凑的任务相关结构表示和扁平化文本。

Details

Motivation: 解决多表问答中如何高效地将关系数据结构化信息融入大语言模型的问题，同时保持大语言模型的通用推理能力不被破坏。

Result: 在关系型问答任务上显著提升了性能，尤其在要求高的多表场景中取得了最大增益，提供了一种高效且原则性的方法将关系深度学习与大语言模型连接。

Insight: 创新点在于通过轻量化的图编码器和潜在桥接模块（仅9100万参数）将图结构信息压缩为少量潜在标记注入冻结的大语言模型，实现了结构信息的高效融合与模型能力的保留。

Abstract: We introduce GRAB, a constructor-encoder-bridge pipeline for table question answering. Our method lifts relational data into an heterogeneous graph, encodes it via message passing, and transfers the signals to an LLM through a small set of query-conditioned latent tokens. This provides the LLM with a compact, task-relevant structural representation together with the flattened text. Crucially, the LLM remains strictly frozen to preserve its general reasoning capabilities; we train only the lightweight graph encoder and latent bridge (91M parameters), allowing the entire pipeline to be trained efficiently. Our pipeline significantly improves performance on relational Question Answering, with the largest gains in demanding multi-table settings, offering an efficient, principled way to connect relational deep learning with LLMs.

[4] Memory-Managed Long-Context Attention: A Preliminary Study of Editable Request-Local Memory cs.CL | cs.LGPDF

Junyi Zou, Avrova Donz

TL;DR: 本文研究了内存管理的长上下文注意力机制，提出了一种将快速循环或稀疏主干网络与显式可编辑的请求本地内存槽及查询时稀疏回退相结合的方法。通过合成任务、自然语言生成和诊断测试，证明了纯固定状态或纯稀疏方法在覆盖、版本控制、抗干扰等场景下存在不足，而混合方法能有效兼顾。

Details

Motivation: 长上下文语言模型通常混淆了将历史压缩为高效状态和保持可靠长期记忆两个目标，现有线性、循环和稀疏注意力方法虽降低了长序列处理成本，但未明确指定事实何时应写入、覆盖、保护或丢弃，因此需要分离内存管理机制。

Result: 在结构化合成任务中，混合方法在2,097,152个token的压力测试中达到50/50的池化准确率；一个274万参数的最小因果事件-令牌模型在轻量写入监督下达到595/600准确率，证明了可训练性。在六族冻结隐藏状态桥接任务中达到1079/1080的受控指针准确率，但依赖于生成器提供的元数据。

Insight: 创新点在于将内存管理从注意力机制中解耦，通过可编辑的请求本地内存槽和查询时稀疏回退实现可控的槽生命周期管理；客观分析表明，该方法验证了可控槽生命周期的可行性，并指出稀疏回退在缺乏未来查询信号时是必要的，而学习开放域选择仍是主要架构瓶颈。

Abstract: Long-context language models often conflate two different goals: compressing history into an efficient state, and maintaining reliable long-term memory. Linear, recurrent, and sparse attention reduce the cost of processing long sequences, but they do not by themselves specify when a fact should be written, overwritten, protected from distractors, or discarded. We study memory-managed long-context attention, a research route that separates a fast recurrent or sparse backbone from explicit editable request-local memory slots and query-time sparse fallback. Across structured synthetic tasks, token/chunk/sequence bridges, generated natural language, and local frozen-model diagnostics, pure fixed-state or pure sparse methods fail some overwrite, version, anti-pollution, or no-write-signal cases, while a hybrid covers both routes. A small 2,097,152-token mechanism stress test reaches 50/50 pooled accuracy with 2-132 active chunks. A 2.74M-parameter minimal causal event-token model reaches 595/600 with lite write supervision, supporting proof of trainability rather than scale. A six-family frozen-hidden-state bridge reaches 1079/1080 controlled pointer accuracy, but it uses generator-provided integer key IDs and separately encoded canonical key strings; it is an oracle-metadata probe, not open-text entity resolution. Local non-leaderboard RULER 4K diagnostics remain close to full context, whereas a 33-record LongBench v1 16K subset shows that naive lexical selection is not general. The evidence separates three claims: controlled slot lifecycle is feasible, sparse fallback is needed when writes lack future-query signals, and learned open-domain selection remains the main architectural bottleneck. We do not claim a final generative architecture, global slot-trajectory convergence, or systems superiority.

[5] EVLA: An Electro-Aware Multimodal Assistant for Physically-Grounded Driving Reasoning and Control cs.CLPDF

Yuxin Liu, Zihan Chen, Haoyu Wang, Mingxuan Zhang, Ruijie Lin

TL;DR: 本文提出了EVLA（Electro-Visual-Language Assistant），一种新型的多模态驾驶助手框架，它通过融合视觉、文本和实时车辆电驱系统状态（如电机扭矩、电池SOC）来生成驾驶决策。该框架包含两个核心创新：统一共态编码器（UCSE）用于多模态信息融合与空间能耗建模，以及电驱感知结构化推理链（ESRC）用于基于物理约束的内部确定性推理。实验表明，EVLA在驾驶QA基准上显著优于现有VLM基线，并实现了更快的推理速度。

Details

Motivation: 现有视觉语言模型（VLM）驾驶助手通常将车辆动力学视为黑盒，导致决策缺乏对车辆实时机电状态的感知。本文旨在通过整合多模态场景理解与实时电驱状态感知，来弥合这一差距，开发出更物理接地、能量感知的下一代驾驶助手。

Result: 在驾驶QA基准上的广泛评估显示，EVLA显著优于经过微调的强VLM基线，最终得分提升+0.0871，准确率提升+5.6%。消融研究验证了各组件必要性，效率分析表明EVLA的推理速度比多阶段流程快36%。

Insight: 主要创新点在于将车辆实时状态（电驱参数）作为关键模态输入，并设计了统一共态编码器（UCSE）进行融合与能耗场建模，以及用内部确定性推理链（ESRC）替代外部思维链提示，实现基于物理约束的端到端优化学习。这为开发物理接地、能量感知的智能体提供了新思路。

Abstract: Modern vision-language models (VLMs) for driving assistants typically treat vehicle dynamics as a black box, resulting in decisions that lack awareness of the vehicle’s real-time electro-mechanical state. To bridge this gap, we introduce the Electro-Visual-Language Assistant (EVLA) – a novel framework that combines multi-modal scene understanding with real-time perception of the electrified powertrain state (e.g., motor torque, battery SOC). Our approach features two key innovations: first, a Unified Co-State Encoder (UCSE) that fuses visual, textual, and vehicle-state inputs into a shared latent representation, augmented with an Energy-Efficiency Field to model spatial energy costs; and second, an Electro-aware Structured Reasoning Chain (ESRC), which replaces external chain-of-thought prompting with an internal, deterministic reasoning process grounded in physical constraints and optimization objectives. Trained end-to-end with a physics-guided joint loss, EVLA learns to generate context-aware and energy-optimal driving decisions. Extensive evaluations on a driving QA benchmark demonstrate that EVLA substantially outperforms strong fine-tuned VLM baselines, improving the final score by +0.0871 and accuracy by +5.6%. Ablation studies validate the necessity of each component, and efficiency analyses show that EVLA achieves 36% faster inference than multi-stage pipelines. This work underscores that integrating vehicle-state awareness and structured physical reasoning is crucial for developing next-generation, physically-grounded driving assistants.

[6] A3M: Adaptive, Adversarial and Multi-Objective Learning for Strategic Bidding in Repeated Auctions cs.CL | cs.LGPDF

Junhan Li, Yuxin Zhang, Haoran Wang, Minghao Chen

TL;DR: 该论文提出了A3M框架，用于解决重复多单位拍卖中的出价学习问题。该框架集成了自适应深度强化学习、显式对抗性推理和多目标奖励设计，以动态平衡探索与利用、应对非平稳对手，并联合优化效用、拍卖者收益和公平性。

Details

Motivation: 现有方法通常依赖僵化的探索-利用调度、假设对手策略是平稳的，并且仅优化投标人效用，这限制了方法的适应性和策略鲁棒性。A3M旨在克服这些局限性。

Result: 在歧视性价格拍卖和统一价格拍卖中，与现有基线相比，A3M在标准设置下将最终遗憾降低了30-40%，在面对对抗性策略变化时保持鲁棒性能，并随着单位数量K的增加而良好扩展，同时实现了可调的多目标权衡。消融研究证实了每个核心组件的必要性。

Insight: 创新点在于将自适应DRL、对手建模（用于虚构博弈）和多目标奖励设计集成到一个统一的框架中，以同时处理探索-利用平衡、非平稳对手和多个优化目标（如效用、收益、公平性），从而为复杂拍卖环境提供了一个强大且灵活的学习框架。

Abstract: Learning to bid in repeated multi-unit auctions with bandit feedback poses a fundamental challenge. Existing methods often rely on rigid explore-then-exploit schedules, assume stationary adversaries, and optimize solely for bidder utility, thereby limiting adaptability and strategic robustness. To address these limitations, we introduce the A3M framework, which integrates adaptive deep reinforcement learning (DRL), explicit adversarial reasoning, and principled multi-objective reward design for online auction strategy optimization. A3M employs an actor-critic DRL backbone to dynamically balance exploration and exploitation, an opponent model for fictitious play against non-stationary adversaries, and a composite reward function to jointly maximize utility, auctioneer revenue, and fairness. We provide the first comprehensive empirical evaluation of this integrated approach against established baselines in both discriminatory and uniform price auctions. Results show that A3M reduces final regret by 30–40% in standard settings, maintains robust performance against adversarial strategy shifts, scales favorably with the number of units $K$, and enables tunable multi-objective trade-offs. An extensive ablation study confirms the necessity of each core component. Our work establishes A3M as a powerful and flexible framework for learning in complex auction environments.

[7] The strength of clinical evidence is recoverable from language model representations but not from their stated grades cs.CL | cs.AI | cs.IR | cs.LGPDF

Soroosh Tayebi Arasteh

TL;DR: 该论文研究了大型语言模型（LLMs）在处理临床证据声明时的表现。研究发现，尽管LLMs的内部表征（激活值）中可解码出关于证据强度的有序信号（与事实真实性不同），但模型直接陈述的证据等级却近乎随机，无法准确反映声明的支持强度。

Details

Motivation: 动机在于评估LLMs是否能够识别并准确传达临床证据的强度等级，因为这在医学应用中至关重要，而现有模型在表达置信度方面存在不足。

Result: 在包含45,134个临床声明的数据集上测试了22个开源LLM。线性估计器从模型表征中恢复证据等级的AUROC中位数为71.8，但模型直接陈述的等级准确率仅比随机猜测高25-27个百分点。该信号主要是词汇性的，且在不同主题或框架间无法迁移。

Insight: 创新点在于揭示了LLMs内部表征包含其未明确陈述的、与真实性不同的证据强度信号。这表明模型“知道”的比其“说出”的更多，其直接输出可能不可靠，而通过分析内部表征可能提取更可靠的信息。

Abstract: Large language models (LLMs) increasingly summarize clinical evidence, where a claim’s weight depends on how strongly it is supported. Yet these models convey confidence poorly, and properties they never state, such as truth, are often readable from their activations. Whether a clinical model registers evidence strength, distinct from truth, and states it when asked is untested, and any such signal could be lexical. We compiled 45,134 clinical claims from six public sources, harmonized 20,611 into a four-level evidence grade under three independent frameworks, and tested 22 local, open-weight LLMs from several developers (0.6-70 billion parameters; general, medical, and reasoning), with lexical, truth, and cross-framework controls. A linear estimator recovered the grade in every model (median AUROC 71.8), yet decodability did not rise with scale and was weakest in reasoning models. The grade the models stated fell to chance, 25-27 percentage points below the estimator. The recoverable signal was largely lexical and did not transfer across topics or frameworks, yet it was distinct from factual truth and still flagged weakly supported claims (AUROC 69.2). Clinical LLMs thus carry an ordered evidence-strength signal they do not express, so their stated grades fail to convey a claim’s support even when it is recoverable from their representations and text.

[8] Masked Diffusion Decoding as $x$-Prediction Flow cs.CLPDF

Weitian Wang, Lianlei Shan, Shubham Rai, Cecilia De La Parra, Akash Kumar

TL;DR: 本文提出了一种针对掩码扩散语言模型（MDLM）的连续解码框架，通过将掩码预测重新解释为干净状态预测（x-预测），在输入嵌入空间中诱导连续流，使令牌能够在每个扩散步骤中积累部分进展并保持可修订。该方法采用基于置信度的异步更新调度和轻量级策略网络，在有限解码预算下显著提升了模型性能。

Details

Motivation: 标准MDLM解码器采用全有或全无的二进制操作，丢弃了丰富的预测信息并迫使过早、不可撤销的承诺，导致在有限解码预算下性能不佳。本文旨在通过连续解码框架解决这一问题，保留部分信念表示以提高解码效率。

Result: 在预训练的LLaDA模型上应用该方法，在HumanEval数据集上仅使用25%的解码预算即可达到其原始性能的97%，表明在有限计算资源下实现了高效的文本生成。

Insight: 创新点包括将掩码预测重新解释为x-预测以诱导连续流，引入基于置信度的异步更新调度以适应语言中不均匀的上下文约束，以及通过强化学习训练轻量级策略网络来优化解码过程。这些方法为扩散模型在文本生成中的高效解码提供了新思路。

Abstract: Masked diffusion language models (MDLMs) generate text by iteratively unmasking tokens, but their standard decoder reduces each step to a binary action: a position is either committed to a single token or left fully masked, with no representation of partial belief in between. This all-or-nothing regime discards rich predictive information and forces premature, irrevocable commitments, leading to poor performance under a limited decoding budget. In this paper, we reinterpret mask prediction as clean-state prediction ($x$-prediction) and show that it can be used to induce a continuous flow in input embedding space. Building on this view, we propose a continuous decoding framework for MDLMs where tokens can accumulate partial progress at each diffusion step and remain revisable. To match the uneven contextual constraints across positions in language, we replace the globally synchronous schedule in image diffusion with a confidence-based asynchronous update in which the diffusion progress is token-wise accumulated. Additionally, we introduce a lightweight policy network and formulate its training as a reinforcement learning problem. Applied to pretrained LLaDA, our continuous decoder reaches 97% of its performance on the HumanEval dataset with 25% of decoding budget.

[9] ThinkProbe: Beyond Accuracy – Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs cs.CLPDF

Mohamed Amine Kerkouri, Simon D. Hernandez, Marouane Tliba, Yann Dauxais, Maha Ben-Fares

TL;DR: ThinkProbe是一个用于分析大语言模型（LLM）推理轨迹结构的框架。它将每个推理轨迹转换为一个包含循环、8种节点类型和6种边类型的‘思想图’，并通过一个结合基于规则的分割和判别式语义链接的非生成式流程，导出一个包含19个指标的五维认知画像（广度、深度、结构、元认知、效率）。该框架应用于来自7个原生推理模型在200个开放性问题上的4,200条轨迹，揭示了推理结构是一种稳定的模型级属性，其模型间差异在五个认知维度中的四个上超过领域间差异达四倍，而‘结构’维度对问题领域表现出真正的敏感性，暴露了基于准确率评估所无法看到的、性质上截然不同的认知画像。

Details

Motivation: 动机在于超越传统的基于准确率的评估，深入分析LLM在开放性问题上的推理过程的结构特性，以揭示模型之间在认知能力上的深层差异。

Result: 在7个模型、200个开放性问题、10个认知领域的4,200条推理轨迹上应用ThinkProbe。结果表明，推理结构是稳定的模型级属性，模型间方差在五个认知维度中的四个上比领域间方差高出四倍，而‘结构’维度对问题领域敏感，揭示了基于准确率的评估无法发现的定性差异。

Insight: 创新点在于提出了一个完全非生成式的、基于思想图和五维认知画像的框架，用于对LLM推理轨迹进行结构化分析，这为模型评估提供了超越准确率的新维度，并能揭示模型固有的、稳定的认知特性差异。

Abstract: We present ThinkProbe, a framework for structural analysis of LLM reasoning traces. ThinkProbe converts each trace into a Thought Graph a directed graph with cycles, 8 node types, and 6 edge types and derives a 19-metric five-dimensional cognitive profile (5D-CP: Breadth, Depth, Structure, Metacognitive, Efficiency) through a fully non-generative pipeline combining rule-based segmentation and discriminative semantic linking. Applied to 4{,}200 traces from 7 native reasoning models across 200 open-ended questions and 10 cognitive domains, ThinkProbe reveals that reasoning structure is a stable, model-level property: between-model variance exceeds between-domain variance by up to fourfold across four of five cognitive dimensions, with Structure showing genuine sensitivity to question domain, exposing qualitatively distinct cognitive profiles invisible to accuracy-based evaluation.

[10] Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks cs.CL | cs.LGPDF

Young-Jun Lee, Seungone Kim, Minki Kang, Alistair Cheong Liang Chuen, Zerui Chen

TL;DR: 本文提出了一种名为进化微调（EFT）的中期训练范式，旨在让大型语言模型（LLM）学会跨任务进化解决方案，而不是为每个新问题从头开始。作者构建了包含156K条轨迹、覆盖10个领域和371个优化任务的Finch Collection数据集，并对2B到9B参数的开源LLM进行微调。实验表明，EFT能赋予模型跨任务泛化能力，在22个保留任务上平均性能提升10.22%，并在特定任务上达到或超越最先进水平。

Details

Motivation: 现有方法通常为每个优化任务单独应用搜索框架，导致模型每次都需要从零开始解决问题，且搜索过程中积累的经验在任务完成后被丢弃，无法让模型自身习得迭代进化解决方案的能力并跨任务复用。

Result: 在22个保留任务上，经过EFT微调的模型比其基础版本平均性能提升10.22%。当与测试时强化学习结合时，模型在两个圆填充任务上达到了最先进（SOTA）性能，并在Erdős最小重叠问题上超越了其基础模型。

Insight: 核心创新在于将进化搜索轨迹转化为监督信号，通过跨大量多样化任务的微调，使LLM本身内化迭代进化的能力（如知道如何突变、何时回溯），从而成为无需从零开始的通用发现智能体。这为构建可复用的、具备跨任务泛化能力的优化模型提供了新范式。

Abstract: Would experience designing faster GPU kernels also help close in on a long-standing open mathematical conjecture? Large Language Models (LLMs) integrated into evolutionary search have recently produced state-of-the-art solutions on optimization tasks, including open mathematical conjectures, GPU kernel design, scientific law discovery, and combinatorial puzzles. To achieve this, prior work applied search scaffolds to one target task at a time, so every new problem is approached from scratch and the experience accumulated during search is discarded once the model finishes its attempt. This leaves the capability of iteratively evolving a solution (e.g., knowing which part to mutate and how, deciding when to backtrack) entirely in the scaffold rather than in the model itself. Whether the model itself could acquire this capability and reuse it across different tasks has been largely unexamined. To address this, we introduce Evolution Fine-Tuning (EFT), a mid-training paradigm that teaches LLMs to evolve solutions across tasks by converting evolutionary search trajectories into supervision. We construct Finch Collection, a 156K-trajectory dataset spanning 10 domains and 371 optimization tasks, and fine-tune open-source LLMs from 2B to 9B parameters. Empirically, EFT confers cross-task generalization: across 22 held-out tasks, our models surpass their base counterparts by 10.22% on average. Furthermore, when paired with test-time RL, our model matches state-of-the-art performance on two circle-packing tasks and outperforms its base-model counterpart on the Erdős minimum-overlap problem. EFT thus serves as a “practice phase” for general-purpose discovery agents that do not solve new problems from scratch.

[11] DistilledGemma: Balanced Efficiency-Accuracy for Person-Place Relation Extraction from Multilingual Historical Articles cs.CLPDF

Youssef Aboelwafa, Ahmed Samir, Nagwa Elmakky, Marwan Torki

TL;DR: 本文提出了DistilledGemma系统，用于从英语、德语和法语的多语言历史报纸文章中提取人物-地点关系。该方法采用三阶段知识蒸馏流程，旨在平衡分类准确性与计算效率。在官方评估中，该系统在标准测试集上排名第三，在二元测试集上排名第二，并通过将知识从26B教师模型蒸馏到2.3B学生模型，在保持强大推理能力的同时显著减小了部署模型规模。

Details

Motivation: 解决从多语言历史文献中提取人物-地点关系这一具有挑战性的任务，并平衡任务性能与计算效率，以提供实际可扩展的历史文档处理解决方案。

Result: 在HIPE-2026共享任务官方评估中，标准测试集准确度平均得分为0.688（排名第三），二元测试集平均得分为0.8156（排名第二）。在平衡效率-准确性指标上，在标准和二元测试集均排名第二。

Insight: 创新点在于设计了一个三阶段知识蒸馏流程，结合了提示工程探索、基于QLoRA的教师模型监督微调以生成思维链数据，以及响应级蒸馏到紧凑学生模型。客观来看，该方法将强大的多语言推理能力高效迁移到小模型，为资源受限场景下的历史文档处理提供了实用方案。

Abstract: We present DistilledGemma, an efficient and accurate system for the HIPE-2026 shared task on person-place relation extraction from multilingual historical newspaper articles in English, German, and French. Our approach adopts a three-stage knowledge distillation pipeline designed to balance classification accuracy with computational efficiency. In the first stage, we systematically explored prompt engineering strategies across eight large language models to identify the most effective reasoning architecture for this challenging task. In the second stage, we applied supervised fine-tuning (SFT) via QLoRA to a Gemma 4 26B A4B teacher model, leveraging its strong multilingual capabilities to generate silver-standard chain-of-thought traces across the training corpus. In the final stage, we performed response-level distillation to transfer these learned reasoning patterns into a compact Gemma 4 E2B student model. In the official evaluation, our team WHEREAMI ranked 3rd on the standard test set with an accuracy profile mean score of 0.688, and 2nd on the binary test set with a mean score of 0.8156. Notably, by distilling knowledge from the 26B teacher to the 2.3B student, we preserved strong reasoning capabilities while reducing the deployed model size to approximately 2.3B effective parameters; the LoRA adapters used during training were merged into the student for inference. This configuration ranked 2nd in the balanced efficiency-accuracy profile across both the standard and binary test sets. These results demonstrate that knowledge distillation provides a practical and scalable solution for historical document processing, achieving competitive performance without excessive computational cost.

[12] Can OCR-VLMs Read Devanagari? A Stress-Test Benchmark and Post-Correction Study cs.CL | cs.CVPDF

Aditya Pratap Singh

TL;DR: 本文评估了十种OCR系统（包括经典引擎、专用OCR-VLM和前沿多模态LLM）在Devanagari（印地语）脚本上的性能，涵盖合成退化条件和真实印刷扫描。研究发现：在干净文本上所有系统性能接近；退化条件下专用OCR-VLM最脆弱；真实扫描中多数系统性能大幅下降，且合成数据严重高估质量；英语OCR性能不能预测印度语OCR表现。作者还提出了错误分类法并测试了后校正方法。

Details

Motivation: 现有OCR系统在英语和中文基准测试中表现强劲，但其在印度文字（如Devanagari）上的性能尚未得到充分表征，因此需要系统评估以揭示其局限性和实际表现。

Result: 在真实扫描基准测试中，系统性能差异巨大（chrF++分数跨度达76点），Gemini和Claude领先（86.3和82.2），而GPT-5.5降至58.5，olmOCR-7B仅40.5。专用OCR-VLM在退化条件下出现灾难性重复错误。

Insight: 研究揭示了合成数据与真实数据性能的巨大差距，强调需用真实扫描评估OCR；提出了针对Devanagari的结构性错误分类法；发现后校正方法仅对同引擎分布有效，缺乏跨引擎泛化能力。

Abstract: OCR systems, ranging from classical engines to specialised OCR vision-language models (OCR-VLMs) and frontier multimodal LLMs, report strong results on English and Chinese document benchmarks, yet their behaviour on Indic scripts is largely uncharacterised. We benchmark ten systems on Devanagari (Hindi): classical EasyOCR; open VLMs (Qwen2.5-VL-3B, Qwen3-VL-8B, olmOCR-7B); specialised OCR-VLMs (DeepSeek-OCR, Unlimited-OCR); and frontier closed models (Gemini 2.5 Flash, Claude Opus 4.7, GPT-5.5, Mistral OCR), across four synthetic degradation conditions and 300 real printed scans. We report four findings. First, on clean rendered text all ten cluster within chrF++ 91 to 98, so synthetic text does not separate them. Second, under degradation the specialised OCR-VLMs are the most fragile: DeepSeek-OCR suffers rare but catastrophic repetition failures (outputs up to 71 the reference length) that wreck its corpus mean even though its median is the best of any system, which is why we report median and catastrophic-rate instead of the mean. Third, on real scans nine of the ten systems collapse (EasyOCR falls from chrF++ 93.6 to 58.3) and the field spreads across a 76-point range, so synthetic renders badly overstate Devanagari quality. Fourth, strong English OCR does not predict Indic OCR: GPT-5.5 drops to chrF++ 58.5 (tying classical EasyOCR) and olmOCR-7B, the model behind olmOCR-Bench, falls to 40.5, while the open Qwen3-VL-8B (75.2, runnable on a single 24 GB GPU) beats GPT-5.5 and approaches Mistral; Gemini and Claude lead at 86.3 and 82.2. An error taxonomy separates surface errors (numerals, punctuation) from structural ones (conjuncts, matras, nukta), and a byte-level (ByT5) post-corrector improves a cheap engine on its own error distribution (chrF++ +1.2 to +1.5) but does not transfer across engines. We release the benchmark, code, and models.

[13] Travel-Oriented Reasoning Large Language Model via Domain-Specific Knowledge Graphs cs.CL | cs.NEPDF

Vignesh Ram Nithin Kappagantula, Shayan Hassantabar, Samuel Simpson, Golnaz Moallem

TL;DR: 本文提出了一种基于领域知识图谱的模块化流程，用于构建面向旅行领域的推理大语言模型。该流程整合了专家设计的旅行知识图谱、自底向上生成多跳问答对的构建方法、利用问答对作为可审计推理轨迹的监督微调阶段，以及用于评估模型准确性和校准性的旅行领域基准数据集。

Details

Motivation: 大语言模型在通用领域展现出广泛的推理能力，但在旅行等专业领域中，其推理依赖于精确的定义、规则和专家定义的概念框架，模型往往因未能内化底层领域图谱（而不仅仅是缺乏领域知识）而产生自信但无根据的输出，导致准确性和可靠性不足。

Result: 在基于Qwen3-4B模型并使用LoRA进行适配的评估中，所提出的推理模型在基准测试上达到了82.4%的精确匹配率，显著优于预训练Qwen3-4B基线的22.4%。校准分析将剩余的17.57%错误分解为两种不同的失败模式。

Insight: 创新点在于提出了一种将专家设计的领域知识图谱显式地融入LLM推理过程的模块化流程，通过生成可审计的多跳问答对进行监督微调，从而提升模型在专业领域的准确性和不确定性解释能力。客观分析认为，该方法证实了基于知识图谱的显式推理能有效改善LLM在专业领域的性能，并明确了逐选项校准和基于轨迹长度的解码是下一步改进的方向。

Abstract: Large language models (LLMs) demonstrate broad reasoning abilities but struggle with accuracy and reliability in specialized domains such as travel, where reasoning depends on precise definitions, rules, and expert-defined conceptual frameworks, and where confident but unfounded outputs arise from a reasoning failure in which the model has not internalized the underlying domain graph rather than from missing domain knowledge alone. We propose a modular pipeline for building a travel-domain reasoning LLM grounded in an expert-designed knowledge graph (KG). Our pipeline integrates a travel KG that encodes domain entities and their relationships, a bottom-up construction procedure that walks the KG to produce multi-hop question answer (QA) pairs, a supervised fine-tuning stage that embeds the domain knowledge into a reasoning-capable LLM using the generated QA pairs as auditable reasoning traces, and a travel-domain benchmark dataset that measures the fine-tuned model’s accuracy and calibration. We evaluate our approach using Qwen3-4B with LoRA adaptation. Our reasoning model achieves an $82.4%$ exact match on the benchmark. This performance significantly outperforms the pretrained Qwen3-4B baseline at $22.4%$. A calibration analysis decomposes the residual $17.57%$ of errors into two distinct failure modes: an over-confident multi-label decoder that predicts both correct answers plus one spurious option on most dual-answer mistakes, and a smaller reasoning failure on single-answer questions where the supporting facts are present in the KG but the model fails to reconstruct the correct multi-hop path. This split confirms that explicit KG-grounded reasoning substantially improves the accuracy and uncertainty interpretation of LLMs in specialized domains, and isolates per-option calibration and trace-length-aware decoding as the next axes of improvement.

[14] MIThinker: A Plug-and-Play Policy-Optimized Thinker For Motivational Interviewing Counseling cs.CLPDF

Yizhe Yang, Palakorn Achananuparp, Heyan Huang, Jing Jiang, Ee-Peng Lim

TL;DR: 本文提出MIThinker，一个轻量级的思维模型，用于生成治疗性思考以指导动机性访谈（MI）咨询代理的策略选择和响应生成。通过引入AugR1-MI自动管道从观察到的响应中逆向工程咨询师的思维，并采用监督微调和强化学习的两阶段训练，MIThinker提升了心智理论评估和策略对齐。评估表明，基于MIThinker的代理MindfulMI在计算量大幅减少的情况下，达到了与最先进系统相当的MI能力。

Details

Motivation: 现有基于大语言模型（LLM）的咨询代理（包括使用动机性访谈的代理）在生成响应时，没有明确将思维与咨询技术对齐，这限制了其有效性。

Result: 在综合评估中，MindfulMI代理实现了与最先进（SOTA）系统相当的MI能力，同时计算量减少了一个数量级。

Insight: 创新点包括提出一个可插拔的策略优化思维模型（MIThinker），以及通过自动管道AugR1-MI逆向工程思维数据以解决标注数据缺乏的问题，结合监督微调和强化学习进行两阶段训练以优化策略对齐。

Abstract: Reasoning large language models (LLMs) have recently made much progress in complex problem-solving, leveraging internal reasoning (or thought) to guide their solution generation. However, existing LLM-based counseling agents, including those using Motivational Interviewing (MI), generate responses without explicitly aligning thoughts with counseling techniques, limiting their effectiveness. We propose MIThinker, a lightweight thinking model that generates therapeutic thoughts to guide MI counseling agents in strategy selection and response generation. To overcome the lack of annotated thought data, we introduce AugR1-MI, an automated pipeline that reverse-engineers counselor’s thoughts from observed responses. Through two-stage training combining supervised fine-tuning and reinforcement learning, MIThinker demonstrates improved theory-of-mind assessment and strategy alignment. Comprehensive evaluations show that MindfulMI, our agent leveraging MIThinker, achieves MI competency comparable to state-of-the-art systems with an order of magnitude less computation.

[15] TriageRA-CCF: Source-Side Clinical Confidence and Coverage Signals for Adaptive Rank Budgeting in Medical LLMs cs.CLPDF

Shucan Ji, Yining Huang, Hongliang Guo

TL;DR: 该论文提出了TriageRA-CCF方法，用于医学大语言模型中的自适应低秩预算分配。该方法根据每个医学问题的置信度、临床覆盖率和跨领域难度，动态决定激活LoRA适配器中不同规模（小、中、大）的秩通道子集，从而实现参数高效的医学问答。

Details

Motivation: 现有医学大语言模型通常采用固定的低秩预算进行适配，但不同医学问题在置信度、临床覆盖范围和跨领域难度上差异很大，固定的预算分配不够灵活高效。

Result: 在匹配的CMB-source训练协议下，TriageRA-CCF在Qwen3-8B和Llama3.1-8B模型上，相比LoRA、DoRA和MoELoRA等基线方法取得了最佳平均准确率。具体而言，在Qwen3-8B上比最强外部基线平均提升0.21个百分点，在Llama3.1-8B上提升0.16个百分点，但增益在不同基准测试中表现不均且较为有限。

Insight: 创新点在于提出了一种基于源端训练数据的自适应秩预算路由器，它结合了三个仅从源数据计算得到的信号：基础模型答案置信度、元数据单元的临床覆盖率以及反事实近似失误代理。这些信号通过直通预算路由器监督秩的选择，并辅以预算成本、熵和秩平衡正则化。组件消融实验表明，这三个信号均能提供有效的预算监督，但它们的组合并非在所有骨干模型上都单调最优。

Abstract: Medical large language models are commonly adapted with a fixed low-rank budget, even though medical questions differ substantially in confidence, clinical coverage, and cross-domain difficulty. We study adaptive rank budgeting for parameter-efficient medical question answering: for each question, the adapter decides whether to activate a small, medium, or large subset of LoRA rank channels. The central challenge is that a naive adaptive budget router can collapse to unstable choices or spend capacity without improving shifted benchmarks. We propose TriageRA-CCF, a source-side teacher for adaptive rank-budgeted LoRA. It combines three signals computed only from source training data: base-model answer confidence, metadata-cell clinical coverage, and a counterfactual close-miss proxy. These signals supervise a straight-through budget router over active ranks {2,4,8}, together with budget-cost, entropy, and rank-balance regularization. Under a matched CMB-source training protocol, TriageRA-CCF achieves the best average accuracy among LoRA, DoRA, and MoELoRA baselines on both Qwen3-8B and Llama3.1-8B. The gains are modest and non-uniform across benchmarks: +0.21 average points over the strongest external baseline on Qwen3-8B and +0.16 on Llama3.1-8B. Component ablations show that confidence, coverage, and counterfactual signals all provide useful budget supervision, but their combination is not monotonically best on every backbone.

[16] EntroRouter: Learning Efficient Model Routing via Entropy Regulation cs.CLPDF

Kaiyi Zhang, Xueliang Zhao, Zhuocheng Gong, Wei Wu, Yankai Lin

TL;DR: 本文提出EntroRouter，一种通过熵调节学习高效模型路由的单轮框架，旨在解决多轮路由框架中因推理与路由深度耦合导致的信任区域崩溃问题。该方法通过软监督初始化策略建立高熵先验，并利用软锚点稳定强化学习以实现受控熵收缩，在保持最强专家模型98.3%准确率的同时降低48.25%计算成本。

Details

Motivation: 针对现有模型路由方法在平衡精度与计算成本时，因推理和路由深度耦合及稀疏监督下强预训练先验主导导致的信任区域崩溃问题，即强大专家被系统性抑制的退化局部最优现象。

Result: 在广泛实验中，EntroRouter在保持最强专家模型98.3%准确率的同时，将计算成本降低48.25%，实现了高效的路由性能。

Insight: 创新点在于将熵调节作为核心目标以解耦推理与路由过程，通过软监督初始化建立高熵探索先验，并利用基于离线能力估计的软锚点实现强化学习的稳定优化，从而避免信任区域崩溃并提升路由效率。

Abstract: Model routing balances solution accuracy and computational cost by selecting among models of varying capabilities. While recent multi-round frameworks interleave reasoning and planning, we identify a structural failure mode termed Trust Region Collapse. We demonstrate that the deep coupling of reasoning and routing, exacerbated by the dominance of strong pre-training priors under sparse supervision, leads to degenerate local optima where capable experts are systematically suppressed. To decouple these processes, we propose $\textbf{EntroRouter}$, a single-round routing framework that treats entropy regulation as a core objective. We first initialize the policy via Soft Supervision, fitting a distribution of suitable models to establish a high-entropy prior for exploration. Subsequently, we stabilize Reinforcement Learning using a Soft Anchor, which utilizes offline capability estimates to orchestrate controlled entropy contraction within a safe trust region. Extensive experiments demonstrate that EntroRouter retains 98.3% of the strongest expert’s accuracy while reducing computational costs by 48.25%.

[17] To Reason or to Fabricate: Reasoning Without Shortcuts via Hint-Anchored Pairwise Aggregation cs.CL | cs.AIPDF

Jiuheng Lin, Chen Zhang, Yansong Feng

TL;DR: 本文提出了HIPPO框架，通过提示注入聚合和成对奖励模型来解决强化学习（RL）中因预训练或SFT数据重叠导致的模型走捷径问题，即模型通过记忆正确答案和编造事后推理来伪造推理过程。该方法利用提示注入触发重叠诱导行为，生成可区分的偏好信号，使轻量级评判模型能可靠区分真实推理与捷径驱动的合理化，从而提取真实、可迁移的推理技能。

Details

Motivation: 动机是解决强化学习在提升LLM推理能力时，因预训练或监督微调数据重叠导致的模型走捷径问题，即模型倾向于记忆答案并伪造推理过程，而非真正学习推理。

Result: 大量实验表明，HIPPO在标准基准上显著优于基线方法，并能有效泛化到分布外的一般任务，证明其提取了真实、可迁移的推理技能，而非表面捷径模式。

Insight: 创新点在于结合提示注入聚合和成对奖励模型，通过故意触发重叠行为生成明确的锚点进行成对比较，提供高区分度的偏好信号，确保稳定优化，从而区分真实推理与捷径合理化，提升推理的鲁棒性和泛化性。

Abstract: While reinforcement learning (RL) significantly enhances LLM reasoning, its efficacy is severely undermined by Pre-RL data overlap, where RL datasets overlap with pretraining or SFT corpora, causing models to exploit shortcuts by memorizing correct answers and fabricating post-hoc reasoning. To address this, we introduce HIPPO, a novel RL framework that integrates hint-injected aggregation with a tailored pairwise reward model. By utilizing hint injection to deliberately trigger overlap-induced behaviors, the resulting traces naturally serve as explicit anchors for pairwise comparison. This provides highly discriminable preference signals, enabling a lightweight judge model to reliably distinguish genuine reasoning deduction from shortcut-driven rationalization, while the pairwise formulation ensures stable and robust optimization compared to standard PRMs. Extensive experiments demonstrate that HIPPO yields substantial improvements over standard baselines and generalizes effectively to out-of-distribution general tasks, showing it extracts authentic, transferable reasoning skills rather than superficial shortcut patterns.

[18] The Verbose Context Problem in Medical Records cs.CL | cs.AIPDF

Shiva Kaul, Min-Gyu Kim, Anjum Khurshid, Sriram Vishwanath

TL;DR: 本文提出了PopMedQA基准，用于解决医疗记录中结构化概念文本表示冗长导致的上下文冗长问题。该基准通过neopatient库生成人工患者记录，并评估了多种方法（如提示策略、压缩和代理分解）在纵向患者记录群体分析中的表现。研究发现，领域无关方法无法有效缓解上下文冗长问题，需利用领域特定结构改进语言模型输入。

Details

Motivation: 解决医疗记录中结构化概念文本表示冗长（上下文冗长问题）导致的推理瓶颈，特别是在群体健康分析中，纵向患者记录常超过40万tokens，影响语言模型效率。

Result: 在PopMedQA基准上，通过广泛消融实验（包括提示策略、提示压缩和代理分解），发现领域无关方法未能有效缓解上下文冗长问题，表明现有方法在群体规模推理上存在局限。

Insight: 创新点包括提出PopMedQA基准和neopatient库，用于隔离和评估上下文冗长问题；客观分析显示，需开发利用医疗领域特定结构的语言模型输入方法，以提升大规模推理能力。

Abstract: The verbose context problem occurs when structured concepts have token-inefficient textual representations. This bottleneck is acute in population health: cohort-level analysis of longitudinal patient records requires reasoning over thousands of medically-coded events, often exceeding 400K tokens in total. We present PopMedQA, a benchmark isolating this problem through computational tasks on groups of longitudinal patient records. We construct the benchmark using neopatient, a new library for language-controlled generation of artificial patient records. Through extensive ablations – including prompting strategies, prompt compression, and agentic decomposition – we find that domain-independent methods fail to alleviate the verbose context problem. There remains significant opportunity to exploit domain-specific structure in language model inputs for population-scale reasoning.

[19] Coverage-Driven KV Cache Eviction for Efficient and Improved Inference of LLM cs.CL | cs.AIPDF

Shuvendu Roy, Mengyao Zhai, Hossein Hajimirsadeghi, Golnoosh Samei

TL;DR: 本文提出了一种名为K-VEC的新型KV缓存驱逐策略，旨在解决大语言模型推理过程中KV缓存内存开销过大的问题。该方法通过引入跨头和跨层覆盖模块，优先保留具有高覆盖率的token，从而在减少内存占用的同时保持模型性能。

Details

Motivation: 现有KV缓存驱逐策略虽然能减少内存占用，但会导致性能显著下降，尤其是在需要长上下文推理的任务中。作者发现性能下降与唯一token覆盖率的降低有关，并理论证明了覆盖率降低会限制输入输出之间的互信息，从而损害预测准确性。

Result: 在LongBench的16个子集上评估，K-VEC在相同驱逐率和内存约束下，相比现有方法最高提升了10.35个点，验证了其有效性。

Insight: 创新点在于首次将token覆盖率作为KV缓存驱逐的核心指标，并设计了跨头和跨层覆盖模块来增强token保留。从客观角度看，该研究将信息论原理（互信息）与工程优化结合，为高效LLM部署提供了新思路。

Abstract: Large language models (LLMs) excel at complex tasks like question answering and summarization, thanks to their ability to handle long-context inputs. However, deploying LLMs is costly, not only due to the high computational demands of quadratic complexity of self-attention and auto-regressive generation, but also because of the significant memory overhead required for storing the key-value (KV) cache during inference. To reduce the memory cost, existing KV-cache eviction strategies leverage the sparsity in attention to selectively store a subset of tokens. While reducing the memory footprint, such approaches show a considerable drop in performance, especially in tasks that require long-context reasoning. We identify that the drop in performance is linked to a reduction in the coverage of unique tokens. Additionally, we theoretically show that reduced coverage limits the mutual information between inputs and outputs, thereby impairing predictive accuracy. To this end, we introduce K-VEC, a novel coverage-aware KV-cache eviction strategy that prioritizes token coverage while evicting tokens in the cache. K-VEC introduces a cross-head and a cross-layer coverage module to enhance token retention across attention heads and model layers, mitigating performance degradation caused by low coverage. Evaluated on 16 LongBench subsets, K-VEC exhibit up to 10.35 points improvement over the existing methods under the same eviction rate and memory constraint. Comprehensive evaluations validate the effectiveness of our approach and demonstrate its potential for efficient LLM deployment in resource-constrained settings.

[20] AURORA: Asymmetry and Update-Induced Rotation for Robust Hallucination Detection in Large Language Models cs.CLPDF

Zishuai Zhang, Hainan Zhang, Zhiming Zheng

TL;DR: 本文提出了AURORA框架，一种新颖的大语言模型幻觉检测方法，通过分析模型权重梯度动态而非静态表示来区分幻觉和忠实回答。该方法利用梯度更新模式的不对称性和结构错位，通过余弦相似度分布的偏度和旋转比两个互补特征进行检测。AURORA在四个模型家族和四个基准数据集上表现出色，并能有效扩展到不同模型规模及跨领域任务。

Details

Motivation: 大语言模型在生成过程中容易产生事实错误或不忠实的幻觉输出，这阻碍了其在关键应用中的部署。现有检测方法通常依赖昂贵的输出级一致性检查或捕捉浅层数据集特定模式的静态隐藏状态探针，导致跨数据集评估性能大幅下降。

Result: AURORA在四个模型家族和四个基准数据集上实现了强大的幻觉检测性能。进一步分析表明，该方法能有效扩展到不同模型规模，并可迁移到数学推理和视觉语言等域外任务。

Insight: 论文的创新点在于将检测焦点从静态表示转向模型权重梯度动态，揭示了幻觉样本会触发不对称和结构错位的梯度更新模式。客观来看，通过余弦相似度分布的偏度和基于SVD的旋转比来量化这种模式，提供了一种可泛化且高效的检测机制。

Abstract: Large Language Models (LLMs) have demonstrated remarkable capabilities across a wide range of natural language processing tasks. However, their tendency to generate hallucinations, namely factually incorrect or unfaithful outputs, poses a critical obstacle to their deployment in high-stakes applications. Although recent hallucination detection methods have made encouraging progress, they typically rely on costly output-level consistency checks or static hidden-state probes that capture shallow dataset-specific patterns, leading to substantial degradation under cross-dataset evaluation. In this work, we propose AURORA, a novel hallucination detection framework that shifts the focus from static representations to the weight-gradient dynamics of LLMs. Our key insight is that hallucinated and faithful answers induce qualitatively different gradient update patterns on the model’s parameters. Specifically, hallucinated samples trigger asymmetric and structurally misaligned gradients, which can be captured through two complementary features: (1) the skewness of the cosine similarity distribution between weight matrices and their gradient update directions, and (2) the rotation ratio, which quantifies how much the gradient update reorients the singular-vector basis of weight matrices via SVD. AURORA achieves strong hallucination detection performance across four model families and four benchmark datasets. Further analyses demonstrate that our method scales effectively across model sizes and transfers to out-of-domain tasks, including mathematical reasoning and vision-language scenarios.

[21] Hybrid Retriever Evolution for Multimodal Document Reasoning Agents cs.CL | cs.AI | cs.LG | cs.MAPDF

Bohan Yao, Shruthan Radhakrishna, Vikas Yadav

TL;DR: 本文提出了一种失败驱动的进化框架，用于学习多模态文档推理中的检索协调策略。该框架通过元代理分析错误推理轨迹，主动诊断根因，并迭代重写任务代理的指令，将检索从固定前端阶段转变为自适应的、逐步的推理决策。

Details

Motivation: 现有系统通常通过固定流程组合词汇、语义和多模态等互补性强的检索器，无法适应单个推理步骤的需求。本文旨在探索是否可以将检索编排本身作为推理过程的一部分来学习。

Result: 在MMLongBench-Doc和DocBench基准测试上，进化后的代理相比未进化的基线提升了高达19.6分，并持续优于MACT、MDocAgent和SimpleDoc等近期系统。

Insight: 核心创新在于将检索协调转化为一个可通过多智能体交互和失败分析来自主学习的决策过程，实现了从狭隘的词汇行为到丰富的多工具协调的渐进转变，而非依赖任何硬编码的检索模式。

Abstract: Different retrievers, including lexical, semantic, and multimodal approaches, provide highly complementary strengths for multimodal document understanding, yet most systems combine them through fixed pipelines that cannot adapt to the demands of individual reasoning steps. In this work, we ask whether retrieval orchestration itself can be learned as part of the reasoning process. We introduce a failure-driven evolution framework in which a meta-agent autonomously discovers how a tool-using task agent should coordinate diverse retrievers during multi-step document question answering. The meta-agent analyzes incorrect reasoning trajectories, actively probes the same tool environment to diagnose root causes, and iteratively rewrites the task agent’s instructions, turning retrieval from a fixed front-end stage into an adaptive, step-wise reasoning decision. The evolved agent learns when to invoke each retriever, how to combine them, and how to compose evidence across modalities and pages. On MMLongBench-Doc and DocBench, the evolved agent achieves gains of up to +19.6 points over the unevolved baseline and consistently outperforms recent systems including MACT, MDocAgent, and SimpleDoc. Detailed retrieval analyses confirm that these improvements arise from adaptive routing and evidence composition rather than reliance on any hard coded retrieval mode, and evolution dynamics reveal a progressive shift from narrow lexical behavior to rich multi-tool coordination. These findings establish autonomous multi-agent coordination as a promising paradigm for multimodal document reasoning.

Aunabil Chakma, Mihai Surdeanu, Eduardo Blanco

TL;DR: 本文提出了一种两阶段提示优化框架，用于小语言模型的少样本关系抽取任务。该框架首先使用基于推理的优化器对自然语言提示进行广泛改进，然后应用提出的GradPO方法，利用损失和梯度信号识别高影响力提示片段并进行局部编辑优化。

Details

Motivation: 针对小语言模型在少样本关系抽取任务中自动提示优化研究不足的问题，旨在结合推理和梯度方法提升提示质量。

Result: 在FS-TACRED和FS-FewRel数据集上的实验表明，局部优化通常能改进第一阶段的提示，且GradPO是最稳定的优化器；该框架在FS-TACRED上使用Qwen3-4B模型达到了SOTA性能，在FS-FewRel上保持竞争力。

Insight: 创新点在于将基于推理的全局搜索与基于梯度的局部细化相结合的两阶段优化框架，以及提出的GradPO方法通过梯度信号指导提示编辑，为小模型提示工程提供了新思路。

Abstract: Automatic prompt optimization is still underexplored for episodic few-shot relation extraction with smaller language models. We propose a two-stage framework that combines reasoning-based prompt optimization with gradient-based prompt optimization. The first stage can use any reasoning-based optimizer to make broadprompt improvements in natural language. The second stage applies our GradPO, which uses loss and gradient signals to identify high-impact prompt spans and refine them with local edits. Experiments on FS-TACRED and FS-FewRel show that local refinement usually improves prompts found by the first stage, and GradPO is the most consistent refiner. Our framework achieves state-of-the-art performance on FS-TACRED with Qwen3-4B and remains competitive on FS-FewRel.

[23] Resolution Thresholds in VLM Detection of Harmful ASCII Art Across Construction Modes and Languages cs.CLPDF

Yikai Hua, Peter West

TL;DR: 本文研究了图像分辨率对大型视觉语言模型（VLM）检测有害ASCII艺术能力的影响。通过评估八种最先进的VLM在英文和中文语料上的表现，发现当图像分辨率超过特定阈值时，检测率会急剧下降，其中基于单词的构建模式最难被检测。

Details

Motivation: 大型视觉语言模型作为内容审核工具部署时，容易受到将有害文本编码为ASCII艺术的越狱攻击，导致有害内容绕过审核系统。本文旨在探究图像分辨率如何影响VLM对此类攻击的检测能力。

Result: 实验结果表明，检测率在超过特定分辨率阈值后会急剧下降；基于单词的构建模式在整个分辨率范围内对检测的抵抗力最强。这些发现揭示了基于VLM的内容审核系统存在系统性漏洞。

Insight: 论文的创新点在于系统性地研究了不同分辨率、字符构建模式（L1-L8）和语言对VLM检测有害ASCII艺术的影响，并发现了可复现的检测失败阈值，这为建立分辨率感知的评估标准提供了动机。从客观角度看，该研究为理解VLM在内容安全方面的盲点提供了量化依据。

Abstract: Large Vision-Language Models (VLMs) are increasingly deployed as content moderation tools, yet they remain vulnerable to jailbreak attacks in which harmful text is visually encoded as ASCII art. This can allow inappropriate or harmful content to bypass moderation systems. To address this vulnerability, this paper investigates how image resolution affects VLM detection of harmful ASCII art across eight character construction modes (L1-L8), ranging from dense block characters to word-embedded designs. We evaluate eight state-of-the-art VLMs on English and Chinese corpora using a pipeline that generates ASCII art images at ten resolution scales, probing whether a consistent detection-failure threshold exists across models, modes, and languages. Results indicate that detection rates decline sharply above certain resolution thresholds, and that word-based modes are the most resistant to detection across the full resolution range. These findings reveal a systematic vulnerability in VLM-based content moderation systems and motivate resolution-aware evaluation standards.

[24] How LLMs See Creativity: Zero-Shot Scoring of Visual Creativity with Interpretable Reasoning cs.CLPDF

William Orwig, Roger E. Beaty

TL;DR: 本研究探讨了多模态大语言模型（LLMs）能否作为视觉创造力的零样本评估者，并分析其推理过程的可解释性。研究测试了六种多模态LLM对AI生成图像和手绘草图的创造力评分，发现模型评分与人类评分高度一致，且其逐步推理过程揭示了模型评估创造力的关注点与权衡逻辑，但推理本身并未提升与人类评分的一致性。

Details

Motivation: 评估视觉图像的原创性是创造力评估中的长期挑战。尽管AI模型在语言领域的自动评分已被证明有效，但在视觉创造力评估方面，如何评估以及理解模型如何得出评分仍是关键问题。

Result: 在两项研究中，所有测试的多模态LLM（包括Gemini 3 Flash、GPT-5.4 Mini等）在AI生成图像数据集（r = .57-.68）和手绘草图数据集（r = .29-.68）上均显示出与人类创造力评分的高度一致性。然而，分析模型的逐步推理过程发现，虽然推理使评估过程可解释，但并未进一步提高与人类评分的一致性。

Insight: 论文的创新点在于首次系统验证了多模态LLMs作为视觉创造力零样本评估者的可行性，并利用其“推理”输出为AI模型的评估过程提供了可解释的窗口，揭示了模型如何权衡原创性与质量等维度。这为自动化、可解释的创造力评估提供了新范式。

Abstract: Evaluating the originality of visual images poses enduring challenges for creativity assessment. Automated scoring using AI models has proven effective in the verbal domain, yet key questions remain about evaluating visual creativity and understanding how models arrive at their ratings. The present research asks whether multimodal large language models (LLMs) can serve as judges of visual creativity zero-shot (without any fine-tuning or examples of human ratings) and whether their “reasoning” output offers an interpretable window into their evaluation process. We tested six multimodal LLMs (Gemini 3 Flash, Gemma 4 31B IT, GPT-5.4 Mini, GLM-5v Turbo, Kimi K2.5, and Qwen 3.6 Plus) on 992 AI-generated images (based on human-written prompts) and 1,500 hand-drawn sketches scored for creativity by human raters. In Study 1, all models showed substantial alignment with human creativity ratings on both datasets (r = .57-.68 on AI-generated images; r = .29-68 on sketches). In Study 2, we analyzed the step-by-step reasoning processes of three LLMs evaluating the same images and drawings. Although reasoning made model evaluations interpretable – showing what they attend to, how they balance originality vs. quality, and how they justify their ratings – reasoning did not improve alignment with human ratings. In sum, our findings indicate that multimodal LLMs can match human judgments of visual creativity without any additional training, and that their reasoning reveals how AI models evaluate creativity. An open scoring app implementing this pipeline is available at https://review-visual-eval-scoring.hf.space.

[25] Can MLLMs Critique Like Humans? Evaluating Open-Ended Aesthetic Reasoning in Multimodal Large Language Models cs.CLPDF

Sajjad Ghiasvand, Maryam Amirizaniani, Haniyeh Ehsani Oskouie, Mahnoosh Alizadeh, Ramtin Pedarsani

TL;DR: 本文评估了多模态大语言模型（MLLMs）在开放式审美评论任务上的表现，发现基于参考相似度的指标会误导性地高估模型与人类评论的接近程度。通过使用Reddit照片评论数据集，在多种提示条件下测试了五个开源MLLMs，研究发现模型生成的评论在长度、内容选择性和多样性方面与人类评论存在系统性差异。

Details

Motivation: 解决现有审美评估主要依赖数值分数或多选题，而缺乏对模型生成的实际书面评论（即开放式审美推理）与人类评论相似度的有效评估问题。

Result: 在Reddit Photo Critique Dataset上，使用基于参考的相似度指标评估发现，严格的词汇和基于学习的指标仅显示与人类评论的弱对齐；而粗略的嵌入余弦相似度虽然显示出广泛的主题重叠，但经图像接地控制实验发现，这主要源于模型稳定的‘固定风格’而非针对图像的特定观察。

Insight: 创新点在于系统性地解构了提示条件（如角色设定、方面提示、长度控制、单次与多次生成）并引入了图像接地控制来诊断模型行为。客观分析表明，基于参考的相似度评估会奖励流畅、全面的评论风格，而非人类评论的选择性和特异性，这对评估和训练开放式多模态生成任务具有重要启示。

Abstract: Open-ended aesthetic critique is a challenge for multimodal large language models (MLLMs): unlike multiple-choice aesthetic benchmarks, it has no single correct answer, and most aesthetic evaluation has measured models against numeric scores rather than the written critiques people actually give. We evaluate MLLM critiques against ranked human references and ask whether they are close to human ones. Using the Reddit Photo Critique Dataset, we score five open-weight MLLMs against multiple ranked human critiques per photo with reference-based similarity metrics, under six prompt conditions that disentangle persona framing, aspect hinting, length control, and single- versus multi-pass generation, and add an image-grounding control that feeds each model the wrong photograph. We find that reference-based similarity gives a misleading picture. Stricter lexical and learned metrics show only weak alignment with human critiques, while a coarse embedding cosine reports broad topical overlap that the grounding control traces to a stable house style rather than image-specific observation. Behaviorally, the models diverge from humans in consistent ways the scores do not surface: even under a length cap they write two to three times as much, cover nearly every aesthetic aspect where humans are selective, engage each aspect more uniformly and at greater depth, and repeat themselves across critiques of the same photo where humans vary. We argue that reference-based similarity rewards a fluent, comprehensive critique style rather than the selectivity and specificity of human critique, and discuss implications for evaluating and training open-ended multimodal generation.

[26] Why Struggle with Continuous Latents? Interpretable Discrete Latent Reasoning via Rendered Compression cs.CLPDF

Shuochen Chang, Qingyang Liu, Shaobo Wang, Bingjie Gao, Qianli Ma

TL;DR: 本文提出了一种名为离散潜在推理（DLR）的新方法，旨在解决连续潜在推理方法训练不稳定和不可解释的问题。该方法通过将连续潜在状态转换为显式的离散标记，实现了在潜在空间中进行高效且可解释的推理。实验表明，DLR在多个推理基准上优于现有基线，并实现了高达20倍的压缩率。

Details

Motivation: 现有基于连续潜在空间的推理方法存在训练不稳定和推理轨迹不可解释的问题，这源于连续状态与离散符号监督之间的不匹配。本文旨在通过引入离散潜在标记来解决这一根本性挑战。

Result: 在五个推理基准和两个模型系列（Qwen3-VL和LLaMA-3）上的实验表明，DLR在性能上超越了先前的潜在推理基线，并实现了高达20倍的压缩。学习到的潜在轨迹保留了可解释的语义结构。

Insight: 核心创新点是将连续潜在状态离散化为可解释的标记，这通过受渲染压缩启发的流程实现，包括将思维链渲染为图像、提取视觉特征以及基于聚类的微调构建词汇表。这为高效、可控且可解释的潜在推理提供了新基础。

Abstract: Large language models achieve high reasoning performance via explicit chain-of-thought and reinforcement learning, but require long output sequences and extended inference time. Latent reasoning reduces this cost by shifting computation into a latent space; however, continuous latent methods are hard to train, suffering from unstable and uninterpretable reasoning trajectories. We argue these issues stem from a misalignment between continuous-space reasoning and discrete symbolic supervision, as continuous states lack explicit anchors for step-by-step alignment. To resolve this, we propose \textbf{Discrete Latent Reasoning(DLR)}, the first method that converts continuous latent states into explicit discrete tokens. Inspired by render-based compression, we render textual chains of thought into images, extract visual features, and construct a discrete latent vocabulary via clustering-based fine-tuning. Expanding the vocabulary and output head enables standard autoregressive modeling over both natural language and latent tokens, supporting pretraining alignment, SFT, and RL. Experiments on five reasoning benchmarks and two model series(Qwen3-VL and LLaMA-3) confirm that \textbf{DLR} outperforms prior latent reasoning baselines with up to \textbf{20$\times$ compression}. Furthermore, the learned latent trajectories retain an interpretable semantic structure. Overall, discrete latent tokens provide a controllable and interpretable basis for efficient latent reasoning.

[27] SEVA: Self-Evolving Verification Agent with Process Reward for Fact Attribution cs.CL | cs.AI | cs.LGPDF

Aojie Yuan, Yi Nian, Haiyue Zhang, Zijian Su, Yue Zhao

TL;DR: 本文提出了SEVA（自演进验证智能体），一种用于事实归因的结构化验证智能体，旨在解决大语言模型（LLM）智能体的幻觉问题。SEVA不仅输出二元标签，还生成证据对齐、逐步推理链、校准置信度和六类错误诊断等结构化输出。通过引入过程奖励（process reward）解决多组件输出的强化学习训练难题，并实现了验证、反思、探测、精炼的自演进循环。

Details

Motivation: 当前基于LLM的智能体存在幻觉问题，而现有的事实归因验证器仅输出不透明的二元标签，这导致智能体无法自我纠正，操作者也无法进行审计。因此，需要一种能提供结构化、可解释和可操作输出的验证器作为最后一道防线。

Result: 在ClearFacts基准测试上，SEVA-3B模型达到了69.0的F1分数，与GPT-4o-mini（69.8 F1）性能相当，同时输出更丰富、可审计的结果。在自演进过程中，模型在HaluEval基准上提升了15个百分点，但在TruthfulQA上有所下降，表明产生了基准专家而非通用模型。训练中，证据对齐从0.917提升至0.997，格式正确率从72%提升至100%，最终F1从64.9提升至69.0。

Insight: 核心创新在于提出了过程奖励机制，将验证质量分解为五个独立组件（70%权重给过程信号，30%给结果），解决了多组件输出强化学习中的优势崩溃（advantage collapse）问题，恢复了梯度并诱导了隐式课程学习。此外，结构化输出支持了验证->反思->探测->精炼的自演进循环，揭示了模型在迭代中会演变为基准专家而非通用模型的重要发现。其原则（奖励粒度必须与输出粒度匹配）可推广至其他多组件生成的RL任务。

Abstract: Hallucination is the reliability bottleneck for LLM-based agents, and fact attribution verifiers are the last line of defense – yet today’s verifiers emit only opaque binary labels, leaving agents unable to self-correct and operators unable to audit. We present SEVA, a structured verification agent that emits evidence alignments, step-by-step reasoning chains, calibrated confidence, and a six-category error diagnosis with actionable fixes. Training such an agent with RL is non-trivial: standard binary reward on multi-component output triggers advantage collapse – within-group reward variance vanishes and the GRPO gradient disappears. We resolve this with a process reward that decomposes verification quality into five independent components weighted 70/30 toward process signals, restoring the gradient and inducing an implicit curriculum – the agent first masters verification behavior (alignment 0.917 -> 0.997, format 72% -> 100%), then outcomes (F1 64.9 -> 69.0). Structured output further enables a Verify -> Reflect -> Probe -> Refine self-evolution loop, which over four rounds on a 7B model surfaces an unexpected structural finding: each round produces a benchmark-specialist, not a generalist (+15 pp on HaluEval, -10 to -14 pp on TruthfulQA in the same model, persistent at 4x data). On ClearFacts, SEVA-3B matches GPT-4o-mini (69.0 vs. 69.8 F1) while producing substantially richer, auditable output – confirming a principle that should generalize: for any RL task with multi-component generation, reward granularity must match output granularity.

[28] KbSD: Knowledge Boundary aware Self-Distillation for Behavioral Calibration in Agentic Search cs.CLPDF

Tao Feng, Xinke Jiang, Chao Wu

TL;DR: 本文提出KbSD（知识边界感知自蒸馏）框架，用于解决智能体搜索中知识边界校准的奖励稀疏性问题。该方法通过构建一个接收明确知识边界信号的教师模型来生成校准推理示范，并结合密集的令牌级监督、稀疏的结果级奖励以及象限自适应优化目标，以指导模型在不同知识状态下做出校准决策。

Details

Motivation: 现有基于强化学习的智能体搜索方法在知识边界校准（即决定何时信任参数记忆、何时依赖检索证据、何时弃权）方面受到奖励稀疏性的限制，二元奖励无法为跨不同知识状态的校准推理过程提供足够指导。

Result: 在多个基准测试上的实验表明，KbSD在任务准确性和幻觉缓解方面持续优于强基线模型，在稀疏奖励信息量最少的挑战性象限中提升最为显著。

Insight: 核心创新点在于信息不对称的自蒸馏框架，它利用内部结构相同的教师-学生架构，通过向教师注入知识边界信号（如参数确定性、检索质量、真实答案）来生成密集监督，而无需外部大模型；此外，针对不同知识状态推理分布的异质性，提出了象限自适应的蒸馏目标（反向KL、前向KL和帕累托最优双向KL），以精细化优化不同校准需求。

Abstract: Agentic search equips large language models with dynamic retrieval abilities, but existing reinforcement learning methods remain limited by reward sparsity in knowledge boundary calibration – deciding when to trust parametric memory, when to rely on retrieved evidence, and when to abstain. Binary rewards can penalize undesirable outcomes, but provide little guidance on the reasoning process required to make calibrated decisions across different knowledge states. To address this, we propose KbSD (Knowledge boundary Self-Distillation), a framework that tackles this limitation through dense token-level supervision, outcome-level sparse rewards, and quadrant-adaptive optimization. KbSD constructs a hint-augmented teacher, architecturally identical to the student, that receives explicit knowledge boundary signals – including parametric certainty, retrieval quality, and ground-truth answers – to generate calibrated reasoning demonstrations. This information-asymmetric self-distillation enables dense supervision without requiring a larger external model. To further account for the heterogeneous reasoning distributions across knowledge states, we introduce a quadrant-adaptive distillation objective: reverse KL for concentrated integration, forward KL for diverse refusal, and Pareto-optimal bidirectional KL for asymmetric quadrants requiring both precision and coverage. Experiments on multiple benchmarks show that KbSD consistently improves both task accuracy and hallucination mitigation over strong baselines, with the largest gains appearing in the challenging quadrants where sparse rewards are least informative.

[29] ARKD: Adaptive Reinforcement Learning-Guided Bidirectional KL Divergence Distillation for Text Generation cs.CL | cs.AIPDF

Zilong Liu, Xuewen Zhang, Jinrui Xing, Juyi Qiao, Huiyong Wang

TL;DR: 本文提出了一种自适应强化学习引导的双向KL散度蒸馏框架（ARKD），用于提升大型语言模型（LLM）在文本生成任务中的知识蒸馏效果。该方法通过强化学习策略网络动态调整前向与反向KL散度的权重，以同时优化主分布拟合与长尾概率建模，从而改善生成质量与泛化能力。

Details

Motivation: 现有基于单一KL目标的蒸馏方法难以平衡主分布拟合与长尾概率建模，限制了生成质量与泛化能力，因此需要一种更精细的分布对齐机制。

Result: 在多个基准测试上的实验表明，该方法在Rouge-L和BertScore指标上持续提升，超越贪婪启发式方法0.4-0.6点，并优于其他基线方法。

Insight: 创新点在于从理论与实证角度分析了前向与反向KL散度的互补作用，并引入强化学习策略进行自适应加权，实现了对主模式与长尾模式的双重对齐，为分布对齐提供了新思路。

Abstract: Knowledge distillation (KD) is a key technique for compressing Large Language Models (LLMs), yet methods relying on a single KL objective often fail to balance primary distribution fitting with long-tail probability modeling, limiting both generation quality and generalization. To address this, we analyze the complementary roles of forward and reverse KL divergence (FKL/RKL) in distribution alignment from theoretical and empirical perspectives. We then propose a reinforcement-learning-based adaptive KL-weighted distillation framework, in which a policy network dynamically assigns weights to FKL and RKL based on teacher-student distributional characteristics, guided by immediate reward signals to achieve dual alignment on principal and long-tail modes. Extensive experiments demonstrate consistent improvements across Rouge-L and BertScore metrics, surpassing greedy heuristics by 0.4-0.6 points and outperforming other baseline methods on diverse benchmarks.

[30] Clinical Reasoning Graphs: Structured Evaluation of LLM Diagnostic Reasoning Reveals Competence Without Consistency cs.CL | cs.AI | q-bio.QMPDF

Nisarg A. Patel

TL;DR: 本文提出临床推理图（clinical reasoning graphs），一种基于领域本体的结构化图表示方法，用于评估大型语言模型（LLM）在临床诊断推理中的过程。研究发现，尽管LLM在复杂临床病例基准上达到60-70%的诊断准确率，但其推理过程缺乏跨病例的一致性模式，表明准确率不能完全反映稳定的临床推理能力。

Details

Motivation: 当前LLM在临床诊断基准上表现出较高的准确率，但仅凭准确率无法区分模型是基于稳定的临床推理还是简单的模式匹配。因此，需要一种结构化方法来评估LLM诊断推理的过程一致性。

Result: 在五个LLM对50个《新英格兰医学杂志》临床病理会议病例的750条诊断轨迹分析中，临床相似病例间的图相似性与不相似病例间几乎相等，且正确诊断与错误诊断的图结构相似度相近（0.488 vs 0.484）。结构化反思提示能增加轨迹内显性特征分析（+33%），但未提高跨病例一致性。

Insight: 创新点在于引入基于本体的临床推理图来结构化提取和评估LLM诊断轨迹，揭示了模型诊断能力与推理过程一致性之间的脱节。该方法强调需要结合过程级评估来补充最终答案准确率，为LLM临床推理评估提供了新工具和资源。

Abstract: Modern large language models (LLMs) reach 60-70% diagnostic accuracy on complex clinical case benchmarks, but accuracy alone cannot distinguish stable clinically-grounded reasoning from pattern matching. We introduce clinical reasoning graphs, structured graph representations extracted from free-text LLM diagnostic traces using a domain-grounded ontology with 5 node types and 7 edge types. We apply this pipeline to 750 traces from five LLMs across 50 New England Journal of Medicine Clinicopathological Conference cases and three prompt conditions, and test whether diagnostic traces show stable structured reasoning patterns, or diagnostic schemas, for clinically similar cases. We operationalize this as higher graph similarity among clinically similar cases than among clinically dissimilar ones. Across 15 model-condition comparisons, within-cluster and between-cluster composite similarity are nearly equal, and no comparison survives multiple-testing correction; a component-level analysis finds any residual content signal far below schema scale. Graph similarity is also nearly identical for pairs of models that are both correct (0.488) and both incorrect (0.484), suggesting that graph structure captures a dimension not reflected in diagnostic accuracy. Structured reflection prompting increases explicit discriminating-feature analysis within traces (+33%) but does not increase cross-case consistency. These results show diagnostic competence without schema-scale reasoning consistency, and indicate that final-answer accuracy should be complemented by process-level evaluation. We release the ontology, extraction pipeline, validation protocol, and the extracted reasoning graphs and similarity artifacts as resources for structured evaluation of LLM clinical reasoning.

[31] Towards Physical Intuitions for Alignment Dynamics: A Case Study With Randomness Crystallization cs.CLPDF

Kunal Samanta, Ari Holtzman, Peter West

TL;DR: 本文提出了一种基于物理科学（特别是热力学相变理论）的框架来理解语言模型对齐过程中的动态变化。作者以材料结晶过程为类比，将随机数生成任务的对齐过程划分为三个阶段：预训练模型的高熵液态阶段、监督微调引发的成核阶段以及强化学习导致的沉降阶段。

Details

Motivation: 当前对齐研究主要依赖能力基准测试，而对模型在训练后如何变化的动态过程理解不足。作者认为物理科学（尤其是热力学相变理论）能为理解这些动态提供原则性且未被充分探索的视角。

Result: 作者提出了直观的度量方法来验证这些相变阶段，并在多种随机任务上验证了这一理念。

Insight: 创新点在于将材料结晶的相变理论引入对齐研究，为理解对齐诱导的结构来源、收敛原因及根本性限制提供了物理直觉框架。这代表了一类更广泛的物理框架，可用于解释对齐动态的本质。

Abstract: The alignment of language models is typically studied through the lens of capability benchmarks, but the dynamics of how models change during post-training remain poorly understood. We argue that the physical sciences, and thermodynamic phase-transition theory in particular, offer a principled and underexplored vocabulary for reasoning about these dynamics. As a case study, we instantiate this position through the lens of material Crystallization, which is a well-studied thermodynamic phase transition. For tasks like random number generation, this breaks into 3 phases: (1) the high entropy liquid phase in the pretrained model, with many distinct sampling distributions promptable from the model; (2) the nucleation phase caused by supervised finetuning, in which behavior collapses onto a single seed distribution present in the pretrained LLM; and (3) a settling phase in which reinforcement learning techniques redistribute probability of the collapsed distribution, but largely keep it concentrated on the same options as the seed distribution. We propose intuitive metrics to verify the transitions between these phases, and validate the idea across a range of random tasks. Crystallization is one instance of a broader class of physical frameworks we believe alignment research should import to answer questions about where alignment-induced structure comes from, why it converges where it does, and what it fundamentally cannot change.

[32] LatentRevise: Learning from Zero-Hit Reasoning cs.CLPDF

Yiqiu Guo, Xueting Han, Qi Jia, Guangtao Zhai, Jing Bai

TL;DR: 论文提出LatentRevise方法，针对强化学习与可验证奖励（RLVR）中零命中提示（即正确轨迹采样概率极低）的问题，通过利用失败轨迹的信息，在词汇嵌入的凸包内优化输入嵌入，使前缀远离错误延续并接近正确答案，从而生成包含自我反思和正确解答的修订轨迹，用于提升监督微调（SFT）和RLVR在数学基准上的性能。

Details

Motivation: 解决RLVR中因硬提示导致正确轨迹采样概率低、在有限预算下无法获取有效训练信号的问题，利用失败轨迹揭示模型推理错误之处，从零命中区域中恢复训练信号。

Result: 在数学基准测试上，使用修订轨迹作为训练数据，相比标准基线，改善了SFT和RLVR的性能，达到了SOTA水平。

Insight: 创新点在于提出一阶潜在修订方法，将失败轨迹与正确答案作为锚点，在词汇嵌入凸包约束下优化输入嵌入，生成可引导正确推理的修订前缀，从而有效利用零命中区域的失败信息进行模型训练。

Abstract: Reinforcement learning with verifiable rewards (RLVR) is bottlenecked by hard prompts on which correct trajectories have low probability, so sampling misses them within a practical budget and leaves the policy update with little useful signal. We frame such zero-hit prompts as RLVR’s sampling frontier, where new reasoning behavior is most valuable yet least likely to be sampled. Importantly, failed rollouts can be informative: they expose where the model’s reasoning went wrong. We introduce LatentRevise, a first-order latent revision method that recovers training signal for this zero-hit regime. Given a failed rollout and the gold answer as an anchor, LatentRevise optimizes the input embeddings of its reasoning prefix under two complementary gradients, moving the prefix away from the failed continuation and toward the gold answer. The optimization is constrained to the convex hull of the model’s vocabulary embeddings, so each update moves the latent toward a real token embedding rather than an arbitrary feature direction. We find that continuations from the revised prefix lengthen, exhibit self-reflection, and reach correct answers missed by the original rollouts. Used as training data, these trajectories improve SFT and RLVR on math benchmarks over standard baselines.

[33] Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning cs.CLPDF

Sangmook Lee, Minbeom Kim, Jeonghye Kim, Dohyung Kim, Sojeong Rhee

TL;DR: 本文针对大语言模型数学推理中的多样性评估问题，指出现有指标主要捕捉表面层面的变化，而非解题策略的差异。作者提出了‘方法层面多样性’的概念，即同一问题正确解法中策略的差异，并通过实验发现现有多样性指标与该方法层面多样性不匹配，且方法多样性候选集能提升测试时的扩展性能，但直接优化方法多样性仍具挑战性。

Details

Motivation: 现有大语言模型数学推理的多样性指标大多只关注表面层面的变化，忽略了解题策略的差异，这限制了模型在探索和推理中的真实多样性表现，因此需要引入更细粒度的‘方法层面多样性’来填补这一空白。

Result: 通过人类校准的LLM评判框架，实验表明先前的多样性指标是方法层面多样性的不可靠代理，且这种不匹配延续到多样性感知的RLVR中；方法多样性候选集能改善测试时的扩展性能，但直接优化方法多样性奖励会导致模型利用评判者偏好而非拓宽方法。

Insight: 创新点在于提出了‘方法层面多样性’这一新概念，揭示了表面层面与方法层面信号之间的系统性差异，这有助于推动大语言模型实现更真实、类人的多样化推理方式；从客观角度看，该研究强调了评估指标设计的重要性，为未来优化模型推理多样性提供了新方向。

Abstract: Diversity in LLM mathematical reasoning is critical for exploration, but common diversity metrics mostly capture surface-level variation rather than differences in how a problem is solved. We address this gap by introducing approach-level diversity: variation in strategies across correct solutions to the same problem. Using a human-calibrated LLM judge framework, we show that prior diversity measures are unreliable proxies for approach-level diversity, and this mismatch carries over to diversity-aware RLVR, where target metrics are preserved while approach-level diversity declines. Investigating when approach-level diversity helps and whether it can be directly induced, we find that approach-diverse candidate sets improve test-time scaling. However, optimizing an LLM judge diversity reward during training causes the policy to exploit judge-specific preferences rather than broaden its approaches, leaving direct optimization of approach-level diversity as an open problem. Together, our work introduces the notion of approach-level diversity and uncovers a systematic divergence between surface- and approach-level signals, marking a step toward LLMs that reason in genuinely diverse, human-like ways.

[34] Not-quite-human tastes: the stylized omnivorousness of LLM survey surrogates cs.CL | econ.GNPDF

Xiangyu Ma, Mengmi Zhang, Shannon Ang, Minne Chen

TL;DR: 本研究探讨了大型语言模型（LLM）作为调查替代者（silicon surrogates）在模拟人类文化品味方面的能力。通过使用OpenAI、Anthropic和DeepSeek的LLM生成大量模拟受访者数据，并与‘公共艺术参与调查’（SPPA）的真实数据对比，发现LLM生成的品味数据存在系统性偏差，包括过度积极的偏好、失去真实品味结构的复杂性，以及扭曲了品味与社会空间（如年龄、阶级、性别和种族）之间的已知关联。

Details

Motivation: 随着市场研究公司开始提供‘合成’调查面板，以及LLM生成响应可能污染标准调查数据，评估LLM在模拟文化品味方面的算法保真度和对齐性变得日益紧迫。本研究旨在实证检验LLM是否能合理近似人类文化品味，以应对这一开放性问题。

Result: 在SPPA基准上，LLM生成的硅样本表现出高度风格化的品味模拟：存在系统性积极偏好偏差，导致品味估计膨胀；真实品味结构的复杂关系性完全丧失；品味与社会空间（年龄、阶级、性别、种族）之间的文化对齐性被严重削弱或扭曲。

Insight: 创新点在于将硅采样研究扩展到文化消费领域，系统量化了LLM作为调查替代者的局限性，揭示了其品味模拟的‘非人类’特征（如积极偏差和关系性丧失）。客观分析表明，这提醒我们LLM在社会科学应用中的风险，需谨慎处理其生成数据的保真度，避免误导性结论。

Abstract: Large-language models have proven to be remarkable if inconsistent parrots of public attitudes and opinions. The extent to which LLMs are able to produce reasonable approximations of cultural taste remains an open empirical question that becomes more urgent by the day, with market research companies already offering provisional `synthetic’ survey panels and the contamination of standard survey data from LLM-generated responses. In this study, we build on past work on silicon sampling by extending considerations of its algorithmic fidelity and alignment to the domain of cultural consumption. We use large-language models from OpenAI, Anthropic, and DeepSeek to each produce 277,470 (30x9249) silicon surrogates of survey respondents from the Survey of Public Participation in the Arts (SPPA). We find these silicon surrogates’ tastes to be highly stylized facsimiles of human tastes. (1) Silicon samples have a systematic postive-bias for liking, resulting in inflated ecological estimates of tastes. The individual-level bias of silicon samples are not well-explained by the WEIRD-bias often discussed in the literature. (2) The complex relationality in real taste structures is completely lost among silicon samples. (3) Finally, very little of the known cultural alignment between tastes and social space are preserved. Silicon samples attenuate age-taste associations, resurrect anachronistic class-taste associations, caricaturize gender- and race-taste associations.

[35] Efficient Retrieval-Augmented Generation via Token Co-occurrence Graphs cs.CL | cs.IRPDF

Gianluca Bonifazi, Christopher Buratti, Michele Marchetti, Federica Parlapiano, Giulia Quaglieri

TL;DR: 本文提出了一种名为TIGRAG的高效检索增强生成框架，通过构建基于词元共现的知识图谱来解决传统RAG方法在多跳推理上的不足。该方法利用滑动窗口共现统计直接建模词元间的拓扑关系，实现了可扩展的图谱构建，并结合基于图谱的语义扩展与神经重排序来检索用于多跳推理的互连证据。

Details

Motivation: 标准RAG方法在处理需要多跳推理的问题时存在困难，而现有的基于图谱的RAG方法通常依赖于计算成本高且容易出错的、基于LLM的抽取流程。本文旨在解决这些效率与可靠性问题。

Result: 在三个广泛采用的多跳问答基准测试上，TIGRAG在检索和下游问答任务中均持续优于密集检索和基于图谱的RAG方法，同时显著减少了索引时间、推理延迟和提示占用空间。

Insight: 核心创新在于提出了一种基于词元共现统计的轻量级知识图谱构建方法，避免了复杂的LLM抽取流程，并引入了迭代的实体驱动检索策略，通过从已检索上下文中提取的桥接实体来渐进式扩展查询，从而高效支持多跳推理。

Abstract: Retrieval-Augmented Generation (RAG) mitigates hallucinations in Large Language Models (LLMs) by grounding the generation process on external knowledge. However, standard RAG approaches struggle with multi-hop reasoning. While recent graph-based RAG methods improve the retrieval of interconnected chunks, they often rely on computationally expensive and error-prone LLM-based extraction pipelines. To address these issues, we propose TIGRAG (Token-Induced GraphRAG), an efficient graph-augmented RAG framework based on a token co-occurrence Knowledge Graph. TIGRAG directly models topological relationships between tokens using sliding-window co-occurrence statistics, thus enabling scalable graph construction. During inference, it combines graph-based semantic expansion and neural reranking to retrieve interconnected evidence for multi-hop reasoning. Specifically, it introduces an iterative entity-driven retrieval strategy that progressively expands the query using bridging entities extracted from previously retrieved contexts. We evaluated TIGRAG on three widely adopted multi-hop Question Answering (QA) benchmarks. Experimental results demonstrated that our framework consistently outperforms dense retrieval and graph-based RAG methods in both retrieval and downstream QA tasks, while substantially reducing indexing time, inference latency, and prompt footprint.

[36] CORTEX: High-Quality Cross-Domain Organization of Web-Scale Corpora through Ontological Corpus Graph cs.CLPDF

Chengtao Gan, Xiaoke Guo, Yushan Zhu, Zhaoyan Gong, Zhiqiang Liu

TL;DR: 本文提出了Cortex框架，首次通过三层异构的本体语料图（OCG）将网络规模语料构建从平面文档过滤提升为结构化知识组织，包含质量精炼内容层、LLM驱动的轻量级本体层和跨领域对齐层。实验验证了其有效性，并基于OCG构建了跨领域搜索与推理基准CortexBench，在八个前沿大语言模型上评估了质量精炼、领域组织和跨领域数据合成的效果。

Details

Motivation: 随着大语言模型发展对数据规模和质量要求不断提高，不同训练阶段需要更定制化的数据需求，现有语料构建流程仅产生扁平、无差别的文档集合，普遍缺乏系统化的知识组织，因此需要一种结构化方法来系统组织高质量语料。

Result: 通过综合实验验证了Cortex框架的有效性，特别是在构建的跨领域搜索与推理基准CortexBench上，对八个前沿大语言模型的评估结果表明了其在质量精炼、领域组织和跨领域数据合成方面的有效性。

Insight: 创新点在于首次提出了本体语料图（OCG）这一三层异构结构，将语料组织从平面过滤转变为结构化知识图谱，并通过LLM驱动实现本体的自动化演进和跨领域任意分类粒度的对齐，为高质量、可系统组织的语料构建提供了新范式。

Abstract: The continuous evolution of large language models drives escalating demands on data scale and quality, and as different training stages impose increasingly tailored data requirements, systematic organization of high-quality corpora becomes indispensable. Existing corpus construction pipelines confine the resulting corpora to flat, undifferentiated document collections, universally lacking systematic knowledge organization. We present Cortex, to our knowledge the first framework that elevates web-scale corpus construction from flat document filtering to structured knowledge organization through an Ontological Corpus Graph (OCG), a three-layer heterogeneous structure unifying a quality-refined content layer, a hierarchical lightweight ontology layer via LLM-driven automated evolution, and a cross-domain alignment layer enabling inter-domain association at arbitrary taxonomic resolution. Comprehensive experiments confirm the effectiveness of Cortex. In particular, we leverage the OCG to synthesize CortexBench, a cross-domain search-and-reasoning benchmark whose evaluation across eight frontier LLMs validates the effectiveness of quality refinement, domain organization, and cross-domain data synthesis. We will publicly release the complete codebase, a 24.14B-token refined corpus with its OCG, and CortexBench.

[37] Forewarned is Forearmed: When Non-Sequential Embedding Turns Into an Anomaly Detector cs.CL | cs.AI | cs.LG | eess.ASPDF

Elys Allesiardo, Antoine Caubrière, Valentin Vielzeuf

TL;DR: 本文深入分析了非序列多模态句子级嵌入，特别是SONAR模型，发现某些嵌入维度对扰动敏感，可作为解码异常的指标。通过利用连续编码与解码之间的一致性，成功构建了一个精确的异常检测器，并尝试修改特定维度以进行纠正。

Details

Motivation: 论文旨在通过理解嵌入本身来增强多模态表示的可靠性，解决解码异常检测问题。

Result: 该方法成功构建了准确的异常检测器，但摘要未提及具体基准测试或定量结果。

Insight: 创新点在于揭示非序列嵌入维度对扰动的敏感性可作为异常检测指标，并探索通过修改嵌入维度来纠正异常，强调了分析嵌入本身的重要性。

Abstract: This paper offers an in-depth analysis of non-sequential multimodal sentence-level embeddings, with a particular focus on the SONAR model. We demonstrate that certain embedding dimensions are sensitive to perturbations and can serve as indicators of decoding anomalies. By leveraging the consistency between successive encoding and decoding, we successfully build an accurate detector. Additionally, we explore modifying specific dimensions of interest to attempt to correct them. This work underscores the importance of understanding and analyzing the embeddings themselves to enhance the reliability of multimodal representations.

[38] DAIN: Dynamic Agent-Based Interaction Network for Efficient and Collaborative Multimodal Reasoning cs.CLPDF

Xinxin Chen, Yuchen Li, Zihan Wang, Haoyu Zhang, Ruixin Liu

TL;DR: 本文提出了一种基于动态代理的交互网络（DAIN），用于高效协作的多模态推理。DAIN将多模态融合重新概念化为一个动态的多代理协作过程，通过上下文感知的元控制器动态调度稀疏激活的专用交互代理，并协调压缩的代理间通信以达成共识。

Details

Motivation: 当前基于静态混合专家（MoE）架构的多模态融合方法难以满足复杂现实应用对自适应和高效协作推理的需求，因此需要一种动态、高效的协作框架。

Result: 在ADNI、MIMIC-IV、MM-IMDB、CMU-MOSI和ENRICO五个多样化基准测试上的综合评估表明，DAIN达到了新的最先进水平（SOTA），例如在ADNI上实现了2.6%的准确率提升。消融研究验证了动态调度和代理通信的关键作用。

Insight: 创新点在于将多模态融合建模为动态多代理协作过程，通过元控制器实现上下文感知的稀疏激活和压缩通信，同时联合优化任务准确性、代理专业化和操作效率，在保持计算效率的同时增强了可解释性。

Abstract: Current multimodal fusion approaches, particularly those based on static Mixture-of-Experts (MoE) architectures, often struggle to provide the adaptive and efficient collaborative reasoning required by complex real-world applications. We introduce the Dynamic Agent-based Interaction Network (DAIN), which reconceptualizes multimodal fusion as a dynamic, multi-agent collaborative process. DAIN employs a context-aware Meta-Controller that dynamically schedules sparse activation of specialized interaction agents and orchestrates compressed inter-agent communication for consensus-building. The framework is guided by a multi-objective loss function that jointly optimizes task accuracy, agent specialization, and operational efficiency through sparse activation and communication regularization. Comprehensive evaluations across five diverse benchmarks – ADNI, MIMIC-IV, MM-IMDB, CMU-MOSI, and ENRICO – establish DAIN as a new state-of-the-art, delivering significant performance improvements including a 2.6% accuracy gain on ADNI. Ablation studies verify the critical roles of both dynamic scheduling and agent communication. Furthermore, DAIN offers enhanced interpretability by exposing context-dependent agent roles and collaboration patterns while maintaining computational efficiency through sample-wise sparse agent activation. Our work demonstrates the promise of dynamic, agent-based paradigms for multimodal reasoning.

[39] Before Thinking, Learn to Decide: Proactive Routing for Efficient Visual Reasoning cs.CLPDF

Yinan Zhou, Haokun Lin, Yichen Wu, Caifeng Shan, Zhenan Sun

TL;DR: 本文提出了一种名为PRP（Proactive Routing Paradigm）的前瞻性路由范式，旨在解决大型多模态模型推理效率低下的问题。该方法通过联合评估草稿模型和目标模型的能力，在推理早期做出路由决策，将查询自适应地分配给合适的模型，从而在不牺牲整体性能的前提下显著加速推理过程。

Details

Motivation: 现有方法在多模态场景下建立可靠的查询难度信号存在瓶颈：基于后验令牌概率的方法不适用，而依赖监督微调的策略则对数据敏感。它们都在完整输出后才进行路由，且忽略了目标模型是否真能解决被路由的实例。

Result: 在多个多模态推理基准测试上的广泛实验验证了该方法的有效性和效率，表明其能大幅加速推理且不影响整体性能。

Insight: 核心创新在于提出了一个前瞻性路由范式，通过草稿模型评级学习和联合评级学习，使模型具备内部置信度估计和预测目标模型处理能力，从而实现细粒度的、实例级的早期路由决策，优化了计算资源的分配。

Abstract: Large multimodal models have achieved strong reasoning on complex visual tasks, but their inference efficiency is often restricted by long chains of thought. A promising solution is to pair a small draft model with a large target model, enabling cooperative inference employing a routing signal that adaptively routes queries to either the draft or target model based on their difficulties for optimal efficiency and accuracy. Yet, the remaining bottleneck is to establish a reliable query difficulty signal under multimodal settings. Existing approaches designed for language models either rely on post-hoc token probabilities, which fall short in multimodal scenarios, or depend on supervised fine-tuning, which is a data-sensitive strategy. Both paradigms perform routing only after a complete output, and ignore whether the target model can actually solve the routed instances. To address this, we propose PRP, a Proactive Routing Paradigm that enables early decision-making by jointly evaluating the competence of both the draft and target models. Our Draft Rating Learning (DRL) equips the draft model with an internal confidence estimator, while Joint Rating Learning (JRL) predicts how well the target model can handle a given query, thereby prioritizing the allocation of samples it excels at rather than the hardest ones. These ratings enable fine-grained, instance-level \textbf{Proactive Routing} and substantially accelerate inference without compromising overall performance. Extensive experiments across multiple multimodal reasoning benchmarks validate our effectiveness and efficiency.

[40] Grounding LLM Reasoning under Incomplete Graph Evidence cs.CLPDF

Jiaqi Li, Fanghui Song

TL;DR: 本文提出了一种理论框架，用于在大语言模型（LLM）推理过程中，对基于不完整知识图谱证据的轨迹进行“软”接地（grounding）。该框架将不完整的证据状态（包括实体锚点、关系残差、路径能量和支持区域）与LLM提供的先验分布相结合，通过KL正则化变形来处理开放世界不完整性，避免了硬性规则带来的问题。

Details

Motivation: 动机在于解决LLM利用知识图谱进行推理时面临的核心问题：系统所见的图谱通常是检索、链接、有时间范围且不完整的证据状态，而非完整的真实情况。现有基于观察状态的硬性规则无法同时拒绝所有错误的、无支持的轨迹并保留所有真实但未被观察到的轨迹。

Result: 论文提出了一个理论框架，并推导了在证据扰动下的稳定性界限。该框架阐明了适用于GraphRAG、KGQA、图智能体、约束解码和忠实生成等任务的约束机制。

Insight: 核心创新点在于将“软接地”形式化为对LLM先验的KL正则化变形，允许有限的松弛以支持未被观察但未被反驳的轨迹，而硬性条件化则被视为无限惩罚的极限。这提供了一种处理证据不完整性的原则性方法，并将知识图谱兼容性视为声明的支持而非事实真相，这是一种证据相对性的主张。

Abstract: Knowledge graphs can guide large language models (LLMs) reasoning, but the graph seen by a system is usually a retrieved, linked, temporally scoped, and incomplete evidence state rather than a complete account of truth. We develop a theoretical perspective on grounding observable LLM trajectories under such incomplete graph evidence.The evidence state induces entity anchors, typed relation residuals, path energies, and support regions, while the language model supplies a prior over candidate trajectories. We show that, under open-world incompleteness, no hard rule based only on the observed state can both reject every false unsupported trajectory and retain every true-but-unobserved one.We then characterize soft grounding as a KL-regularized deformation of the LLM prior: finite slack preserves support for unsupported but non-contradicted trajectories, whereas hard conditioning appears as an infinite-penalty limit.The framework also yields stability bounds under evidence perturbations and clarifies the constraint regimes appropriate for GraphRAG, KGQA, graph agents, constrained decoding, and faithful generation. The claims are evidence-relative: KG compatibility is treated as declared support, not factual truth.

[41] MOPD: Multi-Teacher On-Policy Distillation for Capability Integration in LLM Post-Training cs.CL | cs.LGPDF

Wenhan Ma, Jianyu Wei, Liang Zhao, Hailin Zhang, Bangjun Xiao

TL;DR: 本文提出了一种名为多教师策略蒸馏（MOPD）的后训练范式，用于将多个领域强化学习教师模型的能力整合到一个学生大语言模型中。该方法首先通过各领域专门的强化学习获得一组教师模型，然后在学生模型自身的轨迹上进行蒸馏，以消除暴露偏差并提供密集的优化信号。

Details

Motivation: 现代大语言模型依赖后训练中的强化学习来提升特定能力，但将多种能力整合到一个模型中仍然很困难。现有方法如离策略微调和Mix-RL存在效率低下或性能损失的问题。

Result: 在Qwen3-30B-A3B模型上，MOPD超越了Mix-RL、Cascade RL、离策略微调和参数合并等基线方法，几乎继承了每位教师的所有能力。该方法已成功部署于工业级前沿模型MiMo-V2-Flash的后训练中，证明了其实际价值。

Insight: 论文的创新点在于提出了一种基于学生自身轨迹的、消除暴露偏差的多教师策略蒸馏方法，实现了能力的有效整合。其核心优势在于允许领域教师模型并行、独立开发，消除了多领域后训练中典型的跨领域耦合，为前沿规模大语言模型的能力集成提供了实用方案。

Abstract: Modern large language models (LLMs) rely on reinforcement learning during post-training to push specific capabilities, yet integrating multiple capabilities into one model remains hard. Existing methods, such as Off-Policy Finetune and Mix-RL, are either inefficient or lose performance. In this work, we propose Multi-teacher On-Policy Distillation (MOPD), a post-training paradigm for combining the capabilities of multiple domain RL teachers: we first run per-domain specialised RL to obtain a set of domain teachers, then distill these teachers into the student on its own rollouts. This eliminates exposure bias and provides a dense optimization signal. On Qwen3-30B-A3B, MOPD outperforms Mix-RL, Cascade RL, Off-Policy Finetune, and Param-Merge baselines, inheriting nearly all of each teacher’s capability. MOPD also enables parallel, independent development of domain teachers, removing the cross-domain coupling typical of multi-domain post-training. MOPD has been deployed in the post-training of MiMo-V2-Flash, an industrial-scale frontier model, demonstrating its practical value for capability integration in frontier-scale LLMs.

[42] Regime-Aware Peer Specialization for Robust RAG under Heterogeneous Knowledge Conflicts cs.CLPDF

Bo Wang, Heyan Huang, Yaolin Li, Yanghao Zhou, Jiahao Teng

TL;DR: 本文提出了一种名为RAPS-DA的机制感知对等专业化框架，用于解决检索增强生成（RAG）中检索上下文与模型参数知识冲突的鲁棒性问题。该方法在样本层面将冲突划分为三种机制，并为每种机制训练一个同规模的专业化对等模型；在token层面则通过双层选择器动态调整监督信号。实验表明，该方法在多种冲突场景和分布外基准测试中超越了现有基线方法。

Details

Motivation: 现有方法在处理检索上下文与模型参数知识之间的异质性冲突时，通常采用机制无关的监督方式，这可能会混淆不同可靠性机制下不兼容的学习信号。本文旨在解决这一问题，以提升RAG在知识冲突下的鲁棒性。

Result: 在五种冲突场景和两个分布外基准测试上的实验结果表明，RAPS-DA超越了所有基于提示、解码、微调、强化学习和单教师模型的基线方法。

Insight: 核心创新点在于提出了机制感知的对等专业化框架，在样本层面进行机制划分与硬路由，在token层面通过双层选择器实现动态、细粒度的监督信号过滤与加权，从而在不增加模型规模或部署开销的情况下实现鲁棒性提升。

Abstract: Retrieval-augmented generation (RAG) improves language models by grounding generation in external context. However, it can be fragile when the retrieved context conflicts with the model’s parametric knowledge. Such conflicts span a reliability spectrum, ranging from reliable and partially reliable evidence to adversarial context. Existing remedies often handle such heterogeneous conflicts with regime-agnostic supervision, which can conflate incompatible learning signals across reliability regimes. To disentangle these signals, we propose RAPS-DA, a regime-aware peer specialization framework that addresses conflict at two complementary granularities. At the sample level, conflicts are divided into three regimes, including Grounding, Arbitration, and Resistance, with one same-scale peer specialist trained per regime from a shared base model. Each sample is then hard-routed to its regime-matched peer for on-policy reverse-KL supervision. At the token level, a dual-layer selector uses inter-teacher disagreement, student-teacher divergence, and student entropy to filter uninformative or unstable tokens, upweight confidently misaligned ones, and gradually focus supervision on high-conflict tokens as the student matures. Gains stem from specialization at a fixed model scale, not from a stronger teacher, and the peer specialists exist only during training, so the deployed student requires no regime labels or peer access. Experiments on five conflict scenarios and two out-of-distribution benchmarks show RAPS-DA surpasses all prompting, decoding, fine-tuning, RL, and single-teacher baselines.

[43] Uncertainty-Aware Generation and Decision-Making Under Ambiguity cs.CL | cs.LGPDF

Nico Daheim, Iryna Gurevych

TL;DR: 本文提出并评估了基于贝叶斯决策理论和风险规避决策的不确定性感知决策算法，应用于LLM驱动的辅导和自动同行评审任务。研究通过考虑辅导策略和评审分数的不确定性，并利用共形预测提供保证，发现这些算法能提升生成内容的效用，但在高模糊性场景下需谨慎实施。

Details

Motivation: 随着LLM能力快速提升并用于复杂现实任务，许多任务具有高度主观性且要求模型输出可信，但决策算法研究相对不足，因此需要开发不确定性感知的决策方法来提升LLM在模糊环境下的决策质量。

Result: 在辅导和自动同行评审任务上的实证研究表明，不确定性感知算法能提高生成效用，但风险规避规则可能因优化通用输出而降低性能，而贝叶斯方法表现更优；研究使用共形预测为策略和分数提供统计保证。

Insight: 创新点在于将贝叶斯决策理论和风险规避决策引入LLM生成过程，明确处理任务模糊性，并通过共形预测提供不确定性量化保证；客观来看，该方法为LLM在主观性任务中的可靠决策提供了可扩展的理论框架。

Abstract: With rapidly improving capabilities, Large Language Models (LLMs) are increasingly used in many complex real-world tasks. Beyond requiring in-depth knowledge and reasoning skills, many of these tasks exhibit a high degree of subjectivity and require that the outputs of the model can be trusted. While a lot of progress has been made to train better models, decision-making algorithms have received less attention. In this work, we present and evaluate various uncertainty-aware decision-making algorithms based on Bayesian decision theory and risk-averse decision making on the tasks of tutoring and automatic peer reviewing. Concretely, we take uncertainty over tutoring strategies and review scores into account when generating a tutor response or review and use conformal prediction to provide guarantees over strategy and score. We find empirically that these algorithms can improve the utility of the generations but need to be carefully implemented when ambiguity is high. For example, risk-averse rules can degrade performance by optimizing for generic outputs, while Bayesian methods tend to perform better. Our work uses techniques from decision theory to improve LLM-based decision-making and outlines open challenges for the community.

cs.CV [Back]

[44] RADIANT-PET: Reasoning-Augmented PET/CT Lesion Segmentation with Large Language Models and Reinforcement Learning cs.CV | cs.AI | cs.LGPDF

Jiasheng Wang, Tanun Jitwatcharakomol, Piyawadee Jongpradubgiat, Simeng Zhu

TL;DR: 本文提出了RADIANT-PET，一个用于PET/CT病灶分割的推理增强框架。该方法首先通过一个高敏感度的体素级分割模型生成候选病灶区域，然后将其转化为结构化的文本描述，并利用大语言模型（LLM）结合放射学报告进行病灶真伪分类。此外，通过强化学习优化LLM，以提升病灶分类和解剖位置分配的准确性。

Details

Motivation: PET/CT中的病灶精准分割对肿瘤学至关重要，但由于生理性示踪剂摄取和伪影可能模仿恶性信号，该任务仍具挑战性。现有方法缺乏病灶级别的语义推理能力。

Result: 在AutoPET和OSU测试队列上，RADIANT-PET持续优于仅使用图像的强基线方法，尤其在提供放射学报告时改进最大，实现了SOTA水平的性能。

Insight: 创新点在于将体素级分割与病灶级LLM推理相结合，为传统分割增加了新颖的语义推理层。通过将图像特征转化为结构化文本描述，并利用LLM和强化学习进行裁决，有效抑制了生理性假阳性，并使分割结果与临床解读保持一致。

Abstract: Accurate lesion segmentation in PET/CT is critical for oncology, yet remains challenging because physiologic tracer uptake and artifacts can mimic malignant signal. We present RADIANT-PET, a reasoning-augmented framework that couples a high-sensitivity voxel-level segmentation model with lesion-level large language model (LLM) adjudication. Candidate uptake regions are generated with a deliberately permissive segmentation stage, then converted into structured textual descriptions that summarize uptake intensity, morphology, and regional and global anatomical context. An LLM classifies each candidate as true lesion vs. false positive, optionally leveraging the radiology report as additional clinical context. To strengthen lesion-level reasoning, we further optimize a local LLM via reinforcement learning using Group Relative Policy Optimization, rewarding correct lesion classification and anatomically concordant site assignment. Across AutoPET and an OSU test cohort, RADIANT-PET consistently outperforms strong image-only baselines, with the largest improvements observed when radiology reports are provided. Overall, these results demonstrate that LLM-based lesion-level reasoning adds a novel reasoning layer beyond conventional segmentation, suppressing physiologic false positives and aligning voxel-level predictions with clinical interpretation. The project repository is available at: https://github.com/jwang-580/RADIANT-PET.

[45] SoccerNet 2026 Player-Centric Ball Action Spotting: Per-Player Attention with Agreement-Based Ensembling cs.CVPDF

Faisal Altawijri, Ismail Mathkour

TL;DR: 本文介绍了针对SoccerNet 2026以球员为中心的球动作定位挑战的解决方案，采用两阶段流程：第一阶段使用Track-Aware Action Detector（TAAD）从广播视频中生成每位球员的动作逻辑值，第二阶段通过Denoising Sequence Transduction（DST）变换器将比赛状态特征和TAAD逻辑值转换为结构化事件序列。通过引入时间变换器增强TAAD的跨帧上下文建模，并采用两阶段逐球员注意力机制优化DST，最终通过加权事件融合集成四个模型变体，显著提升了挑战赛的宏观F1分数。

Details

Motivation: 解决足球视频中球员为中心的球动作（如传球、射门）的精准定位问题，旨在从广播视频中自动识别并结构化记录每位球员的关键动作事件。

Result: 在SoccerNet 2026挑战赛上，最终系统将宏观F1分数从基线48.6提升至58.94，验证集上通过空间优先的注意力排序（先跨球员注意力再时间注意力）使宏观F1提高了1.87%。

Insight: 创新点包括：在TAAD中引入时间变换器以增强跨帧上下文；在DST阶段设计两阶段逐球员注意力机制，并发现空间优先的注意力排序更有效；采用加权事件融合集成方法，通过一致性过滤抑制单模型误报，同时针对罕见的铲球类动作设计专用例外处理，以平衡精度和召回率。

Abstract: We present our submission to the SoccerNet 2026 Player-Centric Ball Action Spotting challenge, which uses a two-stage pipeline: a Track-Aware Action Detector (TAAD) produces per-player action logits from broadcast video, and a Denoising Sequence Transduction (DST) transformer converts game-state features and TAAD logits into structured event sequences. We improve the TAAD with a temporal transformer that adds cross-frame context, alongside several training fixes. For the DST stage, we introduce a two-stage per-player attention mechanism operating on game-state features, and show that a spatial-first attention ordering (cross-player attention before temporal attention) improves validation Macro-F1 by 1.87%. To exploit architectural diversity, we train four model variants and combine them with a Weighted Event Fusion ensemble that applies agreement filtering to suppress single-model false positives while preserving recall, plus a dedicated exception for the rare tackle class. Our final system improves the challenge Macro-F1 from a baseline of 48.6 to 58.94.

[46] Transition-Aware best-of-N sampling for Longitudinal Chest X-ray Reports cs.CVPDF

Halil Ibrahim Gulluk, Max Van Puyvelde, Wim Van Criekinge, Olivier Gevaert

TL;DR: 本文提出了一种无需训练的transition-aware best-of-N采样方案，用于改进预训练的胸部X光报告生成器在纵向临床场景下的表现。该方法将报告拆分为句子并嵌入为无序集合，通过设计集合间距离来编码先验与当前检查之间的变化，并利用余弦距离对候选报告进行评分。

Details

Motivation: 在纵向临床实践中，放射科医生解读胸部X光片时需参考患者先前的检查，并重点报告两次检查间的变化。现有预训练报告生成器缺乏对这种纵向过渡的显式建模，本文旨在解决这一问题。

Result: 在包含多次就诊的AP-PA队列数据集上评估，使用三种提示和三种视觉语言生成器进行推理。该方法在所有评估中均优于随机选择，在报告“印象”部分取得了最大的相对性能提升。

Insight: 创新点在于首次提出了无需训练、显式建模纵向过渡的best-of-N采样框架，并引入了四种方向性集合距离（均值偏移、新颖性残差、定向豪斯多夫锚点和成本加权最优传输）来量化报告间的变化。

Abstract: In longitudinal clinical practice, every chest X-ray is read in the context of the patients prior exam, and much of what the radiologist communicates is the change from one visit to the next. To the best of our knowledge, we present the first training-free best-of-N sampling scheme for pre-trained chest X-ray report generators that is explicitly aware of this longitudinal prior to current transition. We call it transition-aware best-of-N sampling, each report is split into sentences and embedded into an unordered set in Rd; each (prior, current) pair is reduced to a fixed-dim directional vector via a set-to-set distance designed to encode the change between the two sets; and candidates are scored by cosine distance from their candidate transition vector to a cached bank of ground-truth training transition vectors, aggregated as min or kNN. We instantiate the framework with four directional set distances (mean-shift, novelty residual, directed-Hausdorff anchor, and cost-weighted optimal transport) and evaluate on a multi-visit AP-PA cohort, running inference under three prompts on three vision-language generators. Transition-aware best-of-N outperforms random selection across the board, with the largest relative gains on the Impression section.

[47] RadarTwin: Scene-Specific mmWave Radar Simulation and Learning for Mobile Indoor Perception cs.CV | cs.LGPDF

Emily Bejerano, Federico Tondolo, Devang Gupta, Aaron Mano Cherian, Taeyoo Kim

TL;DR: 本文提出RadarTwin框架，用于在真实数据采集前生成特定场景的毫米波雷达训练数据。该方法利用目标空间的3D重建，通过视觉语言模型推断雷达相关表面材料，并基于物理的射线追踪器合成包含多径传播的原始FMCW雷达测量值。研究表明，仿真与真实雷达共享相同的对象判别性形状和材料特征，且建模环境多径对于匹配真实测量至关重要。

Details

Motivation: 毫米波雷达感知受限于数据稀缺性：在现有雷达数据集上训练的模型难以泛化到新对象、新环境和新传感轨迹。

Result: 仅使用仿真数据训练的表示模型，在没有任何真实雷达标签的情况下，识别真实对象的准确率是随机猜测的2.5倍；加入少量标注示例后，在12类识别任务上准确率达到95.3%。

Insight: 创新点在于提出了一个端到端的场景特定雷达仿真与学习框架，将视觉语言模型用于材料推断，并结合物理射线追踪来合成逼真的多径雷达数据，实现了在真实数据采集前为新空间训练雷达感知模型的能力。

Abstract: Millimeter-wave (mmWave) radar perception is limited by data scarcity: models trained on existing radar datasets fail to generalize to new objects, environments, and sensing trajectories. We present RadarTwin, a framework for generating deployment-specific radar training data before real data collection. Given a 3D reconstruction of a target space (phone LiDAR, robot-mounted sensing, or RGB-to-3D), RadarTwin uses a vision-language model to infer radar-relevant surface materials and a physics-based ray tracer to synthesize raw frequency-modulated continuous-wave (FMCW) radar measurements with multi-bounce propagation. To study what transfers from simulation to reality, we collect a paired real-simulated dataset spanning household objects, material classes, distances, rotations, translations, and mobile sensing trajectories. We show that simulated and real radar share the same object-discriminative shape and material features, and that modeling the environment’s multipath is essential to matching real measurements. A representation trained on simulation alone recognizes real objects at 2.5 times chance with no real radar labels, and a few labeled examples raise this to 95.3% on a 12-way recognition task. RadarTwin enables training radar perception for a new space before any real radar data is collected there.

[48] GPU-Accelerated Inverse Structural Anastylosis from Block Collapse Dynamics cs.CV | cs.LGPDF

L. A. Muñoz

TL;DR: 本文提出Jenga Inverse Predictor (JIP-2)，一个GPU加速的深度学习框架，将倒塌建筑的结构复原问题转化为逆预测任务。该框架通过结合刚体物理引擎、摩擦力学分析和双流ResNet-18网络，从倒塌的积木块图像中重建最可能的原始塔楼结构，并生成逐块反向重建的3D视频。

Details

Motivation: 传统建筑遗迹的物理复原（结构重建）高度依赖考古专家的主观判断和手动操作，过程费力且主观。受Jenga游戏组合复杂性的启发，本文旨在开发一个自动化、数据驱动的计算机辅助解决方案来应对这一挑战。

Result: 论文在模拟的450个场景中，应用了Ziglar分析力阈值（考虑三个摩擦系数水平）进行训练和评估。模型能够联合预测积木移除数量、每个位置的移除概率、质心不平衡度以及扭矩风险，但摘要未明确提及在特定基准测试上的定量结果或与SOTA的比较。

Insight: 创新点在于将结构复原问题形式化为一个基于物理的逆预测任务，并构建了一个集成了GPU加速刚体物理模拟（使用Numba JIT和CuPy CUDA）、摩擦力学先验知识和深度神经网络（双流ResNet-18）的完整端到端框架。这为文化遗产的计算机辅助修复提供了新的技术路径。

Abstract: The physical anastylosis of collapsed architectural monuments – the meticulous reassembly of fallen stone elements into their original structural configuration – represents one of the most intellectually demanding challenges in conservation science. Traditional approaches depend heavily on expert archaeologist judgement and manual block-by-block correspondence, a process that is both labour-intensive and inherently subjective. Inspired by the combinatorial complexity of this problem as manifested in the game of Jenga, we present Jenga Inverse Predictor , a GPU-accelerated deep learning framework that addresses structural anastylosis as an inverse prediction task. Given an image of a collapsed block assembly, JIP-2 reconstructs the most probable prior tower configuration by: (1) implementing a complete rigid-body physics engine with OBB/SAT collision detection and a Projected Gauss-Seidel (PGS) contact solver accelerated with Numba JIT and CuPy CUDA; (2) applying the analytical force thresholds of Ziglar (CMU, 2006) – F_app = 3mu_smg (Y-axis, torque-free) and F_app = 4mu_smg (X-axis, torque risk) – over three friction levels (mu_s in {0.25, 0.40, 0.60}) across 450 simulated episodes; (3) training a dual-stream ResNet-18 that injects a friction one-hot vector and jointly predicts block removal count, per-position removal probabilities, centre-of-mass imbalance, and Ziglar torque risk; and (4) generating a smooth 3-D video of the block-by-block reverse reconstruction. We discuss implications for computer-assisted anastylosis at the UNESCO Maya site of Uxmal, Yucatan, and provide a detailed technical description of the full pipeline, architecture, and loss formulation.

Shaoxuan Li, Xiangyu Dong, Xiaoguang Ma, Junfeng Chen, Haoran Zhao

TL;DR: 本文提出了一种名为CLOSER-VLN的闭环自验证检索增强推理方法，用于解决空中视觉语言导航任务中传统开环决策方法因中间动作错误累积而导致目标丢失的问题。该方法在具体动作执行前，以闭环方式顺序执行动作推理、可靠性验证、目标检索和动作校正。

Details

Motivation: 现有基于大模型的视觉语言导航方法大多采用开环的决策-执行方式，生成的候选动作在执行前很少被验证或纠正，这在空中导航等场景中容易导致微小的中间动作错误快速累积成大的轨迹偏差，从而丢失目标。

Result: 在CityNav基准测试的test-unseen划分上，CLOSER-VLN取得了32.01%的成功率和21.28%的路径长度加权成功率，验证了闭环推理的有效性。

Insight: 创新点在于将开环决策转变为包含推理、验证、检索、校正的闭环流程，并引入了多维动作验证器和验证触发的多模态检索器，仅在验证失败时从记忆库中检索针对性示例进行校正，从而在不训练特定导航策略的情况下提升了导航的鲁棒性和准确性。

Abstract: Vision-language navigation (VLN) has recently advanced with large language and multimodal models, enabling agents to follow natural-language instructions in unseen environments without training a task-specific navigation policy. However, most existing VLN methods relying on large models still adopt an open-loop decision-execution approach, where candidate actions are generated from instructions and observations but are rarely verified or corrected before execution. This causes critical issues in aerial VLN, where minor errors in intermediate actions may quickly accumulate into large trajectory deviations and lead to target loss. To address this issue, we propose Closed-loop Self-verified Retrieval-augmented Reasoning (CLOSER), a training-policy-free method that sequentially performs action reasoning, reliability verification, targeted retrieval, and action correction in a closed-loop manner before executing concrete actions. We instantiate the CLOSER in aerial VLN tasks and develop a CLOSER-VLN framework, which is composed of three components: a hierarchical reasoner for generating candidate actions based on available information, a multidimensional action verifier for assessing the reliability of actions generated by the reasoner, and a verification-triggered multimodal retriever for retrieving targeted exemplars from a memory bank only when verification fails. We conduct experimental evaluations on the CityNav benchmark, where CLOSER-VLN achieves 32.01% SR and 21.28% SPL on the test-unseen split, confirming the effectiveness of closed-loop reasoning.

[50] Meta-learning as a principle for human-like visual representations cs.CV | cs.LG | q-bio.NCPDF

Can Demircan, Marcel Binz, Alireza Modirshanechi, Eric Schulz

TL;DR: 该论文提出人类视觉表征的灵活性源于元学习（学习如何学习）的压力，这种压力使表征能够从少量观察中快速适应新任务。作者通过训练一个序列模型在数千个语义丰富的任务上进行元学习，发现其表征比预训练编码器更能预测人类相似性判断、语义规则学习和高级视觉皮层活动。

Details

Motivation: 预训练神经网络虽然能成功建模人类视觉表征，但仍存在较大差距，作者认为这是因为这些网络优化的是单一固定目标，而人类表征需要支持开放式的任务。

Result: 元学习表征在预测人类相似性判断、语义规则学习和高级视觉皮层活动方面优于预训练编码器，行为增益依赖于解耦的高级任务分布，而大脑对齐主要由学习到学习的压力驱动。

Insight: 创新点在于将元学习作为人类视觉表征灵活性的原理，并通过无监督的跨任务训练验证了其有效性，表明人类表征的灵活性反映了动态学习新语义关系的功能需求。

Abstract: The structure of human visual representations underpins our capacity for adaptive behaviour. While pretrained neural networks model human visual representations with unprecedented success, a large discrepancy remains. We propose one reason: these networks optimise a single fixed objective, whereas human representations must support open-ended tasks. We hypothesise this flexibility arises from meta-learning (learning to learn), a pressure shaping representations to acquire new tasks from few observations. To test this, we train a sequence model, without any supervision from human data, across thousands of semantically rich tasks mapping images to high-level concepts. Compared to their pretrained base encoders, meta-learned representations better predict human similarity judgements, semantic rule learning, and high-level visual cortex. Behavioural gains depend on disentangled, high-level task distributions, while brain alignment is driven primarily by the learning-to-learn pressure. Our results suggest the flexibility of human visual representations reflects the functional demand to learn new semantic relationships on the fly.

[51] Vision-driven Preference Synthesis for Mitigating Hallucinations in VLMs cs.CV | cs.LGPDF

Yunhun Nam, Jongheon Jeong

TL;DR: 本文提出了ViPSy（Vision-driven Preference Synthesis）框架，用于构建既与策略分布对齐又基于视觉信息的偏好数据，以缓解视觉语言模型（VLM）中的幻觉问题。该框架通过从语义对齐的图像变体中提取视觉线索，并基于此引导策略生成候选响应，从而在减少幻觉的同时保持与策略响应分布的接近。实验表明，使用ViPSy构建的偏好数据对齐的VLM在幻觉缓解方面达到了新的最先进水平。

Details

Motivation: 视觉语言模型（VLM）在视觉理解方面表现出色，但仍存在幻觉问题，即生成与图像内容无关的内容。偏好对齐是提高视觉忠实度的有前景的方法，但其成功很大程度上取决于偏好对的构建方式。现有方法存在两个关键限制：基于干预的方法常导致策略分布显著偏离，而基于采样的方法在构建过程中未能充分利用视觉信息。

Result: 使用ViPSy构建的偏好数据对齐的VLM在幻觉缓解方面达到了新的最先进水平（SOTA）。与之前的最先进方法相比，在AMBER和Object HalBench基准上，幻觉率分别降低了35.7%和24.5%。该模型还在通用视觉基础基准（如MMStar、MMVP和CV-Bench）上有所提升，同时在语义分割和ImageNet线性探测任务中取得了增益。

Insight: 论文的创新点在于提出了一个两阶段的ViPSy框架，通过从语义对齐的图像变体中提取对象级视觉线索来指导偏好对的构建，确保候选响应既基于视觉信息又接近策略分布。这解决了现有方法在视觉信息利用不足和策略分布偏离方面的局限性，从而有效提升了模型的视觉忠实度和泛化能力。

Abstract: Vision-Language Models (VLMs) have shown strong performance in visual understanding, yet they still suffer from hallucinations, generating content that is not grounded in the image. Preference alignment is a promising approach to improve visual faithfulness, but its success depends heavily on how preference pairs are constructed. Existing methods exhibit two key limitations; (a) intervention-based methods often introduce significant deviation from the policy distribution, and (b) sampling-based methods often underuse visual information during the construction. In this paper, we propose ViPSy (Vision-driven Preference Synthesis), a framework for constructing preference data that are both policy-aligned and visually grounded. Our framework consists of two stages; in the first stage, ViPSy derives a visual cue from recurring object-level content across semantically aligned image variants, so preference construction can rely on visual information rather than language priors. In the second stage, ViPSy conditions the policy’s own rollouts on this cue, allowing candidates to be guided by visually grounded content while staying close to the policy’s response distribution. The resulting candidates remain close to the policy’s response distribution while better leveraging visual information from the image. Experiments show that the resulting VLM, preference-aligned with ViPSy-constructed preference pairs, achieves a new state-of-the-art in hallucination mitigation. Compared with the previous state-of-the-art method, it reduces hallucination rates on AMBER and Object HalBench by 35.7% and 24.5%, respectively. The resulting model further improves on general visual grounding benchmarks, e.g., MMStar, MMVP, and CV-Bench, while also yielding gains in semantic segmentation and ImageNet linear probing, underscoring the effectiveness of our framework in enhancing the model’s visual capabilities.

[52] AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors cs.CV | cs.AIPDF

Maher Boughdiri, Mounira Msahli, Albert Bifet

TL;DR: 本文提出了AEGIS框架，一个结合了语义感知和证据学习的鲁棒对抗样本检测方法，用于视觉传感器网络中的图像分类任务。该框架通过SemantiGAN模块进行初步语义过滤，再通过随机增强和手工设计的不稳定性度量提取特征，最后利用证据深度学习分类器进行检测和不确定性估计。

Details

Motivation: 深度神经网络在视觉识别任务中表现出色，但易受对抗性攻击和微小扰动的影响，导致错误预测。本文旨在设计一个鲁棒的对抗样本检测框架，以增强视觉传感器网络中的图像分类系统的安全性。

Result: 在Tiny ImageNet数据集上，针对六类攻击（干净样本、FGSM、PGD、基于补丁的攻击、功能攻击和几何攻击）进行评估，AEGIS框架取得了92.1%的AUROC、90.2%的AUPRC和90.7%的准确率，在检测性能、鲁棒性、可解释性和不确定性校准方面均优于传统的基于softmax的检测器。

Insight: 创新点在于将语义生成对抗网络（SemantiGAN）作为多类语义判别器进行初步过滤，并结合随机增强与手工设计的不稳定性度量（如FlipScore、预测不一致性等）提取特征，最后采用证据深度学习（EDL）进行不确定性建模，实现了端到端的鲁棒对抗检测与可解释的不确定性估计。

Abstract: Deep neural networks (DNNs) have shown outstanding performance in visual recognition tasks within vision sensor networks; however, they are still vulnerable to adversarial manipulations and imperceptible perturbations that can lead to erroneous predictions. To address that, this paper presents AEGIS, a semantic aware and uncertainty guided adversarial detection framework designed for robust image classification in vision sensors pipelines. At its core, a SemantiGAN module functions as a multi class semantic discriminator, identifying and filtering visually inconsistent adversarial inputs before they propagate further in the pipeline. For inputs that pass this stage, a stochastic augmentation process generates test time variations, from which handcrafted instability metrics FlipScore, Prediction Inconsistency, Layerwise Cosine Similarity (early and mid layers), and Entropy are computed. These features are aggregated into a compact five dimensional vector and processed by an Evidential Deep Learning (EDL) classifier, which models output evidence using a Dirichlet distribution to yield both class predictions and calibrated uncertainty estimates. Evaluations on the Tiny ImageNet dataset across six categories clean, FGSM, PGD, patch based, functional, and geometric attacks demonstrate the effectiveness of AEGIS. The proposed framework achieves an AUROC of 92.1%, an AUPRC of 90.2%, and an accuracy of 90.7%, outperforming conventional softmax-based detectors in terms of detection performance, robustness, interpretability, and uncertainty calibration.

[53] CLEAR-MoE: Shared-Basis Expert Extraction from Frozen Vision Transformers via Calibration-Driven Layer Selection cs.CV | cs.DCPDF

Md Irtiza Hossain, Humaira Ayesha, Junaid Ahmed Sifat

TL;DR: CLEAR-MoE是一个四阶段后训练流程，旨在将冻结的预训练视觉Transformer（ViT）转换为稀疏的专家混合（MoE）模型，而无需更新主干网络权重。该方法通过评估前馈网络层的稀疏性、可聚类性和输出敏感性来选择层，将其分解为共享的低秩SVD基和基于聚类的残差专家，并训练轻量级路由器进行令牌分发。实验表明，该方法在多个ViT架构上能保持与原始密集模型相当的精度（差异≤0.10个百分点），同时揭示了共享SVD基是保持精度的关键因素。

Details

Motivation: 解决如何在不更新预训练权重的情况下，将密集的ViT模型高效地转换为稀疏的MoE架构，以探索模型压缩和推理加速的潜力。

Result: 在Imagenette数据集上使用DeiT-Small，CLEAR-MoE保持了密集模型99.9%的准确率（86.70% vs 86.73%）。该方法在五个ViT骨干网络（DeiT-Tiny到ViT-Base）上均表现稳定，与密集模型的准确率差异不超过0.10个百分点，但当前实现因路由开销导致FFN部分比密集版本慢1.3-1.7倍。

Insight: 创新点在于提出了一个无需微调权重的后训练MoE转换流程，并实证发现共享SVD基是维持精度的核心；客观分析表明，路由机制对性能影响极小，这为未来优化（如融合分发内核）提供了明确方向，以解决内存瓶颈问题。

Abstract: We present CLEAR-MoE, a four-phase post-training pipeline that converts a frozen pretrained Vision Transformer (ViT) into a sparse Mixture-of-Experts (MoE) model without updating backbone weights. The pipeline (i) scores feed-forward network (FFN) layers by sparsity, clusterability, and output sensitivity; (ii) decomposes selected layers into a shared low-rank SVD basis and per-cluster residual experts using k-means clustering; (iii) trains lightweight routers supervised by cluster labels; and (iv) dispatches tokens through pluggable CUDA backends. On Imagenette with DeiT-Small, CLEAR-MoE retains 99.9% of the dense model’s accuracy (86.70 +/- 0.02% versus 86.73%). Extensive ablation studies reveal a consistent empirical finding: the shared SVD basis is the primary factor responsible for preserving accuracy. Random routing, learned routing, and three different router architectures produce nearly identical performance, with accuracy varying by at most 0.06 percentage points (86.62%-86.68%). Accuracy also remains stable across different SVD ranks, expert counts (2-8), calibration set sizes (50-500), and random seeds. This behavior generalizes across five ViT backbones (DeiT-Tiny, DeiT-Small, DeiT-Base, ViT-Small, and ViT-Base), covering models from 5.7M to 86.6M parameters, with accuracy differences <= 0.10 percentage points from their dense counterparts. On a GTX 960 GPU, routing and scatter-gather overhead make the CLEAR-MoE FFN 1.3-1.7x slower than the dense implementation. A dispatch microbenchmark further shows that routing is an order of magnitude more memory-bound than expert matrix multiplications, identifying fused dispatch kernels as a promising direction for future optimization.

[54] Detecting Clinical Hallucinations in LVLMs via Counterfactual Visual Grounding Uncertainty cs.CV | cs.CLPDF

Xiao Song, Haonan Qin, Zhaoxu Zhang, Jiong Zhang, Yuqi Fang

TL;DR: 本文提出了一种可追溯视觉证据的幻觉检测框架，用于检测大型视觉语言模型（LVLMs）在临床图像理解中产生的幻觉。该方法通过提取模型响应中的视觉可验证实体，并使用医学领域适配的Qwen-VL定位器在图像上定位每个实体，结合反事实实体扰动来估计视觉证据的不确定性，从而进行二元幻觉决策。

Details

Motivation: 大型视觉语言模型在临床图像理解中容易产生幻觉，即生成图像不支持的文本发现或属性，这限制了其在医疗领域的可靠应用。现有方法通常需要修改模型或访问其内部状态，因此需要一种无需侵入模型、可追溯视觉证据的检测框架。

Result: 在多种医学成像模态和LVLM骨干网络上的实验表明，该方法在幻觉检测性能上持续优于近期基线，同时提供了可解释的定位证据和强大的跨模型可迁移性。

Insight: 创新点在于提出了一个无需修改或访问LVLM内部状态的可追溯视觉证据检测框架，并通过引入反事实实体扰动来估计视觉证据不确定性，增强了检测的鲁棒性和可解释性。从客观角度看，该方法将反事实推理与视觉定位不确定性结合，为模型幻觉检测提供了一种新颖且通用的评估思路。

Abstract: Large vision-language models (LVLMs) are increasingly used for clinical image understanding, yet they remain vulnerable to \emph{hallucinations}–producing textual findings or attributes not supported by the image. We present a vision-traceable hallucination detection framework that audits arbitrary LVLM responses via visual evidence grounding, requiring neither modification nor internal access to the hidden states of LVLMs. Given an LVLM response, we extract visually verifiable entities and use a medical-domain-adapted Qwen-VL grounding verifier to localize each entity on the input image. To enhance the robustness of our detection method, we introduce a counterfactual entity perturbation method and estimate visual evidence uncertainty by contrasting factual and counterfactual grounding results. Specifically, we compute an entity-level uncertainty score from the positive confidence, counterfactual confidence, and their grounding overlap for binary hallucination decision-making. Experiments on multiple medical imaging modalities and LVLM backbones demonstrate that our method consistently improves hallucination detection performance over recent baselines, while providing interpretable localization evidence and strong cross-model transferability. Code and dataset are available at https://github.com/Agentic-CliniAI/CounterVHD.

[55] Digitizing Coaching Intelligence: An Agentic Framework for Holistic Athlete Profiling using VLM and RAG cs.CV | cs.AI | cs.MAPDF

Deep Ghosal, Ishani Sen, Wazib Ansar, Amlan Chakrabarti

TL;DR: 本文提出了一种基于LLM的混合智能体框架，用于自动化、全面的运动员评估。该框架结合了计算机视觉的几何精度和视觉语言模型的语义推理能力，通过创新的时间分块策略和自主校正机制，实现了对运动员生理指标的定性定量分析，并利用双持久化RAG管道支持自然语言查询。

Details

Motivation: 传统运动员评估方法依赖主观的人工观察或仅能进行定量重复计数的基本计算机视觉系统，缺乏评估姿势退化、脊柱活动性和疲劳等定性生理指标的’教练智能’，且难以规模化。

Result: 实验结果表明，该多智能体方法显著缩小了原始生物特征追踪与可操作的教练洞察之间的差距，为国家级人才识别提供了可扩展的客观解决方案。

Insight: 创新点包括：1) 结合CV几何精度与VLM语义推理的双管道架构；2) 3×3’智能网格’时间分块策略，降低88%计算开销；3) 自主的’LLM-as-a-Judge’自校正循环；4) 支持自然语言复杂语义查询的双持久化RAG管道。

Abstract: Athlete assessment is a critical process for tracking physical progress and identifying elite talent. However, during mass recruitment drives, traditional methods rely on manual observation, which is inherently subjective and unscalable, or basic computer vision (CV) systems limited to quantitative repetition counting. These standard approaches lack the “coaching intelligence” required to evaluate qualitative physiological markers such as form degradation, spinal articulation, and fatigue. This paper presents a novel, LLM-based hybrid agentic framework for automated, holistic athlete profiling that strictly aligns with the Sports Authority of India (SAI) assessment protocols. Orchestrated via LangGraph, our dual-pipeline architecture synthesizes the geometric precision of CV (MediaPipe) for kinematic tracking with the semantic reasoning of Vision-Language Models (Llama-4-scout). To overcome the latency and token constraints associated with multimodal video processing, we introduce a 3 X 3 “Smart Grid” temporal chunking strategy, reducing computational overhead by over 88% while preserving critical temporal continuity. To ensure data integrity and mitigate hallucination, the framework pioneers an autonomous “LLM-as-a-Judge” self-correction loop that cross-references quantitative and qualitative metrics before persistence. Finally, we implement a dual-persistence Retrieval-Augmented Generation (RAG) pipeline utilizing a vector search engine (ChromaDB). This enables coaches to bypass rigid SQL databases and perform complex semantic queries (e.g., “Identify athletes with high endurance but poor core rigidity”) using natural language. Experimental results demonstrate that this multi-agent approach significantly bridges the gap between raw biometric tracking and actionable coaching insights, offering a scalable, objective solution for national talent identification.

[56] DataComp-VLM: Improved Open Datasets for Vision-Language Models cs.CV | cs.CL | cs.LGPDF

Matteo Farina, Vishaal Udandarao, Thao Nguyen, Selim Kuzucu, Maximilian Böther

TL;DR: 本文介绍了DataComp for VLMs (DCVLM)，一个用于评估和改进视觉语言模型训练数据策展策略的基准测试。它整合了160个数据集，涵盖四种数据类型，总计6T多模态token，并允许在1B-8B模型和6.25B-200B token预算下测试数据过滤、混合、格式化和采样等策略。通过大量实验发现，数据混合（而非过滤）是构建高质量训练集的关键，特别是富含指令的混合数据在更大规模下表现更优。基于此构建的DCVLM-Baseline数据集，在200B token训练下，能使8B VLM在33个核心任务上达到63.6%的准确率，比当前最优的开源VLM训练数据集FineVision提升了5.4个百分点。

Details

Motivation: 当前社区缺乏系统性的基准来评估视觉语言模型训练数据的策展策略，这阻碍了构建高性能VLMs。本文旨在通过引入一个受控的数据中心实验基准DCVLM来解决这一问题。

Result: 在DCVLM基准上进行的实验表明，使用其构建的DCVLM-Baseline数据集，能够训练一个8B参数的VLM，在33个核心下游任务上达到63.6%的准确率（使用200B训练token）。这比当前最先进的开源VLM训练数据集FineVision的性能提升了5.4个百分点。

Insight: 论文的主要创新点在于提出了一个系统化的、受控的基准DCVLM，用于科学评估VLM数据策展策略。一个关键的实证发现是，数据混合（尤其是富含指令的数据）比数据过滤更能有效提升模型性能，且这种优势在更大规模下更为明显，这为未来VLM训练数据构建提供了重要指导。

Abstract: Building performant Vision-Language Models (VLMs) requires carefully curating large-scale training datasets, yet the community lacks systematic benchmarks for evaluating such curation strategies. We introduce DataComp for VLMs (DCVLM), a benchmark for controlled data-centric experiments to improve VLM training. As part of DCVLM, we collect 160 datasets spanning four data types – image-caption pairs, multimodal interleaved documents, text-only, and instruction-tuning data – into a corpus of 6T multimodal tokens. DCVLM allows participants to test curation strategies (filtering, mixing, formatting, sampling) across 1B-8B models and 6.25B-200B token budgets. Models are then evaluated on a carefully selected suite of up to 52 downstream benchmarks across 9 domains. We conduct extensive experiments on DCVLM and find that data mixing, not filtering, is key to a high-quality training dataset: instruction-heavy mixtures scale better than caption-heavy ones, with gains widening at larger scales. The resulting dataset, DCVLM-Baseline, enables training an 8B VLM to 63.6% accuracy on our 33-task core suite with 200B training tokens. Compared to FineVision, the state-of-the-art open VLM training dataset, this represents an improvement of +5.4pp. DCVLM and all accompanying artifacts will be made publicly available at https://www.datacomp.ai/dcvlm/.

[57] SatSplat: Geometrically-Accurate Gaussian Splatting for Satellite Imagery cs.CVPDF

Shuang Song, Jiyong Kim, Rongjun Qin

TL;DR: 本文提出了SatSplat框架，首次将2D高斯泼溅（2DGS）技术应用于卫星摄影测量，通过在线相机调整、仿射相机模型近似和几何阴影映射等方法，旨在解决多时相、高海拔卫星图像在3D重建中因光照变化和几何不准确导致的质量下降问题。

Details

Motivation: 高分辨率卫星影像需要兼顾速度和几何精度的3D重建方法。现有的3D高斯泼溅（3DGS）方法在卫星影像上效率高，但在多时相、高海拔采集（小交会角）的不同光照条件下，重建质量会下降，限制了其在遥感和视觉任务中的应用。

Result: 在DFC2019和IARPA2016基准测试中，SatSplat实现了强大的几何精度，显著优于先前的3DGS基线方法。在处理后的DFC2019基准上，SatSplat将平均绝对误差降低了11.93%，并将峰值视频内存减少了31%，达到了最先进水平（SOTA）。

Insight: 创新点包括：将2D高斯泼溅（而非3D）首次适配到卫星摄影测量；采用仿射相机模型近似和在线相机调整（delta参数化）来优化几何精度；集成几何阴影映射和逐相机颜色校正以处理时变阴影和光照变化，从而在保持计算效率的同时实现大规模数字表面建模。

Abstract: High-resolution satellite imagery demands 3D reconstruction methods that deliver both speed and geometric accuracy. Recent adaptations of 3D Gaussian Splatting (3DGS) to satellite imagery demonstrate strong efficiency, but reconstruction quality often degrades under diverse illumination across multi-date, high-altitude acquisitions (with small intersection angles), limiting applicability to remote sensing and vision tasks. We present SatSplat, the first framework to adapt 2D Gaussian Splatting (2DGS) to satellite photogrammetry, with online camera adjustment. We approximate satellite cameras with an affine model and learn a minimal delta parameterization for in-splat camera refinement from dense observations. The formulation is implemented with a 2DGS scene representation. To handle time-varying shadows and illumination changes, we integrate geometric shadow mapping and per-camera color correction during training. Across the evaluated DFC2019 and IARPA2016 benchmark sites, SatSplat achieves strong geometric accuracy while significantly outperforming prior 3DGS-based baselines. On our processed DFC2019 benchmark, SatSplat reduces mean absolute error by 11.93% and peak video memory by 31% relative to the previous state of the art. Our approach enables large-scale digital surface modeling with practical computational efficiency. The project page is available at https://gdaosu.github.io/satsplat/.

[58] IMU-HOI: A Symbiotic Framework for Coherent Human-Object Interaction and Motion Capture via Contact-Conscious Inertial Fusion cs.CVPDF

Lizhou Lin, Songpengcheng Xia, Zengyuan Lai, Lan Sun, Jiarui Yang

TL;DR: 本文提出了IMU-HOI框架，通过身体和物体上的稀疏IMU传感器，联合估计全身人体姿态和物体的6自由度轨迹。该方法首先从IMU数据流中推断出手与物体接触的概率，并以此作为高级信号来协调运动学和惯性推理，通过一个三阶段融合流程，最终生成抗漂移、连贯的人体与物体运动轨迹。

Details

Motivation: 解决传统视觉方法在人体与物体交互（HOI）运动捕捉中因遮挡和捕获范围受限而面临的挑战，同时弥补现有基于IMU的运动捕捉方法忽略物体接触和动力学的不足。

Result: 在具有挑战性的人体-物体交互场景实验中，该方法在准确性上相比先前的惯性运动捕捉方法有显著提升。

Insight: 创新点在于明确建模人体-物体交互，并利用从IMU直接推断的接触概率作为高层信号来引导融合流程；其框架设计具有可扩展性，能以最小改动集成到现有的稀疏IMU运动捕捉骨干网络中，从而将纯惯性运动捕捉的范围从孤立人体扩展到完整的人-物交互与联合运动估计。

Abstract: Capturing full-body human motion with object interactions is crucial for AR/VR and robotics applications, yet it remains challenging for conventional vision-based methods due to occlusions and constrained capture volumes. Inertial measurement units (IMUs) offer a compelling alternative without line-of-sight requirements, but existing IMU-based motion capture assumes an isolated human and ignores object contacts and dynamics. To bridge this gap, we present IMU-HOI, a novel framework that jointly recovers full-body human pose and 6-DoF object trajectory from sparse IMUs on the body and object, explicitly modeling human-object interaction. Our approach first infers probabilistic hand-object contacts directly from IMU streams and uses them as a high-level signal to route between kinematic and inertial reasoning. These contact cues drive a three-stage fusion pipeline that refines human pose and root translation, and fuses hand-based forward kinematics with object-IMU integration for object motion, yielding coherent, drift-resilient trajectories for both human and object. Experiments on challenging human-object interaction scenarios demonstrate substantial accuracy gains over prior inertial motion capture methods. Moreover, IMU-HOI can be plugged into existing sparse-IMU mocap backbones with minimal changes, effectively extending the scope of purely inertial motion capture from isolated humans to full human-object interaction and joint motion estimation.

[59] Animation2Code: Evaluating Temporal Visual Reasoning in Video-to-Code Generation cs.CV | cs.AI | cs.CLPDF

Anya Ji, Abhijith Varma Mudunuri, David M. Chan, Alane Suhr

TL;DR: 本文提出了Animation2Code基准，用于评估视觉语言模型在视频到代码生成任务中的时序视觉推理能力，特别是从网页动画视频重建可执行HTML/CSS/JavaScript代码。该基准包含1,069个具有多样化视觉外观和运动模式的动画视频及其对应代码实现，并引入了外观相似性和时序相似性两个人工对齐的评估指标。实验表明，当前最先进的视觉语言模型在重建时难以保持时序一致性，即使在外观相似性很高的情况下也是如此。

Details

Motivation: 尽管当前视觉语言模型在静态视觉到代码任务（如生成网页、图表或SVG代码）上取得了显著进步，但其在存在运动的情况下能否恢复时序动态仍不清楚。因此，需要建立一个基准来专门评估模型在时序视觉推理方面的能力。

Result: 在该基准上对最先进的视觉语言模型进行评测，结果显示，即使在外观相似性得分很高的情况下，现有模型（包括经过微调和迭代优化的设置）在重建动画时仍难以保持时序一致性，表明其在时序对齐方面存在显著不足。

Insight: 论文的创新点在于构建了首个专注于评估时序视觉推理的视频到代码生成基准，并提出了可分离视觉保真度和时序对齐的评估指标。从客观角度看，这项工作揭示了当前视觉语言模型在理解和生成动态视觉内容时序逻辑方面的核心弱点，为未来模型在动态场景理解方向的发展提供了重要的评估工具和方向指引。

Abstract: While recent vision-language models (VLMs) have achieved significant improvements on static visual-to-code tasks such as generating code for webpages, charts, or SVGs, it remains unclear whether they can recover temporal dynamics when motion is present. To this end, we introduce Animation2Code, a benchmark for evaluating temporal visual reasoning via reconstructing executable web animation code from videos. Animation2Code consists of 1,069 web animation videos with diverse visual appearances and motion patterns, paired with corresponding HTML/CSS/JavaScript implementations. We propose two human-aligned metrics, appearance similarity and temporal similarity, which allow us to disentangle visual fidelity from temporal alignment when comparing rendered animations against ground-truth samples. Benchmarking state-of-the-art VLMs on this dataset shows that current VLMs struggle to maintain temporal consistency in reconstruction, even when achieving high appearance similarity, including under finetuning and iterative refinement settings. Code and data are available at https://anya-ji.github.io/animation2code-website .

[60] Obliviate: Erasing Concepts from Autoregressive Image Generation Models cs.CVPDF

Hossein Shakibania, Jonas Henry Grebe, Tobias Braun, Ege Aktemur, Saleh Aslani

TL;DR: 本文提出了Obliviate，一种基于引导的概念擦除方法，专门用于自回归图像生成模型。该方法通过KL散度监督、轨迹级更新和对齐的视觉前缀三个关键设计，有效移除模型生成有害或不安全图像的能力。实验在Liquid、Emu3-Gen和Janus-Pro等先进模型上进行，成功擦除了色情、暴力及品牌图像等内容。

Details

Motivation: 随着生成式AI的广泛应用，其可能被滥用于生成不安全或令人不安的图像引发了担忧。尽管已有针对多模态生成模型的概念擦除方法，但自回归图像生成模型中的概念擦除研究仍很缺乏，而这类模型在统一多模态架构趋势中日益重要。

Result: 在defensive RAB基准测试上，Obliviate将裸露内容的生成率从91.58%大幅降低至3.15%，同时保持了模型的整体实用性。该方法在Liquid、Emu3-Gen和Janus-Pro三个最先进的自回归文本到图像模型上均优于现有替代方案。

Insight: 论文宣称的创新点在于首次系统性地探索了自回归图像生成模型的概念擦除问题，并提出了结合KL监督、轨迹级更新和对齐视觉前缀的引导式方法。从客观角度看，其将概念擦除技术适配到序列生成范式，并为未来统一多模态模型的安全部署提供了可借鉴的框架。

Abstract: The widespread adoption of generative AI models has intensified concerns about misuse, including the creation of unsafe or disturbing imagery. To mitigate such issues, several concept erasure approaches have been proposed to remove harmful content from multimodal generative models. Yet concept erasure for autoregressive image generation remains largely unexplored, despite the growing relevance of these models in recent trends toward unified multimodal architectures. In this work, we fill this gap by introducing Obliviate, a guidance-based concept erasure method for autoregressive image generation. Our method builds on three key design choices: KL-based supervision over visual token distributions, trajectory-level updates over full autoregressive rollouts, and aligned visual prefixes for stable target construction. We evaluate Obliviate on three state-of-the-art autoregressive text-to-image models, Liquid, Emu3-Gen, and Janus-Pro, covering the erasure of explicit content, graphic violence, and branded imagery. Obliviate consistently outperforms current alternatives, reducing nudity on the defensive RAB benchmark from 91.58 to 3.15 while preserving overall model utility.

[61] FedLAS: Feature-Modulated Bidirectional Label Smoothing for Neural Network Calibration cs.CV | cs.AI | cs.LGPDF

Thiru Thillai Nadarasar Bahavan, Sachith Seneviratne, Saman Halgamuge

TL;DR: 本文提出FedLAS，一种用于神经网络校准的特征调制双向标签平滑算法。该方法通过特征范数置信度指示器动态控制平滑程度，并利用双向校准门控模块同时检测过自信和欠自信样本，从而提升模型校准性能。

Details

Motivation: 现有标签平滑方法（包括基于边缘的LS）依赖预定义的均匀平滑规则，仅能处理过自信问题，而实际训练中样本具有不同难度/模糊性，且可能同时存在过自信和欠自信状态，需要一种能根据样本特性动态调整平滑程度的机制。

Result: 在标准和高分辨率细粒度视觉基准测试上的大量实验表明，FedLAS相比现代基线方法持续改善了校准性能，降低了预期校准误差和自适应校准误差，同时保持了Top-1准确率。

Insight: 创新点在于提出了基于特征范数的置信度指示器和双向校准门控模块，实现了对每个样本在训练过程中特定置信状态的动态识别与平滑程度调整，这是一种可插拔的、能同时处理过自信和欠自信问题的校准增强机制。

Abstract: Deep Neural Network (DNN) classifiers suffer from poor calibration when their softmax outputs (predictive confidence) deviate from the empirical likelihoods. This manifests itself as either overconfident incorrect predictions or under-confident correct predictions. Label smoothing (LS) enhances model calibration by introducing entropy regularization during training through redistributing probability mass from the ground-truth label to the remaining classes. LS, including Margin-based LS (MbLS), have restrictive assumptions: they rely on predefined, uniform smoothing rules and only tackle overconfidence. In reality, samples exhibit diverse characteristics, such as difficulty/ambiguity, that interact with the evolving nature of the model being trained. In training, samples may have various degrees of under- or overconfidence. To overcome this, a mechanism that identifies the specific confidence state of each sample and determines the appropriate degree of smoothing in each training step is needed, tailoring the adjustment to the individual sample. We propose FedLAS: Feature-Modulated Bidirectional Label Smoothing, a plug-and-play algorithm for label smoothing-based losses. In FedLAS, we introduce a Feature Norm-based Confidence Indicator (NCI) to control smoothing and a Bidirectional Calibration Gating (BCG) module to detect both over and under-confidence. Our algorithm can be integrated with LS and MbLS based losses when applied to standard DNNs, enhancing performance. Extensive experiments on standard and fine-grained high-resolution vision benchmarks show that FedLAS consistently improves calibration compared to modern baselines, reducing Expected Calibration Error (ECE) and Adaptive ECE while maintaining Top-1 accuracy. Code: github.com/nadarasarbahavan/FEDLAS

[62] BackTranslation2.0 – A Linguistically Motivated Metric to Assess Sign Language Production cs.CVPDF

Oliver Cory, Maksym Ivashechkin, Karahan Sahin, Oline Ranum, Jianhe Low

TL;DR: 本文提出了BackTranslation2.0，一种用于手语生成评估的、基于语言学原理的度量方法。该方法采用一个确定性流程来协调多个专用工具，从语法正确性、音系准确性、动作流畅性和生成保真度四个维度进行评估，并通过LLM驱动的交叉参考模块进行一致性检查。在一个人工标注的英国手语数据集上验证，该方法在所有维度上都与人类判断高度相关。

Details

Motivation: 现有手语生成评估指标过于简单，且与人类判断一致性差，无法满足手语作为聋人主要沟通方式的高质量评估需求。

Result: 在一个由语言学家和聋人专家共同制定协议、人工标注的英国手语数据集上，与六个基线指标相比，BackTranslation2.0在所有四个质量维度上都显示出与人类判断的强相关性。

Insight: 创新点在于超越了简单的回译，构建了一个基于代理的、工具化的多维度评估框架，并引入LLM进行跨工具输出的交叉参考和一致性推理，从而实现了更全面、可解释且符合语言学原则的评估。

Abstract: Sign Languages (SLs) are the primary means of communication for millions of deaf individuals, yet existing evaluation metrics for generated SL remain simplistic and poorly aligned with human judgements. We introduce BackTranslation2.0, a linguistically grounded evaluation metric for text-to-sign translation that moves beyond naïve backtranslation. Our approach adopts an agentic framework in which a deterministic pipeline orchestrates a suite of specialised tools to assess four scoring dimensions - grammatical correctness, phonological accuracy, motion fluency, and generation fidelity - aligned with human rater assessments. Tool outputs are not treated independently: a set of large language model (LLM)-based cross-referential comparison modules evaluates consistency across tools and checks outputs against linguistic expectations, enabling structured reasoning over grammatical, phonological, and motion-level evidence. Final dimension scores are computed through deterministic weighted formulas over validated tool outputs. To validate BackTranslation2.0, we introduce and evaluate on a British Sign Language (BSL) dataset rated in a human rater study across the same quality dimensions, following a protocol developed in collaboration between linguists and deaf experts, benchmarking against six baseline metrics. Our method demonstrates strong correlation with human judgements across all dimensions, providing a more comprehensive, interpretable, and linguistically principled evaluation framework for sign language production systems.

[63] Predicting Metastatic Risk from Primary Tissue Architecture via Distance-Aware Spatial Modeling cs.CV | cs.AIPDF

Sandesh Pokhrel, Hamid Manoochehri, Bodong Zhang, Beatrice S Knudsen, Tolga Tasdizen

TL;DR: 本文提出了一种名为DTMf-MIL的新方法，用于从原发性肿瘤组织学中预测远处转移风险。该方法通过引入带符号距离函数来显式建模肿瘤微环境中不同细胞类型（如肿瘤细胞、成纤维细胞和淋巴细胞）之间的空间几何关系，从而克服了传统多示例学习忽略空间布局的局限性。

Details

Motivation: 传统多示例学习（MIL）方法在预测转移风险时将组织图像块视为无序集合，丢弃了定义转移潜力的关键空间布局信息。作者认为转移风险本质上由肿瘤微环境在界面处的几何排列决定，因此需要显式捕获细胞间的空间关系。

Result: DTMf-MIL在原发性肿瘤组织转移预测任务上显著优于忽略空间布局的现有SOTA方法。在公开基准测试上的进一步验证表明，空间感知能力持续提升了多种临床任务的诊断准确性。

Insight: 创新点在于将带符号距离函数作为空间先验融入多示例学习框架，将视觉特征与显式几何关系相结合，从而识别转移风险的结构特征。这为计算病理学中利用空间建模提升预测性能提供了新思路。

Abstract: Predicting the risk of distant metastasis from primary tumor tissue histology is a critical yet challenging task in computational pathology. Multiple Instance Learning (MIL) approaches can attend to subdomains in tumor regions that harbor features of metastatic cancer progression. However MIL models treat tissue patches as unordered bags, discarding the spatial layout that defines the metastatic potential. We propose that metastatic risk is inherently dictated by the geometric arrangement of the tumor microenvironment at the interface with tumor cells. Our model is designed to explicitly capture the spatial relationships between tumor cells, tumor associated fibroblasts and infiltrating lymphocytes. For this purpose, we propose Distance aware Tissue Modeling for Multiple Instance Learning(DTMf-MIL), a novel method that reinforces visual features with explicit spatial priors. By computing signed distance functions (SDF) relative to tissue phenotypes, our model learns to recognize structural signatures of metastatic risk. This geometric awareness translates directly to superior clinical performance as DTMf-MIL significantly outperforms state-of-the-art methods that ignore spatial layout on metastasis prediction from tissue in the primary tumor. We further validate our approach on public benchmarks, demonstrating that spatial awareness consistently improves diagnostic accuracy across diverse clinical tasks.

[64] SATB-VR: Training Few-Step Video Restoration Diffusion Model using SNR-Aware Trajectory Blending cs.CVPDF

Haoran Bai, Xiaoxu Chen, Xiaoyu Liu, Zongsheng Yue, Sibin Deng

TL;DR: 本文提出了SATB-VR，一种用于视频修复的少步扩散模型训练范式。该方法通过一个辅助预测器来跳过早期低信噪比（SNR）的降噪步骤，从而加速推理。为了解决预测器与降噪器联合训练时存在的训练-推断差异，作者提出了信噪比感知轨迹混合（SATB）策略和降噪器驱动一致性（DDC）损失。实验表明，在少于等于5步的灵活推理设置下，该方法在合成、真实世界和AIGC基准测试中优于现有方法。

Details

Motivation: 扩散模型在视频修复方面表现出色，但其依赖大量迭代步骤限制了效率。而激进的单步蒸馏方法又常常会损害精细纹理的恢复。因此，研究旨在实现效率与质量之间的最佳平衡，开发一种只需少数步骤的高效视频修复方法。

Result: 在合成、真实世界和AIGC（人工智能生成内容）基准测试上进行的大量实验表明，在灵活的少步推理机制下（例如≤5步），SATB-VR的性能优于现有方法。

Insight: 主要创新点在于提出了信噪比感知轨迹混合（SATB）策略，通过在正向过程中根据信噪比动态混合预测器输出与真实轨迹来构建噪声输入，使降噪器能鲁棒地补偿初始预测误差并平滑收敛到干净数据流形。此外，降噪器驱动一致性（DDC）损失利用并更新的降噪器作为动态评估器，显式地对齐内部特征并提升预测器精度，这为解决联合训练中的不一致性问题提供了新思路。

Abstract: While diffusion models excel in video restoration, their reliance on extensive iterative steps limits efficiency. Conversely, aggressive single-step distillation often compromises fine texture recovery. To achieve an optimal balance, we present SATB-VR, a few-step paradigm that jump-starts the denoising process via an auxiliary predictor, explicitly bypassing early low signal-to-noise ratio (SNR) steps. However, naive joint training of the predictor and the denoiser inherently introduces a severe train-inference discrepancy. To resolve this, we propose the SNR-Aware Trajectory Blending (SATB) strategy. During the forward process, SATB constructs the noisy input by dynamically blending the predictor’s output with the ground-truth trajectory based on the SNRs. This forces the denoiser to robustly compensate for initial prediction errors while smoothly converging to the clean data manifold. Furthermore, we introduce a Denoiser-Driven Consistency (DDC) loss, leveraging the concurrently updated denoiser as a dynamic evaluator to explicitly align internal features and boost predictor accuracy. Extensive experiments demonstrate that, under flexible few-step inference regimes (\eg, $\le 5$ steps), SATB-VR performs favorably against existing approaches on synthetic, real-world, and AIGC benchmarks.

[65] Mitigating Batch Effects in Histopathology via Language-Mediated Robust Embedding Generation cs.CV | cs.CLPDF

Yishu Zhang, Shushan Wu, Zhenzhong Zhang, Didong Li, Huaxiu Yao

TL;DR: 本文提出GLMP框架，通过利用预训练的多模态大语言模型将组织病理学图像转换为文本描述作为中间表示，再生成数值嵌入，以减轻跨机构数据中的批次效应，提升模型的泛化能力。

Details

Motivation: 病理学基础模型在临床应用中常受批次效应（即不同组织来源机构引入的非生物变异）影响，导致特征表示失真和泛化性能下降，传统方法如染色归一化效果有限。

Result: GLMP通过文本中间表示有效抑制了机构特异性伪影，增强了跨机构泛化性能，在病理学任务中展现了未开发的潜力。

Insight: 创新点在于首次使用组织学特征的文本描述作为中间表示来生成病理图像的数值嵌入，为构建通用、可泛化且鲁棒的病理学模型提供了新范式。

Abstract: Pathology foundation models (PFMs) have demonstrated strong potential across clinical and scientific applications, yet their performance is often hindered by batch effects, which are non-biological variations across tissue source institutions (TSIs) that distort learned feature representations and impair generalization. Conventional mitigation strategies, such as stain normalization, offer limited success in addressing these high-dimensional, complex artifacts. We present GLMP (General-purpose LLM-Mediated Pathology model), a novel framework that generates robust numerical embeddings from histology image patches through an intermediate textual representation. By leveraging pretrained general-purpose multimodal large language models (MLLMs) and text encoders, GLMP effectively prioritizes biologically meaningful signals over TSI-specific artifacts, thereby improving cross-institutional generalization. To our knowledge, GLMP is the first pathology model to use text descriptions of histological features as an intermediate representation for generating numerical embeddings from histology images. Our results highlight the untapped potential of broad-domain, non-specialized MLLMs in computational pathology and introduce a new paradigm for building versatile, generalizable, and robust pathology models.

[66] CCRC: A Change-Aware Captioning and Reasoning Chain for Image Change Captioning and Segmentation cs.CV | cs.AIPDF

Jinhong Hu, Xiaoping Wang, Shuyin Huang, Guojin Zhong, Kaitai Liu

TL;DR: 本文提出了一种名为CCRC的双链框架，用于解决图像变化描述与分割（ICCS）这一新多模态任务，该任务要求同时生成结构化变化描述和像素级定位。CCRC通过解耦语义推理和空间分割，利用变化感知的注意力机制和令牌细化器，在合成和真实世界变化检测基准上实现了最先进的性能。

Details

Motivation: 传统图像变化描述（ICC）方法缺乏空间定位能力，限制了其精确性，因此需要一种能够联合进行结构化描述和像素级定位的新任务（ICCS）及相应方法。

Result: 在具有像素级监督的合成和真实世界变化检测基准上的实验表明，CCRC实现了最先进的（SOTA）性能。

Insight: 创新点在于提出了ICCS新任务和CCRC双链框架，通过解耦语义推理（CCC链）与空间分割（CCS链），并引入多头部变化感知注意力机制和变化感知令牌细化器，实现了对细微变化的精细感知与精确定位。

Abstract: Understanding and localizing subtle changes between paired images is critical for tasks such as surveillance and image editing. However, traditional Image Change Captioning (ICC) methods lack spatial grounding, limiting their precision. We introduce Image Change Captioning and Segmentation (ICCS), a new multimodal task that jointly requires structured change description and pixel-level localization. To address ICCS, we propose the Change-aware Captioning and Reasoning Chain (CCRC), a dual-chain framework that decouples semantic reasoning from spatial segmentation. The first chain, Chain-of-Change-Captioning (CCC), enhances fine-grained change perception via a visual fusion module based on Multi-Head Change-aware Attention inserted between the visual and language components of a Multimodal Large Language Model (MLLM). CCC also determines whether a change is segmentable. If not, it alone generates the caption. Otherwise, the second chain, Chain-of-Change-Segmenting (CCS), is activated, leveraging spatial priors from CCC and refining masks with a Change-aware Token Refiner for accurate boundary localization. We evaluate CCRC on both synthetic and real-world change detection benchmarks with pixel-level supervision. Experiments show CCRC achieves state-of-the-art performance.

[67] A Physics-Grounded Benchmark for Multi-Agent Dynamics in World Models cs.CV | cs.ROPDF

Nuo Chen, Lulin Liu, Zihao Li, Ziyao Zeng, Zihao Zhu

TL;DR: 本文提出了CrashTwin，一个基于物理的评估框架，用于测试生成式世界模型在模拟多智能体交互（特别是碰撞场景）时的物理可信度。该框架包含一个多样化的碰撞数据集和一个无需校准的重建流程，能够从模型生成的视频中恢复3D物理属性，并系统评估时空一致性、动量与动能守恒以及世界动力学完整性三个维度。

Details

Motivation: 当前生成式世界模型的评估范式过于侧重视觉保真度和语义对齐，无法可靠地量化生成的动力学是否符合基本物理定律，这限制了其作为自主系统可靠模拟器的潜力。

Result: 通过对多个最先进模型进行广泛基准测试，发现高感知质量经常掩盖复杂交互中严重的物理违规现象，从而暴露了现有模型的失败模式。

Insight: 论文的主要创新点在于提出了首个专注于物理可信度的评估框架CrashTwin，其包含一个无需校准即可从视频中重建3D物理属性的流程，以及一个系统的物理诊断套件，为开发物理可靠的世界模型提供了关键工具。

Abstract: Generative world models hold immense promise as scalable simulators for autonomous systems, particularly for synthesizing rare but safety-critical multi-agent interactions, such as vehicle collisions. However, current evaluation paradigms index heavily on visual fidelity and semantic alignment, leaving a critical blind spot: they cannot reliably quantify whether generated dynamics actually obey the fundamental physical laws required for reliable simulation. Assessing this physical plausibility is inherently difficult due to a lack of physical metrics and the challenge of extracting metric-scale kinematics from uncalibrated video rollouts. To bridge this gap, we introduce CrashTwin, a physics-grounded evaluation framework designed to stress-test the physical trustworthiness of world models. CrashTwin couples a diverse dataset of multi-agent collision scenarios, comprising 25K controllable synthetic and 12K in-the-wild real-world collision sequences with a novel calibration-free reconstruction pipeline, enabling the recovery of 3D physical attributes directly from world model rollouts. We propose a diagnostic suite that systematically evaluates three dimensions: spatio-temporal consistency, momentum and kinetic energy conservation, and world-dynamics integrity. Extensive benchmarking of state-of-the-art models reveals a crucial insight: high perceptual quality frequently masks severe physical violations during complex interactions. By quantitatively exposing these failure modes, CrashTwin provides a vital diagnostic tool for developing physically grounded world models capable of reliable real-world simulation.

[68] X-Mind: Efficient Visual Chain-of-Thought via Predictive World Model for End-to-End Driving cs.CV | cs.AIPDF

Bohao Zhao, Chengrui Wei, Guangfeng Jiang, Ruixin Liu, Xuejie Lv

TL;DR: 本文提出X-Mind框架，旨在为视觉-语言-动作模型赋予前瞻性推理能力。通过将预测世界模型内化为视觉思维链，模型在行动前先预测未来状态演化，从而生成更鲁棒且考虑未来后果的驾驶策略。为解决效率挑战，方法引入了融合鸟瞰图与抽象驾驶先验的紧凑草图表示，并采用循环块扩散方案加速生成，实现了在资源受限平台上的低延迟部署。

Details

Motivation: 当前视觉-语言-动作模型缺乏预测未来状态的能力，仅依赖反应式的感知-动作映射，而现有整合预测世界模型的方法要么导致过高延迟，要么作为浅层终端任务未能深度嵌入前瞻性推理。

Result: 在大规模真实世界数据上训练和验证，X-Mind实现了有竞争力的端到端驾驶性能，成为一个高度实用、低延迟的解决方案，成功将大规模认知推理直接部署到资源受限的车载平台。

Insight: 创新点在于将预测世界模型内部化为视觉思维链，强制模型先进行未来推演再行动；并提出了紧凑的草图表示（融合BEV与抽象先验）和循环块扩散方案，以极低token数表示未来推演并将去噪步骤折叠到主干网络的一次前向传播中，显著提升了效率。

Abstract: Predicting future states is essential for autonomous agents, yet current Vision-Language-Action (VLA) models fundamentally lack this capability, relying instead on reactive perception-action mapping. While integrating Predictive World Models (PWMs) addresses this gap, existing approaches either incur prohibitive cascaded latency or act as shallow terminal tasks that fail to deeply embed forward-looking reasoning. To endow VLA models with this reasoning capability, we propose X-Mind. Rather than treating PWMs as an external auxiliary module, this framework internalizes them as the Visual Chain-of-Thought (Visual CoT). By enforcing a world rollout prior to action, the model is constrained to imagine future evolution first, yielding a driving policy that is robustly grounded in environmental dynamics and aware of the future consequences its actions will unfold. The challenge here is efficiency, and we tackle it on two fronts. First, we introduce a compact representation of visual thinking: an abstract sketch that fuses a Bird’s-Eye-View (BEV) layout with abstract driving priors (e.g., navigation intents and traffic rules). Rather than rolling out dense future frames, the model reasons over this sketch as a mental canvas; aided by a Deep Compression Autoencoder (DC-AE), a 12-frame future rollout is reduced to merely 96 tokens, alleviating the long-context computational bottleneck. Second, to accelerate generation further, we propose a recurrent block diffusion scheme that unrolls the denoising steps across the layers of the large drive model, folding iterative refinement into the backbone’s one forward pass. Trained and validated on large-scale real-world data, X-Mind achieves competitive end-to-end driving performance, which makes it a highly practical, low-latency solution that successfully deploys large-scale cognitive reasoning directly onto resource-constrained vehicle platforms.

[69] ViPSim: Collaborating Visual and Parameter Spaces for Consistent Long-Horizon Embodied World Models cs.CV | cs.ROPDF

Longyu Chen, Heng Li, Wei Yang, Manqi Zhao, Dongsheng Jiang

TL;DR: 本文提出了ViPSim框架，通过协同视觉空间和参数空间来解决具身世界模型在长时程生成中的几何不一致问题。视觉空间整合了显式空间先验（如末端执行器位姿、深度信息等），而参数空间则注入原始动作序列和相机矩阵等数值驱动。该方法实现了对几何边界的锚定和数值命令的引导，从而生成长时程一致的状态序列。

Details

Motivation: 当前具身世界模型作为评估基准和基础模拟器时，其可靠性受到低维动作与高维视频合成之间表征差距的阻碍，导致长时程推演中产生累积轨迹漂移和机器人-物体交互不一致的问题。

Result: 大量实验表明，ViPSim与主干网络无关，能显著提升轨迹一致性。该方法在生成与可变形物体（如布料折叠）的复杂交互中展现出涌现能力，并在分布外和跨具身场景中保持鲁棒性能。

Insight: 核心创新在于将视觉空间（提供密集结构基础）和参数空间（提供精确运动指导）协同统一，确保生成状态同时受几何边界锚定和数值命令引导。这为具身智能体的自动化评估和预测控制提供了一个高保真基础。

Abstract: Embodied World Models (EWMs) have emerged as a scalable and risk-free paradigm for advancing embodied intelligence, enabling the safety-critical evaluation of Vision-Language-Action systems. However, their reliability as evaluation benchmarks and foundational simulators is often hindered by the representation gap between low-dimensional actions and high-dimensional video synthesis. This gap results in a lack of geometric correspondence, manifesting as accumulated trajectory drift and inconsistent robot-object interactions during long-horizon rollouts. To bridge this gap, we propose ViPSim, a framework that achieves consistent long-horizon generation through the synergistic collaboration of Visual and Parameter Spaces. We define the Visual Space as a domain of explicit spatial priors, integrating pixel-aligned projections of end-effector pose, camera perspectives, depth-informed scene geometry, and robotic morphological masks to provide dense structural grounding. Concurrently, the Parameter Space serves as a domain of numerical drivers, injecting raw action sequences and camera matrices to provide precise motion guidance. By unifying these two spaces, ViPSim ensures that the generated states are simultaneously anchored by geometric boundaries and steered by numerical commands. Extensive experiments demonstrate that ViPSim is backbone-agnostic and significantly enhances trajectory consistency. Notably, our approach exhibits emergent capabilities in generating complex interactions with deformable objects (e.g., cloth folding) and maintains robust performance in out-of-distribution and cross-embodiment scenarios, providing a high-fidelity foundation for the automated evaluation and predictive control of embodied agents.

[70] CoGS: Compositional Dynamic Human-Object Scenes Gaussian Splatting from Monocular Video cs.CVPDF

Jerrin Bright, John Zelek

TL;DR: CoGS提出了一种基于单目视频的动态人-物交互场景重建框架，通过组合式高斯泼溅技术将场景分解为人体、物体和背景三个分支，并采用六阶段优化策略进行独立稳定与融合，从而解决现有方法中运动纠缠和单目重建约束不足的问题。

Details

Motivation: 现有动态辐射场和高斯泼溅方法在重建单目视频中的动态人-物交互场景时，常因人体、物体和背景共享像素而运动模型不同，导致运动泄漏或重建不准确，尤其是在观测稀少的区域。

Result: 在HOSNeRF和NeuMan数据集上的实验表明，CoGS在完整帧和人体聚焦评估中均提升了人-物交互重建和野外人-场景渲染的保真度与感知质量，实现了更强的性能。

Insight: 创新点在于将场景分解为三个协调分支（人体、物体、背景），并结合六阶段优化策略，通过独立初始化、运动约束和延迟正则化，确保各组件几何与运动的独立性，同时利用光度证据进行最终合成校正，有效避免了运动纠缠。

Abstract: Reconstructing dynamic human–object interaction scenes from monocular video is difficult because the human, manipulated object, and background obey different motion models while sharing the same pixels. Existing dynamic radiance-field and Gaussian-splatting methods often entangle these components, causing object motion to leak into the human or static scene, and monocular human reconstruction remains underconstrained in regions that are rarely observed. We present CoGS, a compositional Gaussian-splatting framework for monocular human–object scene reconstruction. CoGS decomposes the video into three coordinated branches: an articulated human initialized from a complete canonical prior, a rigid object field driven by an estimated object trajectory, and a static scene field regularized by weak scene-only planar primitives when available. A six-stage optimization schedule first stabilizes the human and object independently, then fuses them with the scene under full-image supervision, visibility-aware human anchoring, object silhouette and motion constraints, and delayed scene regularization. This design keeps each component responsible for its own geometry and motion while allowing photometric evidence to correct the final composite. Experiments on HOSNeRF and NeuMan show that CoGS improves both human–object interaction reconstruction and in-the-wild human–scene rendering, achieving stronger fidelity and perceptual quality across full-frame and human-focused evaluations. Code will be released upon publication.

[71] DLGStream: Dynamic Language-embedded Guassian Splatting for Open-vocabulary Enabled Free-viewpoint Video Streaming cs.CVPDF

Zhihui Ke, Yuyang Liu, Xiaobo Zhou, Tie Qiu

TL;DR: 本文提出了DLGStream，一种用于开放词汇自由视点视频流式传输的动态语言嵌入高斯泼溅表示方法。该方法通过双不透明度动态语言高斯表示和基于插值的变形场，在传输颜色属性的同时流式传输时变语言特征，支持4D环境交互、场景编辑和空间智能任务。

Details

Motivation: 基于3D高斯泼溅的自由视点视频通常缺乏用户交互和编辑能力，而现有将CLIP语言特征集成到3DGS的方法无法满足自由视点视频对低帧大小和高帧率的严格要求。

Result: 实验结果表明，DLGStream在开放词汇分割和重建质量方面均取得优异性能，平均帧大小仅为43KB，并能通过变形场实现4D帧插值，将低帧率序列提升至高帧率。

Insight: 创新点包括：1) 双不透明度动态语言高斯表示，分别处理颜色和语言特征以解决联合优化时的性能下降问题；2) 基于插值的变形场，减少时间冗余并支持帧率提升。

Abstract: 3D Gaussian Splatting~~(3DGS) has emerged as a promising paradigm for reconstructing streamable free-viewpoint video~~(FVV) from multi-view videos. However, 3DGS-based FVVs typically lack user interaction and editing capabilities, which diminishes the immersive experience. Recent research has integrated language features from CLIP into 3DGS via distillation, enabling open-vocabulary queries and supporting many downstream applications. Nevertheless, the stringent requirements of FVV, low frame size and high FPS, make current language Gaussian representations unsuitable for language-embedded FVV. In this paper, we propose DLGStream, a novel language-embedded FVV representation that streams time-varying language features alongside Gaussian attributes to support 4D environment interaction, scene editing, and spatial intelligence. Specifically, we propose a dual-opacity dynamic language Gaussian representation, which maintains two opacity attributes for color and language features to deal with performance degradation that occurs when colors and features are jointly optimized. Furthermore, we introduce an interpolation-based deformation field to reduce temporal redundancy. This deformation field can also be used for 4D frame interpolation, boosting FVV sequences from low to high FPS. Experimental results demonstrate that DLGStream achieves superior performance in both on open-vocabulary segmentation and reconstruction quality with an average frame size of merely 43 KB. The code is available on \href{https://github.com/kkkzh/DLGStream}{https://github.com/kkkzh/DLGStream}.

[72] Ground4D: Consistency-Aware 4D Reconstruction from Monocular Video cs.CVPDF

Qing Zhao, Weijian Deng, Pengxu Wei, Liang Lin

TL;DR: Ground4D是一个从单目视频进行一致性感知的4D重建框架。它通过两个阶段实现：首先利用3D基础模型（如VGGT）以无训练方式初始化多视角一致的3D几何和相机位姿；然后通过动态高斯泼溅进行几何一致性感知的优化，在保持多视角几何一致性的同时实现逼真的动态新视角合成。

Details

Motivation: 解决从单目视频学习4D场景表示的挑战，即如何在支持动态新视角合成的同时，保持随时间变化的几何一致性。现有方法如动态高斯泼溅虽渲染性能强但缺乏显式的多视角几何一致性约束，而3D基础模型能恢复连贯几何但不适用于逼真渲染。

Result: 通过将基础级别的几何先验整合到动态高斯优化中，Ground4D实现了更强的重建保真度和渲染性能，强调了基于几何的约束在鲁棒4D场景建模中的作用。

Insight: 创新点在于提出一个两阶段的几何接地框架，将3D基础模型的几何一致性优势与动态高斯泼溅的渲染能力相结合。客观来看，其无训练的几何初始化和在优化中显式保持多视角几何一致性的方法，为4D重建提供了更可靠的结构化初始化与约束。

Abstract: Learning a 4D scene representation from a single monocular video that supports dynamic novel-view synthesis while maintaining faithful geometry over time remains challenging. Dynamic Gaussian Splatting achieves strong rendering performance through photometric optimization, yet does not explicitly enforce multi-view geometric consistency. In contrast, 3D foundation models recover coherent scene geometry and camera motion, but their point-based outputs are not designed for photorealistic rendering. We propose Ground4D, a geometry-grounded framework built on two stages. First, we perform geometry initialization via 3D foundation models, leveraging VGGT in a training-free manner to reconstruct multi-view-consistent 3D geometry and camera poses from monocular video. The recovered geometry provides a structured and reliable initialization for dynamic Gaussian representations. Second, we conduct geometry-consistency-aware refinement via dynamic Gaussian Splatting, optimizing the representation through differentiable rendering while maintaining multi-view geometric consistency across both observed and synthesized viewpoints. Furthermore, Ground4D inherently models the continuous 4D dynamics of the scene, naturally supporting rendering at arbitrary timestamps. By integrating foundation-level geometric priors into dynamic Gaussian optimization, Ground4D achieves stronger reconstruction fidelity and rendering performance, underscoring the role of geometry-grounded constraints in robust 4D scene modeling.

[73] Personalizing MLLMs via Reinforced Multimodal Reference Game cs.CVPDF

Deepayan Das, Davide Talon, Yiming Wang, Massimiliano Mancini, Elisa Ricci

TL;DR: 本文提出了一种名为强化参考游戏（RRG）的学习框架，用于个性化多模态大语言模型（MLLMs），旨在生成准确、具有区分性且无干扰细节的概念描述，以更好地识别用户的独特概念。

Details

Motivation: 现有MLLMs在个性化任务中生成的概念描述常包含状态、上下文等无关信息，这些信息可能干扰对目标概念的独特识别，因此需要一种方法生成更具区分性的描述。

Result: 在三个个性化基准测试的多个任务上，RRG均达到了最先进的性能水平，并且能够泛化到未见过的领域，超越了基于概念描述和特定个性化强化学习框架的现有方法。

Insight: 创新点在于通过一个新颖的强化多模态参考游戏，让MLLM在对比游戏设置中同时扮演说话者和倾听者角色，并利用硬正例和硬负例制定可验证的对比奖励，以促进生成区分性描述。

Abstract: Personalizing Multimodal Large Language Models (MLLMs) aims to recognize users’ unique concepts from visual data and provide personalized responses. Although prior work has shown the benefit of concept descriptions and reasoning for this task, MLLM descriptions often include information, such as state and context, that does not help and may in fact hinder the unique identification of the target concept among other visually similar items. Effective descriptions of personal concepts should instead be accurate, discriminative, and free of distracting details. To achieve such descriptions, we introduce Reinforced Reference Game (RRG), a learning framework that promotes discriminative descriptions through a novel reinforced multimodal reference game. The MLLM plays both the roles of speaker and listener in a contrastive game setting, whose goal is to effectively communicate discriminative information about a target concept. Our approach formulates a verifiable contrastive reward over hard positives (dissimilar views of the same concept) and hard negatives (visually similar but different concepts). Empirically, RRG achieves state-of-the-art across multiple tasks on three personalization benchmarks. RRG generalizes to unseen domains and outperforms existing methods based on concept descriptions and personalization-specific RL frameworks. We will release code and models in the project page.

[74] On Test-Time Scaling for Vision-Language Models cs.CVPDF

Fawaz Sammani, Tzoulio Chamiti, Nikos Deligiannis

TL;DR: 本文首次对大型视觉语言模型（LVLMs）的测试时扩展方法进行了全面研究，探讨了传统为LLMs开发的测试时扩展方法是否可直接应用于LVLMs。研究发现，小型高性能模型从测试时扩展中受益最大，性能提升可达约30%，甚至能超越大型模型；同时，LVLMs在获得超出必要的计算资源时会失去焦点，且视觉信息在推理链早期被编码，之后推理主要由纯文本主导。

Details

Motivation: 测试时扩展是一种在推理时使用额外计算资源以提升性能而不改变模型权重的范式，在LLMs中已被广泛研究，但其在LVLMs中的适用性、时机和程度尚缺乏深入探索和分析。本文旨在回答一个基础问题：为LLMs开发的传统测试时扩展方法能否直接应用于LVLMs？

Result: 研究覆盖了多个模型和模型规模、九种测试时扩展方法以及六个多样化基准测试。主要结果显示，小型高性能模型通过测试时扩展可实现高达约30%的性能提升，达到甚至超越大型模型的水平。

Insight: 创新点在于首次对LVLMs的测试时扩展进行全面实证分析，揭示了小型模型受益最大、LVLMs存在计算过载导致注意力分散的现象，以及视觉信息在推理链中早期编码后贡献显著下降的规律。这些发现为LVLMs的研究和工业部署提供了实用指导和深入见解。

Abstract: Test-time scaling is a paradigm where large models use additional compute at inference to achieve better performance, without changing model weights. While it has been widely studied for Large Language Models (LLMs), its applicability to Large Vision-Language Models (LVLMs) remains less explored and analyzed, with limited analysis of whether, when, and to what extent these approaches transfer to LVLMs. In this work, we ask a simple but fundamental question: can conventional test-time scaling methods developed for LLMs be directly applied to LVLMs? We present the first comprehensive study of test-time scaling for LVLMs, spanning multiple models and model sizes, nine test-time scaling methods, and six diverse benchmarks. Our main findings is that 1) different from previous findings, small, well-performing models benefit the most from test-time scaling, enabling performance improvements of up to around 30%, reaching large models performance, and often outperforming them, 2) LVLMs lose focus when given more compute than necessary, and 3) Visual information is encoded early in the reasoning chain, after which the chain is dominated by text-only reasoning and the contribution of image tokens drops significantly. Finally, we also provide a global and fine-grained analysis on the quality and information sufficiency of the reasoning chains produced. Overall, our findings and analysis provide practical guidance and insights into LVLMs and their deployment in research and industry.

[75] HKVLM: Faithful Reasoning Grounding by Binding Language Queries to a Frozen Detector cs.CVPDF

Bo Ma

TL;DR: HKVLM提出了一种新的视觉语言模型架构，通过将语言查询与冻结的检测器绑定来解决视觉推理任务中的定位与绑定错误问题。该方法使用冻结的语言对齐检测器生成类别无关的区域提议，并使用冻结的语言模型将推理指令编码为查询嵌入，然后通过一个轻量级的对齐钩子（alignment hook）在共享嵌入空间中进行对比检索和二部图分配来绑定查询与区域。该方法还引入了基于感知的忠实性否决机制，防止模型命名未被任何区域支持的对象。

Details

Motivation: 现有视觉语言模型（VLMs）在处理需要推理的视觉请求时，常出现“看到但说错”的问题，即关注到正确区域但返回错误的边界框或标签。作者认为这是绑定失败导致的，因为定位过程与语言生成过程耦合，或者模型的意图被压缩为单个类别字符串。

Result: 在RefCOCO/RefCOCO+/RefCOCOg和POPE基准上，仅训练对齐钩子即可将基础准确率提升50-90倍；忠实性否决机制将POPE准确率从接近随机（0.50）提升至0.66-0.76，并将幻觉率从约0.99降低至0.23-0.43。增加提议区域数量（从M=50到M=300）无需重新训练即可将基础准确率提升19-24%，表明剩余误差主要是感知误差（SeeErr）而非绑定误差（SayErr）。

Insight: 核心创新在于将定位任务从语言路径中解耦，通过冻结预训练模型和训练轻量级对齐钩子来实现高效绑定，这在小数据冷启动场景下具有优势。提出的“说-看”分解（say-vs-see decomposition）为分析模型错误提供了新视角，而忠实性否决机制有效减少了幻觉现象。

Abstract: Many visual requests – the object to open this bottle'', the person not wearing a helmet’’ – require reasoning, not just category matching. Pure open-vocabulary detectors need an explicit phrase; vision-language models (VLMs) can reason yet ``see but mis-speak’’, attending to the right region but returning the wrong box or label. We argue this is a \emph{binding} failure: in coordinate-as-text VLMs localization passes through the autoregressive head, coupling it to language generation; in two-stage pipelines the model’s intent is squeezed through a single class string. We present HKVLM, which removes localization from the language path. A frozen, language-aligned detector emits class-agnostic region proposals; a frozen language model encodes reasoning instructions as referential query embeddings; a lightweight \emph{alignment hook} binds queries to regions by contrastive retrieval and bipartite assignment in a shared embedding space. A perception-grounded faithfulness veto forbids naming an object that no region supports. Only the hook is trained, targeting small-data cold-start settings where monolithic VLM tuning struggles. We formalize a \emph{say-vs-see} decomposition separating localization error (SeeErr) from binding error (SayErr), and evaluate on RefCOCO/RefCOCO+/RefCOCOg and POPE. With frozen Grounding DINO and Qwen2.5-VL, training only the hook lifts grounding accuracy by $50$–$90\times$ over untrained cross-space matching; the faithfulness veto raises POPE accuracy from near-chance ($0.50$) to $0.66$–$0.76$ and reduces hallucination from ${\sim}0.99$ to $0.23$–$0.43$, with gains from $200$ expressions. Increasing proposals from $M{=}50$ to $M{=}300$ improves grounding by $19$–$24%$ without retraining, confirming that residual error is perceptual (SeeErr) rather than binding (SayErr).

Zixiao Zhang, Lingling Li, Pei He, Xu Liu, Licheng Jiao

TL;DR: 本文提出了一种名为ExACT的训练免费视觉定位框架，用于遥感图像中的开放词汇视觉定位。该框架通过单样本视觉提示机制，利用给定的示例图像提取细粒度视觉对应关系，以校准冻结多模态大语言模型（MLLMs）的粗略跨模态先验，并结合结构感知细化器生成几何提示，引导Segment Anything Model（SAM）实现精确的像素级定位。

Details

Motivation: 遥感视觉定位（RSVG）旨在使用自然语言描述定位高分辨率遥感图像中的特定对象，但现有训练免费的多模态大语言模型（MLLMs）在适应时面临抽象语言语义与细粒度视觉线索之间的模态鸿沟，导致在杂乱场景中定位漂移严重。

Result: 大量实验证实，ExACT在现有训练免费和弱监督方法中表现出优越性，在遥感视觉定位基准上实现了更精确的像素级预测。

Insight: 创新点包括提出基于视觉示例的校准器（VEC）来提取细粒度视觉对应以抑制背景伪影，以及结构感知细化器（SAR）通过迭代聚类策略生成高质量几何提示，结合SAM实现无需训练的精确定位，有效弥合了模态鸿沟。

Abstract: Remote sensing visual grounding (RSVG) aims to locate specific objects in high-resolution RS imagery using free-form natural language descriptions. While recent advances in multimodal large language models (MLLMs) show great potential for such open-vocabulary RSVG, their training-free adaptation is hindered by the modality gap between abstract linguistic semantics and fine-grained visual cues. In cluttered RS scenes, this gap inevitably causes severe localization drift. To bridge this gap, we propose Exemplar-driven Calibrated Refinement (ExACT), a novel training-free framework driven by a one-shot visual prompting mechanism to explicitly provide discriminative structural guidance for precise pixel-level localization. Specifically, we propose a Vision Exemplar-based Calibrator (VEC) that extracts fine-grained visual correspondences from the given exemplar to rectify the rough cross-modal priors from frozen MLLMs, effectively suppressing background artifacts and accurately outlining target boundaries. Subsequently, a Structure-Aware Refiner (SAR) employs an iterative merge-and-select clustering strategy to consolidate the calibrated priors into high-quality positive and negative geometric prompts. These prompts then guide the Segment Anything Model (SAM) to achieve precise pixel-level predictions. Extensive experiments confirm the superiority of ExACT over existing training-free and weakly-supervised methods.

[77] Self-Evolving Agentic Image Restoration via Deliberate Planning and Intuitive Execution cs.CVPDF

Shuang Cui, Fan Ji, Guanglong Sun, Yufei Guo, Xiongxin Tang

TL;DR: 本文提出了一种名为SEAR的自进化智能体图像复原框架，将复原任务建模为序列决策问题。该框架受双过程理论启发，包含一个遵循快速思考的直觉执行器和一个遵循慢速思考的审慎规划器。审慎规划器采用剪枝感知的蒙特卡洛树搜索进行长程推理，并结合混合无参考奖励与基于多模态大语言模型的锦标赛来防止指标利用；直觉执行器则利用基于退化感知状态指纹索引的自进化情景记忆，将昂贵的搜索轨迹提炼为自适应专业知识，从而克服情景遗忘并逐步分摊冷启动探索成本。

Details

Motivation: 解决现实世界图像复原中复杂且耦合的退化问题。现有基于大语言模型的智能体复原框架存在两个关键局限：一是搜索策略上过度依赖贪婪策略，无法平衡探索与利用；二是现有智能体系统未能充分利用信息，表现出情景遗忘。

Result: 在合成和真实世界基准测试上的大量实验表明，该方法在感知质量和定量指标上均表现出色。

Insight: 核心创新点在于受双过程理论启发的直觉执行器与审慎规划器协同框架，以及结合了剪枝感知蒙特卡洛树搜索、混合无参考奖励、MLLM锦标赛的审慎规划机制，和基于退化感知状态指纹索引的自进化情景记忆机制。这为解决智能体系统中探索-利用权衡和情景遗忘问题提供了新思路。

Abstract: Real-world image restoration (IR) remains challenging due to complex and coupled degradations. While recent agentic IR frameworks leverage Large Language Models for flexible tool planning, they face two critical limitations. First, from a search scheme perspective, excessive reliance on greedy strategies fails to balance exploration and exploitation. Second, existing agentic systems underutilize information, exhibiting episodic amnesia. To address these challenges, we propose \textbf{Self-Evolving Agentic Image Restoration (SEAR)}, which formulates restoration as a sequential decision-making problem. Inspired by the dual-process theory, SEAR comprises an Intuitive Executor and a Deliberate Planner, respectively following the fast-thinking \textit{System 1} and slow-thinking \textit{System 2} principles. The Deliberate Planner employs Pruning-Aware Monte Carlo Tree Search for long-horizon reasoning, utilizing a hybrid no-reference reward and a Multimodal Large Language Model (MLLM)-based tournament to prevent metric exploitation. Complementarily, the Intuitive Executor leverages a self-evolving episodic memory indexed by degradation-aware state fingerprints. This mechanism distills expensive search trajectories into adaptive expertise, overcoming episodic amnesia while progressively amortizing cold-start exploration costs through memory reuse. Extensive experiments on synthetic and real-world benchmarks demonstrate its strong perceptual and quantitative performance.

[78] SciFlow: Semantic Cross Interference for Self-Supervised Optical Flow Domain Generalization cs.CVPDF

Jamie Menjay Lin, Jisoo Jeong, Hong Cai, Kai Wang, Fatih Porikli

TL;DR: 本文提出了SciFlow方法，用于解决光流估计模型从合成域到真实域的泛化问题。该方法通过自监督学习，在训练过程中将真实图像的语义干扰引入合成图像，并结合几何一致性约束，使模型能够适应真实世界场景。

Details

Motivation: 由于像素级光流标注成本高且稀缺，模型通常在合成数据上训练，但在真实场景中部署，因此需要解决合成到真实域的泛化挑战。

Result: 实验结果表明，SciFlow显著提升了模型在域变化下的鲁棒性，并实现了无需真实域标注的合成到真实域泛化。

Insight: 创新点在于通过语义干扰和几何一致性自监督，实现网络无关的域泛化训练，可借鉴其跨域特征混合策略用于其他视觉任务的域适应。

Abstract: Motions of objects and scenes carry essential intelligence in video understanding, offering rich cues for interpreting dynamic settings and interactions. Due to the cost and scarcity of high-quality annotation or ground truth of pixel-wise optical flow, however, motion estimation models are typically trained in synthetic domains while deployed in real-world domains. Addressing synthetic-to-real domain generalization challenges has been crucial for developing practical solutions in diverse open-world use cases. This paper introduces SciFlow, a simple yet effective, network-agnostic, training-based approach that leverages self-supervised learning to generalize motion estimation across synthetic and open-world domains. Specifically, SciFlow imposes semantic interference from open-world images onto synthetic images during training, blending indomain features with cross-domain interference, which enables the network to adapt to the real-world domains. Additionally, SciFlow utilizes geometric consistency to ensure validity of the self-supervision. Our experiment results show that SciFlow not only significantly enhances model robustness amidst domain variations, but also remarkably enables synthetic-to-real domain generalization without requiring any ground truth in the open world.

[79] Mural: Transferring LLM knowledge to image generation via Mixture-of-Transformers cs.CVPDF

Achin Jain, Jie An, Siddharth Chaudhary, Davide Modolo

TL;DR: 这篇论文提出了一种名为Mural的方法，通过混合专家（Mixture-of-Transformers, MoT）架构，将冻结的大型语言模型（LLM）的知识迁移到基于扩散模型的文本到图像（T2I）生成任务中。该方法仅使用标准文本-图像对进行训练，无需额外的多模态数据或显式推理监督，旨在探索LLM的内在知识在图像生成中的可访问性和由此产生的新能力。

Details

Motivation: 研究动机是探索如何有效利用冻结LLM的知识来增强文本到图像生成，特别是在仅使用标准文本-图像对训练的情况下，以解决传统T2I模型可能缺乏深层语义理解和推理能力的问题。

Result: 在多个基准测试中，该方法取得了强劲性能：在GenEval上得分为0.85，在DPG-Bench上得分为86.75，在WISE上通过推理时推理得分为0.66，这些结果在统一的理解-生成系统中表现突出。

Insight: 创新点在于通过MoT架构共享注意力机制，将冻结LLM与扩散生成器集成，实现了LLM知识向图像生成的迁移，并发现了训练数据中未出现的涌现行为，如跨语言图像生成、颜色引导构图等，这为资源受限的多模态学习提供了新途径。

Abstract: Leveraging capabilities of large language models (LLMs) in text-to-image (T2I) synthesis is an important research direction. In this work we investigate whether the knowledge of a frozen LLM can be effectively utilized in T2I generation when trained exclusively on standard text-image pairs. We integrate a frozen, reasoning-capable LLM with a diffusion-based image generator via shared attention within the Mixture-of-Transformers (MoT) architecture. Our experiments span two critical questions: (1) what degree of the LLM’s intrinsic knowledge remains accessible during T2I training, and (2) what novel capabilities emerge in the resulting system. Across established benchmarks, our models achieve strong performance among unified understanding-generation systems: 0.85 on GenEval, 86.75 on DPG-Bench, and 0.66 on WISE with inference-time reasoning, using only text-image data. Remarkably, we uncover emergent behaviors absent from training data, including cross-lingual image generation, color-guided composition, emoji / ASCII scene construction, and generation directed by world knowledge. These results demonstrate that pretrained LLM knowledge can guide image synthesis under standard text-to-image training paradigms, without interleaved multimodal signals or explicit reasoning supervision. Our findings open new avenues for harnessing frozen model capabilities in resource-constrained multimodal learning.

[80] Evidence-Based Text-Conditioned 3D CT Synthesis for Ovarian Cancer cs.CV | cs.AI | cs.LGPDF

Francesca Pia Panaccione, Eugenio Lomurno, Francesca Fati, Carlotta Pecchiari, Marina Rosanu

TL;DR: 本文提出OvESyn框架，用于卵巢癌的文本条件3D CT合成。该框架直接从CT衍生的影像描述符和常规临床元数据构建标准化的‘发现与印象’部分，无需原始放射学报告，并利用这些文本条件驱动针对493名高级别浆液性卵巢癌患者调整的潜在扩散模型。这是首个适应于腹盆腔肿瘤环境的文本条件3D CT合成框架。

Details

Motivation: 卵巢癌常在晚期确诊，术前增强CT对分期和手术规划至关重要，但标注影像数据稀缺且受隐私法规限制，阻碍了该领域通用计算模型的发展。现有文本条件3D CT合成方法依赖于配对的放射学报告，且仅在胸部CT上评估，无法直接应用于腹盆腔肿瘤场景。

Result: 在针对高级别浆液性卵巢癌患者的实验中，完整OvESyn框架在分布和强度保真度上达到最佳（FID2.5D 29.35，精确度0.671，Wasserstein-1 0.044）。仅生成器微调的变体则最大化覆盖率（召回率0.645），体现了编码器调整所控制的保真度/覆盖率权衡。若无生成器领域适应，合成结果仍锚定在胸部预训练域，性能指标崩溃（精确度和召回率降至零，FID2.5D超过140）。

Insight: 创新点在于提出了一种无需原始放射学报告的文本条件合成框架，通过CT衍生描述符和临床元数据自动构建文本条件，首次将文本条件3D CT合成适应于腹盆腔肿瘤成像。关键洞察是：跨越领域差距（从胸部到腹盆腔）的核心机制是生成器的领域适应（微调），而非视觉-语言编码器的对齐；编码器对齐主要优化强度和细节保真度。这为报告稀缺环境下的可迁移性和合成队列生成提供了基础。

Abstract: Ovarian cancer is frequently diagnosed at an advanced stage, making preoperative contrast-enhanced computed tomography (CT) central to staging and surgical planning; yet the scarcity of annotated imaging data, compounded by privacy regulations, limits the development of generalizable computational models in this domain. Text-conditioned 3D CT synthesis has shown promise, but existing pipelines depend on paired radiology reports and have been evaluated only on chest CT. We propose OvESyn (Ovarian Evidence-based Synthesis), a framework that constructs standardized Findings and Impression sections directly from CT-derived imaging descriptors and routine clinical metadata, without any original radiology report, and uses them to condition a latent diffusion model adapted to 493 high-grade serous ovarian carcinoma patients. This is the first text-conditioned 3D CT synthesis framework adapted to an abdomino-pelvic oncologic setting. A systematic ablation over two adaptation axes, vision-language encoder alignment and generator fine-tuning, identifies generator domain adaptation as the operative mechanism for crossing the domain gap and establishing the target anatomy: without it, synthesis remains anchored to the thoracic pretraining domain, with Precision and Recall collapsing to zero and FID2.5D exceeding 140, regardless of encoder alignment. Encoder alignment instead refines intensity and fine detail. The full OvESyn attains the best distributional and intensity fidelity (FID2.5D 29.35, Precision 0.671, Wasserstein-1 0.044), while the generator-only variant maximizes coverage (Recall 0.645), reflecting a fidelity/coverage trade-off governed by encoder adaptation. Requiring only automatic segmentations and routine preoperative metadata, OvESyn supports transferability to report-scarce settings and provides a foundation for synthetic cohort generation in abdomino-pelvic oncologic imaging.

[81] Learning from Acquisition: Metadata-driven Multimodal Pre-training for Cardiac MRI cs.CV | eess.IVPDF

Xueyi Fu, Liwei Hu, Zi Wang, Guang Yang

TL;DR: 本文提出MetaCLIP-CMR，一个基于对比语言-图像预训练（CLIP）的元数据驱动框架，用于心脏磁共振成像（CMR）的表征学习。该框架将成像模态、解剖视图、扫描仪厂商、场强和扫描仪型号等结构化采集元数据转换为文本监督信号，以增强预训练。

Details

Motivation: 大多数CMR基础模型主要依赖纯图像预训练，而忽略了常规记录的结构化采集元数据这一天然的弱语义监督源。本文旨在利用这些未被充分探索的元数据来改进CMR表征学习。

Result: 在成像模态分类和电影视图分类任务上，MetaCLIP-CMR分别达到86.8%和86.5%的准确率，明显优于ImageNet和掩码重建初始化方法。在下游心脏分割任务中，在ACDC和M&Ms电影短轴（SAX）数据集的全数据和20%微调设置下，均获得最高的Dice分数，与近期大规模CMR预训练模型在ACDC分割上性能相当，但预训练图像规模仅需其不到1%。

Insight: 主要创新点在于将常规采集元数据（如模态、视图、扫描仪信息）转化为文本监督，用于构建多模态（图像-文本）对比学习框架。这为利用医疗影像中天然存在的弱监督信息提供了一种自然且易于实施的策略，显著降低了大规模预训练的数据需求。

Abstract: Cardiac magnetic resonance imaging (CMR) routinely records structured acquisition metadata, yet most CMR foundation models rely primarily on image-only pre-training and leave this naturally available source of weak semantic supervision largely underexplored. We propose MetaCLIP-CMR, a metadata-driven framework based on Contrastive Language–Image Pre-training (CLIP), which converts imaging modality, anatomical view, scanner vendor, field strength, and scanner model into textual supervision for CMR representation learning. The pretrained image encoder is evaluated on imaging modality classification, cine view classification, and cardiac segmentation. MetaCLIP-CMR achieves 86.8% modality accuracy and 86.5% cine view accuracy, clearly outperforming ImageNet and masked reconstruction initialisations. For downstream cardiac segmentation, MetaCLIP-CMR consistently obtains the highest Dice score across the evaluated ACDC and M&Ms cine short-axis (SAX) settings under both full-data and 20% fine-tuning regimes. Compared with recent image-focused large-scale CMR pre-training models, MetaCLIP-CMR achieves comparable ACDC segmentation performance, while requiring less than 1% of their pre-training image scale. These results suggest that metadata learning offers a natural and easy-to-use strategy for transforming routinely recorded acquisition information into effective supervision for foundation-level CMR representation learning, highlighting the promise of metadata-driven multimodal pre-training.

[82] Adaptive Spectrum-Aware Feature Disentangled Network for Small Object Detection cs.CVPDF

Yang Guo, Zihan Yang, Feifei Kou, Yulan Hu, Ran Zhang

TL;DR: 本文提出了一种名为SFDNet的自适应频谱感知特征解耦网络，用于解决小目标检测中因空间分辨率有限和视觉线索弱导致的性能下降问题。该框架通过自适应频谱解耦模块将骨干网络特征分解为多个互补的频谱分量，以消除背景干扰，并结合类原型蒸馏过程增强同类目标的语义一致性。

Details

Motivation: 小目标检测因空间分辨率低和视觉特征弱而具有挑战性，现有方法在不同频谱中的背景干扰仍会降低检测性能，因此需要构建更具判别性的目标相关表示。

Result: 在多个具有挑战性的基准测试上进行的大量实验表明，SFDNet大幅优于现有的最先进方法，实现了显著的性能提升。

Insight: 创新点包括自适应频谱解耦模块，通过分解特征频谱来消除背景干扰，以及类原型蒸馏过程，通过建立类原型并执行高效蒸馏来增强语义一致性，从而提升小目标检测的鲁棒性和准确性。

Abstract: Small Object Detection (SOD) is a fundamental yet challenging problem in computer vision due to its limited spatial resolution and weak visual cues. Although recent approaches have achieved remarkable advances, the background distractors in different frequency spectra still degrade the performance. In this paper, we propose a novel small object detection framework termed SFDNet, which is capable of detecting small objects via efficient spectrum-aware feature disentanglement. Specifically, we propose an Adaptive Spectrum Disentanglement (ASD) module that decomposes backbone features into multiple complementary spectral components, aiming to construct discriminative object-relevant representations by discarding the background distractors for each component. Afterwards, to strengthen the semantic consistency of the similar objects in the same class, we propose a Class-Wise Prototype Distillation (CPD) procedure, which establishes class prototypes for the object instances and enforces the compact representation by efficient prototype distillation. Extensive experiments on multiple challenging benchmarks show that SFDNet outperforms existing state-of-the-art methods by a large margin. Our code is available at https://github.com/ManOfStory/SFDNet.

[83] Efficient Spatio-Temporal Grounding with Multimodal Large Models via Second-Level Tracking and RL Verification cs.CV | cs.AIPDF

Tianshu Zhang, Yan Wang, Ji Qi, Lijie Wen

TL;DR: 本文提出了一种高效的长视频时空定位方法，通过秒级跟踪和强化学习验证来优化基于自然语言查询的时空定位。该方法将帧级推理转换为秒级跟踪并进行跨秒平滑，以减少计算开销并保持连续性。同时，利用多模态模型合成思维链式轨迹进行监督，并通过基于t_IoU+mv_IoU的验证器进行强化学习优化，实现了效率与定位质量之间的良好平衡。

Details

Motivation: 解决长视频时空定位中直接应用逐帧推理计算成本高且不稳定的问题，旨在提升定位的效率和鲁棒性。

Result: 在多种FPS设置下的实验中，该方法在效率和定位质量之间取得了较强的权衡，但未明确提及是否达到SOTA或与特定模型相当。

Insight: 创新点包括从帧级到秒级的跟踪转换、跨秒平滑技术、利用合成思维链轨迹进行监督，以及结合强化学习和验证器进行策略优化，这些方法可借鉴于视频理解任务中以降低计算复杂度并提升性能。

Abstract: Spatio-temporal grounding in long videos requires precise temporal localization and robust object tracking conditioned on natural-language queries. While recent vision-language models (VLMs) show strong reasoning ability, directly applying frame-by-frame inference to long sequences is computationally expensive and unstable. We propose a practical pipeline that shifts from frame-level to second-level tracking and performs cross-second smoothing to preserve continuity while reducing sequence length. To improve reasoning supervision, we synthesize chain-of-thought style trajectories using advanced multimodal models for temporal localization and target selection, and replace generated spatio-temporal coordinates with ground-truth annotations to avoid noisy supervision. We further optimize the policy with reinforcement learning using a verifier based on $t_\mathrm{IoU}+mv_\mathrm{IoU}$. Experiments across multiple FPS settings show that our method achieves a strong trade-off between efficiency and localization quality.

[84] Semantic-Aware, Physics-Informed, Geometry-Grounded Weather Video Synthesis cs.CV | cs.AI | cs.ET | cs.MMPDF

Chenghao Qian, Nedko Savov, Lingdong Kong, Yeying Jin, Rui Song

TL;DR: 本文提出了一种语义感知、物理驱动、几何约束的天气视频合成框架，旨在为输入视频添加多样且可控的天气效果。该方法将合成过程分解为语义、动态和几何三个独立的引导信号，通过引导现成的视频编辑器来生成具有真实感的全局外观和粒子动态。实验表明，该方法能合成多样且物理真实的天气效果，并能有效提升自动驾驶语义分割模型在恶劣天气下的鲁棒性。

Details

Motivation: 现有天气合成方法存在天气外观多样性不足、对天气动态（如时间演化和粒子运动）控制力弱的问题，且依赖的文本提示过于笼统，难以生成细节特征；同时，通用视频编辑器倾向于抑制强烈的天气现象，难以生成密集的粒子效果。

Result: 实验表明，该方法能生成多样、物理和视觉上真实的天气效果。此外，合成的数据显著提升了自动驾驶语义分割模型在恶劣天气条件下的鲁棒性。

Insight: 创新点在于将天气合成分解为语义（外观）、物理（动态）和几何（位置）三个独立且稳定的引导信号，并具体引入了语义感知的外观锚定、物理驱动的动态模拟（基于高斯粒子场）和几何约束的视频合成。这提供了对天气外观、动态和空间分布的精细可控性，超越了仅依赖文本提示的方法。

Abstract: Weather synthesis aims to add weather effects to input videos while preserving scene identity, structure, and motion. The key limitation of existing methods is the lack of diversity in weather appearance and effective control over weather dynamics (e.g., temporal evolution and particle motion). Most approaches rely on text prompts, which are inherently underspecified and often fail to produce detailed weather characteristics. Additionally, general-purpose video editors optimized for clean and aesthetic outputs tend to suppress heavy weather phenomena, making dense particle effects difficult to generate. To address these, we propose a Semantic-Aware, Physics-Informed, and Geometry-Grounded framework that steers an off-the-shelf video editor to synthesize diverse global appearances and detailed particle dynamics. We factorize the synthesis into three conditional signals, so that each provides a distinct and stable source of guidance: semantics specifies what the weather should look like, dynamics governs how it evolves over time, and geometry determines where it should appear in the scene. Specifically, we introduce (1) semantic-aware appearance anchoring to establish the target appearance from scene semantics and user input; (2) physics-informed dynamic simulation to generate particle effects by simulating a Gaussian-represented particle field under gravity, wind, and turbulence; and (3) geometry-grounded video synthesis to align the simulated particles with target scene geometry and synthesize the final video. Experiments demonstrate that our method produces diverse, physically and visually realistic weather effects. Furthermore, we show that our synthesized data significantly improves the robustness of autonomous driving semantic segmentation under adverse weather conditions. Project page: https://jumponthemoon.github.io/w-crafter/.

[85] Flow Matching in Feature Space for Stochastic World Modeling cs.CV | cs.AIPDF

Francois Porcher, Nicolas Carion, Karteek Alahari, Shizhe Chen

TL;DR: 本文提出FlowWM，一种在预训练特征空间（如DINOv3）中直接进行流匹配的随机世界模型，旨在同时提升未来预测的准确性和多样性，并保持对下游感知任务有用的信息。

Details

Motivation: 现有视觉世界模型难以兼顾预测不确定性和感知性能：基于VAE的随机模型在低维重建潜在空间中操作，限制了感知性能；而使用强预训练特征的确定性预测器则将多模态未来坍缩为单一的模糊均值。

Result: 在合成基准测试和真实世界基准测试FuturePerception上，FlowWM在感知性能、模态覆盖和预测时域鲁棒性方面均得到提升，验证了其在高效训练和任务驱动目标下的有效性。

Insight: 创新点在于将流匹配直接应用于高维预训练特征空间，并引入了可微分的一步投影机制以实现高效训练，解决了高维特征空间中标准扩散方法次优的挑战，为随机世界建模提供了新的设计思路。

Abstract: World modeling requires forecasting uncertain futures while preserving information useful for downstream perception. Existing visual world models often struggle to satisfy both goals: VAE-based stochastic models operate in low-dimensional reconstruction latents, which can limit perception performance, while deterministic predictors using strong pretrained features collapse multimodal futures into a single blurry mean. In this work, we propose FlowWM, a stochastic world model that performs flow matching directly within pretrained feature space (e.g., DINOv3). This is challenging because pretrained features are substantially high-dimensional, making standard diffusion recipes suboptimal. To address this, we investigate the design choices needed for feature-space flow matching and introduce a differentiable one-step projection mechanism that enables efficient training with temporal consistency and task-driven objectives. We evaluate FlowWM on two benchmarks: a synthetic benchmark for systematic evaluation of accuracy and diversity, and a real-world benchmark FuturePerception. FlowWM improves perception performance, mode coverage, and horizon robustness, validating our proposed design for stochastic world modeling in high-dimensional feature spaces.

[86] HorizonRelight: Relighting Long-horizon Videos Consistently via Diffusion Transformers cs.CV | cs.GR | cs.LGPDF

Jing Yang, Mayoore Jaiswal, Zian Wang, Steven Zeng, Rochelle Pereira

TL;DR: 本文提出HorizonRelight框架，通过将长时视频重光照任务重构为时序条件化的隐空间域转换问题，利用跨块传播目标域隐变量和掩码目标域自条件训练来提升时间一致性，并引入基于可控生成模型的暖启动提示来建立初始目标域状态。

Details

Motivation: 现有基于扩散的视频重光照方法在短片段上训练，通过分块滑动窗口推理应用于长时视频时，常导致块边界处的时间不连续性问题。

Result: 在真实世界长时视频上的实验表明，该方法显著提升了时间一致性，大幅减少了块边界伪影并抑制了跨块的不必要外观变化。

Insight: 创新点包括将长时重光照重构为时序条件隐空间域转换、引入掩码目标域自条件训练实现跨块连续性学习，以及通过暖启动提示建立目标域初始状态并提供基于提示的重光照通用接口。

Abstract: Diffusion-based video relighting enables controllable relighting from a single input video, but modern video diffusion backbones are trained on short clips and applied to long-horizon videos through chunked sliding-window inference, often causing temporal discontinuities at chunk boundaries. We address this by reframing long-horizon relighting as \emph{temporally conditioned latent domain translation}. Our framework enforces cross-chunk continuity by propagating target-domain latents across boundaries and makes this behavior learnable using \emph{masked target-domain self-conditioning}, training the model to continue from temporally masked propagated context. We further introduce \emph{warm-start prompting} with a relit prompt anchor from a controllable generative model, which establishes the initial target-domain state and creates a general interface for prompt-based relighting. Experiments on in-the-wild long-horizon videos show markedly improved temporal consistency, with chunk-boundary artifacts largely reduced and unwanted appearance changes across chunks greatly suppressed.

[87] From Fog Chamber to Aircraft Window: Pixel-Registered Imaging and Synthetic Fine-Tuning Enable Cross-Domain Defogging cs.CV | physics.opticsPDF

Alexander Ingold, Sabina D. Menon, Manya Yellepeddy, Alec Ikei, John D. Hodges

TL;DR: 该论文提出了一种跨域去雾方法，通过实验室可控雾图像预训练和域随机化合成雾微调，实现了从实验室环境到真实场景（如飞机舷窗视频）的泛化，无需目标域训练。该方法解决了真实世界双目去雾中的迁移限制问题，并展示了在多种分布外场景下的有效性。

Details

Motivation: 解决真实世界去雾任务中模型从受控实验室环境迁移到未知传感器、场景和光路时的泛化能力不足问题，特别是针对先前研究中报告的开源迁移限制。

Result: 在552张保留测试集上，NAFNet达到最佳性能（24.33 dB / 0.7912 SSIM）；在未配对的飞机舷窗视频上，微调后NIQE从6.22降至4.97，输出时间稳定；在O-HAZE/NH-HAZE分割集上达到20.71 dB / 0.683 SSIM。

Insight: 创新点包括使用像素对齐的实验室雾/清晰图像对进行精确配准训练，避免对抗性幻觉；以及通过域随机化合成雾进行微调以增强跨域泛化。客观分析认为，该方法通过结合物理精确数据和合成增强，有效提升了模型在未知真实场景中的鲁棒性。

Abstract: A deep defogging pipeline pretrained on controlled laboratory fog and fine-tuned with domain-randomized synthetic fog applied to clear outdoor scenes generalizes across a graded sequence of out-of-distribution settings with no target-domain training, from chamber-free free-flowing fog to iPhone video recorded through an aircraft cabin window in flight, an entirely unseen sensor, scene, and optical path. This directly addresses an open transfer limitation reported for real-world binocular defogging. Two design choices support the transfer. First, a single-camera fog imager photographs a flat-panel display through an artificial-fog enclosure with a fixed 114mm scattering path, producing 5{,}495 pixel-aligned foggy/clear pairs. Exact registration permits a paired Laplacian ratio that predicts per-image restoration quality far better than single-image proxies (Spearman $ρ= 0.632$ versus $0.399$) and supports pixel-exact $L_1$ reconstruction training that avoids adversarial hallucination. Second, the fog-chamber checkpoint is fine-tuned on Mapillary Vistas crops overlaid with on-the-fly randomized synthetic fog spanning a broad range of strengths, spatial variations, airlights, and noise conditions. On a 552-image held-out split, a uniform comparison of 30 restoration backbones places NAFNet at the top (24.33dB/0.7912~~SSIM), with a compact alternative within 1.29~~dB at 3% of the parameter count, and a ResNet-50 classifier confirms that the restoration preserves semantic content rather than only pixel-level structure. On unpaired aircraft-window video, NIQE decreases from a mean of 6.22 to 4.97 after fine-tuning, with temporally stable output across full-motion sequences. The same backbone, under paired supervision, also reaches 20.71dB/~~0.683~~SSIM on a non-overlapping O-HAZE/NH-HAZE split (a transferability check rather than a competitive ranking).

[88] TrafficAlign: Aligning Large Language Models for Traffic Scenario Generation cs.CVPDF

Zhi Tu, Liangkun Niu, Tianyi Zhang

TL;DR: 本文提出了TrafficAlign框架，通过利用真实驾驶视频合成交通场景并进行数据验证，以对齐大型语言模型（LLMs）生成符合真实世界交通分布的交通场景。评估表明，TrafficAlign生成的场景在三个自动驾驶模型上平均能揭示比现有方法多10.8%的碰撞，且用这些场景微调模型可将碰撞率降低36.1%。定性研究显示，该框架生成的场景与六个不同地理区域的交通分布高度一致。

Details

Motivation: 预训练的大型语言模型（LLMs）在生成交通场景时往往与真实世界交通分布不一致，这限制了其在自动驾驶测试中的有效性，因此需要一种自动化框架来对齐LLMs以生成更真实的场景。

Result: 在三个自动驾驶模型上，TrafficAlign生成的场景平均比现有方法多揭示10.8%的碰撞；用这些场景微调模型后，碰撞率比原始模型降低了36.1%。定性分析基于六个不同地理区域的交通数据集，显示生成场景与区域交通分布强对齐。

Insight: 创新点在于自动化框架TrafficAlign，它通过合成真实驾驶视频场景并验证数据来对齐LLMs，从而生成更符合现实分布的交通场景，这为LLMs在自动驾驶领域的应用提供了数据驱动的对齐方法，有助于提升场景生成的真实性和测试效果。

Abstract: Recent research has investigated the use of large language models (LLMs) to generate traffic scenarios for autonomous driving. However, pretrained LLMs often fail to align with real-world traffic distributions. In this work, we present TrafficAlign, an automated framework that synthesizes traffic scenarios based on real-world driving videos, performs data validation, and aligns LLMs with the synthesized scenarios. The evaluation shows that traffic scenarios generated by TrafficAlign are highly effective, revealing up to 10.8% more collisions on average across three autonomous driving models than state-of-the-art methods. Furthermore, fine-tuning these driving models with TrafficAlign-generated scenarios significantly reduced collision rates by 36.1% compared with the original models. A qualitative study using traffic datasets from six geographically diverse regions shows that TrafficAlign-generated scenarios exhibit strong alignment with corresponding traffic distributions in these regions.

[89] CMTFormer: Marrying Transformer with Hierarchical Information Interaction for RGB-Event Object Detection cs.CV | cs.AIPDF

Yu Li, Yuenan Hou, Yingmei Wei, Jiangming Chen, Yanming Guo

TL;DR: 本文提出了一种名为CMTFormer的跨模态信息交互Transformer，用于RGB-事件目标检测。该方法通过浅层到深层的信息交互方案，分层整合RGB图像和事件流信息，以解决模态异构性导致的噪声放大或冗余特征融合问题。

Details

Motivation: 事件相机具有高时间分辨率和高动态范围，能弥补传统RGB帧的不足，但现有多模态融合技术难以处理RGB帧与事件流之间的固有异质性，容易在跨模态融合中导致噪声放大或冗余特征整合。

Result: 在DSEC-Detection和PKU-DAVIS-SOD两个主流事件目标检测基准上进行了广泛实验，CMTFormer在单模态和多模态设置下均超越了现有检测方法，证明了其有效性。

Insight: 创新点在于设计了浅层对齐模块（SAM）、跨模态增强模块（CEM）和可学习的深度融合模块（LDFM）的分层交互方案，以及利用全局空间信息的空间先验模块，实现了高效稳定的多模态协作与自适应融合。

Abstract: Event cameras capture sparse brightness changes with high temporal resolution and high dynamic range, compensating for the deficiencies of the conventional RGB frames. However, previous multi-modal fusion techniques typically fail to handle the inherent heterogeneity between RGB frames and event streams, thus easily leading to noise amplification or redundant feature integration during cross-modal fusion. In this paper, we propose a Cross-Modal information inTeraction transFormer, coined as CMTFormer, which hierarchically integrates RGB and event information to achieve efficient and stable multimodal collaboration. Specifically, we design a shallow-to-deep information interaction scheme. In the shallow stage, we present the Shallow Alignment Module (SAM) to achieve an efficient fusion of RGB and event low-level features, which mitigates attribute disparities and prevents noisy information. In the middle stage, we devise the Cross-modal Enhancement Module (CEM) that utilizes texture and edge information to produce mutually reinforced middle-level features. In the deep stage, we present the Learnable Deep Fusion Module (LDFM) which performs high-level information aggregation through learnable weights, thus enabling the network to adaptively fuse RGB and event clues. A Spatial Prior Module is further designed to utilize global spatial information to enhance localization accuracy. Extensive experiments are conducted on two prevalent event-based object detection benchmarks, i.e., DSEC-Detection and PKU-DAVIS-SOD. Our CMTFormer consistently surpasses the detection counterparts in both uni-modal and multi-modal settings, strongly demonstrating the effectiveness of our paradigm. Codes will be available upon publication.

[90] Beyond Backscatter: AlphaEarth Land-Cover Priors for Rapid SAR Flood Segmentation Across Foundation Backbones cs.CVPDF

Sanjay Thasma, Yu-Hsuan Ho, Ali Mostafavi

TL;DR: 本研究探讨了在没有配准的季节匹配灾前SAR图像的情况下，如何利用稳定的土地覆盖先验信息（AlphaEarth）来改进单时相SAR图像的洪水快速分割。通过对比四种不同预训练机制的基础模型在多种配置下的表现，发现结合AlphaEarth或DEM先验均能提升分割性能，其中AlphaEarth在更具挑战性的飓风事件上表现更优，而DEM则更稳定。

Details

Motivation: 解决在重大洪水期间光学影像不可用，而单时相SAR图像因新生淹没区、永久水体及其他光滑表面产生相似后向散射导致分割模糊的问题，探索稳定的土地上下文先验信息能否提升洪水事件后SAR分割的可靠性。

Result: 在ImpactMesh-Flood的CONUS子集上，使用四种基础模型（从头训练的CNN UNet、ImageNet预训练的UNet、SAR预训练的TerraMind ViT、光学卫星预训练的DINOv3 ViT）进行评估。所有模型在添加AlphaEarth或DEM先验后，均优于仅使用SAR的基线。AlphaEarth在更难的飓风Florence事件上对所有模型均优于DEM，并取得了最佳IoU；而DEM在Louisiana洪水事件上更具竞争力。种子分析表明DEM更稳定，而AlphaEarth在困难事件上能提供更高的峰值性能和召回率。

Insight: 将单时相SAR洪水分割重新定义为雷达观测与稳定地表先验信息之间的对齐问题，其中学习到的上下文（如AlphaEarth）与物理上下文（如DEM）为更可靠的快速洪水制图提供了互补的途径。这表明结合领域特定的先验知识可以显著提升模型在复杂、模糊场景下的分割性能，尤其是在缺乏匹配灾前图像的情况下。

Abstract: Rapid flood mapping is critical for emergency response, yet optical imagery is often unusable during major flooding and single-temporal SAR is ambiguous, since new inundation, permanent water, and other smooth surfaces produce similar backscatter. This study evaluates whether stable land-context priors can improve post-event SAR flood segmentation when a registered, seasonally matched pre-event acquisition is unavailable. Using the CONUS (Continental United States) subset of ImpactMesh-Flood, we compare four backbones spanning distinct pretraining regimes-a from-scratch CNN UNet, an ImageNet-pretrained UNet, the SAR-pretrained TerraMind Vision Transformer, and the optical-satellite-pretrained DINOv3 Vision Transformer-in SAR-only, SAR+DEM, and SAR+AlphaEarth configurations under an identical fusion design, training protocol, and event-stratified split. Models are selected on a validation flood event and evaluated separately on two held-out events, Hurricane Florence and the Louisiana floods, with three-seed reporting for auxiliary configurations. Both auxiliary priors improve over the observed SAR-only baselines across all backbones and test events. AlphaEarth exceeds DEM on the harder Florence event for every backbone and achieves the best Florence IoU, while DEM is competitive on Louisiana and produces the best result there. The seed analysis reveals a trade-off: DEM is more stable across initializations, whereas AlphaEarth offers higher peak performance and higher recall on the harder event. Cross-event differences track flood-class prevalence and similarity to the training distribution, underscoring the need for per-event evaluation. We reframe single-temporal SAR flood segmentation as an alignment between radar observations and stable land-surface priors, where learned and physical context offer complementary pathways to more reliable rapid flood mapping.

[91] GPC: Large-Scale Generative Pretraining for Transferable Motor Control cs.CV | cs.AI | cs.GR | cs.LGPDF

Yi Shi, Yifeng Jiang, Chen Tessler, Xue Bin Peng

TL;DR: 本文提出了生成式预训练控制器（GPC）框架，通过令牌化和下一个令牌建模，从大规模运动数据中创建通用、可重用的生成式控制器。该框架利用端到端强化学习联合优化基于有限标量量化的运动词汇表及相应的控制策略，并训练GPT风格的自回归Transformer来建模词汇结构，从而生成物理模拟角色的控制信号。训练后，通过一系列适应技术对控制器进行微调，以应用于新的下游任务。

Details

Motivation: 开发能够以自然、逼真方式完成广泛任务的控制器，是实现基于物理的角色动画实际应用的关键挑战。

Result: 该框架相比之前的令牌化方法大大简化了训练过程，在重现大量运动片段上达到了99.98%的成功率。生成控制器展现出多种自然涌现行为，如对扰动的响应行为和跌倒后的恢复行为，从而为各种下游应用提供了高度鲁棒的通用控制器。

Insight: 创新点在于将大规模生成式预训练（如GPT）的思想引入物理角色控制，通过联合学习离散运动词汇和基于词汇的生成策略，构建了可迁移的通用控制器。其提出的适应技术简化了下游任务微调，实现了从大规模数据中学习鲁棒、多样化行为的能力。

Abstract: Developing controllers capable of completing a wide range of tasks in a natural and life-like manner is a key challenge in enabling practical applications of physics-based character animation. In this work, we introduce Generative Pretrained Controllers (GPC), which leverage tokenization and next-token modeling to create general-purpose, reusable generative controllers from large-scale motion datasets. Our framework utilizes end-to-end reinforcement learning to jointly optimize a “motion vocabulary”, modeled via Finite Scalar Quantization (FSQ), along with a corresponding control policy that can map the discrete codes to physics-based controls. After the “codebook” has been learned, the underlying structure of this large vocabulary is modeled by training a GPT-style autoregressive transformer, leading to a powerful generative controller that generates controls for a physically simulated character by performing next-token prediction. Once the generative controller has been trained, we propose a suite of adaptation techniques for finetuning the controller for new downstream tasks. Our proposed framework greatly simplifies the training process compared to previous tokenized methods, and achieves a 99.98% success rate in reproducing a vast corpus of motion clips. The generative controller exhibits a variety of natural emergent behaviors, such as responsive behaviors to perturbations and recovery behaviors after falling. This results in highly robust general purpose controllers for a variety of downstream applications.

[92] Spatially Localized Image Degradation Embeddings for Image Quality Assessment cs.CV | eess.IVPDF

Krishna Srikar Durbha, Hassene Tmar, Ping-Hao Wu, Ioannis Katsavounidis, Alan C. Bovik

TL;DR: 本文提出了一种名为SLIDE-IQA的自监督学习方法，用于无参考图像质量评估。该方法通过一个双分支视觉Transformer框架，在对比预训练目标中注入空间有界的图像退化，以解决现有方法对真实世界中空间局部和共现退化不敏感的问题。

Details

Motivation: 现有自监督学习方法在无参考图像质量评估中表现优异，但它们通常在整个图像上均匀应用合成失真，这限制了其对真实内容中空间局部和共现退化的敏感性。本文旨在弥补这一表征盲点。

Result: SLIDE-IQA仅使用合成数据进行预训练，显著提升了对局部失真的敏感性，同时在无参考图像质量评估基准测试中，与现有的自监督学习模型相比，取得了具有竞争力的性能。

Insight: 论文的核心创新点是引入了空间局部图像退化嵌入和双分支ViT框架，并提出了阈值有界排除机制，以解决空间局部失真带来的结构冲突，确保潜在空间同时尊重退化类型和空间尺度。这为处理复杂空间退化模式提供了一种新的表征设计思路。

Abstract: Self-supervised learning (SSL) currently drives state-of-the-art performance in no-reference image quality assessment (NR-IQA). However, standard SSL pipelines uniformly apply synthetic distortions across the entire image field, which can limit their sensitivity to spatially localized and co-occurring degradations encountered in real-world content. In this work, we empirically expose this representational blind spot across existing state-of-the-art encoders, demonstrating their reduced sensitivity to spatially bounded image degradations. To bridge this gap, we introduce Spatial Localized Image Degradation Embeddings for Image Quality Assessment (SLIDE-IQA). SLIDE-IQA employs a dual-branch Vision Transformer framework that injects spatially bounded degradations into a contrastive pretraining objective. To handle the spatial complexity of these degradations, we introduce a Threshold-Bounded Exclusion Mechanism, a representational design choice that resolves structural conflicts arising from spatially localized distortions to ensure the latent space respects both degradation type and spatial scale. Finally, we show that SLIDE-IQA’s synthetic-only pretraining significantly improves sensitivity to localized distortions, while achieving competitive performance on NR-IQA benchmarks against existing SSL NR-IQA models.

[93] DTI: Dynamic Trajectory Initialization for Generative Face Video Super-Resolution cs.CVPDF

Yingwei Tang, Chen Yan, Wendi Liu, Qiang Hu, Xiaoyun Zhang

TL;DR: 本文提出了一种名为动态轨迹初始化（DTI）的新范式，用于生成式人脸视频超分辨率（GFVSR）。该方法将GFVSR重新定义为输入驱动的方向性恢复，通过一种新颖的增强-注入条件机制和判别性引导（DG）来动态设置采样起点，在显著提升保真度的同时不牺牲感知质量。

Details

Motivation: 现有基于预训练扩散模型的生成式人脸视频超分辨率方法存在采样固定、推理成本高、过度追求通用感知指标导致保真度低等问题，本文旨在解决这些局限性。

Result: 该方法仅需少量模型适配和微调，便在多种指标和基准测试中实现了最先进的综合性能。分析表明，LPIPS是评估其性能最可靠的指标。

Insight: 核心创新在于将GFVSR重新定义为输入驱动的方向性恢复，并提出了增强-注入条件机制和基于信噪比对齐的判别性引导来动态初始化采样轨迹，从而在感知-失真权衡中取得更好平衡。

Abstract: As the most perceptually powerful Face Video Super-Resolution (FVSR) method, existing works in Generative FVSR (GFVSR) mainly exploit the generative prior of pretrained diffusion models. However, viewed as full generation, they suffer from fixed sampling and expensive inference costs if without large-scale auxiliary training. Furthermore, an excessive pursuit of generic perceptual metrics often results in low fidelity. To address these issues, we present Dynamic Trajectory Initialization (DTI) paradigm for GFVSR, which reformulates GFVSR as an input-driven directional restoration. With a novel enhancement-and-injection conditioning mechanism for pretrained DiT backbone, fidelity of our model has been significantly improved without compromising perceptual quality. To dynamically set the starting sampling point, we propose a Discriminative Guide (DG) trained via objective Signal-to-Noise Ratio (SNR) alignment. With only minor model adaptation and fine-tuning, our method achieves a SOTA overall performance across diverse metrics and benchmarks. An analysis of relationship between actual comprehensive quality and common metrics is also conducted, which demonstrates the perception-distortion trade-off and that the LPIPS is the most convincing metric in our case.

[94] Articulating then Matching: Zero-Shot Shape Matching for Uncurated Data cs.CVPDF

Qilong Liu, Qinfeng Xiao, Chenyuan Yi, Liying Zhang, Kit-lun Yick

TL;DR: 本文提出了一种名为ATM的零样本三维形状匹配框架，通过‘先参数化后匹配’的范式，无需针对对应关系进行专门训练，即可处理未经过整理的极端高分辨率数据。该方法利用预训练的视觉基础模型和参数化形状先验，从多视角渲染中估计参数化形状模型，并通过多视角几何一致性进行系统化验证，从而在共享的规范参数空间中建立鲁棒的粗对应关系，再通过谱细化得到精确的密集映射。

Details

Motivation: 解决在现实世界环境中，三维形状之间寻找密集对应关系的根本性挑战，这些挑战包括缺乏训练时间和足够样本、未经过整理的极端高分辨率数据普遍存在拓扑失真，以及需要处理多样化的三维表示形式。

Result: 在非等距基准测试（TOPKIDS和SMAL）上取得了强结果，平均测地误差分别为2.4和3.8，相比基线URSSM分别减少了73%和37%的误差；在野外原始扫描数据上表现出前所未有的鲁棒性，每个形状最多可达20万个顶点，同时保持近乎恒定的计算时间和一致的高精度。

Insight: 创新点在于提出了一种零样本框架，通过结合预训练视觉基础模型和参数化形状先验，将多样化输入映射到共享规范参数空间，从而绕过拓扑噪声并建立鲁棒对应；从客观角度看，该方法无需对应训练数据、天然免疫于连接性伪影，并能无缝处理多种三维模态（如网格、点云、3D高斯），具有很高的实用性和泛化能力。

Abstract: Finding dense correspondences between 3D shapes is a fundamental yet unresolved challenge, especially in real-world environments. These environments present severe challenges, including the lack of time and sufficient samples for training, the prevalence of uncurated extreme-high resolution data with topological distortions, and the need to handle diverse 3D representations. In this paper, we present ATM, a zero-shot framework that requires no correspondence-specific training and robustly addresses these issues at once through an articulate-then-match paradigm. Rather than relying on intrinsic geometric properties, we leverage powerful pretrained vision foundation models and parametric shape priors to estimate parametric shape models from multi-view renderings, and systematically ground these estimations via multi-view geometric consistency. By mapping diverse inputs into a shared canonical parametric space, we inherently establish robust coarse correspondences that bypass topological noise, which are then refined into precise dense mappings via spectral refinement. Operating purely on test-time optimized parametric reconstructions, ATM requires no correspondence training data, is naturally immune to connectivity artifacts, and seamlessly handles diverse 3D modalities, including meshes, point clouds, and 3D Gaussians. Extensive experiments demonstrate that our method achieves strong results on non-isometric benchmarks (average geodesic errors of 2.4-TOPKIDS, 3.8-SMAL), reducing errors by 73% and 37% respectively compared to the baseline URSSM. Furthermore, it exhibits unprecedented robustness on in-the-wild raw scans of up to 200k vertices per shape while maintaining near-constant computation time and consistent superior accuracy.

[95] Zero-Gated Language-conditioned Human Motion Prediction cs.CVPDF

Guanhui Qiao, Lu Zhou, Ding Jiang, Jinqiao Wang

TL;DR: 本文提出了一种轻量级的语言条件化人体运动预测模型ZGL，它利用观测动作的文本描述作为语义先验，同时保留强大的运动主干网络作为主要动态来源。该方法通过冻结的CLIP文本编码器将动作描述编码为少量条件令牌，并通过零门控的紧凑交叉注意力适配器注入到基于DCT的时空Transformer中，实现了在初始化时与纯姿态基线数值相同，仅在减少预测误差时学习使用语言信息。

Details

Motivation: 现有的人体运动预测方法主要依赖姿态历史作为运动学证据，但缺乏明确的高层语义指导，导致预测结果可能不符合语义合理性。

Result: 在Human3.6M数据集上，ZGL相比代表性基线模型降低了整体MPJPE；在CMUMocap上的结果进一步表明，紧凑的文本描述条件化能够迁移到第二个基准数据集，并为3D人体运动预测提供了实用的语义线索。

Insight: 创新点在于提出了零门控初始化策略，使模型在训练初期等同于纯姿态模型，从而稳定训练并让模型自适应地学习何时利用语言信息；同时采用轻量级的适配器注入语义条件，保持了主干网络的完整性并实现了高效的语义融合。

Abstract: Pose histories provide the core kinematic evidence for 3D human motion prediction, but they lack explicit high-level semantic guidance. This paper introduces ZGL, a lightweight language-conditioned predictor that uses captions of the observed motion as a semantic prior while preserving a strong motion backbone as the main source of dynamics. We render only the observed poses, generate a one-sentence description with a vision-language model, encode the caption with a frozen CLIP-L text tower, and project it into a small set of conditioning tokens. These tokens are injected into a DCT-based spatial-temporal Transformer by compact crossattention adapters with zero gates: each adapter output is multiplied by a learnable gate initialized to zero, so the full network is numerically identical to the pose-only baseline at initialization and can learn to use language only when it reduces prediction error. On Human3.6M, ZGL improves overall MPJPE over representative motion-prediction baselines in our comparison. Results on CMUMocap further show that compact caption conditioning transfers to a second benchmark and provides a practical semantic cue for 3D human motion prediction.

[96] When Does Synthetic CT Transfer? A Label-Free Donor/Host Diagnostic for Medical Vision-Language Model Routing on Real Lung CT cs.CVPDF

Fakrul Islam Tushar

TL;DR: 这篇论文探讨了在医学影像中，如何无标签地预测合成CT数据到真实CT数据的模型能力迁移。研究发现，模型能力迁移可分为供体驱动（移植结节属性）和宿主驱动（周围解剖结构属性），前者在合成到真实数据的转换中保持稳定，后者则不一定。作者在三个肺部CT视觉语言任务、五个VLMs、四种指导条件和七个真实数据集上验证了这一预测，并通过TrialCouncil方法实现了与最佳固定模型匹配的无训练路由。

Details

Motivation: 医学影像缺乏真实标签来验证模型能力，因此需要一种无标签的方法来预测合成CT数据到真实数据的模型迁移效果，以解决模型在真实场景中的适用性问题。

Result: 在实验中，供体驱动的能力迁移（如结节存在性和大小排序）在合成到真实数据转换中保持高相关性（R2 >= 0.96），而宿主驱动的能力（如肺叶位置）则未迁移。TrialCouncil方法在预测迁移的区域中与最佳固定模型完全匹配，验证了预测的有效性。

Insight: 创新点在于发现模型能力迁移本身可以从合成数据中无标签地预测，这为医学影像中的模型路由提供了新机制；客观来看，研究通过区分供体与宿主驱动能力，为合成数据在真实场景的适用性评估提供了理论框架，可借鉴于其他医学视觉任务。

Abstract: A synthetic measurement of model competence is useful only if it survives the move to real data, yet the real labels that would verify it are exactly what medical imaging lacks. We ask whether transfer can be predicted in advance, label-free, and answer with a mechanism: on synthetic digital twins, competence that is donor-driven (a property of the transplanted nodule) survives the synthetic to real change of host, while host-driven competence (a property of the surrounding anatomy) need not. We test this on three lung CT vision-language tasks chosen to span that axis, across five public VLMs, four guidance conditions, and seven real datasets. The prediction holds in every case: presence and size orderings transfer (R2 >= 0.96), lobe does not; the split survives leave-source-out calibration, and the diagnostic names that boundary before any real label. TrialCouncil, a training-free council calibrated only on synthetic CT, confirms it by matching the best fixed model exactly where transfer is predicted. The contribution is not the router but the finding that transfer itself is predictable, label-free, from synthetic data alone.

[97] Again-Pose: Anchor-Guided Adaptive Inter-Frame Motion Cues Propagating for High-quality Human Pose Reconstruction cs.CVPDF

Shuaikang Zhu, Yiding Sun, Yang Yang

TL;DR: 本文提出了一种名为Again-Pose的框架，用于从无约束视频中重建高质量的3D人体姿态，特别是在存在严重运动模糊和遮挡的极端运动场景下。该方法通过显式地识别高质量锚定帧，并传播可靠的运动学线索来“修复”退化中间帧的姿态，从而解决了现有方法在视觉退化严重时特征崩溃和重建错误的问题。

Details

Motivation: 当前最先进的方法通常依赖隐式时间注意力来聚合跨帧特征，但在严重视觉退化（如运动模糊、遮挡）下，输入特征会崩溃并与噪声难以区分，导致隐式聚合失效和灾难性重建错误。本文旨在解决这一鲁棒性差距。

Result: 在标准基准测试（Human3.6M、3DPW、PoseTrack）和具有挑战性的FineDiving数据集上的大量实验表明，Again-Pose在鲁棒性和稳定性方面显著优于最先进的方法，能在其他方法失败时有效恢复合理的姿态。

Insight: 创新点在于将退化帧的姿态估计重新定义为运动引导的恢复任务，而非盲目平滑特征。具体包括：基于特征显著性显式识别高质量锚定帧；通过双路径运动感知模块捕获细粒度帧间动态；以及通过差异加权融合模块自适应传播运动线索以抑制漂移。这是一种简单而有效的显式运动传播策略。

Abstract: Reconstructing continuous 3D human poses from unconstrained videos is challenging, especially in extreme motion scenarios involving severe motion blur and occlusion. Current state-of-the-art methods typically rely on implicit temporal attention to aggregate features across frames. However, under severe visual degradation, input features often suffer from collapse, rendering them indistinguishable from noise. In such cases, implicit aggregation fails to distinguish valid signals, leading to catastrophic reconstruction errors. To address this robustness gap, we propose a simple yet effective framework called Anchor-guided adaptive inter-frame motion cues propagating (Again-Pose), reformulating pose estimation in degraded frames as a motion-guided recovery task. Instead of blindly smoothing features, we explicitly identify high-quality Anchor Frames based on feature saliency and propagate reliable kinematic cues to “inpaint” the poses of degraded intermediate frames. Specifically, a Dual-path Motion-aware Module captures fine-grained inter-frame dynamics, while a Difference-weighted Fusion Module adaptively propagates these cues to suppress drift. Extensive experiments on standard benchmarks (Human3.6M, 3DPW, PoseTrack) and the challenging FineDiving dataset demonstrate that Again-Pose significantly outperforms state-of-the-art methods in robustness and stability, effectively recovering plausible poses where other methods fail.

[98] Confidence-feedback-weighted graph matching network: online-offline laser-induced damage site matching under complex interference cs.CV | cs.AIPDF

Yueyue Han, Guanhua Chen, Hangcheng Dong, Kang Zhang, Fengdong Chen

TL;DR: 本文提出了一种置信度反馈加权的图匹配网络，用于高功率激光装置最终光学元件在线检测图像与离线真实损伤点之间的匹配。该方法仅需损伤点质心坐标作为输入，通过多轮匹配置信度反馈抑制干扰点传播，并引入几何一致性约束和难例挖掘损失提升匹配性能。在复杂场景数据集上的实验表明，该方法实现了96.36%的F1分数，具有鲁棒高效的性能。

Details

Motivation: 解决高功率激光装置在线检测图像中伪损伤点与真实损伤点难以区分的问题，现有方法主要通过损失函数隐式抑制干扰点，但面临特征判别性有限、局部几何畸变和大量干扰点的挑战。

Result: 在自建的Complex-Scene数据集上，所提方法达到了96.36%的匹配F1分数，表现出鲁棒且高效的性能。

Insight: 创新点在于通过置信度反馈机制显式抑制干扰点传播，结合几何一致性校准和难例挖掘损失增强跨图判别性；客观来看，该方法将匹配置信度动态反馈至特征聚合过程，为图匹配任务提供了一种可解释的干扰抑制策略。

Abstract: Online inspection images of final optics in high-power laser facilities contain pseudo-damage sites that closely resemble true damage sites. Determining the authenticity of online-detected sites is therefore difficult and requires accurate matching to offline ground-truth sites. However, this matching remains highly challenging due to limited match-discriminative features, local geometric distortions, and numerous distractor sites. Existing matching models mainly suppress distractors implicitly through loss-function supervision. We propose a confidence-feedback-weighted graph matching network that requires only damage-site centroid coordinates as input. It estimates node matchability confidence from each round of matching scores and feeds it back as a reliability weight to guide subsequent edge-feature aggregation, thereby suppressing distractor propagation and enhancing cross-graph discriminability. Within this framework, a geometric consistency constraint calibrates spurious high-confidence matchability estimates, while a hard-example mining loss improves discrimination between structurally similar sites. Experiments on our Complex-Scene dataset show that the proposed method achieves a matching F1-score of 96.36$%$ with robust and efficient performance.

[99] Enhancing Part-Level Point Grounding for Any Open-Source MLLMs cs.CVPDF

Jin-Cheng Jhang, Fu-En Wang, Xin Yang, Nan Qiao, Lu Xia

TL;DR: 本文提出了一种通用方法，使任何开源多模态大语言模型（MLLM）具备精确的2D部件级点定位能力，为细粒度视觉任务提供更直接的定位表示。该方法通过提出的Q-Synth模块在模型中间层合成文本条件化、定位感知的查询，利用MLLM固有的注意力机制捕获目标相关模式，并通过轻量级Attention-to-Point解码器将其转换为点中心热图进行最终预测，且无需微调原始模型参数。

Details

Motivation: 现有MLLMs在视觉定位任务中主要擅长物体级定位，但在部件级定位（如机器人操作等细粒度任务所需）上表现不佳，本文旨在解决这一局限性。

Result: 实验表明，该方法在不同数据集上持续提升了部件级定位的准确性，并可无缝集成到任何开源MLLMs中，验证了其通用性和有效性。

Insight: 创新点在于提出了一种冻结原始MLLM参数的通用增强框架，通过合成中间查询和注意力到点的解码，将MLLMs的注意力机制直接用于细粒度点定位，为开放模型赋予了新的细粒度视觉理解能力。

Abstract: Visual grounding aims to associate free-form textual queries with specific regions in an image. While recent Multimodal Large Language Models (MLLMs) have demonstrated promising capabilities in this domain, they primarily excel at object-level grounding and often struggle with part-level grounding-an essential requirement for fine-grained tasks such as robotic manipulation. In this work, we introduce a general approach that equips any open-source MLLMs with accurate 2D part-level point grounding, offering a more direct alternative to conventional grounding representations. Our method leverages the attention mechanisms inherently present in MLLMs. By synthesizing text-conditioned, grounding-aware queries within intermediate layers via the proposed Q-Synth Module, we capture target-relevant attention patterns and refine them with a lightweight Attention-to-Point Decoder, which converts these patterns into a point-centric heatmap for final prediction. Notably, all original MLLM parameters are frozen, ensuring full preservation of their pre-trained capabilities. Experiments show that our design consistently improves part-level grounding accuracy across datasets and can be seamlessly integrated into any open-source MLLMs.

[100] Pointer-CAD v2: Plan-Then-Construct CAD Generation with Dimension-Aware Parametric Precision cs.CVPDF

Dacheng Qi, Chenyu Wang, Jingwei Xu, Yi Ma, Shenghua Gao

TL;DR: 本文提出了Pointer-CAD v2，一个用于计算机辅助设计（CAD）生成的统一框架。它采用“先规划后构建”的范式，将参数推理与几何构造解耦，首先生成包含明确度量尺度参数的结构化设计规划，然后通过指针机制在序列生成中直接引用这些连续值参数，从而避免了量化误差。该方法在几何精度上显著优于现有基线，适用于对精度要求苛刻的工程应用。

Details

Motivation: 现有基于大语言模型的CAD生成方法主要关注视觉相似性，忽略了精确的几何参数和正确的度量尺度。它们通常将数值参数进行激进的量化以方便预测，这导致了微小的数值偏差，可能违反工业公差要求。本文旨在解决这一问题，实现高精度的CAD生成。

Result: 大量实验表明，Pointer-CAD v2在几何精度上持续优于现有基线，并取得了实质性改进。评估使用了新引入的三个分层几何精度指标（顶点、边、面级别），证明了其在参数保真度方面的优越性。

Insight: 核心创新点是“先规划后构建”的范式以及指针机制的应用。该方法将设计规划（包含连续参数）与构造过程解耦，通过指针直接引用规划中的精确参数，从根本上消除了离散化/量化误差，确保了维度一致的执行。同时，构建了带有规划级标注的新大规模数据集和分层评估指标，以更好地衡量参数精度。

Abstract: Computer-aided design (CAD) plays a fundamental role in modern manufacturing by providing the high precision required for industrial production. Recent large language model based approaches formulate CAD generation as a sequence prediction problem and have achieved promising results. However, existing methods and evaluation protocols primarily emphasize visual similarity, while overlooking precise geometric parameters and correct metric scale. Small numerical deviations that are negligible at the shape-level may still violate industrial tolerance requirements, a problem further compounded by current autoregressive paradigms that utilize command sequence representations, aggressively quantize numerical parameters to ease LLM prediction. In this work, we present Pointer-CAD v2. Compared with v1 (arXiv:2603.04337), this version directly predicts continuous values, bypassing the need for quantized numerical parameters and thereby eliminating quantization errors. Specifically, we propose a unified framework that decouples parameter reasoning from geometric construction through a Plan-Then-Construct paradigm. Our method first produces a structured design plan with explicit metric scale parameters. These parameters are organized into a dictionary and directly referenced during sequence generation via a pointer mechanism, eliminating discretization errors and ensuring dimensionally consistent execution. In addition, we construct a new large-scale dataset with plan-level annotation and introduce three hierarchical geometry accuracy metrics to evaluate parametric fidelity at the vertex, edge, and face levels. Extensive experiments demonstrate that Pointer-CAD v2 consistently outperforms existing baselines and achieves substantial improvements in geometric accuracy, enabling reliable CAD generation for precision-critical engineering applications.

[101] Occlusion-Robust Multi-Object Decoupling for Physics-Based Interaction cs.CVPDF

Xin Dong, Wenfeng Deng, Yansong Tang

TL;DR: 本文提出了一种无需掩码的、从稀疏且遮挡的真实世界视图中实现无损多物体3D重建的方法，该方法通过基于物理的Material Point Method (MPM)模拟实现物理上合理的交互。核心思路是将多物体解耦问题视为稀疏视图重建问题，利用3D高斯泼溅作为基础表示，结合SAM2训练的分割场获得粗略实例分割，并通过联合Score Distillation Sampling (SDS)过程整合2D/3D扩散先验来重建碎片化几何，同时引入几何感知先验进行正则化。

Details

Motivation: 解决从稀疏、遮挡的真实世界视图中进行多物体3D重建时，由于遮挡和视角有限导致的物体耦合问题，旨在实现无需手动掩码即可获得完整、可用于物理模拟的3D物体。

Result: 实验结果表明，该方法在合成和真实世界数据集上都能生成完整、可用于模拟的3D物体，并实现逼真的动态交互。

Insight: 创新点在于将多物体解耦形式化为稀疏视图重建问题，并提出了一个不依赖掩码的联合SDS框架，整合了参考视图监督、新颖视图合成以及2D/3D扩散先验，同时利用几何感知先验（如物体内和物体间相似性）来规范几何推理，从而实现了遮挡鲁棒的无损重建。

Abstract: We propose a mask-free method for lossless multi-object 3D reconstruction from sparse and occluded real-world views, enabling physically plausible interaction via Material Point Method (MPM) simulation. Our key insight is that object coupling stems from occlusion and limited viewpoints, which we address by formulating multi-object decoupling as a sparse-view reconstruction problem. Using 3D Gaussian Splatting as base representation, we first obtain coarse instance partitions with a SAM2-trained segmentation field. Rather than relying on masks, we reconstruct fragmented geometries by leveraging a joint Score Distillation Sampling (SDS) process, which integrates reference-view supervision with novel-view synthesis guided by 2D and 3D diffusion priors to enforce both texture fidelity and 3D consistency. Furthermore, we incorporate geometry-aware priors such as intra-object and inter-object similarity to regularize geometric reasoning. Experimental results demonstrate that our method produces complete, simulation-ready 3D objects without requiring manual masks, enabling realistic dynamic interactions on both synthetic and real-world datasets.

[102] Multi-scale Object-Aware Gaze Estimation via Geometric Reasoning cs.CVPDF

Jiajie Mi, Xinyu Liu, Mengke Song, Chenglizhao Chen

TL;DR: 该论文提出了一种基于对象语义引导的两阶段注视目标估计框架，将注视目标估计重新定义为层次化推理过程。该方法在特征编码中融入对象级表示以对齐图像特征与离散语义实体，并引入多尺度特征融合以及来自头部姿态和注视方向的几何约束，以实现细粒度定位和对象级判别。

Details

Motivation: 现有方法通常将注视目标估计建模为从全局特征到注视热图的直接映射，本质上将其视为像素级回归问题，未能将注视对象明确表示为独立实体，导致在复杂场景中难以产生稳定且语义一致的预测。

Result: 在GazeFollow、VideoAttentionTarget、ChildPlay和GOO-Real数据集上的大量实验表明，该方法分别实现了0.961、0.948、0.987和0.977的AUC，在所有基准测试中均表现出色，同时保持了7.1M的紧凑参数量。

Insight: 创新点在于通过对象级表示和层次化推理过程，将注视目标估计从像素级回归问题转化为对象感知任务，并结合多尺度特征融合与几何约束来提升语义一致性和定位精度。

Abstract: Gaze target estimation aims to predict the semantic object an observer fixates upon within an image, a task deeply rooted in the object-oriented nature of human gaze. Observers tend to select a specific semantic entity as the attentional target, rather than responding randomly across arbitrary regions of the image. However, existing methods typically model this task as a direct mapping from global features to gaze heatmaps, essentially treating it as a pixel-level regression problem. This approach fails to explicitly represent the gazed object as a distinct entity, making it difficult to produce stable and semantically consistent predictions in complex scenes. To address this, we propose a two-stage gaze estimation framework guided by object semantics, reformulating gaze target estimation as a hierarchical reasoning process. Our method incorporates object-level representations during feature encoding to align image features with discrete semantic entities, then introduces multi-scale feature fusion and geometric constraints from head pose and gaze direction for fine-grained localization and object-level discrimination. Extensive experiments on GazeFollow, VideoAttentionTarget, ChildPlay, and GOO-Real demonstrate that our method achieves AUC of 0.961, 0.948, 0.987, and 0.977 respectively, delivering strong performance across all benchmarks while maintaining a compact parameter size of 7.1M.

[103] HiReFF: High-Resolution Feedforward Human Reconstruction from Uncalibrated Sparse-View Video cs.CVPDF

Yiming Jiang, Hanzhang Tu, Wenfeng Song, Siyou Lin, Liang An

TL;DR: HiReFF是一种前馈方法，用于从无标定的稀疏视角视频（如四个相隔90度的视图）中重建2K分辨率的360度人体视频。该方法将问题分解为两个关键任务：从稀疏视频中进行前景3D高斯重建，以及计算高效的高分辨率合成。通过提出的尺度同步相机标定和高斯前景掩码技术解决了稀疏视图下的重建问题，并通过高分辨率侧调优技术实现了2K渲染，同时大幅降低了计算开销。

Details

Motivation: 解决无标定稀疏视角视频流式人体重建中的时间一致性和计算效率挑战，现有方法依赖逐场景优化或标定相机，而前馈模型仅限于低分辨率单帧合成。

Result: 实验表明，HiReFF在高分辨率流式体积视频重建方面显著优于现有方法，实现了2K分辨率的重建。

Insight: 创新点包括尺度同步相机标定解决尺度歧义、高斯前景掩码调制高斯参数以重建干净前景，以及高分辨率侧调优通过补充特征增强高斯头部实现高效2K渲染，同时保持主干网络在0.5K以降低计算成本。

Abstract: Uncalibrated volumetric video streaming for human reconstruction is essential for holographic communication and AR/VR, yet remains challenging due to the need for temporal consistency and computational efficiency from sparse-view inputs. Existing methods rely on per-scene optimization or calibrated cameras, while recent feed-forward models are limited to low-resolution (0.5K) single-frame synthesis. We present HiReFF, a feed-forward method for 2K-resolution 360° human video reconstruction from uncalibrated sparse-view videos. Our framework decomposes the problem into two key tasks: foreground 3D Gaussian reconstruction from sparse-view videos (four views separated by 90°) and computationally efficient high-resolution synthesis. To enable the former, we propose Scale-synchronized Camera Calibration to resolve scale ambiguity for multi-view supervision, and Gaussian-wise Foreground Masking to reconstruct clean foregrounds by modulating Gaussian parameters. For efficient high-resolution synthesis, our High-resolution Side-tuning achieves 2K rendering by augmenting the Gaussian head with supplementary features while keeping the backbone at 0.5K, drastically reducing computational overhead. Experiments demonstrate that HiReFF significantly outperforms existing methods in high-resolution streaming volumetric video reconstruction. https://iridescentjiang.github.io/HiReFF

[104] Fast Enough to Act: Spatio-Temporal Visual Token Merging for Low-Latency Robotic VLMs and VLAs cs.CV | cs.AIPDF

Junzhou Chen, Jindong Wang, Gang Zhou

TL;DR: 本文提出了一种名为ST-Merge的即插即用、无需训练的框架，旨在解决视觉语言模型（VLMs）和视觉语言动作模型（VLAs）在处理视频和高分辨率图像时因视觉令牌数量庞大而导致的高推理延迟问题。该方法通过在视觉编码阶段直接融合冗余令牌，显著提升了模型推理速度，同时保持了较高的任务精度。

Details

Motivation: 动机是解决机器人应用中视觉语言模型因处理视频和高分辨率图像产生海量视觉令牌而导致的极高推理延迟问题，这严重阻碍了机器人的实时控制能力。

Result: 在主流VLM模型Qwen2.5-VL的视频问答任务中，ST-Merge实现了2倍的推理加速，精度损失仅为1%。在π_{0.5} VLA策略上部署时，在1024×1024分辨率下实现了8.3倍加速，并在此高分辨率设置下达到了基线成功率；在较低分辨率下，精度有小幅下降。

Insight: 创新点在于提出了一种基于3D时空坐标构建的多队列并行匹配与加权聚合机制，以实现跨帧冗余令牌的高效且几何一致的融合，并引入了合并后位置校正机制，通过动态重新评估视觉令牌加权质心的旋转位置编码来消除合并导致的空间偏差，从而确保灵巧操作所需的高精度空间感知能力。该方法无需训练，具有即插即用特性。

Abstract: Vision-language models and vision-language action models endow the robot with unprecedented capabilities. However, the input of video and high-resolution images yields a massive number of visual tokens, leading to extremely high inference latency and severely hindering the robot’s real-time control. To break through this computational bottleneck, we propose ST-Merge, a plug-and-play, training-free framework that efficiently fuses redundant tokens directly during the visual encoding phase. By explicitly constructing 3D spatiotemporal coordinates, it employs a multi-queue parallel matching and weighted aggregation mechanism to achieve efficient and geometrically consistent fusion of redundant tokens across frames. In addition, we introduce a post-merge positional correction mechanism that effectively eliminates spatial deviation caused by merging by dynamically re-evaluating the rotational position code of the weighted centroid of the vision token, thereby ensuring the high-precision spatial awareness required for dexterous operation. In the Video Question Answering task on the mainstream VLM, Qwen2.5-VL, ST-Merge achieves a 2$\times$ inference speedup with only a tiny 1% loss in precision. When deployed on the $π_{0.5}$ VLA policy, ST-Merge achieves an 8.3$\times$ speedup at 1024 $\times$ 1024 resolution and matches the baseline success rate at this high-resolution setting. At lower resolutions, it introduces a small drop in accuracy.

[105] Dynamic Parsing and Updating Natural Language Specification using VLMs for Robust Vision-Language Tracking cs.CV | cs.AI | cs.LGPDF

Xiao Wang, Liye Jin, Dan Xu, Yuehang Li, Lan Chen

TL;DR: 本文提出了一种基于视觉语言模型（VLM）的动态解析与更新自然语言规范的鲁棒视觉语言跟踪方法。该方法通过新颖的语言依赖解析机制，从语言描述中精确提取目标对象、语义概念和背景上下文等核心跟踪要素，并利用预训练VLM Qwen-VL的强大跨模态理解能力进行组件感知的自适应文本描述更新，以缓解跟踪过程中的语义-视觉失配问题。

Details

Motivation: 现有基于序列模型或大语言模型直接生成文本信息的主流方法，在视觉语言跟踪中存在错误目标更新、过度背景干扰和普遍幻觉伪影等固有缺陷，导致语义-视觉失配。本文旨在解决这些问题，提升跟踪的准确性和鲁棒性。

Result: 该方法在多个大规模视觉语言跟踪基准测试（包括TNL2K、LaSOT、TNLLT和OTB-LANG）上取得了一致且优越的跟踪性能。

Insight: 创新点在于提出了一个语言依赖解析机制来精确提炼跟踪核心组件，并利用预训练VLM进行组件感知的自适应更新，这比直接生成文本更精准，能有效减少错误和幻觉，增强对动态外观变化的鲁棒性。

Abstract: Vision-language tracking guided by natural language specifications leverages high-level semantic cues of target objects to substantially boost tracking accuracy and robustness. Existing studies have verified that adaptively optimizing textual descriptions throughout the tracking process can effectively mitigate the semantic-visual mismatch induced by dynamic variations in target appearance, position, and other inherent attributes. Nevertheless, mainstream methods that directly generate textual information via sequence models or large language models inevitably suffer from inherent defects, including erroneous target updating, excessive background distraction, and pervasive hallucination artifacts. To address the aforementioned limitations, this paper proposes a novel language dependency parsing mechanism to precisely distill core tracking principal components, encompassing target objects, semantic concepts, and background contextual information. On this basis, we perform component-aware adaptive textual description updates by exploiting the powerful cross-modal understanding capability of the pre-trained vision-language model Qwen-VL. By integrating the proposed elaborately designed modules into the baseline framework, our method achieves consistent and superior tracking performance on multiple large-scale vision-language tracking benchmarks, including TNL2K, LaSOT, TNLLT, and OTB-LANG. The source code and pre-trained models will be released at https://github.com/Event-AHU/Open_VLTrack.

[106] L2D2-GS: Learning to Densify for Feedforward Dynamic Gaussian Scene Reconstruction cs.CV | cs.GRPDF

Zetian Song, Chenming Wu, Junnan Liu, Chitian Sun, Liangliang He

TL;DR: L2D2-GS是一个用于动态高斯场景重建的统一前馈框架，它将泛化重建重新定义为优化和致密化的鲁棒迭代过程。该方法通过自监督的致密化策略和几何正则化机制，解决了现有方法在高分辨率下内存消耗大、难以融合多视角观测的问题。

Details

Motivation: 动态城市环境的高保真重建是自动驾驶仿真和大规模世界建模的基石。现有基于3D高斯泼溅（3DGS）的方法依赖昂贵的逐场景优化，可扩展性有限；而前馈推理方法速度快，但在高分辨率下内存消耗巨大，且难以一致地融合密集多视角观测。

Result: 在PandaSet和Waymo数据集上的大量实验表明，该方法在重建保真度上达到了最先进水平（SOTA），并具有很强的零样本泛化能力，同时使用的基元数量少于对比基线。

Insight: 主要创新点在于将泛化重建重新定义为迭代优化和致密化过程，并提出了一个从全局重建增益中获取显式奖励信号以指导局部致密化的自监督致密化策略。此外，通过利用重参数化来约束优化流形、防止收敛到不良局部最优的几何正则化机制，缓解了不可逆的早期伪影。

Abstract: High-fidelity reconstruction of dynamic urban environments is a cornerstone of autonomous driving simulation and large-scale world modeling. While 3D Gaussian Splatting (3DGS) has established a new standard for real-time rendering, its reliance on expensive per-scene optimization limits scalability. Conversely, recent feedforward methods that infer Gaussian parameters offer faster speed but face fundamental bottlenecks: they are memory-prohibitive at high resolutions and struggle to fuse dense multi-view observations consistently. This paper presents L2D2-GS, a unified framework that reformulates generalizable reconstruction not as a one-shot regression, but as a robust iterative process of optimization and densification. To resolve the ambiguity of supervision in primitive generation, we propose a self-supervised densification policy that derives explicit reward signals from global reconstruction gains to guide local densification. Furthermore, we mitigate irreversible early-stage artifacts through a geometric regularization mechanism, utilizing reparameterization to constrain the optimization manifold and prevent convergence to poor local optima. Extensive experiments on the PandaSet and Waymo datasets demonstrate that our method achieves state-of-the-art reconstruction fidelity and strong zero-shot generalization, while using fewer primitives than competing baselines.

[107] SAD-GS: Learning Reliable 3D Semantic Gaussian Fields via Dynamic Geo-Semantic Anchoring cs.CVPDF

Yufei Zhang, Chenlu Zhan, Gaoang Wang, Hongwei Wang

TL;DR: 论文提出了SAD-GS框架，旨在解决开放词汇3D语义高斯场学习因依赖不可靠的多视角2D语义监督（如语义漂移、边界泄漏）而导致的误差累积问题。该方法通过动态地理语义锚定，包括语义锚点蒸馏（SAD）和地理语义反馈循环（GSFL），来建立视角不变的语义一致性并主动过滤异常，从而学习更可靠的3D语义表示。

Details

Motivation: 开放词汇3D语义高斯场学习依赖于多视角2D监督，但这些监督的语义目标和空间分配往往不可靠（如视角依赖特征导致的语义漂移、跟踪器掩码传播引起的边界泄漏和身份切换），直接优化会迫使3D表示吸收多视角矛盾，导致严重的误差累积。

Result: 在LERF-OVS、3D-OVS和Mip-NeRF360等基准上的广泛评估表明，SAD-GS在开放词汇定位和语义分割任务中始终实现了最佳的整体性能，验证了其有效性和鲁棒性。

Insight: 创新点在于提出了动态地理语义锚定机制：语义锚点蒸馏（SAD）将每视角视觉嵌入提炼为共识文本锚点以建立视角不变的语义身份；地理语义反馈循环（GSFL）利用演化的3D场通过保守的三门更新规则主动过滤跟踪器异常并优化空间掩码分配。这为解决多视角语义不一致提供了系统性的优化策略。

Abstract: Open-vocabulary 3D semantic Gaussian field learning relies on multi-view 2D supervision, whose semantic targets and spatial assignments are often unreliable. Across varying viewpoints, view-dependent features cause semantic identity drift, while propagated tracker masks introduce boundary leakage and identity switches. Directly optimizing against these unreliable 2D targets forces the 3D representation to absorb multi-view contradictions, leading to severe error accumulation. To resolve this limitation, we propose SAD-GS, a framework for learning reliable 3D semantic Gaussian fields via dynamic geo-semantic anchoring. Specifically, Semantic Anchor Distillation (SAD) distills per-view visual embeddings into consensus text anchors to establish a viewpoint-invariant semantic identity. Concurrently, the Geo-Semantic Feedback Loop (GSFL) leverages the evolving 3D field to actively filter tracker anomalies and refine spatial mask assignments via a conservative three-gate update rule. Extensive evaluations on LERF-OVS, 3D-OVS, and Mip-NeRF360 show that SAD-GS consistently achieves the best overall performance in both open-vocabulary localization and semantic segmentation. These comprehensive improvements validate the effectiveness and robustness of dynamic geo-semantic anchoring for reliable 3D semantic Gaussian field learning.

[108] Event-VLA: Action-Conditioned Event Fusion for Robust Vision-Language-Action Model cs.CV | cs.ROPDF

Jiaxin Liu, Xun Xu, Zhenhao Zhang, Hanqing Wang, Ruiqi Chen

TL;DR: 本文提出了Event-VLA，一种增强型视觉-语言-动作模型，通过融合事件相机流来提高机器人在不同光照条件下的操作鲁棒性。它采用动作查询路由机制，将事件信息选择性地注入到VLA推理过程中，从而在保持预训练RGB-语言语义先验的同时，有效利用事件流进行稳健的动作预测。

Details

Motivation: 现有VLA模型通常假设光照良好且稳定的室内环境，而真实世界机器人操作常面临光照变化导致的RGB观测质量下降问题，这给鲁棒操作带来了关键挑战。

Result: 在仿真和真实世界部署的实验中，Event-VLA在正常光照下保持强劲的操作性能，并在低光照退化及近黑暗的真实世界设置中提高了任务成功率。

Insight: 创新点在于将事件流作为光照鲁棒、运动敏感的互补观测引入VLA框架，并设计了基于动作查询路由的融合机制，而非直接将事件特征合并到全局语义令牌空间，这有助于保留预训练先验并有效聚合任务相关事件信息。

Abstract: Vision-Language-Action (VLA) models have become an important paradigm of embodied AI. However, existing VLA models typically assume well-lit and stable indoor settings, while real-world embodied manipulation may involve degraded RGB observations caused by illumination shifts, posing critical challenges for robust robotic manipulation. To address this gap, we propose \textbf{Event-VLA}, an event-enhanced VLA framework for generalizable manipulation across varying illumination conditions. We formulate VLA-based manipulation under degraded visibility as a practical robustness problem for RGB-centric policies, and introduce event streams as an illumination-robust, motion-sensitive complementary observation to improve robustness across visibility levels. Specifically, unlike conventional multimodal fusion that directly merges event features into the global semantic token space, Event-VLA injects event information through an action-query routing pathway. It uses learnable action queries to extract task-relevant semantics from the VLA reasoning process, and selectively aggregates event tokens via gated cross-attention to construct event-aware action representations. This design preserves the pretrained RGB-language semantic priors while effectively leveraging event information for robust action prediction. Experiments in simulation and real-world deployment show that Event-VLA maintains strong manipulation performance under normal lighting and improves success rates under low-light degradation and near-dark real-world settings.

[109] NaLA: A 3D Native LLM Layout Agent for High-quality 3D Scene Generation cs.CVPDF

Cheng Wan, Yongsen Mao, Wenzheng Wu, Yuxuan Xie, Chucheng Xiang

TL;DR: 本文提出NaLA，一种用于高质量3D场景生成的原生3D大语言模型布局智能体。该方法直接将3D场景边界和3D资产编码到LLM中，避免了现有方法将3D信息转换为文本描述导致的信息损失问题。NaLA采用从粗到细的预测机制，先自回归预测离散位姿，再通过连续回归进行细化，从而在生成质量和推理效率上均优于现有布局智能体。

Details

Motivation: 现有基于LLM的3D场景布局智能体通常将3D资产和布局转换为文本描述进行输入输出，这导致文本模态与3D资产/布局之间存在模态鸿沟，造成严重信息损失，从而生成不合理的布局。

Result: 实验表明，NaLA在生成质量和推理效率上均优于先前的布局智能体。通过全面的消融研究验证了每个组件的有效性。

Insight: 创新点在于提出了一种原生3D LLM布局智能体，直接编码3D几何信息，保留了细粒度几何特征，并支持对碰撞、表面支撑、包含等关系的显式推理。其从粗到细（离散自回归预测后接连续回归细化）的位姿预测机制也是一个关键设计。

Abstract: Recently, Large Language Models (LLMs) have emerged as promising layout agents for 3D scene generation. Existing layout agents still suffer from implausible layout generation because most of them convert 3D assets and 3D layouts into textual descriptions as inputs and outputs, which involves severe information loss due to the modality gap between texts and 3D assets and 3D layouts. We propose NaLA, a native 3D LLM layout Agent for high-quality 3D scene generation by placing 3D assets in the scene. For the inputs, NaLA encodes 3D scene boundaries and 3D assets directly into the LLM, preserving fine-grained geometry and enabling explicit reasoning over relationships like collisions, surface supporting, and containment. To accurately output the positions and orientations of assets, NaLA adopts a coarse-to-fine prediction mechanism that first predicts discrete poses in an autoregressive manner and then refines the discrete poses with a continuous regression. Trained on diverse layout datasets, NaLA attains strong geometric perception and layout coherence. Experiments demonstrate that NaLA outperforms prior layout agents in both generation quality and inference efficiency, with comprehensive ablation studies to verify each component’s effectiveness.

[110] Bit-ViP: Leveraging Bit-planes to Preserve Visual Privacy in Images through Obfuscation cs.CV | cs.CR | cs.ETPDF

Vishesh Kumar Tanwar, Ashish Gupta, Sanjay Madria, Sajal K. Das

TL;DR: 本文提出了一种名为Bit-ViP的新型图像混淆方案，旨在保护视觉隐私同时保持图像可用性。该方法基于位平面技术，通过结合Lorenz混沌系统和差分隐私生成不可逆噪声，防止恶意攻击者从混淆图像中重建原始图像。实验在UCF101和HMDB51两个活动识别数据集上进行，验证了其安全性和有效性。

Details

Motivation: 随着计算机视觉应用（如监控系统和社交媒体）的快速增长，数据存储在云服务器上引发了安全和视觉隐私问题。现有混淆方案要么易受恶意攻击（如通过模型反转重建原始图像），要么生成不可训练的混淆图像，导致无法在保持合理准确性的同时保护隐私。

Result: 在UCF101和HMDB51数据集上的实验表明，Bit-ViP在面对重建攻击、像素频率、信息熵和像素互相关攻击时，通过严格的安全分析展示了相对于现有方案的实质性改进。

Insight: 创新点在于利用位平面进行图像混淆，并设计了一个结合Lorenz混沌系统和差分隐私的端到端混淆函数，生成不可逆噪声以增强安全性，同时保持混淆图像的可训练性，适用于基于图像的识别任务。

Abstract: The unprecedented growth of computer vision applications, such as surveillance systems and social media, raises security and visual privacy concerns, especially when data is stored on cloud servers. Image obfuscation offers a way to preserve visual privacy while maintaining an adequate level of usability; thus, it has been a topic of great interest in recent years. However, prior obfuscation schemes are either vulnerable to malicious attacks, such as model inversion to reconstruct original images from obfuscated images, or generate non-trainable obfuscated images, making them unusable for achieving reasonable accuracy. This paper proposes a novel bit-plane-based image obfuscation scheme, {\em Bit-ViP}, to preserve visual privacy for image-based recognition tasks. The Bit-ViP scheme produces secure, usable images by incorporating an innovative end-to-end obfuscation function. While doing so, the obfuscated image would contain non-invertible noise (generated by Lorenz’s chaotic system and differential privacy), making it hard for an adversary to reconstruct the original image. We conduct extensive experiments on two popular activity recognition datasets, namely UCF101 and HMDB51, to validate the effectiveness of Bit-ViP. In the face of attacks on reconstruction, pixel frequency, information entropy, and pixel inter-correlation, we present a rigorous security analysis demonstrating tangible improvements over existing schemes.

[111] Can Machines Really See Objects in Images? A Study Based on Syntactic Distance and Visual Self-Referential Instances cs.CV | cs.AI | cs.LGPDF

Xingyu Peng, Junran Wu, Yue Hou, Zhongliang Qiao, Jiaheng Liu

TL;DR: 该论文探讨了视觉模型是否真正理解图像中的物体，还是仅仅拟合了表面视觉线索。通过引入句法距离的概念，作者构建了一个视觉自指任务，其中正负样本在全局语义上不同但句法距离为零，从而消除了局部统计捷径。实验发现，当图像尺度超过临界点时，ResNet和Vision Transformer等模型的准确率会骤降至随机猜测水平，揭示了当前架构在处理全局概念任务上的结构能力边界。

Details

Motivation: 动机是质疑当前视觉模型是否真正’看到’物体，还是仅仅依赖局部统计线索进行分类。作者认为模型的认识能力受限于其学习的描述系统，因此探究当局部线索无法提供稳定区分依据时，模型是否仍能正确分类。

Result: 在基于最大方差二值噪声构建的视觉自指任务上，ResNet和Vision Transformer模型表现出一致的相变现象：一旦图像尺度超过临界点，准确率便崩溃至随机猜测水平，且在测试范围内无法恢复。更大的训练集和模型仅能延迟崩溃，而具有全局注意力的ViT模型崩溃得更早。

Insight: 创新点在于形式化地提出了’句法距离’来衡量类别可分性，并设计了零句法距离的视觉自指任务来剥离局部统计捷径。客观分析认为，该研究揭示了当前主流视觉架构在依赖全局语义的任务上存在根本性能力边界，暗示通用智能可能需要创造新的’语言’，而非复用现有表征系统。

Abstract: Can a vision model truly see an object, or does it only fit surface-level visual cues? Following Wittgenstein’s view that the limits of language are the limits of the world, we view a model’s recognition ability as bounded by the descriptive system it has learned. In current vision models, this system is often realized through learned feature representations that exploit local statistical cues. We therefore ask whether a model can still classify correctly when such local cues provide no stable basis for distinction. We formalize this question with syntactic distance, which measures class separability through the symmetry of the operations mapping one class to the other: positive distance exposes exploitable local features, whereas zero distance requires global semantics rather than local rules. We construct a visual self-referential task in maximum-variance binary noise: positive samples contain a closed square, while negative samples contain an otherwise identical square with one flipped boundary pixel. The two classes differ in global semantics but have zero syntactic distance, making local statistical shortcuts unreliable. Experiments on ResNets and Vision Transformers reveal a consistent phase-transition phenomenon, with accuracy collapsing to random guessing once the image scale crosses a critical point and does not recover within the tested range. Larger training sets and models only delay this collapse, while globally attentive ViTs reach it earlier. These results reveal a structural capability boundary of current architectures on global-concept tasks, suggesting that general intelligence may require creating new language, not reusing an existing one.

[112] Bridging VideoQA and Video-Guided Agentic Tasks via Generalized Keyframe Extraction cs.CV | cs.AIPDF

Sunqi Fan, Qingle Liu, Runqi Yin, Meng-Hao Guo, Shuojin Yang

TL;DR: 本文针对当前视频理解任务中多模态大语言模型（MLLMs）在视频问答（VideoQA）基准上表现良好，但缺乏对从视频教程中学习深层知识或程序技能并泛化至下游长视野智能体任务能力的评估这一问题，提出了VG-GUIBench基准。同时，作者观察到模型在VideoQA和视频引导的智能体任务上的性能高度依赖于有效的关键帧提取，因此提出了TASKER算法，该算法联合考虑任务相关性和场景动态性来选择信息帧。实验表明TASKER在多个基准上取得了显著性能提升。

Details

Motivation: 现有VideoQA基准主要评估模型对浅层视觉线索的感知能力，而很少检验MLLMs能否从视频教程中学习更深层的知识或程序技能，并将其泛化到下游长视野的智能体任务中。为了填补这一空白，作者旨在构建一个更全面的评估基准，并提升视频理解任务中关键帧提取的有效性。

Result: 在EgoSchema完整数据集上，TASKER比最佳基线模型性能提升了2.0%；在NExT-QA数据集上提升了1.8%。这些结果表明TASKER在VideoQA和视频引导的智能体任务基准上均取得了显著的性能改进。

Insight: 论文的核心创新点在于提出了一个专门用于评估视频引导GUI智能体任务的新基准VG-GUIBench，以及一个通用的、任务驱动且场景感知的关键帧提取算法TASKER。TASKER的创新之处在于将任务相关性与场景动态性联合优化，而非仅依赖传统的均匀采样或基于变化的检测，这为提升视频理解任务的性能提供了一个有效且可泛化的技术路径。

Abstract: Video understanding is a fundamental capability for multimodal intelligence, and recent Multimodal Large Language Models (MLLMs) have achieved remarkable performance on Video Question Answering (VideoQA) benchmarks. However, existing benchmarks primarily evaluate whether models can perceive shallow visual cues, while rarely examining whether MLLMs can learn deeper knowledge or procedural skills from video tutorials and generalize them to downstream long-horizon agentic tasks. To address this gap, we introduce VG-GUIBench (Video-Guided GUI Benchmark), a new benchmark designed to evaluate whether MLLM-based GUI agents can follow video tutorials to complete corresponding GUI interactive tasks. Furthermore, we observe that the performance of models on both VideoQA and video-guided agentic tasks critically depends on effective keyframe extraction. Based on this observation, we propose TASKER (Task-driven And Scene-aware Keyframe searchER), a keyframe extraction algorithm that jointly considers task relevance and scene dynamics to identify informative frames. Experimental results demonstrate that TASKER achieves significant performance improvements on both VideoQA and video-guided agentic task benchmarks, outperforming the best baseline by 2.0% on the EgoSchema fullset and 1.8% on the NExT-QA dataset, respectively. These results further highlight the potential of generalized keyframe extraction methods for video understanding tasks. Our code and data are available at https://github.com/VG-GUI-TASKER/VG-GUI-TASKER.

[113] MIRROR: Aligning Semantic Relations from Language to Image via Gromov–Wasserstein cs.CVPDF

Hong-Han Wang, Yuntao Wang, Hu Ding

TL;DR: 本文提出MIRROR框架，通过基于最优传输的几何正则化方法，将语言模型中丰富的语义关系先验对齐到视觉表示中，以解决多模态大语言模型在视觉场景中应用语义关系时的失败问题。

Details

Motivation: 多模态大语言模型虽然继承了语言主干中的丰富关系先验，但在视觉上下文中应用这些关系时经常失败，这源于投影对齐方法只关注单个视觉令牌的语义正确性，而忽略了概念间关系的跨模态保持。

Result: MIRROR在保持通用视觉-语言任务性能的同时，提高了关系一致性，无需额外参数或推理成本，适用于长令牌序列的高效计算。

Insight: 创新点在于提出了半逆Gromov-Wasserstein问题及其闭式解，通过几何正则化显式对齐语言与视觉的关系结构，并在Transformer中设计了层、头和令牌级别的针对性策略以实现稳定提取。

Abstract: Multimodal Large Language Models (MLLMs) inherit rich relational priors from their language backbones, yet often fail when asked to apply these relationships in visual contexts. We trace this failure to a structural blind spot: projection-based alignment trains each visual token to carry the right semantics, but never asks whether the relationships between concepts survive the crossing from language to vision. To address this, we propose MIRROR (Mapping Inter-concept Relations from language to visual Representation via Optimal-transport-based Regularization), a geometric regularization framework that transfers relational priors from language to vision by exploiting the rich relational structure encoded in language representations. Specifically, we derive a surrogate loss from the proposed Semi-Inverse Gromov-Wasserstein (SI-GW) problem, an inverse geometric problem that aligns visual representations with language-derived relational priors. We show that this formulation admits a unique closed-form solution that prescribes the ideal visual relational structure implied by language geometry and cross-modal coupling. The structure of the formulation also enables efficient computation, making it applicable to long token sequences. Applying SI-GW inside decoder-only Transformers requires careful design. We introduce targeted strategies at the layer, head, and token levels to ensure stable extraction without additional parameters or inference cost. MIRROR improves relational consistency while preserving performance on general vision-language tasks.

[114] MAVIN: Multi-Shot Audio-Visual Generation with Narrative Control cs.CVPDF

Kaiqi Liu, Yunyao Mao, Ziqi Cai, Zheng Geng, Jing Wang

TL;DR: MAVIN是首个支持定制化叙事控制的多镜头音视频生成框架。它通过边界感知注意力解决时序错位问题，利用ID感知传播提升多主体场景的可控性，并采用多智能体脚本管道将用户输入转化为分层描述。

Details

Motivation: 现有生成模型难以实现连贯的多镜头音视频生成所需的复杂叙事控制，存在时序错位、可控性有限和脚本不完整等问题。

Result: 在构建的MAVINSet数据集上进行的大量实验表明，MAVIN实现了最先进的性能。

Insight: 创新点包括边界感知注意力机制、ID感知传播方法以及多智能体脚本管道，为生成模型融入专业电影制作流程开辟了新途径。

Abstract: While recent generative models produce high-fidelity videos, they struggle with the complex narrative control required for coherent multi-shot audio-visual generation. Existing methods suffer from temporal misalignment, limited controllability, and incomplete scripting. In this paper, we propose MAVIN, the first framework for multi-shot audio-visual generation with customized narrative control. To resolve temporal misalignment, we propose boundary-aware attention, which leverages hierarchical captions and boundary-aware token routing to render audio-visual elements within their respective temporal boundaries. To improve the controllability for multi-subject scenarios, we propose ID-aware propagation, utilizing identity embeddings and an identity-aware mask to bind specific identities to consistent visual appearances and vocal timbres. To provide comprehensive audio-visual narratives, we present a multi-agent scripting pipeline to transform free-form user inputs into hierarchical captions. Furthermore, we construct MAVINSet, a multi-shot audio-visual dataset for robust training and evaluation. Extensive experiments demonstrate that MAVIN achieves state-of-the-art performance, opening up a new avenue for integrating generative models into professional filmmaking workflows.

[115] Rank-Aware Hyperbolic Alignment for Vision-Language Dataset Distillation cs.CV | cs.AIPDF

Jongoh Jeong, Sun-Kyung Lee, Kuk-Jin Yoon

TL;DR: 本文提出了一种名为RAHA的秩感知双曲对齐方法，用于视觉语言数据集蒸馏。该方法将多模态表示提升到双曲空间，通过非对称目标优化蒸馏后的合成图像-文本对，在共享语义的低维范围内强制测地线对齐，同时正则化残差子空间以保留模态私有多样性。

Details

Motivation: 现有视觉语言数据集蒸馏方法通常在欧几里得嵌入空间中强制全维对齐，但由于图像-文本相关性存在秩不足问题，共享语义集中在低维范围，而剩余变化分布在弱相关的残差子空间中，这种全维对齐过于严格。

Result: 在基准测试上的实验表明，RAHA在固定数据量和计算预算下，展示了具有竞争力的跨模态检索性能，并改善了迁移鲁棒性指标。

Insight: 创新点在于结合了分层几何（双曲空间）与显式的对齐容量控制，通过非对称目标分别处理共享语义子空间和模态私有残差子空间，从而更有效地进行数据集蒸馏。

Abstract: Vision-language dataset distillation (VLDD) compresses a large image-text paired dataset into a small set of synthetic pairs that can efficiently train contrastive vision-language models under strict data and compute budgets. Most existing methods match expert trajectories or cross-modal statistics, yet still enforce full-dimensional alignment in a Euclidean embedding space. This is often overly restrictive due to rank-deficient image–text correlation, with shared semantics concentrated in a low-dimensional range and remaining variation spread across a weakly correlated residual subspace. LoRS relaxes alignment at the similarity level by low-rank factorization, but does not explicitly control dominant alignment capacity and structure in the representation space. We thus propose a rank-aware hyperbolic alignment (RAHA) that combines hierarchical geometry with explicit alignment-capacity control. RAHA lifts multimodal representations to hyperbolic space and optimizes distilled pairs with asymmetric objectives that enforce geodesic alignment in the shared range while regularizing the residual subspace to preserve modality-private diversity and improve transfer robustness. Experiments on benchmarks show that RAHA demonstrates competitive cross-modal retrieval and improved transfer indicators under fixed budgets.

[116] VCS-SLAM: Geometry-Validated Semantic Evidence Fusion for 3D Gaussian SLAM cs.CVPDF

Raman Jha, Shuaihang Yuan, Yi Fang

TL;DR: 本文提出了VCS-SLAM，一种用于RGB-D 3D高斯SLAM的几何验证语义证据融合框架。该方法通过评估语义观测的几何可靠性（如可见性一致性、表面支持的边界证据和光线级冲突不确定性），来抑制不可靠的语义更新，从而减少语义伪影并提升地图质量。在Replica和ScanNet数据集上的实验验证了其在语义一致性、边界保持和重建质量方面的改进。

Details

Motivation: 现有的语义3D高斯SLAM通常使用均匀优化权重将2D语义先验融合到持久3D地图中，但在线建图时，由于遮挡、无支持的语义边界和模糊的光线几何等因素，这些先验并不可靠，会导致全局高斯地图中出现持续的语义伪影。

Result: 在Replica数据集上，VCS-SLAM展示了改进的语义一致性、边界保持和重建质量。在ScanNet数据集上的结果表明，该方法在真实RGB-D输入下保持了具有竞争力的跟踪性能。

Insight: 核心创新在于提出了一个几何验证的语义证据融合机制，通过可靠性感知的目标函数动态评估和加权语义观测，而非统一处理。这有效抑制了被遮挡的语义更新、减少了无支持的语义渗色，并延迟了模糊区域的过早标签分配，从而提升了语义建图的鲁棒性和准确性。

Abstract: Visual SLAM performance often deteriorates in complex real-world applications. Semantic 3D Gaussian SLAM commonly fuses 2D semantic priors into a persistent 3D map using uniform optimization weights. However, such priors are not equally reliable in online mapping: occlusions, unsupported semantic boundaries, and ambiguous ray geometry can introduce persistent semantic artifacts into the global Gaussian map. We propose VCS-SLAM, a geometry-validated semantic evidence fusion framework for RGB-D 3D Gaussian SLAM. Instead of treating all semantic observations as uniformly valid supervision, VCS-SLAM evaluates their geometric reliability through visibility consistency, surface-supported boundary evidence, and ray-level conflict uncertainty. The resulting reliability-aware objective suppresses occluded semantic updates, reduces unsupported semantic bleeding, and delays premature label assignment in ambiguous regions. Experiments on Replica demonstrate improved semantic consistency, boundary preservation, and reconstruction quality. Results on ScanNet further show that VCS-SLAM maintains competitive tracking performance under real RGB-D inputs

[117] Learning Where and When: Patch-Based Spatiotemporal Localization in Weakly Supervised Video Anomaly Detection cs.CVPDF

Hamza Karim, Nghia Nguyen, Lokman Bekit, Yasin Yilmaz

TL;DR: 本文提出了一种基于patch的时空弱监督视频异常检测框架，能够同时定位异常发生的时间和空间位置。该方法在网格级patch特征上操作，通过多示例学习范式学习区域级异常分数，并引入了一种邻近感知的Top-k时空选择策略，无需训练时的边界框监督即可生成细粒度空间异常图。

Details

Motivation: 现有弱监督视频异常检测方法主要关注时间定位（何时发生异常），而忽略了异常在帧内的空间范围（何处发生）。空间定位对于实际部署的可解释性至关重要，因此需要开发能够联合建模时空异常位置的方法。

Result: 该方法在多个基准测试中超越了现有的最先进方法，在时空定位精度上取得了显著提升。作者还发布了两个广泛使用数据集的测试集帧级边界框标注，为未来研究提供了新资源。

Insight: 核心创新点在于将弱监督异常检测从单纯的时间定位扩展到联合时空定位，并提出了无需边界框监督即可生成细粒度空间异常图的邻近感知Top-k选择策略。这为弱监督设置下的精细异常解释提供了新思路。

Abstract: Weakly supervised video anomaly detection (WSVAD) has predominantly focused on temporal localization, identifying when anomalies occur while largely neglecting their spatial extent within frames. Yet, spatial localization is essential for interpretability and practical deployment in real-world settings. We introduce a patch-based spatiotemporal framework for weakly supervised anomaly localization that jointly models where and when anomalies occur. Our approach operates on grid-level patch features and learns region-level anomaly scores under a multiple instance learning paradigm. We further propose a Proximity-Aware Top-k spatiotemporal selection strategy that enables the model to generate fine-grained spatial anomaly maps without requiring bounding-box supervision during training. Our method surpasses existing state-of-the-art approaches across multiple benchmarks, yielding substantial gains in spatiotemporal localization accuracy. In addition, we release frame-level bounding-box annotations for the test sets of two widely used datasets, along with our code and pretrained models, providing new resources to facilitate future research in spatially grounded WSVAD.

[118] MotionAtlas: Detailed Region Captioning for Motion-Centric Videos cs.CV | cs.AIPDF

Weisong Liu, Haochen Wang, Kuan Gao, Yuhao Wang, Yikang Zhou

TL;DR: 本文提出了MotionAtlas系统，用于对以运动为中心的视频进行细粒度的区域感知描述。该系统包含一个精心标注的基准测试集、一个可扩展的高质量训练数据构建流程，以及一系列基于视频多模态大语言模型的强大模型。

Details

Motivation: 传统全局运动描述数据集存在视觉杂乱和运动纠缠问题，难以进行可靠评估。本文旨在解决区域感知运动描述任务，即给定视频和时空掩码，模型需精确描述目标区域内的运动。

Result: 在构建的MotionAtlas-Bench基准上，所提方法在通用运动基准测试中显著提升了基线模型的性能。例如，MotionAtlas-4B模型平均比Qwen3-VL-4B高出5.2个百分点。

Insight: 创新点在于提出了区域感知运动描述任务范式，并设计了包含自引导精炼的数据流水线来抑制细粒度幻觉，以及针对性的训练数据组合策略，这些方法可泛化提升多种视频多模态大语言模型的性能。

Abstract: We propose MotionAtlas, a system for detailed captioning of motion-centric videos, comprising (1) a dedicated human-annotated benchmark, (2) a scalable, high-quality pipeline to construct training samples, and (3) a family of powerful Video-MLLMs. Unlike conventional global motion captioning datasets, we focus on region-aware motion captioning: given a video and a spatiotemporal mask, the model generates precise descriptions of motion within the target region, thereby alleviating visual clutter and motion entanglement and enabling reliable, quantifiable evaluation. Concretely, we first build MotionAtlas-Bench, a comprehensive benchmark comprising 2,073 multiple-choice questions, meticulously annotated for a curated set of high-quality, motion-centric videos, to evaluate fine-grained motion understanding of the objects in question. Second, we design a rigorous and scalable data pipeline that leverages self-bootstrap refinement to suppress fine-grained hallucinations, yielding 159k high-quality motion captioning data. Third, we design a tailored training data composition strategy, which achieves consistent and substantial performance gains across diverse baseline Video-MLLMs, including Molmo2 and Qwen3-VL. For instance, MotionAtlas-4B surpasses Qwen3-VL-4B by an average of 5.2 percentage points across general motion benchmarks. The benchmark, dataset, and code have been released.

Mohammadreza Rashidi

TL;DR: 本文对一种用于肩扛式视频监控的多模态触摸检测框架进行了实证评估，该框架旨在通过物理手指交互重建移动设备键盘界面的击键事件。系统整合了四种并行检测模态：手部关键点、肤色过滤、帧间差分运动检测和边缘分析。评估表明，在受控场景下运动检测和边缘检测配置表现尚可，但多模态组合性能有限，且在真实无控视频中泛化能力差，无法可靠重建击键序列。

Details

Motivation: 研究动机是探索在不直接录制屏幕的情况下，通过肩扛式视频监控（VIDINT）来监控人机交互模式，特别是从物理手指交互中重建移动设备上的击键事件。

Result: 在120帧受控的第一人称密码输入视频上评估，仅运动检测和仅边缘检测配置的F1分数分别为18.5%和18.2%，而多模态组合配置的F1分数为16.7%，序列相似度仅为3.0%。在5个真实第三方手机视频上的泛化测试中，检测器产生了大量误报（中位数每帧57个触摸点），远高于真实点击率，表明其在非受控环境下不可靠。

Insight: 论文宣称的创新点在于提出并实证评估了一个整合四种并行模态的BEHINT触摸检测框架。从客观角度看，其研究价值在于系统地揭示了此类基于计算机视觉的间接击键重建方法在受控场景下的性能局限，以及在真实、无控视频中因肤色过滤器响应整个手部而非指尖接触导致的严重泛化问题，这为类似监控技术的实际可行性提供了重要的经验性洞见。

Abstract: Video Intelligence Surveillance (VIDINT) on over-the-shoulder footage is a proposed vector for monitoring human-computer interaction patterns without direct screen recording access. In this paper, we evaluate a Behavioral Intelligence (BEHINT) touch-detection framework designed to reconstruct keystroke events on mobile keypad interfaces from physical finger interactions. Our system integrates four parallel detection modalities: (1) anatomical hand landmarks via MediaPipe, (2) HSV skin color filtering, (3) temporal frame differencing for motion detection, and (4) shape-guided Canny edge analysis. We map relative touch coordinates to a reference screen layout to reconstruct typing sequences. Evaluation on a 120-frame first-person staged video of passcode entry reveals that while MediaPipe and Skin Detection fail to run autonomously due to partial hand occlusion and ambient noise, Motion-Only and Edge-Only configurations achieve F1-scores of 18.5% and 18.2%, respectively. The combined multi-modal configuration achieves an F1-score of 16.7% and a sequence similarity of 3.0% when mapped to the iOS passcode layout. We conduct ablation, resolution decay, noise sensitivity, and proximity threshold tuning to characterize the system’s operational envelope. We then audit generalization on 5 real, publicly licensed third-person phone videos and find that the detector emits a median of 57 touch points per frame (peaking at 205), one to three orders of magnitude more than the rate of real taps, because the skin filter responds to the whole hand rather than to fingertip contact. The staged keystroke result does not survive contact with uncontrolled footage; the system does not achieve reliable keystroke reconstruction outside the calibrated staged setting.

[120] Benchmark AUC Is Not Deployable Reliability: A Cross-Dataset Audit of Off-the-Shelf Features for Surveillance Video Anomaly Detection cs.CV | cs.CRPDF

Mohammadreza Rashidi

TL;DR: 这篇论文对监控视频异常检测中常用的基准测试AUC指标进行了跨数据集审计，发现当模型在一个场景上训练并在另一个场景上测试时，其性能会急剧下降至随机水平。研究使用现成的特征提取器（如CLIP、DINOv2）和简单检测器（如最近邻距离）在四个真实数据集上进行实验，揭示了基准测试结果与部署可靠性之间存在巨大差距。

Details

Motivation: 当前AI监控领域在标准视频异常检测基准上报告了高帧级ROC-AUC，但这些指标通常基于同一摄像头和场景的训练与测试。论文旨在审计当这一假设被打破时，即模型跨数据集部署时的实际性能表现。

Result: 实验显示，同数据集测试的平均AUC为0.704，但跨数据集测试的平均AUC降至0.499（随机水平），其中DINOv2在同数据集上表现最佳（Ped2上达0.901），但跨数据集下降最大。使用PaDiM式马氏距离检测器也得到类似结果（跨数据集差距0.202 vs 0.208），且每小时误报率高达约31,931次。

Insight: 论文的创新点在于通过跨数据集审计揭示了监控异常检测基准指标的局限性，强调基准AUC不能代表部署可靠性。从客观角度看，这提醒社区需要更注重模型泛化能力和真实场景评估，而非仅依赖实验室校准设置下的高指标。

Abstract: Automated “suspicious behavior” flagging is a headline promise of AI surveillance, and the field reports high frame-level ROC-AUC on standard video anomaly detection benchmarks. Those numbers are measured by training and testing on the same camera and scene. We audit what happens when that assumption is dropped. We build an unsupervised normality model from the all-normal training frames of one dataset, using frozen off-the-shelf embeddings (CLIP, DINOv2, ResNet-50, EfficientNet-B0) and a nearest-neighbour distance, and score the test frames of the same and of other datasets. Across 4 real datasets (UCSD Ped1, UCSD Ped2, CUHK Avenue, ShanghaiTech) and 4 backbones, same-dataset AUC averages 0.704 but cross-dataset AUC averages 0.499, which is chance: a detector calibrated on one scene is no better than a coin flip on another, and in several pairs it is below chance. The strongest backbone makes this worse, not better: DINOv2 has the best same-dataset AUC (up to 0.901 on Ped2) and the largest cross-dataset drop. The collapse is not an artefact of the scoring rule: replacing the nearest-neighbour detector with a PaDiM-style Mahalanobis detector reproduces it almost exactly (cross-dataset gap 0.202 versus 0.208). Even at a favourable operating point the false-alarm rate is on the order of 31,931 per hour. We conclude that the benchmark numbers quoted for surveillance anomaly detection describe a calibrated laboratory setting and overstate deployable reliability by a wide margin, and we release the code that reproduces every number.

[121] Reliability-Prioritized Fine-Grained Generation in Multimodal Large cs.CVPDF

Xiaomeng Fan, Wu Wei, Yuwei Wu, Zhi Gao, Shiyu Luo

TL;DR: 本文针对多模态大语言模型生成细粒度视觉描述时可靠性下降的问题，提出了GranFact基准和一种基于直接偏好优化的可靠性优先优化方法，以在保持可靠性的前提下提升细粒度生成能力。

Details

Motivation: 研究发现细粒度生成比粗粒度生成更容易出错，因此模型应生成在可靠前提下最细粒度的描述，而非一味追求更具体的输出。

Result: 在GranFact基准上的实验表明，所提方法在保持可靠性的同时，改善了细粒度生成性能。

Insight: 创新点在于提出了一个粒度感知的评估基准和一种层次感知的评估算法，以及一种通过偏好优化来权衡可靠性与特异性的训练方法，为解决MLLM中细粒度生成的可靠性权衡问题提供了系统框架。

Abstract: Multimodal large language models (MLLMs) are increasingly expected to generate fine-grained descriptions of visual content. However, we observe and theoretically show that generating fine-grained responses poses a reliability challenge, \textit{i.e.}, fine-grained generation is more error-prone than coarse-grained generation. This phenomenon suggests that models should generate the finest description that remains reliable rather than simply produce more specific outputs. To investigate this problem, we develop \textsc{GranFact}, a granularity-aware benchmark consisting of expert-verified multi-object images with coarse-to-fine category annotations. Then, we design a hierarchy-aware evaluation algorithm, which assesses both whether model predictions are visually correct and how specific the correct predictions are. We also propose a reliability-prioritized preference optimization method based on Direct Preference Optimization, which penalizes unreliable fine-grained claims while rewarding reliable specificity. Experiments on \textsc{GranFact} show that our method improves fine-grained generation while preserving reliability. Code and data are available \href{https://github.com/WeiWu2025/GranFact}{here}.

[122] Scenes as Objects, Not Primitives: Instance-Structured 3D Tokenization from Unposed Views cs.CV | cs.GRPDF

Mijin Yoo, In Cho, Subin Jeon, Jiwoo Lee, Eunbyung Park

TL;DR: 本文提出了一种前馈框架，直接从无位姿多视角图像中将3D场景分解为实例结构化的3D令牌组。这些令牌组是面向对象的紧凑单元，包含捕获实体级身份的实例令牌和编码局部几何与外观的锚点令牌，可解码为一组3D高斯。该表示将物体身份与局部外观解耦，使物体实例成为表征的原生接口，从而直接支持重建、分割和操作。

Details

Motivation: 现有前馈重建方法输出的是密集、无结构的点或高斯集合，物体级结构需要事后恢复。本文动机是直接从图像中学习具有实例结构的3D表示，使物体成为场景理解的基本单元，而非事后推导的产物。

Result: 在无需3D标注的情况下，该前馈模型在类无关实例分割任务上超越了基于每场景优化的基线方法，同时在新视角合成任务上保持竞争力。此外，该表示可直接支持实例级场景编辑（如移除、平移、插入物体）和高效的开放词汇3D实例检索，其检索复杂度与实例数量而非基元数量成正比。

Insight: 核心创新在于提出了一个两级因子化表示：将场景分解为实例结构化的令牌组，将物体身份（实例令牌）与局部几何外观（锚点令牌）解耦。这使得物体实例成为表征的原生接口，而非派生结果，从而统一支持分割、重建和编辑等多种下游任务，且无需3D监督。

Abstract: A 3D scene is understood through its objects, not the primitives that compose them. Yet feed-forward reconstruction methods output dense, unstructured sets of points or Gaussians, leaving object-level structure to be recovered after the fact. We propose a feed-forward framework that decomposes a scene into instance-structured 3D token groups directly from unposed multi-view images – compact object-centric units from which reconstruction, segmentation, and manipulation all follow. Each token group pairs an instance token capturing entity-level identity with anchor tokens that encode local geometry and appearance, which are decoded into a set of 3D Gaussians. This two-level factorization decouples object identity from local appearance, making object instances a native interface of the representation rather than a derived product. The token groups are learned through differentiable rendering with joint reconstruction and segmentation supervision, requiring no 3D annotations. Our feed-forward model surpasses per-scene optimization baselines in class-agnostic instance segmentation while remaining competitive in novel view synthesis. Beyond these metrics, the same token groups directly unlock instance-level scene editing – removing, translating, or inserting objects by operating on their groups – as well as efficient open-vocabulary 3D instance retrieval, where retrieval complexity scales with the number of instances rather than primitives.

[123] ScAle: Attention Head Scaling as a Minimal Adapter for Spatial Reasoning in Vision Language Models cs.CV | cs.AI | cs.LG | cs.MMPDF

Rahul Chowdhury, Timothy A Rupprecht, Xuan Shen, Pu Zhao, Yanzhi Wang

TL;DR: 本文提出了一种名为ScAle的超轻量级适配方法，用于提升视觉语言模型在空间推理任务上的性能。该方法通过在冻结的骨干网络中学习一组标量系数来调制最后一层注意力机制和多层感知机的激活值，仅需1K可训练参数即可在多个基准测试上取得显著性能提升。

Details

Motivation: 空间推理是许多视觉语言模型面临的持续挑战，传统改进方法通常需要大量额外参数进行微调。初步分析发现，仅对选定Transformer层的激活值进行重新缩放（不修改预训练权重）就能显著影响下游性能，这启发了本文研究一种极简的适配方法。

Result: 在合成空间推理基准SpatialEval以及真实世界VQA数据集（COCOQA和VGQA）上，ScAle方法仅使用1K参数就实现了最高134.1%的相对准确率提升，恢复了标准参数高效微调方法（如LoRA）的大部分性能，同时保持了非空间VQA任务的准确率。

Insight: 论文的创新点在于提出了一种通过有界激活重加权来适配预训练VLMs的简单、架构无关且参数高效的方法。从客观角度看，其核心洞察是模型的空间推理能力可以通过极低成本的激活缩放来有效调控，这为模型适配提供了新的轻量化思路。

Abstract: Spatial reasoning remains a persistent challenge for many vision language models (VLMs), and improving it typically requires fine-tuning with substantial additional parameters. Our preliminary analysis reveals that rescaling activations in selected transformer layers-without modifying pretrained weights-can significantly influence downstream performance. Motivated by this observation, we propose ScAle, an ultra-lightweight adaptation method that learns a small set of scalar coefficients to modulate last-token attention and MLP activations in a fully frozen backbone. We evaluate our method on the synthetic spatial reasoning benchmark SpatialEval and on real-world VQA datasets (COCOQA and VGQA) across multiple model families. Our method, ScAle, achieves up to 134.1% relative accuracy gains using only 1K trainable parameters without requiring millions of trainable parameters as in standard PEFT methods such as LoRA. Despite its extreme compactness, our approach recovers a substantial fraction of standard PEFT performance while preserving strong non-spatial VQA accuracy. These results demonstrate that bounded activation reweighting provides a simple, architecture-agnostic, and highly parameter-efficient alternative for adapting pretrained VLMs.

[124] One Scene, Two Depths: Probing Geometric Ambiguity in Monocular Foundation Models cs.CV | cs.AIPDF

Xiaohao Xu, Feng Xue, Xiang Li, Haowei Li, Shusheng Yang

TL;DR: 本文探讨单目深度估计模型在透明场景中面临的几何歧义问题，即单个像素射线可能对应多个有效深度层。作者提出了MultiDepth-3k（MD-3k）基准来量化模型的深度层偏好和多层空间关系准确性，并发现不同模型对同一场景的深度层解析存在差异。通过拉普拉斯视觉提示（LVP）这一无需训练的频谱输入变换，可以显著改变冻结模型的深度层预测结果，其中DAv2-L模型在RGB/LVP组合下达到了75.5%的ML-SRA准确率。

Details

Motivation: 解决单目深度估计在透明场景中因几何歧义导致的监督目标不一致问题，即传统方法将多层几何结构简化为每像素单一深度值，而透明物体（如玻璃）使同一射线可能穿透前景并观测背景，使得深度标注成为数据与训练惯例而非场景固有真值。

Result: 在提出的MD-3k基准上，主流深度基础模型在标准RGB输入下表现出多样化的深度层偏好；使用拉普拉斯视觉提示（LVP）可改变冻结模型的预测层，最佳模型DAv2-L在RGB/LVP组合下达到75.5%的多层空间关系准确率（ML-SRA）。

Insight: 创新点在于提出首个稀疏双层序数基准MD-3k来量化深度模型的层偏好与多层几何关系准确性，并引入无需训练的频谱变换LVP来揭示模型潜在的互补几何假设；客观而言，该研究推动社区以歧义感知视角重新思考深度监督与评估，将多重有效3D解释视为可测量、保留和表达的几何结构。

Abstract: A faithful 3D world representation should account for layered geometry, where a single camera ray may contain multiple visible and geometrically valid surfaces. Monocular depth estimation, however, reduces this structure to one scalar depth per pixel. Transparent scenes make this ambiguity measurable: the same ray can pass through foreground glass and observe the background, turning the supervised target into a convention of annotation, data, and training rather than a scene-intrinsic truth. A learned predictor exposes this convention as its depth-layer preference. We introduce MultiDepth-3k (MD-3k), a sparse two-layer ordinal benchmark for measuring depth-layer preference and multi-layer spatial relationship accuracy (ML-SRA). On MD-3k, leading depth foundation models exhibit diverse layer preferences under standard RGB input, showing that the same layered geometry can be resolved differently across models. We further find that Laplacian Visual Prompting (LVP), a training-free spectral input transformation, can substantially change the reported layer for certain frozen models. The strongest RGB/LVP pair, DAv2-L, reaches 75.5% ML-SRA. These results suggest that depth foundation models may express complementary geometric hypotheses that standard RGB inference leaves unexpressed. We invite the community to rethink depth supervision and evaluation through an ambiguity-aware lens, where multiple valid 3D interpretations are treated as geometric structure to be measured, preserved, and expressed.

[125] SonoCLIP: Mask-Guided Region-Aware Vision-Language Pretraining for Fetal Ultrasound Analysis cs.CV | cs.AIPDF

Hang Su, Chao Sun, Zhaofan Li, Wei Hu, Juhua Liu

TL;DR: 本文提出了SonoCLIP，一个百万级规模的、区域可控的胎儿超声视觉-语言基础模型。该模型在视觉编码器中集成了分割掩码作为掩码通道视觉提示，实现了联合的全局-局部对比表征学习，并引入了一种基于Sigmoid的成对对比损失以提高大规模监督下的稳定性。研究还构建了一个包含144万张图像、涵盖24个标准切面的多模态胎儿超声数据集用于大规模预训练。

Details

Motivation: 现有基于CLIP的模型主要依赖全局图像-文本对齐，对临床决策关键的局部结构敏感性不足。胎儿超声领域存在严重的斑点噪声、采集变异性和细微解剖边界，导致观察者间差异大，因此需要更精细、可控的基础模型。

Result: 广泛的跨中心评估表明，SonoCLIP在全局和掩码引导推理下均实现了卓越的零样本迁移性能，为胎儿超声分析建立了一个可控且面向临床的基础模型。

Insight: 核心创新点在于将分割掩码作为视觉提示集成到视觉编码器中，实现了区域可控的视觉-语言对齐，并设计了新的对比损失函数以支持大规模区域-文本对齐。这为医学影像分析提供了一种结合全局语义与局部解剖结构敏感性的新范式。

Abstract: Vision-language foundation models have shown strong potential in medical image analysis. Although foundation models for ultrasound imaging have recently emerged, the domain remains particularly challenging due to severe speckle noise, acquisition variability, and subtle anatomical boundaries, leading to high inter-observer variability. Existing CLIP-based models rely primarily on global image-text alignment, limiting their sensitivity to clinically decisive local structures. We propose SonoCLIP, the first million-scale region-controllable fetal ultrasound vision-language foundation model that integrates segmentation masks as mask-channel visual prompts within the vision encoder, enabling joint global-local contrastive representation learning. To support scalable region-text alignment, we introduce a sigmoid-based pairwise contrastive loss that improves stability under large-scale supervision. We further curate a 1.44M-image multimodal fetal ultrasound dataset spanning 24 standard planes for large-scale pretraining. Extensive cross-center evaluations demonstrate that SonoCLIP achieves superior zero-shot transfer performance under both global and mask-guided inference, establishing a controllable and clinically oriented foundation model for fetal ultrasound analysis. Our code and data are available at https://github.com/Harrison-one/SonoCLIP.

[126] Unlocking the Visual Record of Materials Science: A Large-Scale Multimodal Dataset from Scientific Literature cs.CV | cond-mat.mtrl-sci | cs.AIPDF

Subham Ghosh, Shubham Tiwari, Mohammad Ibrahim, Abhishek Tewari

TL;DR: 本文提出了MatMMExtract端到端开源流程，用于从材料科学文献中大规模提取结构化多模态数据。该方法通过分解复合图表为独立子图，并利用大语言模型生成基于材料科学分类法的标注，构建了包含39.1万面板级图文对的MatSciFig数据集。同时发布了MaterialScope领域特定检测数据集，并验证了数据在视觉-语言检索任务中的有效性。

Details

Motivation: 解决材料科学文献中大量复合图表因单一标题描述多个子图而导致图像-文本配对不可靠的问题，解锁被锁定的视觉实验知识以供AI大规模使用。

Result: 在自建的MaterialScope检测数据集上，微调的YOLO12-m检测器达到mAP_50为0.9227；在MatSciFig数据集上的双编码器检索基线比零样本CLIP的R@1提升了4.4倍；在标注生成任务中，Gemini 3.1 Flash Lite模型以82%的良好输出率和4.8%的幻觉率取得了最佳性价比。

Insight: 创新点在于提出针对科学文献复合图的结构化解构流程，并结合领域分类法指导LLM生成结构化标注；客观来看，其构建的领域专用检测数据集和规模化、细粒度标注的多模态数据集，为材料科学领域的视觉-语言学习提供了关键资源。

Abstract: The materials science literature encodes decades of experimental knowledge in figures, yet this visual record remains locked away and inaccessible to AI at scale. The core difficulty is structural: most scientific figures are compound, with a single caption describing multiple sub-panels simultaneously, making direct image-text pairing unreliable. We present MatMMExtract, an end-to-end open-source pipeline that resolves this by decomposing compound figures into individual sub-panels and generating structured, grounded annotations using a large language model guided by a curated materials science taxonomy. Applied to 14,810 open-access articles, MatMMExtract produces MatSciFig; 391,606 panel-level image-text pairs from 180,571 figures, each annotated with a sub-caption, a two-level visualisation category spanning 19 classes and over 100 subtypes, and a scientific summary. To enable accurate panel localisation, we introduce MaterialScope, a domain-specific detection dataset of 2,811 manually annotated materials science figures, on which a fine-tuned YOLO12-m detector achieves mAP_50 of 0.9227. Among six benchmarked language models, Gemini 3.1 Flash Lite delivers the best cost-quality trade-off for annotation generation, with 82% of outputs rated good and a hallucination rate of 4.8%. A dual-encoder retrieval baseline on MatSciFig achieves a 4.4 times improvement in R@1 over zero-shot CLIP, demonstrating the dataset’s immediate utility for vision-language learning. All resources are released openly to the community.

[127] LEIQ-Assessor: Multi-dimensional Quality Assessment of Low-light Enhanced Images via Multi-task Learning cs.CV | cs.MMPDF

Wei Sun, Yanwei Jiang, Dandan Zhu, Jinqiu Sang, Jikai Xu

TL;DR: 本文提出了LEIQ-Assessor，一个基于多任务学习的低光照图像增强质量评估模型。该模型使用预训练的SigLIP2 Vision Transformer作为主干网络，同时预测图像的整体平均意见得分（MOS）以及六个感知子属性，包括亮度、色彩保真度、噪声水平、曝光质量、自然度和内容恢复。

Details

Motivation: 低光照图像增强算法在提升图像可见度的同时，常会引入噪声放大、色彩偏移、结构损伤和过曝等伪影，从而降低增强图像的感知质量。因此，需要一个可靠的图像质量评估指标来评估增强效果，这对于LIEAs的开发和实际应用至关重要。

Result: 在MLE基准测试上的实验表明，LEIQ-Assessor显著优于现有的无参考IQA模型和手工制作的质量描述符。该方法在QoMEX 2026低光照增强图像质量评估大挑战中获得了第二名。

Insight: 创新点在于采用多任务学习框架，通过联合优化整体MOS和六个相关的感知子属性（使用PLCC损失），使得共享的表征能够捕获比单任务模型更丰富的质量感知特征。这为图像质量评估提供了一种更细粒度和信息更丰富的建模方法。

Abstract: Low-light image enhancement algorithms (LIEAs) aim to improve the visibility of images captured under poor illumination. However, the enhancement process often introduces artifacts such as noise amplification, color shift, structural damage, and over-exposure, which degrade the perceptual quality of the enhanced images. Therefore, a reliable image quality assessment (IQA) metric for evaluating enhancement effects is of great importance for both the development of LIEAs and their practical applications. In this paper, we present \textbf{LEIQ-Assessor}, a multi-dimensional quality assessment model for low-light image enhancement based on multi-task learning, developed for the QoMEX 2026 Grand Challenge on Low-light Enhanced Image Quality Assessment. Specifically, our method leverages a pre-trained SigLIP2 Vision Transformer as the backbone and simultaneously predicts the overall Mean Opinion Score (MOS) together with six perceptual sub-attributes: lightness, color fidelity, noise level, exposure quality, naturalness, and content recovery. By jointly optimizing these correlated objectives via the PLCC loss, the shared representation captures richer quality-aware features than its single-task counterpart. Experiments on the MLE benchmark demonstrate that LEIQ-Assessor significantly outperforms existing no-reference IQA models and hand-crafted quality descriptors. Our method achieved second place in the QoMEX 2026 Grand Challenge on Low-light Enhanced Image Quality Assessment. The code is available at https://github.com/sunwei925/LEIQ-Assessor.

[128] HTC-SGA Former: A Hybrid Transformer-CNN Network with Self-Guided Attention and a New Boundary-Weighted Adaptive Loss for Coronary DSA Vessel Segmentation cs.CVPDF

Rayan Merghani Ahmed, Marwa Omer Mohammed Omer, Mohamed Elmanna, Shijie Li, Bin Li

TL;DR: 本文提出了一种名为HTC-SGA Former的轻量级混合Transformer-CNN网络，用于冠状动脉数字减影血管造影（DSA）的血管分割。该方法结合了CNN编码器提取局部血管形态和Transformer解码器建模上下文特征，并引入了多尺度全局-局部窗口注意力模块和自引导特征注意力模块来增强对微弱血管的响应。此外，还提出了一种边界加权自适应复合损失函数来强调薄血管边界并自适应地平衡血管恢复与边界细化。

Details

Motivation: 冠状动脉DSA血管分割对于计算机辅助诊断和治疗规划至关重要，但现有方法难以在全局上下文推理与保留微弱血管、血管连续性和精细边界之间取得平衡，主要挑战包括薄弱的低对比度血管、背景干扰以及严重的类别不平衡。

Result: 在私有的左右冠状动脉DSA子集上的实验表明，HTC-SGA Former在仅0.81M参数的紧凑架构下，性能优于14种最先进的分割方法。所提出的边界加权自适应复合损失函数在四种编码器-解码器架构上均优于二元交叉熵和Dice损失，展示了强大的跨骨干网络适用性。

Insight: 创新点在于通过互补的全局-局部上下文建模、针对血管的细化以及自适应优化，有效改善了薄血管恢复、血管连续性和边界定位。具体技术贡献包括多尺度全局-局部窗口注意力块、自引导特征注意力模块以及边界加权自适应复合损失函数，这些设计共同支持了可靠且计算高效的冠状动脉血管分析。

Abstract: Accurate coronary Digital Subtraction Angiography (DSA) vessel segmentation is essential for computer-aided diagnosis and treatment planning of coronary artery disease (CAD). However, thin low-contrast vessels, background interference, and severe vessel-background class imbalance make reliable segmentation of weak distal branches and vessel boundaries challenging. Existing methods struggle to balance global contextual reasoning with preservation of weak vessels, vessel continuity, and fine boundaries. To address these limitations, we propose HTC-SGA Former, a lightweight hybrid Transformer-CNN framework for coronary DSA vessel segmentation. It employs a CNN encoder for local vessel morphology extraction and a Transformer decoder for contextual feature modeling. A Multi-Scale Global-Local Window Attention (MS-GLWA) block performs efficient global-local contextual modeling, while a Self-Guided Feature Attention (SGFA) module enhances weak-vessel responses. In addition, a Boundary-Weighted Adaptive Compound Loss (BWACL) emphasizes thin-vessel boundaries and adaptively balances vessel recovery and boundary refinement. Experiments on private right and left coronary artery DSA subsets show that HTC-SGA Former outperforms 14 state-of-the-art segmentation methods while maintaining a compact architecture with only 0.81M parameters. BWACL also improves performance over binary cross-entropy and Dice losses across four encoder-decoder architectures, demonstrating strong cross-backbone applicability. HTC-SGA Former improves thin-vessel recovery, vessel continuity, and boundary localization through complementary global-local contextual modeling, vessel-focused refinement, and adaptive optimization, supporting reliable and computationally efficient coronary vessel analysis for future computer-assisted cardiovascular interventions.

[129] Early Warning Signals for OpenVLA Failure under Visual Distribution Shift cs.CV | cs.AIPDF

Dipesh Tharu Mahato, Rachel Ren

TL;DR: 本文研究了在视觉分布偏移下OpenVLA模型的故障预警信号。通过在LIBERO操作任务中分析模型前馈激活，发现其内部状态包含可线性解码的、关于近期任务失败的信息，尤其是在特定层（如第16层）的激活对遮挡导致的失败具有高预测能力（AUROC达0.972）。

Details

Motivation: 视觉语言动作模型将感知、语言接地和控制集成在单一策略中，但其在视觉条件变化下的故障难以诊断。本文旨在探索OpenVLA模型的前馈激活是否包含可解码的、关于近期任务失败的信号，以帮助诊断模型在分布偏移下的失效。

Result: 在遮挡作为主要压力测试下，OpenVLA成功率从57%降至17%。在层16拟合的逻辑回归探针，在15步预测范围内，AUROC达到0.972，AUPRC为0.352，优于均值差异方向和动作不一致基线。层间解码能力不均，层16最强，层8仍有信息量，层10较弱。针对颜色偏移和相机抖动（无需重新拟合）的评估表明，训练于遮挡的监控器在相机抖动下仍优于随机猜测。

Insight: 创新点在于提出并验证了从固定策略的OpenVLA内部激活中线性解码故障预警信号的可行性，并发现故障相关信号在模型深度上分布不均。这为诊断VLA模型在视觉分布偏移下的失败提供了一种轻量级监控思路，但研究明确指出了其局限性（如未建立因果机制、任务外泛化或可部署恢复系统）。

Abstract: Vision Language Action models combine perception, language grounding, and control in a single policy, but their failures are hard to diagnose once visual conditions shift. We test whether OpenVLA feedforward activations contain linearly decodable information about near term task failure in LIBERO manipulation rollouts. The policy is fixed throughout. We log internal activations during execution and fit lightweight monitors after the rollouts are collected. Occlusion is the main controlled stress test. It reduces OpenVLA success from $57%$ to $17%$ over $100$ episodes per condition. Under this shift, a logistic probe at layer 16 reaches AUROC $0.972$ and AUPRC $0.352$ for predicting failure within a $15$ step horizon. It outperforms both a mean difference direction and an action disagreement baseline. A sparse layer sweep finds uneven decodability across depth: layer 16 is strongest among the tested layers, layer 8 remains informative, and layer 10 is weaker. To check whether the monitor is just an occlusion detector, we also evaluate color shift and camera jitter without refitting. Color shift produces no failures in this setting, so it is a benign control rather than a failure benchmark. Camera jitter does induce failures, and the occlusion trained monitor remains above random. The result is deliberately limited: OpenVLA internal states contain failure relevant structure under controlled perceptual shift, but these experiments do not establish a causal mechanism, task held out generalization, or a deployable recovery system.

Yuan Li, Youyuan Lin, Zitang Sun, Yung-Hao Yang, Kiyofumi Miyoshi

TL;DR: 本文提出了一种名为MR-IQA的盲图像质量评估（BIQA）框架，该框架通过统一回归和排序两种学习范式的视角，将质量边际（quality margin）作为共同桥梁，并采用强化学习直接优化质量边际误差。

Details

Motivation: 现有BIQA方法通常基于回归（校准绝对分数）和排序（从序数关系恢复质量结构）两种范式，但两者关系多基于经验且未被充分探索。本文旨在从理论层面统一这两种范式，以更显式地建模质量结构。

Result: 在六个BIQA基准测试上的实验表明，MR-IQA取得了具有竞争力的综合性能；控制性对比显示，在基于回归或排序的强化学习方法中，MR-IQA实现了最高的平均PLCC/SRCC（皮尔逊线性相关系数/斯皮尔曼等级相关系数）。

Insight: 核心创新点在于将回归和排序统一为对质量边际的优化：回归拟合由分数端点诱导的边际，而排序则通过偏好概率拟合变换或符号级边际。这为理解BIQA中的质量结构建模提供了新的理论视角，并启发了直接优化边际的强化学习框架。

Abstract: Blind image quality assessment (BIQA) is commonly built on two basic learning paradigms: regression and ranking. Regression calibrates absolute scores, whereas ranking recovers quality structure from ordinal relations. Although joint regression-ranking supervision often improves BIQA, the relation between the two paradigms remains largely empirical and underexplored. In this work, we revisit what underlies regression and ranking and identify pairwise relational distance, termed quality margin, as their common bridge. Our derivation shows that, at the objective-optimization level, both paradigms fit quality margins: regression fits margins induced by score endpoints, while ranking fits transformed or sign-level margins through preference probabilities. Motivated by this insight, we propose MR-IQA, a direct quality-margin optimization framework for reinforcement learning (RL)-based BIQA. MR-IQA samples quality scores and optimizes pairwise margin errors as policy rewards, thereby modeling quality structure more explicitly. Experiments on six BIQA benchmarks show competitive general performance, and controlled comparisons demonstrate that MR-IQA achieves the strongest average PLCC/SRCC over regression- or ranking-based RL methods. Our findings provide a new insight into unifying regression and ranking, offering a theoretical basis for understanding quality-structure modeling in BIQA and beyond.

[131] TopoAgent: An Agentic Framework for Automated Topology Learning in Medical Imaging cs.CV | cs.AIPDF

Guangyu Meng, Pengfei Gu, Xueyang Li, Yiyu Shi, Erin Wolf Chambers

TL;DR: TopoAgent是一个基于大语言模型的智能体框架，用于医学影像分析中的自动化拓扑学习。它通过感知-推理-行动-反思循环，利用领域特定工具和双记忆系统，自动为给定数据集选择最合适的拓扑描述符并生成拓扑特征向量，无需针对特定任务进行训练。

Details

Motivation: 动机在于解决传统像素级深度学习方法忽视医学影像几何结构特性的问题，以及现有拓扑数据分析方法通常默认使用单一固定描述符，未能充分探索拓扑表示多样性的局限性。目前缺乏能够自动为给定数据集确定最合适拓扑描述符的LLM智能体框架。

Result: 该框架的技能集源自对15种拓扑描述符在26个数据集上使用六种分类器的系统性评估。TopoAgent能够分析输入图像及其拓扑特征，推理出最适合输入的拓扑描述符，并确定最优描述符及其配置。

Insight: 创新点在于首次提出了一个基于LLM的智能体框架来自动化医学影像的拓扑学习过程，通过智能体循环和工具集成实现无需任务特定训练的拓扑描述符选择与配置。其双记忆系统和从大规模评估中提炼的技能集是值得借鉴的设计思路。

Abstract: Topological data analysis (TDA), particularly persistent homology (PH), captures geometric structural properties in medical images (e.g., connected components, loops, shape characteristics), which conventional pixel-level deep learning approaches often neglect. While many topological descriptors are known for converting persistence diagrams (PDs) or raw images into topological feature vectors, existing methods mostly default to a single fixed descriptor (e.g., persistence images), leaving the diversity of topological representations largely unexplored. To the best of our knowledge, there is no known large language model (LLM)-based agentic framework that can automatically determine the most suitable topological descriptors for a given image dataset and produce the corresponding topological feature vectors for downstream tasks. To fill this gap, we propose \textbf{TopoAgent}, an LLM-based agentic framework that automates topology learning for medical image analysis.TopoAgent operates through a Perception–Reasoning–Action–Reflection loop supported by 21 domain-specific tools and dual memory that accumulates experience across runs. Its skill set is distilled from systematic evaluation of 15 topological descriptors across 26 datasets with six classifiers. TopoAgent analyzes input images and their topological characteristics, reasons about which topological descriptors best suit the input, and determines the optimal descriptor and its configuration, all without task-specific training.

[132] UrbanCDNet: Appearance-Robust and Boundary-Aware Bitemporal Change Detection for Korean Urban Building Monitoring cs.CVPDF

Abdirashid Omar, Jonghyuk Park

TL;DR: 本文提出UrbanCDNet，一种专门用于韩国城市建筑变化检测的孪生卷积神经网络。该网络通过外观鲁棒的多线索比较、对齐感知的中尺度差分、轻量级上下文细化、场景校准和辅助边界监督，解决了城市建筑变化检测中变化区域稀疏、外观差异大和边界精度要求高的问题。在修正的AIHub韩国基准数据集上，UrbanCDNet在F1分数和IoU指标上均优于现有基线模型和最强外部竞争者。

Details

Motivation: 韩国城市建筑变化检测面临三大挑战：变化区域稀疏、不同采集日期间外观差异显著、以及输出结果需要精确的建筑轮廓而非粗略斑块。现有方法在这些场景下性能不足，因此需要设计专门针对这些挑战的解决方案。

Result: 在锁定的测试集上，UrbanCDNet取得了0.7511的F1分数和0.6014的IoU，优于孪生U-Net基线（0.7108 F1, 0.5514 IoU）和最强的外部竞争者ChangeFormer-MIT-B0（0.7107 F1, 0.5512 IoU）。诊断性分析显示，在变化区域少于5%的稀疏子集上，F1从0.4765提升至0.6175；在高光度差异子集上，从0.6349提升至0.7285。边界检测和目标检测性能也有显著提升。

Insight: 论文的创新点在于针对特定任务（韩国城市建筑监测）设计了集成的解决方案，包括外观鲁棒的多线索比较和边界感知的监督。客观来看，其核心洞察是：在特定应用场景下，针对任务特性（如稀疏变化、大外观差异、精确边界需求）定制化的时序比较和边界监督，比单纯增加通用模型的规模更为有效。

Abstract: Urban building change detection from bi-temporal aerial imagery is important for redevelopment monitoring, infrastructure management, and unauthorized-construction screening, but Korean urban scenes remain difficult because changed regions are often sparse, appearance varies strongly between acquisition dates, and useful outputs must follow building footprints rather than coarse blobs. This paper presents UrbanCDNet, a task specific Siamese CNN that combines appearance-robust multi-cue comparison, alignment-aware middle-scale differencing, lightweight context refinement, scene calibration, and auxiliary boundary supervision. Experiments use a corrected AIHub-based Korean benchmark with 3,998 training, 503 validation, and 499 test pairs, and report changed-class precision, recall, F1, and IoU. On the locked test split, UrbanCDNet achieves 0.7335 precision, 0.7696 recall, 0.7511 F1, and 0.6014 IoU, outperforming a strong Siamese U-Net baseline (0.7108 F1, 0.5514 IoU) and the strongest external competitor, ChangeFormer-MIT-B0 (0.7107 F1, 0.5512 IoU). Additional diagnostic slicing shows that the gain is concentrated in the operating regimes that motivated the design: on the sparse-change subset with less than 5% changed area, F1 improves from 0.4765 to 0.6175, and on the high photometric-gap subset it improves from 0.6349 to 0.7285. Boundary F1 at 3-pixel tolerance rises from 0.3445 to 0.4447, while object F1 at IoU 0.3 rises from 0.0690 to 0.2258. These results indicate that, on this Korean benchmark, task-shaped temporal comparison and boundary-aware supervision matter more than generic model scale alone

[133] OP3DSG: Open-Vocabulary Part-Aware 3D Scene Graph Generation for Real-World Environments cs.CV | cs.ROPDF

Yirum Kim, Ue-Hwan Kim

TL;DR: OP3DSG是一个开放词汇、部件感知的3D场景图生成框架，旨在构建统一图结构，联合建模物体、交互部件、空间关系、功能关系和可供性。该框架通过整合物体-部件知识引导的检测与部件感知的3D融合来保留小尺寸且与交互相关的部件，并利用几何初始化的先验图结合基于LLM的优化来减少虚假关系预测，实现高效图构建。

Details

Motivation: 现有基于基础模型的开放词汇3D场景图生成方法仍以物体为中心，编码的关系信息有限，限制了其在需要细粒度理解的现实世界场景（如机器人任务）中的适用性。

Result: 在作者新提出的UniGraph3D基准（专为部件感知和多层次关系推理设计）上，OP3DSG取得了最先进的性能，并证明了其作为感知主干在多种现实世界机器人任务中的有效性。

Insight: 创新点在于提出了一个统一建模物体与部件及其多层次关系的3D场景图生成框架，并引入了部件感知融合与LLM辅助的关系优化策略。其核心是将细粒度的部件级感知与高层次的关系推理相结合，以支持更复杂的现实世界交互任务。

Abstract: 3D scene graphs (3DSGs) provide a compact and structured abstraction of 3D environments. Although advances in foundation models have enabled open-vocabulary 3DSG generation, existing approaches remain object-centric and encode limited relational information – restricting their applicability in real-world scenarios that require fine-grained understanding. We propose OP3DSG, an open-vocabulary part-aware 3DSG generation framework that constructs unified graphs that jointly model objects, interactive parts, spatial relations, functional relations, and affordances. OP3DSG integrates object-part knowledge-guided detection with part-aware 3D fusion to preserve small and interaction-relevant components, and employs a geometry-initialized prior graph with LLM-based refinement to reduce spurious relational predictions while enabling efficient graph construction. To systematically evaluate unified 3D scene graph construction, we introduce UniGraph3D, a benchmark designed for part-aware perception and multi-level relational reasoning. Experimental results show that OP3DSG achieves state-of-the-art performance and demonstrates its effectiveness as a perception backbone in diverse real-world robotics tasks.

[134] Clearer Sight, Fewer Lies: Oriented Pickup Preference Optimization for Multimodal Hallucination Mitigation cs.CVPDF

Xin Zou, Haolin Deng, Yibo Yan, Shuliang Liu, Zhiwei Jin

TL;DR: 本文针对多模态大语言模型（MLLMs）容易产生幻觉的问题，提出了一种名为定向拾取偏好优化（OPPO）的证据感知对齐目标。该方法通过对比同一忠实回答在不同视觉证据强度下的表现，引导模型学习对视觉证据强度的偏好，而非仅关注回答质量，并结合细粒度的正则化来稳定训练。

Details

Motivation: 动机源于一个实证观察：当使用模型自身的注意力机制显式增强查询相关的视觉证据时，生成结果会更准确。这表明许多失败并非仅源于感知缺失，而是源于模型对其已关注到的证据的信任倾向不足。因此，需要一种方法来校准模型的生成偏好，使其更倾向于依赖视觉证据。

Result: 在幻觉和通用基准测试上的广泛评估表明，OPPO方法在减轻幻觉方面持续优于基线方法，取得了更好的性能。

Insight: 主要创新点在于提出了OPPO这一新颖的对齐目标，它将朴素的视觉偏好转化为有序的视觉证据对齐，通过对比不同证据强度的视图来学习偏好。此外，理论分析表明有序的证据边界会诱导局部视觉敏感性的正下界，为方法提供了理论支撑。细粒度的跨度级和令牌级正则化也是稳定训练的关键技术点。

Abstract: Multimodal Large Language Models (MLLMs) are prone to hallucination as their generation preferences are insufficiently calibrated to visual evidence, causing them to fall back on linguistic priors, rather than faithful grounding. In this work, we start from an empirical observation: when query-relevant visual evidence is explicitly strengthened using the model’s own attention, generation becomes more accurate, suggesting that many failures do not arise solely from missing perception, but from an insufficient tendency to trust the evidence the model has already attended to. Motivated by this finding, we propose Oriented Pickup Preference Optimization (\texttt{OPPO}), an evidence-aware alignment objective that learns preferences over the strength of visual evidence, rather than only response quality. Concretely, \texttt{OPPO} contrasts the same faithful response under stronger, anchored, weaker-evidence views, turning naive visual preference into ordered visual-evidence alignment. We further combine this objective with fine-grained span-level and token-level regularization to stabilize the training. Besides, we provide a theoretical analysis showing that ordered evidence margins induce a positive lower bound on local visual sensitivity. Extensive evaluations across hallucination and general-purpose benchmarks demonstrate that \texttt{OPPO} consistently outperforms baseline methods.

[135] Consistency as Inductive Bias: Learning Cross-View Invariance for Robust Multimodal Reasoning cs.CVPDF

Xin Zou, Haolin Deng, Yibo Yan, Shuliang Liu, Kening Zheng

TL;DR: 本文提出ConsistRoll方法，通过跨视图一致性归纳偏置增强多模态大语言模型（MLLMs）的鲁棒性。该方法将语义不变的视图对置于同一生成组中，仅当配对补全结果均正确且一致时才分配联合奖励，从而将一致性转化为在线信用分配信号。理论分析表明跨视图一致性是有效的归纳偏置，实验在数学、通用和幻觉领域基准上验证了其稳健提升。

Details

Motivation: 现有MLLMs训练中缺乏跨视图一致性归纳偏置，即同一实例的语义不变视图应产生相同答案。标准RLVR目标仅对每个视觉输入分配点状奖励，即使使用数据增强，变换视图也通常被独立奖励，一旦视图内奖励饱和则信号有限。

Result: 在数学、通用和幻觉领域的综合基准测试中，ConsistRoll实现了多模态推理的稳健改进，证实了其有效性。

Insight: 创新点在于将跨视图一致性作为归纳偏置注入RLVR训练，通过重用GRPO的组采样机制，无需额外生成开销和标注即可将一致性转化为在线信用分配信号，理论分析揭示了其引入的跨视图校正项能惩罚视图依赖性并缓解优势崩溃。

Abstract: Inductive biases steer learning toward generalizable solutions by encoding task structure. In this work, we identify a crucial missing bias in MLLMs: cross-view consistency, \textit{i.e.}, semantically invariant views of the same instance should lead to the same answer. Standard reinforcement learning with verifiable rewards (RLVR) objectives do not impose this constraint, but instead assign pointwise rewards to each visual input. Even with data augmentation (DA), transformed views are typically rewarded independently, providing little signal once within-view rewards saturate. We propose \textbf{ConsistRoll}, a simple but effective method that injects cross-view consistency into RLVR training by reusing the group-sampling mechanism of GRPO. Specifically, ConsistRoll places original and semantically invariant transformed views in the same generation group, and assigns a joint reward only when paired completions are both correct and consistent. In this way, ConsistRoll turns consistency into an online credit-assignment signal, \textbf{without extra generation overhead and annotations}. Theoretically, we show that cross-view consistency is a valid inductive bias, and ConsistRoll introduces a cross-view correction term absent from DA, penalizing view dependence and alleviating advantage collapse. Comprehensive benchmarks across math, general-purpose, hallucination domains confirm that ConsistRoll achieves robust improvements in multimodal reasoning.

[136] Learning Cross-view Correspondences for Geo-localization on Planetary Surfaces cs.CVPDF

Hong Minh Nguyen, Marcus Märtens, Tat-Jun Chin

TL;DR: 本文提出了一个用于行星表面跨视角地理定位的新基准数据集，该数据集基于高分辨率月球地形模型生成，包含10438个地面全景图与对应顶视图图像。作者评估了基于Transformer的先进地理定位方法在该数据集上的性能，结果表明基于学习的跨视角定位方法可有效应用于行星表面，为全球导航卫星系统提供视觉替代方案。

Details

Motivation: 解决行星表面探索中因缺乏卫星定位系统和里程计漂移导致的全局位置感知难题，以及由于视角差异大、纹理低、地形重复、光照和地形变化引起的表观剧烈变化，使得轨道地图产品与地面观测对齐困难的问题。

Result: 在构建的月球表面跨视角定位基准上，从头训练基于Transformer的先进地理定位方法，并报告了检索准确率，结果表明该方法可成功应用于行星表面定位任务。

Insight: 创新点在于构建了首个基于物理渲染的行星表面跨视角地理定位基准数据集，包含精确对齐的地面全景图和顶视图瓦片，并验证了学习型跨视角定位方法在该领域的可行性，为行星导航提供了视觉基础解决方案。

Abstract: Maintaining global position awareness is a fundamental challenge for planetary surface exploration, since satellite-based positioning systems are unavailable and onboard odometry drifts over time. Although orbital mapping products, such as overhead imagery and terrain-derived maps, provide global context, aligning them with surface observations is challenging due to large viewpoint differences, low texture, repetitive terrain, and drastic changes in appearance caused by varying illumination and topography. We introduce a new cross-view geo-localization benchmark built from physically rendered surface panoramas and overhead tiles derived from a high-resolution lunar terrain model. Our dataset contains 10438 ground views rendered as 360$^\circ$ surface panoramas with matching overhead images precisely centered at the same location. Additionally, a set of overlapping tiles is provided to study off-center localization with multiple plausible candidates per panorama. We study the performance of a state-of-the-art transformer-based geo-localization method on our data, by training it from scratch and reporting retrieval accuracy. Our results demonstrate that learning-based cross-view localization methods can be successfully applied to the domain of planetary surfaces, providing a vision-based alternative to global navigation satellite systems.

[137] Robust Trajectory Distillation: Hybrid Reweighting Meets Teacher-Inspired Targets cs.CVPDF

Kaifeng Chen, Lechao Cheng, Jiyang Li, Shengeng Tang, Fan Zhang

TL;DR: 本文提出了一种基于轨迹的数据集蒸馏框架，旨在解决在噪声监督下进行数据集蒸馏时，容易将噪声与有效信号一同压缩的问题。该框架包含两个互补组件：选择性引导重加权（SGR）和教师启发的辅助目标（TIAT），共同作用以抑制噪声并保留可迁移知识，而无需重新标注或依赖干净数据子集。

Details

Motivation: 动机在于传统数据集蒸馏方法在存在噪声标签时，会同时压缩损坏的关联和有用信号，损害鲁棒性。而现有的噪声标签处理方法通常与模型优化紧密耦合，需要干净数据作为锚点，这与数据集蒸馏追求紧凑、即插即用监督的目标不符。

Result: 在对称、非对称和真实世界噪声设置下，该方法在多个基准测试中均优于最先进的数据集蒸馏基线，取得了稳定的性能提升。

Insight: 创新点在于将全局遗忘模式与局部邻域一致性融合为渐进式重加权方案（SGR），以及从中间教师动态中提取辅助残差引导作为辅助目标（TIAT）。这提供了一种无需干净数据、计算轻量且能保持标签完整性的鲁棒蒸馏新思路。

Abstract: Dataset distillation (DD) condenses large corpora into compact, information-rich subsets for efficient training and reuse. However, under noisy supervision, DD risks condensing corrupted associations together with useful signals, degrading robustness. Conventional noisy-label remedies (sample selection, loss weighting, label correction) tightly couple noise estimation with model optimization, often require clean anchors, and can amplify confirmation bias-assumptions that are misaligned with DD’s goal of compact, plug-and-play supervision. We therefore propose a trajectory-based DD framework that jointly suppresses noise and preserves transferable knowledge without relabeling or clean subsets. It comprises two complementary components: Selective Guidance Reweighting (SGR), which fuses global forgetting patterns (second-split forgetting) with local neighborhood consistency into a progressive reweighting scheme that prioritizes clean supervision along the teacher trajectory; and Teacher-Inspired Auxiliary Targets (TIAT), which inject auxiliary residual guidance distilled from intermediate teacher dynamics to reinforce informative signals while remaining internally consistent. Together, SGR and TIAT produce distilled datasets with cleaner and richer representations under noisy supervision. The framework is robust, label-preserving, computationally lightweight, and broadly applicable, yielding consistent gains over state-of-the-art DD baselines across symmetric, asymmetric, and real-world noise.

[138] See Only When Needed: Context-Aware Attention Intervention for Mitigating Hallucinations in LVLMs cs.CVPDF

Yuqing Lei, Wenbo Lyu, Yingjun Du, Xiantong Zhen, Cees G. M. Snoek

TL;DR: 本文提出了一种名为上下文感知注意力干预（CAI）的训练无关推理时机制，旨在缓解大型视觉语言模型（LVLM）中的物体幻觉问题。该方法遵循’仅在需要时查看’的原则，通过双轴选择性（看哪里、何时干预）来定位语义对齐区域，并仅对深层中不确定性激增的token进行保守的、基于熵和深度门控的注意力倾斜，从而在增强视觉基础的同时保持语言流畅性。

Details

Motivation: 现有无需训练的缓解方法通常统一增强视觉信号，这可能会放大不相关区域并引入虚假证据，损害流畅性。因此，需要一种更精细的干预机制，仅在模型不确定时针对性地加强视觉基础。

Result: 在多个LVLM主干和基准测试上的广泛实验表明，CAI在缓解幻觉方面达到了最先进的水平，并且即使不使用对比解码作为可选的辅助偏置抑制模块，也能带来一致的改进。

Insight: 创新点在于提出了一个基于双轴选择性的推理时注意力干预机制，通过早期层表示获取token特定的视觉相关性来定位语义对齐区域，并仅在深层视觉基础退化时对不确定性激增的token进行门控干预。这实现了KL最小化的注意力重加权，并在非活动门或小倾斜下具有有限干扰，是一种高效且针对性的幻觉缓解方法。

Abstract: Large Vision-Language Models (LVLMs) excel at multimodal tasks but remain prone to object hallucinations. Prior training-free remedies often uniformly strengthen visual signals, which may also amplify irrelevant regions and introduce spurious evidence, harming fluency. We propose Context-aware Attention Intervention (CAI), a training-free inference-time mechanism that enforces a see only when needed principle via two-axis selectivity: where to look and when to intervene. At each decoding step, CAI derives token-specific visual relevance from early-layer representations to localize semantically aligned regions, and applies a conservative, entropy- and depth-gated attention tilt only for uncertainty-spiking tokens in deeper layers where visual grounding degrades, leaving confident tokens and irrelevant regions largely unchanged. This targeted intervention strengthens visual grounding while preserving linguistic fluency, and it yields consistent improvements even without contrastive decoding, which remains optional as an auxiliary bias-suppression module. Extensive experiments across multiple LVLM backbones and benchmarks show that CAI achieves state-of-the-art hallucination mitigation, and our analysis characterizes CAI as a KL-minimal attention reweighting with bounded interference under inactive gates or small tilts. Code is available at https://github.com/Iris1946/CAI.

[139] Efficient Visual Pointing for Embodied AI:Agent-Driven Data Synthesis, Cross-Block Attention, and Iterative Correction cs.CVPDF

Zijian Hong, Qi Lv, Yuxiang Xie, Jianming Xing, Xiang Deng

TL;DR: 本文提出了一个用于具身AI视觉指向任务的解决方案PointArena 2026，旨在将语言指令映射到像素坐标。该方法通过智能体驱动的数据合成、确定性的可操控数据流水线以及两个模型端模块（AttnRes和ABC校正）来针对性地解决三种失败模式，最终在基准测试中取得了77.2%的整体准确率，排名第二。

Details

Motivation: 视觉指向是具身AI的核心技能，但现有方法存在多种失败模式。本文旨在通过系统性的方法解决这些失败模式，提升模型在复杂语言指令下的坐标定位能力。

Result: 该方法在PointArena 2026基准测试中取得了77.2%的整体准确率，排名第二。在本地验证中，不同任务类别的专家模型表现优异，其中可操作性任务达到93.9%，空间关系任务达到82.6%。

Insight: 创新点包括：1) 智能体驱动的数据合成方法，构建了大规模、多样化的候选池；2) 确定性的、可操控的数据流水线，确保了高质量的训练数据；3) 模型端引入门控跨块注意力模块（AttnRes）增强可操控性，以及结合视觉特征的坐标扰动校正模块（ABC）来提升通用坐标定位能力；4) 采用类别感知路由来组合互补的专家模型。

Abstract: Visual pointing maps a language instruction to pixel co ordinates, a core skill for embodied AI. We describe our PointArena 2026 solution, which achieves 77.2% overall accuracy and ranks second on the benchmark. The ap proach targets three failure modes. First, agent-driven syn thesis builds large semantic and anchor-relative candidate pools; the server inventory contains 55,372 processed out puts, 53,772 de-duplicated sample IDs, and 37,574 train able completed or accepted rows. Second, a determinis tic steerable-data pipeline creates a verified 10,000-sample main set, plus reserve samples, using masks, templates, and path verification. Third, two model-side modules address complementary errors: AttnRes adds gated cross-block at tention for steerability, while ABC correction encodes per turbed coordinates with visual features for general coordi nate grounding. Category-aware routing combines comple mentary specialists; local validation used to select experts records 93.9% Affordance, 82.6% Spatial Relation, 78.2% Reasoning, 70.4% Counting, and 63.0% Steerability.

[140] SUMO: Segment and Track Any Motion with Nonlinear State Space Models cs.CV | cs.AIPDF

Kexin Tian, Sixu Li, Keshu Wu, Yang Zhou, Zhengzhong Tu

TL;DR: 论文提出SUMO框架，一种零样本、无需训练的统一方法，通过非线性状态空间模型结合视觉分割，实现精确且一致的视觉目标跟踪（VOT）和运动目标分割（MOS）。该方法利用机器人学原理建模复杂目标动态，并引入选择性无迹滤波器进行状态估计，通过联合评分机制动态融合多源预测。实验表明SUMO在VOT和MOS任务上达到最先进性能。

Details

Motivation: 现有方法主要依赖视觉线索，在现实世界中目标运动复杂且非线性的场景下容易失效，因此需要整合非线性动力学以提升VOT和MOS的准确性和一致性。

Result: SUMO在视觉目标跟踪和运动目标分割任务上实现了最先进（SOTA）性能，通过广泛实验验证了其有效性。

Insight: 创新点包括：基于机器人学原理的非线性状态空间模型来捕获复杂动态；选择性无迹滤波器（SUF）结合联合评分和动态多源预测融合，提升状态估计精度；以及内存选择机制评估帧可靠性。这些设计提供了零样本、训练免费的统一框架，可推广到其他动态建模任务。

Abstract: Visual Object Tracking (VOT) and Moving Object Segmentation (MOS) are two fundamental tasks in computer vision that involve both spatial and temporal object dynamics. Existing methods rely predominantly on visual cues and thus often falter in real-world scenarios where object motions are inherently complex and nonlinear. To address this limitation, we propose SUMO, a zero-shot, training-free, unified framework integrating nonlinear dynamics with vision-based segmentation for accurate and consistent VOT and MOS. Specifically, we develop a nonlinear State Space Model (SSM) inspired by robotics principles to capture the complex object dynamics. Building on this model, we propose a Selective Unscented Filter (SUF) for accurate state estimation, which features a joint scoring mechanism and dynamically fuses multi-source predictions to identify the most plausible object state over time. Furthermore, we apply a memory selection mechanism to evaluate the reliability of memory frames. Our extensive experimental results show that SUMO achieves state-of-the-art performance on both VOT and MOS tasks.

[141] LWDrive: Layer-Wise World-Model-Guided Vision-Language Model Planning for Autonomous Driving cs.CV | cs.AIPDF

Chen Yang, Yuhao Wei, Ze Xu, Ziheng Zou, Shuang Liang

TL;DR: 本文提出了LWDrive框架，一种用于端到端自动驾驶规划的层式世界模型引导的视觉语言模型（VLM）规划方法。该方法将VLM生成的粗略轨迹作为意图感知的初始规划，通过前瞻级联规划器（FCP）利用多层VLM特征、历史状态、动作查询表示和多视角鸟瞰图特征，对候选轨迹进行由粗到细的渐进式精炼，最终选择最优轨迹作为规划输出。

Details

Motivation: 现有基于视觉语言模型（VLM）的端到端自动驾驶规划方法生成的轨迹往往只编码了粗略的驾驶意图，在几何精度、未来感知和多视角场景基础方面存在不足。

Result: 实验表明，LWDrive在NAVSIM基准测试上取得了92.0分，在NAVSIM-v2上取得了89.6分，展现了其有效性。

Insight: 核心创新点在于提出了层式世界模型引导的规划框架，通过引入未来帧生成监督使VLM学习前瞻性场景表示，并设计了前瞻级联规划器（FCP）来渐进式地精炼候选轨迹，从而将高层驾驶意图与几何精确、多视角基础的低层轨迹优化相结合。

Abstract: Vision-Language Models (VLMs) provide powerful semantic understanding and commonsense reasoning for End-to-End Autonomous Driving (E2E-AD) planning. However, trajectories directly generated by VLMs often encode only coarse driving intentions and remain insufficient for geometrically accurate, future-aware, and multi-view-grounded planning. To address these limitations, we develop the Layer-Wise World-Model-Guided Driving framework (LWDrive). LWDrive is a VLM planning framework that refines coarse trajectories through layer-wise world-model guidance. Instead of treating the VLM output as the final trajectory, LWDrive uses it as an intent-aware coarse plan, expands a diverse candidate space around it, and progressively refines the candidates through a Foresight Cascade Planner (FCP). Specifically, we introduce future-frame generation supervision to encourage the VLM to learn forward-looking scene representations, thereby injecting planning-relevant predictive dynamics into its internal hidden states. Built upon these world-model-supervised representations, FCP exploits VLM features across multiple layers and integrates historical temporal states, Action-Query representations, and current-frame multi-view Bird’s-Eye-View (BEV) features to refine candidate trajectories in a coarse-to-fine manner. This design enables progressive correction of spatial positions and motion trends while grounding trajectory refinement with multi-view scene cues and preserving the high-level driving intention produced by the large model. Finally, a score head evaluates the refined candidates and selects the best trajectory as the final planning output. Experiments show that LWDrive achieves a score of 92.0 on the NAVSIM benchmark and 89.6 on NAVSIM-v2. Code and models will be made publicly available.

[142] Building artificial intelligence virtual tissue (AIVT) for tissue state representation, feature prediction, and dynamic simulation cs.CVPDF

Qiqi Lu, Qianjin Feng, Shaoqun Zeng, Shenghua Cheng

TL;DR: 本文提出了人工智能虚拟组织（AIVT）的概念，这是一个基于空间多模态数据的AI框架，旨在对健康和疾病状态下的组织进行建模。AIVT旨在学习统一、空间解析且动态可操纵的组织状态表示，从而实现组织状态表征与分析、分子与形态特征预测以及时空组织动态模拟。

Details

Motivation: 传统计算建模方法难以捕捉组织作为空间组织化、多尺度生物系统的复杂性，而AI在表征复杂系统方面展现出卓越能力，为描述组织状态及其转变创造了新机会。

Result: 论文未在摘要中提及具体的定量实验结果或基准测试，而是提出了一个概念性框架及其核心能力。

Insight: 创新点在于提出了一个整合空间多模态数据的AI驱动组织建模统一框架，其核心是学习可动态操纵的空间解析表示，以实现从表征、预测到模拟的多功能集成，为理解组织稳态和病理重塑提供了新的计算范式。

Abstract: Modeling tissue states and their transitions is essential for understanding tissue homeostasis in health and pathological remodeling in disease. However, conventional computational modeling approaches are inadequate to capture the complexity of tissues as spatially organized, multiscale biological systems. Artificial intelligence (AI) has shown a remarkable ability for representing intricate systems, creating new opportunities to characterize tissue states and their transitions. Here, we propose the concept of AI virtual tissue (AIVT), an AI framework grounded in spatial multimodal data for modeling tissues in health and disease. AIVT is designed to learn unified, spatially resolved, and dynamically manipulatable representations of tissue state, enabling tissue state representation and analysis, molecular and morphological feature prediction, and simulation of spatiotemporal tissue dynamics. We outline the fundamental assumptions, core capabilities, architectural components, as well as data and algorithm foundations of AIVT as a framework for AI-driven tissue modeling.

[143] LLM-based Multimodal Personality Recognition via Facial Action Unit-Text Semantic Fusion cs.CV | cs.AIPDF

Tianyi Zhang, Wei Shan, Yuan Zong, Tianhua Qi, Wenming Zheng

TL;DR: 本文提出了一种基于大语言模型（LLM）的多模态人格识别框架，通过将面部动作单元（AU）序列转化为可解释的文本描述，并与面试者的文本回答进行语义融合，以提升异步视频面试（AVI）中人格评估的准确性。该方法在AVI-6基准测试上取得了优于多数基线模型的结果，预测误差更低，与人工评分相关性更强。

Details

Motivation: 现有方法要么仅依赖LLM分析文本回答，丢失了面部线索信息；要么采用全脸图像或稀疏采样帧的多模态方法，忽略了对于准确人格评估至关重要的细粒度时间动态。本文旨在克服这些局限，充分利用面部动作单元提供的非语言线索。

Result: 在AVI-6基准测试上的实验表明，该方法在多个特质上相比大多数基线模型取得了持续改进，具有更低的预测误差和与人类评分更强的相关性。

Insight: 创新点在于将面部动作单元序列语义化（转化为文本描述）并与文本回答在LLM中进行融合，这提供了与文本互补的非语言线索。此外，在LLM内部将语义理解与回归预测解耦，提高了训练稳定性和模型可解释性，形成了一个心理依据充分且计算高效的框架。

Abstract: Personality recognition in asynchronous video interviews (AVIs) has become increasingly important due to their widespread adoption in modern recruitment. Existing approaches often rely on large language models (LLMs) to analyze textual responses of interviewees in AVI. However, unimodel methods often suffer from information loss (e.g., ignore facial cues). In contrast, multimodal methods that employ full-face images or sparsely sampled frames can discard fine-grained temporal dynamics critical for accurate personality assessment. To overcome these limitations, we propose an LLM-based framework that semantically fuse facial action units (AUs) with textual responses of AVI. AU sequences are first converted into interpretable textual descriptions, which are then fused with participants’ textual responses through an LLM. A lightweight regression head transforms the resulting embeddings into continuous personality scores without disrupting the underlying semantic space. Experiments on the AVI-6 benchmark demonstrate consistent improvements over most baselines, with lower prediction errors and stronger correlations with human-rated scores across multiple traits. Further analysis reveals that AU-derived semantic representations offer complementary non-verbal cues to textual responses. Decoupling semantic understanding from regression prediction within the LLM also leads to greater training stability and clearer interpretability. Overall, these findings demonstrate that AU-text fusion provides a psychologically grounded and computationally efficient framework for personality recognition in AVIs.

[144] StrucTab: A Structured Optimization Framework for Table Parsing cs.CVPDF

Gengluo Li, Shangpin Peng, Chengquan Zhang, Binghong Wu, Hao Feng

TL;DR: 本文提出StrucTab，一个用于表格解析的结构化优化框架。该框架通过将解析过程分解为行列计数、合并单元格分析等子任务进行顺序推理，并引入Uni-TabRL强化学习框架，利用分解的奖励（有效性、结构和内容）提供稳定的优化信号。此外，论文还发布了TableVerse-5K大规模基准数据集。实验表明，StrucTab在多个公开基准上达到了最先进的性能。

Details

Motivation: 现有基于视觉语言模型（VLM）的端到端表格解析方法通常直接监督最终输出，忽略了理解复杂表格结构所需的中层显式推理。同时，使用强化学习（RL）优化这些模型时，往往因奖励设计不稳定或不明确而受限。

Result: 广泛的实验表明，StrucTab在所有评估的公开基准测试中均取得了最先进的（SOTA）性能，并在新提出的大规模、多样化真实场景基准TableVerse-5K上取得了显著改进。

Insight: 主要创新点在于将表格解析任务分解为人类启发的子任务进行顺序推理建模，并设计了统一的强化学习框架（Uni-TabRL），通过分解奖励来解决奖励设计不稳定的问题。从客观角度看，其提出的结构化建模思路和奖励分解策略对于需要复杂空间和文本联合理解的任务具有借鉴意义。

Abstract: Table parsing aims to convert table images into structured, machine-readable representations, a task requiring the joint perception of complex spatial layouts and textual content. While recent vision-language models (VLMs) enable end-to-end parsing, they typically rely on direct supervision of the final output, thereby bypassing the explicit intermediate reasoning that is crucial for understanding complex table structures. Furthermore, attempts to optimize these models using reinforcement learning (RL) are often hindered by unstable or ambiguous reward designs, limiting potential performance gains. To address these limitations, we propose StrucTab, a table parsing model learned through intermediate structural supervision and reward decomposition. At the modeling level, by decomposing the parsing process into human-inspired subtasks, such as row-column counting and merged-cell analysis, StrucTab progressively unifies them through a sequential reasoning strategy. At the optimization level, we introduce Uni-TabRL, a unified RL framework that leverages decomposed rewards (validity, structure, and content) to provide stable and informative optimization signals. Finally, at the evaluation level, we present TableVerse-5K, a large-scale, challenging benchmark encompassing diverse, real-world table scenarios. Extensive experiments demonstrate the state-of-the-art performance of StrucTab across all evaluated public benchmarks and significant improvements on TableVerse-5K, validating the effectiveness of explicit structural modeling and decomposed reward optimization. Code and benchmark are publicly available at https://github.com/VirtualLUOUCAS/StrucTab.

[145] Traffic-CBM: A Structurally Interpretable Multimodal Framework for Encrypted Traffic Classification cs.CVPDF

Honglei Jin, Wenshuo Chen, Shaofeng Liang, Haozhe Jia, Menshuo Zhao

TL;DR: 本文提出Traffic-CBM，一种结构可解释的多模态加密流量分类框架。该框架将异构流量证据（如流统计、数据包序列和字节级表示）组织到统一的分层概念空间中，而非直接融合为黑盒表示，从而提升了分类决策的可解释性。

Details

Motivation: 现有加密流量分类方法通常将多种流量证据融合为不透明的潜在特征，导致预测过程难以解释，无法明确何种证据驱动了分类决策。

Result: 在多个加密流量基准测试上，Traffic-CBM取得了有竞争力且均衡的分类性能，同时比传统的端到端融合模型提供了更清晰的结构化解释接口。

Insight: 创新点在于通过预定义的流量证据组约束，将异构证据映射为标量证据摘要（概念），构建了分层概念空间，使不同层次的流量证据变得易于分析，实现了结构化的可解释性，而非依赖手动标注的语义属性。

Abstract: Encrypted traffic classification has achieved strong performance, but its decision process remains difficult to interpret. Existing methods usually combine flow statistics, packet sequences, and byte-level representations into opaque latent features, making it unclear which type of evidence actually drives the prediction. In this paper, we propose Traffic-CBM, a structurally interpretable multimodal framework for encrypted traffic classification. Instead of directly fusing heterogeneous traffic signals into a black-box representation, Traffic-CBM organizes them into a unified hierarchical concept space. These concepts are not manually annotated semantic attributes; rather, they are scalar evidence summaries constrained by predefined traffic evidence groups. More specifically, grouped flow statistics are mapped to statistical concepts, dedicated temporal encoders learn temporal concepts from disjoint feature subspaces, and byte-level evidence is further organized into packet-level and cross-packet concepts. This design turns heterogeneous traffic evidence into an explicit concept representation and makes different levels of traffic evidence easier to analyze. We evaluate Traffic-CBM on multiple encrypted traffic benchmarks. Results show that it achieves competitive and balanced classification performance while providing a clearer structural interpretation interface than conventional end-to-end fusion models. Further analyses suggest that the learned concept space is actively used in the prediction process and provides a clearer structural explanation of multimodal traffic evidence.

[146] SHOVIR: A Benchmark for Evaluating Vision Shortcut Learning in Radiology Report Generation cs.CV | cs.CLPDF

Filippo Ruffini, Marco Salmé, Rosa Sicilia, Valerio Guarrasi, Paolo Soda

TL;DR: 该论文提出了SHOVIR基准，用于评估放射学报告生成（RRG）中视觉语言模型（VLMs）的‘视觉捷径’学习行为。SHOVIR通过扩展两个带空间标注的胸部X光数据集（MIMIC-CXR和PadChest-GR），并设计图像级和疾病级的遮挡实验，来测试模型预测是否真正基于图像中的病理证据，而非依赖先验知识或虚假关联。

Details

Motivation: 当前RRG评估协议依赖于报告级指标（如词汇重叠或临床正确性聚合），这些指标无法验证单个诊断陈述是否源自图像中实际可见的病理证据，导致模型可能通过利用学习到的先验或虚假相关性（即‘视觉捷径’）获得有竞争力的分数。

Result: 在SHOVIR基准上对八个SOTA VLM进行测试，发现不同架构和数据集间的捷径行为差异显著。基线报告质量最高的模型在空间基础性（即依赖视觉证据）方面不一定排名最高，表明临床流畅的生成可能与对视觉证据的浅层依赖共存。

Insight: 主要创新点在于提出了首个专门评估RRG中视觉捷径行为的基准，通过引入基于空间标注的遮挡实验，能够在疾病类别层面分离出‘直接捷径’和‘上下文捷径’两种失败模式。这揭示了当前RRG评估中的一个盲点，并推动了面向区域的评估协议的发展。

Abstract: Current evaluation protocols for Vision-Language Models (VLMs) in Radiology Report Generation (RRG) rely on report-level metrics that measure lexical overlap or aggregate clinical correctness. However, such metrics do not test whether individual diagnostic statements stem from the actual pathological evidence visible in the image. This allows models to achieve competitive scores by exploiting learned priors or spurious correlations, a failure mode we refer to as vision shortcut. We introduce SHOVIR, a benchmark for evaluating vision shortcut behavior in RRG. SHOVIR extends two spatially annotated chest X-ray datasets, MIMIC-CXR and PadChest-GR, with per-box CheXpert labels, and defines image-level and disease-level occlusion experiments that contrast baseline performance on clean images against localized, region-specific perturbations. Comparing predictions across these conditions isolates two failure modes at the disease-class level: direct shortcuts, where a finding persists after its visual evidence is removed, and contextual shortcuts, where detection degrades once co-occurring pathologies are occluded despite the target region remaining intact. Benchmarking eight state-of-the-art VLMs, we find that shortcut behavior varies substantially across architectures and datasets. Models achieving the highest baseline report quality do not necessarily rank highest in spatial grounding, revealing that clinically fluent generation can coexist with shallow reliance on visual evidence. These findings expose a blind spot in current RRG evaluation and motivate region-aware assessment protocols.

[147] H-GRPO: Permutation-Invariant Reinforcement Learning for Grounded Visual Reasoning cs.CVPDF

Eric Peh, Debaditya Roy, Basura Fernando

TL;DR: 本文提出H-GRPO框架，通过分解式证据定位来提升视觉语言模型在视觉推理任务中的性能和可解释性。该方法将全局查询分解为一系列原子子问题，要求模型为每个子问题提供明确的子答案和定位证据框，从而构建结构化的、类似人类演绎的推理路径。

Details

Motivation: 解决当前视觉语言模型在基准测试中表现良好但仍是’黑盒’、容易产生幻觉或依赖表面捷径的问题，旨在增强模型的性能和可解释性。

Result: 摘要中未提及具体的定量结果、基准测试或与SOTA的比较。

Insight: 创新点在于提出了分解式证据定位方法，强制模型将推理过程分解为可验证的视觉事实步骤，并通过定位框进行证据锚定，这增强了推理过程的透明度和可靠性。

Abstract: Vision-Language Models (VLMs) often achieve high performance on benchmarks while remaining “black boxes”, yet they remain prone to hallucination or rely on superficial shortcuts. In this work, we propose a framework designed to enhance both performance and interpretability through De-compositional Evidence Grounding. Unlike monolithic inference approaches, our approach forces the model to decompose a global query into a sequence of atomic sub-questions, each requiring an explicit sub-answer and critically a localized evidence bounding box. By grounding intermediate logical steps (e.g. identifying a container, analyzing liquid properties, and assessing environmental context) in specific visual regions, we construct a structured reasoning path that mirrors human-like deduction. This allows the final answer to emerge as a logical consequence of verified visual facts rather than a statistical guess.

[148] Latent-CURE for Breast Cancer Diagnosis cs.CV | cs.AIPDF

Weiyi Zhao, Xiaoyu Tan, Lu Gan, Liang Liu, Xihe Qiu

TL;DR: 本文提出了一种名为Latent-CURE的新型乳腺癌超声诊断框架，旨在解决现有多模态大模型在流行病学数据极度不平衡的现实场景中，容易陷入捷径学习、忽略关键恶性指标的问题。该框架基于潜在空间推理，采用非对称加权思维链方法，引导模型先顺序推断标准化的BI-RADS形态学描述符，再得出最终诊断，并结合双非对称优化策略来保护稀有恶性特征。

Details

Motivation: 现有基于多模态大模型的自动化乳腺癌超声诊断框架多为不透明的端到端范式，过度依赖全局统计相关性而缺乏结构化的临床推理，在现实世界数据极端不平衡的情况下，容易学习到捷径（例如，偏向于占主导地位的良性模式），从而忽略罕见但决定性的恶性指标。

Result: 综合评估表明，该知识注入方法在数据不平衡的医疗队列中，不仅提供了透明的临床证据，而且实现了稳健、准确的诊断性能。

Insight: 创新点在于将潜在空间推理与非对称加权思维链相结合，构建隐式推理轨迹，强制模型进行结构化的临床推理（先形态描述后诊断），并设计了双非对称优化策略来动态调整边界和权重，以保护高特异性的恶性描述符不被常见的良性先验所淹没，这增强了模型在数据不平衡下的鲁棒性和可解释性。

Abstract: Multimodal Large Models have significantly advanced automated breast ultrasound diagnosis. However, most existing frameworks utilize opaque, end-to-end paradigms prioritizing global statistical correlations over structured clinical reasoning. Consequently, these models remain susceptible to shortcut learning amid extreme real-world epidemiological imbalances, often bypassing rare but decisive malignant indicators for dominant benign patterns. To address this disconnect, we propose Latent-CURE, a novel diagnostic framework driven by asymmetric weighted chain-of-thought methodology grounded in latent space reasoning. Unlike traditional approaches, our framework constructs an implicit reasoning trajectory forcing the model to sequentially infer standardized BI-RADS morphological descriptors before converging on a final diagnosis. Furthermore, to combat the extreme scarcity of critical malignant features, we couple this architecture with a dual-asymmetric optimization strategy. By dynamically adjusting margins and weights, this strategy safeguards high-specificity malignant descriptors from being overshadowed by common benign priors. Comprehensive evaluations demonstrate that our knowledge-injected approach provides transparent clinical evidence while achieving robust, accurate diagnostic performance in imbalanced medical cohorts.

[149] Learning Efficient 4D Gaussian Representations from Monocular Videos with Flow Splatting cs.CVPDF

Shengjun Zhang, Jinzhao Li, Xin Fei, Yueqi Duan

TL;DR: 本文提出Flow Splatting方法，用于从单目视频中学习高效的4D高斯表示以重建动态3D场景。该方法通过构建速度场，并扩展传统的溅射技术来渲染光流，从而监督从单目视频中学习动态信息的过程。与现有方法相比，该方法在减少训练时间、提高渲染速度的同时，实现了更好的图像质量。

Details

Motivation: 从单目视频重建动态3D场景面临场景复杂性和时序动态的挑战。现有基于3D高斯溅射扩展至4D域的方法存在训练时间长、渲染速度慢或内存消耗高的问题，且未能充分利用密集动态信息。

Result: 在多个基准测试上的实验表明，与最先进方法相比，该方法在更少的时间消耗和更高的渲染速度下，实现了更好的图像质量。

Insight: 核心创新点在于提出Flow Splatting，通过构建和近似速度场，并扩展体积渲染策略来溅射速度场（考虑相机运动影响），从而高效监督动态学习过程。这提供了一种利用光流监督来更有效建模4D时空动态的新思路。

Abstract: Reconstructing dynamic 3D scenes from monocular videos is challenging due to scene complexity and temporal dynamics. With the advancement of 3D Gaussian Splatting in novel view synthesis, existing methods extend 3D Gaussians to 4D domain with deformation fields, trajectories or spatiotemporal 4D volumes to model scene element deformation. However, these methods suffer from long training time, low rendering speed or high memory consumption for per-frame reconstruction of 4D volumes, without fully exploiting dense dynamic information. To address this issue, we propose Flow Splatting, which constructs the velocity field and enables the conventional splatting technique to render optical flow from the velocity field to supervise dynamics learning process from monocular videos. Specifically, we extend 4D volumes with time varying means and covariance to represent complex dynamics. Then, we construct and approximate the velocity field naturally based on this representations. While conventional volume rendering techniques support to render color fields, we extend the volume rendering strategy to splat the velocity field by considering the influence of camera motions. We conduct experiments on various benchmarks to demonstrate the efficiency and effectiveness of our method. Compared to the state-of-the-art methods, our model achieves better image quality with less time consumption and higher rendering speed.

[150] Rigel: Self-Distilled Score Adaptation for Image and Video Captioning Evaluation cs.CVPDF

Shuitsu Koyama, Kazuki Matsuda, Yuiga Wada, Shinnosuke Hirano, Daichi Yashima

TL;DR: 本文提出了Rigel，一种用于图像和视频描述自动评估的指标，它基于自蒸馏分数适应方法。该方法通过从冻结的大型语言模型（LLM）中蒸馏出专门用于评估的评分头，在任务对齐的空间中捕获判断信号，避免了依赖大规模词汇集的问题，并使用人类判断数据对LLM主干进行微调。

Details

Motivation: 现有的自动评估指标与人类判断的一致性有限，而近期基于LLM的评估方法（LLM-as-a-Judge）虽然有所改进，但仍存在大规模语言建模与小型标签集评估之间的不匹配问题。

Result: 在多个基准测试上的实验表明，Rigel优于最先进的评估指标，在无参考设置下的ActivityNet-Fact基准上取得了超过10分的提升。

Insight: 创新点在于提出了自蒸馏分数适应框架，通过任务特定的评分头将LLM的通用知识适配到评估任务上，从而更有效地利用人类标注数据进行微调，提升了评估与人类判断的一致性。

Abstract: Automatic evaluation of image and video captioning is essential for benchmarking multimodal systems, although standard evaluation metrics show limited alignment with human judgments. Recent approaches using large language models (LLMs), commonly referred to as LLM-as-a-Judge, have improved alignment with human judgments but still suffer from a mismatch between large-vocabulary language modeling and evaluation over a small label set. To address this, we propose Rigel, an automatic evaluation metric for image and video captioning, based on self-distilled score adaptation. The metric employs an evaluation-specific scoring head distilled from a frozen LLM, which captures judgment signals in a task-aligned space without relying on large-vocabulary token sets. We then refine the LLM backbone with human judgment data. To train Rigel, we constructed the Vid-Lepus dataset, which contains 3,338 video clips, 33,380 reference captions, and 5,637 candidate captions. Experiments on multiple benchmarks show that Rigel outperforms state-of-the-art metrics, achieving over 10-point improvements on ActivityNet-Fact in the reference-free setting.

[151] SICAGE: Speaker-Independent Culture-Aware Gesture Generation using TED4C-L Dataset cs.CV | cs.GR | cs.HC | cs.SDPDF

Ariel Gjaci, Antonio Sgorbissa, Vittorio Murino

TL;DR: 本文提出了SICAGE框架，用于生成与文化背景相符且独立于说话者的伴随语音手势。该框架通过对抗学习和Fishr正则化等方法，从音频和文本中学习独立于说话者的文化表征，并利用这些表征条件化一个名为ALaDiT的实时扩散手势生成器。为验证方法，作者构建了包含764位演讲者、106小时的TED4C-L多模态数据集。

Details

Motivation: 现有伴随语音手势生成方法常忽视文化差异，且文化条件化模型很少在说话者不相交的数据分割下评估，导致观察到的’文化’行为可能与说话者特定的手势风格相混淆。

Result: 在TED4C-L数据集上的实验表明，SICAGE在运动真实性、多样性、节拍同步性、语义相关性和文化一致性方面均有提升。

Insight: 创新点在于提出了一个模块化框架，通过域泛化技术学习独立于说话者的文化表征，并设计了一个高效的实时扩散生成器ALaDiT来整合这些表征，从而解耦文化风格与说话者身份。

Abstract: Recent co-speech gesture generation methods often overlook cultural differences, limiting their effectiveness in human-agent interaction. Moreover, culture-conditioned models are rarely evaluated under speaker-disjoint splits, so apparent “cultural” behavior may be confounded with speaker-specific gesturing style. We introduce SICAGE, a modular framework for culture-aware co-speech gesture generation that conditions motion synthesis models on speaker-independent cultural representations. SICAGE learns these representations from audio and text by treating each speaker as a separate domain while imposing invariance across speakers. This encourages representations to remain culture-discriminative while reducing dependence on speaker identity. The resulting cultural embeddings condition a multimodal generator to produce culturally appropriate gestures. We instantiate this idea with two domain generalization approaches: adversarial learning and Fishr regularization. We further introduce ALaDiT, a real-time diffusion-based gesture generator designed to efficiently incorporate the learned cultural embeddings. To validate our method, we built TED4C-L, a 106-hour multimodal dataset of 764 TED speakers from four cultural groups. Experiments show that SICAGE improves motion realism, diversity, beat synchronization, semantic relevance, and cultural consistency.

[152] Shell-Supervised Gaussian Splatting for Urban Real-to-Sim Reconstruction cs.CVPDF

Yuan Yang, Peijun Lu, Fangzhou Lu, Sai Fan, Siqi Yan

TL;DR: 本文提出了一种名为壳监督高斯泼溅的方法，用于城市真实到仿真重建。该方法利用外部建筑立面结构壳作为轻量级几何监督，通过视频驱动的高斯重建框架，优化重建几何的稳定性和一致性，同时保持渲染质量。

Details

Motivation: 面向具身AI的真实到仿真重建需要能够支持碰撞推理、导航和智能体-环境交互的几何结构，而不仅仅是逼真的新视角合成。然而，近景城市立面由于玻璃、反射、重复窗户和弱纹理等问题，容易导致视觉上合理但几何不稳定的重建结果。

Result: 在匿名近景城市立面场景上的实验表明，该方法在立面朝向和可见表面点云一致性方面优于仅使用照片、单目线索和面向表面的高斯基线方法，同时保持了相当的保留渲染质量。

Insight: 核心创新点在于引入外部结构壳作为几何监督，并通过掩码门控损失在优化过程中仅对可见的、有壳支撑的立面区域进行正则化，从而在保持RGB驱动外观的同时，有效提升几何稳定性。这为视频到3D重建提供了一种结合轻量级先验知识的新思路。

Abstract: Real-to-sim reconstruction for embodied AI requires geometry that is useful for collision reasoning, navigation, and agent-environment interaction, not only photorealistic novel-view synthesis. However, close-range urban facades are difficult for video-to-3D reconstruction: glass, reflections, repeated windows, and weak texture can produce visually plausible renderings with unstable surface geometry. We introduce shell-supervised Gaussian Splatting, a reconstruction-stage framework that uses an external facade structural shell as lightweight geometric supervision for video-driven Gaussian reconstruction. The method aligns an exterior shell to the video reconstruction frame, renders per-view depth, camera-space normal, and valid-mask maps, and applies these cues through mask-gated losses during Gaussian optimization. This design preserves RGB-driven appearance while regularizing only visible shell-supported facade regions. Experiments on anonymized close-range urban facade scenes show improved facade orientation and visible-surface point-cloud consistency over photo-only, monocular-cue, and surface-oriented Gaussian baselines, while maintaining comparable held-out rendering quality.

[153] GeoEdit: Geometry-Aware Object Editing via Dual-Branch Denoising cs.CVPDF

Yi He, Jiangming Wang, Xinyu Wang, Mark Fong, Songchun Zhang

TL;DR: GeoEdit提出了一种无需训练的Lift-Manipulate-Render-Denoise流程，用于在单张照片中精确操控物体（平移、旋转、缩放）并遵守3D物理约束。该方法通过将场景和物体在3D空间解耦、对齐，并渲染出带有结构深度图的几何对齐代理，然后利用双分支去噪阶段（结合视频扩散主干和方差同质注入的3D约束）来优化结果，解决了现有2D方法缺乏空间感知和产生透视违规的问题。

Details

Motivation: 解决基于扩散的编辑器在单张照片中精确操控物体（如平移、旋转、缩放）时，难以遵守3D物理约束的问题。现有2D方法缺乏空间感知，导致透视违规，而将结构代理强行嵌入潜在空间会破坏方差同质性并引起自注意力泄漏，产生重影和背景模糊。核心难点在于不对称性：重定位的物体必须遵循刚性几何，而露出的背景则需要自由合成合理内容。

Result: 论文在提出的GeoEditBench（一个涵盖物体平移、物体旋转和相机运动的姿态感知基准）上进行了实验，结果表明在几何精度、身份保真度和背景质量方面均取得了持续提升。

Insight: 主要创新点包括：1) 无需训练的Lift-Manipulate-Render-Denoise流程，在3D空间解耦场景与物体以实现几何对齐；2) 双分支去噪阶段，结合视频扩散主干保持物体身份，并通过方差同质注入（在匹配噪声方差的狭窄去噪窗口内）将3D约束注入前景，避免干扰自注意力；3) 引入姿态感知的基准GeoEditBench及相应评估指标，用于系统评测物体操控任务。

Abstract: Precisely manipulating objects in a single photograph (translation, rotation, scaling) while obeying 3D physical constraints remains unsolved for diffusion-based editors. Current 2D methods lack spatial awareness and produce perspective violations. Forcing structural proxies into the latent space also disrupts variance homogeneity, and the resulting self-attention leakage leads to ghosting and background blur. The core difficulty is asymmetric: the relocated object must follow a rigid geometry, yet the uncovered background needs freedom to synthesize plausible content. We present GeoEdit, a training-free Lift-Manipulate-Render-Denoise pipeline that satisfies both constraints. We decouple scene and object in 3D, align them through point correspondence, and render a geometry-aligned proxy with a structural depth map. A Dual-Branch Denoising stage then refines this proxy: a video diffusion backbone preserves object identity, while 3D constraints are injected into the foreground within a narrow denoising window at matching noise variance (variance-homogeneous injection). The background denoises freely. Because the injected signal matches the native latent statistics, self-attention stays undisturbed. We also introduce GeoEditBench, a pose-aware benchmark covering object translation, object rotation, and camera movement with pose-aware evaluation metrics. Experiments confirm consistent gains in geometric accuracy, identity fidelity, and background quality. Our codes are available at https://github.com/Heey731/GeoEdit.

[154] OmniDance: Multimodal Driven Dance Video Generation with Large-scale Internet Data cs.CVPDF

Kaixing Yang, Jiashu Zhu, Xulong Tang, Ziqiao Peng, Xiangyue Zhang

TL;DR: 本文提出了OmniDance框架，用于解决音乐驱动舞蹈视频生成中数据集规模小、音乐信号难以有效融入视频生成基础模型的问题。通过构建大规模高质量数据集CIPE-Dance，并设计一种不牺牲原有可控性和视觉保真度的框架级方案，实现了文本、音乐及两者结合驱动的视频生成。

Details

Motivation: 现有音乐驱动舞蹈视频生成方法面临两大局限：缺乏大规模高质量舞蹈视频数据集，以及缺乏将音乐作为补充条件信号有效整合到视频生成基础模型中的原则性框架。

Result: 在自建的CIPE-Dance数据集上进行的大量实验表明，OmniDance在文本到视频、音乐到视频以及多模态文本音乐到视频三个任务上均达到了最先进的性能，并展现出强大的多模态整合能力。

Insight: 创新点包括：1) 构建了目前最大的舞蹈视频生成数据集CIPE-Dance；2) 提出了一种框架级方案，通过深度感知专业化架构、锚定式由易到难课程学习策略和模态专业化时间相关CFG策略，将音乐作为高频时序动态信号与文本低频语义信号互补地整合到TI2V基础模型中，实现了统一的多模态生成。

Abstract: Music-driven dance video generation aims to synthesize expressive human motion that is temporally aligned with music while maintaining high visual fidelity. Despite recent progress, existing methods still face two key limitations: the lack of large-scale, high-quality dance video datasets, and the absence of principled frameworks for integrating music as a complementary conditioning signal into Video Generation Foundation Models. To address these limitations, we introduce CIPE-Dance, a large-scale Internet-sourced dance video dataset with choreography-informed text annotations, constructed via a progressive expert pipeline. To the best of our knowledge, CIPE-Dance is the largest dataset for dance video generation to date, comprising 300k high-quality clips over 400 hours and covering diverse dancers, environments, and dance genres. We further propose OmniDance, a framework-level recipe for integrating music into a TI2V foundation model without sacrificing its original controllability or visual fidelity. Motivated by the complementary roles of text as low-frequency semantics and music as high-frequency temporal dynamics, OmniDance co-designs a depth-aware specialization architecture, an anchored easy-to-hard curriculum learning strategy, and a modality-specialized time-dependent CFG strategy, enabling unified TI2V, MI2V, and MTI2V generation. Extensive experiments on CIPE-Dance demonstrate that OmniDance achieves state-of-the-art performance across all three tasks and exhibits robust multimodal integration capability. Project is available at https://github.com/AMAP-ML/OmniDance.

[155] Uncertainty Estimation in Pathology Foundation Models via Deep Mutual Learning cs.CVPDF

Gbègninougbo Aurel Davy Tchokponhoue, Sevda Öğüt, Ali Idri, Dorina Thanou, Pascal Frossard

TL;DR: 本文提出了一种名为DICE的即插即用框架，用于提升病理学基础模型（PFMs）在临床决策中的可信度。该框架通过集成多个冻结的PFMs，并利用深度互学习来对齐这些模型，从而将模型间的分歧建模为不确定性估计的代理。该方法在多个全切片图像（WSI）基准测试中，不仅提供了可靠的不确定性估计以识别易失败案例，还在分类、校准和定位任务上达到或超越了现有最佳水平。

Details

Motivation: 病理学基础模型虽然为全切片图像分析提供了通用表征，但其临床采用仍受限，主要因为其预测缺乏可靠的可信度估计，且没有单一模型在所有任务上都是最优的，这严重影响了医疗环境中的信任度。

Result: 在三个具有挑战性的WSI基准测试中，DICE框架提供了可靠的不确定性估计，能够准确识别分布内和分布外设置下的易失败案例，同时在分类、校准和定位任务上匹配或超越了SOTA基线方法。

Insight: 创新点在于提出了一种无需训练基础模型、通过集成和深度互学习来对齐多个冻结PFMs，从而将模型间分歧作为不确定性代理的方法。该方法理论上证明了该目标上界了模型不确定性，并且集成共识能够在无显式监督的情况下实现补丁级别的异常定位，为将PFMs转化为不确定性感知的决策支持系统迈出了关键一步。

Abstract: Pathology foundation models (PFMs) offer generalizable representations for whole-slide image (WSI) analysis, yet their clinical adoption remains limited. Specifically, their predictions lack reliable confidence estimates, and no single PFM is universally best across tasks, which severely undermines trust in medical settings. To overcome this, we propose $\mathtt{DICE}$, a plug-and-play framework that ensembles $K$ frozen PFMs and models their disagreement as a proxy for uncertainty estimation. To ensure this proxy yields meaningful estimates, we align the ensemble members via deep mutual learning, and theoretically show that this objective upper-bounds the model uncertainty. Additionally, we demonstrate that the ensemble’s consensus localizes abnormalities at the patch level without any explicit supervision. We evaluate $\mathtt{DICE}$ on three challenging WSI benchmarks. Notably, our framework provides reliable uncertainty estimates that accurately flag failure-prone cases under in- and out-of-distribution settings, while matching or outperforming SOTA baselines in classification, calibration, and localization. Overall, $\mathtt{DICE}$ takes a crucial step toward translating PFMs into uncertainty-aware decision-support systems.

Yongchang Gao, Junjie Pang, Shuaiyu Yang, Yusheng Yang, Xichao Jia

TL;DR: 本文提出了一种名为IDNet的多模态框架，用于缺血性心脏病（IHD）筛查，该框架通过跨模态蒸馏聚合器（CDA）顺序整合左眼、右眼和临床特征，以解决视网膜图像与稀疏临床变量融合的挑战。同时，研究构建了一个基于UK Biobank的可复现基准数据集，包含25,205名受试者的50,410张图像。在该基准上，IDNet超越了仅使用图像、仅使用临床数据以及多种多模态基线方法。

Details

Motivation: 当前基于彩色眼底摄影（CFP）的IHD筛查研究受限于公开基准稀缺，以及视网膜图像与稀疏临床变量融合效果不佳的问题。

Result: 在构建的UK Biobank基准上，IDNet超越了仅图像、仅临床以及多种多模态基线方法；CDA作为一个即插即用的融合模块，能持续提升多种视觉编码器的性能。

Insight: 创新点在于提出了一种使用可学习查询进行顺序特征整合的跨模态蒸馏聚合器（CDA），有效缓解了高维视觉特征与低维表格输入之间的不平衡问题；同时，通过开源流程构建了一个大规模、可复现的基准数据集，为领域研究提供了重要资源。

Abstract: Color Fundus Photography (CFP) offers a low-cost and non-invasive route for ischemic heart disease (IHD) screening, but current studies are limited by scarce public benchmarks and ineffective fusion of retinal images with sparse clinical variables. We propose IDNet, a multimodal framework with a Cross-Modal Distillation Aggregator (CDA) that uses learnable queries to sequentially integrate left-eye, right-eye, and clinical features, mitigating the imbalance between high-dimensional visual features and low-dimensional tabular inputs. We also construct a reproducible UK Biobank benchmark with open-source curation and quality-control pipelines, yielding 50,410 images from 25,205 subjects. On this benchmark, IDNet outperforms image-only, clinical-only, and several multimodal baselines, and CDA consistently improves multiple visual encoders as a plug-in fusion module.

[157] MuseBench: Benchmarking Intent-Level Audiovisual Arts Understanding in MLLMs cs.CV | cs.AIPDF

Yuxuan Fan, Gyusik Seo, Jing Hao, Jaemin Cho, Mohit Bansal

TL;DR: 本文提出了MuseBench基准，用于评估多模态大语言模型（MLLMs）在视听艺术领域的深层意图理解能力。该基准涵盖电影、视觉艺术、舞台表演和游戏艺术四大领域，包含4016个问题，旨在超越简单的感知识别，专注于对艺术创作意图的推理。对28个先进MLLMs的零样本评估显示，最佳模型准确率仅为48.29%，远低于人类专家水平的87.18%，揭示了模型在创造性领域理解的显著不足。

Details

Motivation: 现有MLLM基准主要衡量感知识别能力，而忽略了艺术理解中至关重要的部分——对创作意图的推理，即理解艺术家为何通过特定的视听和叙事元素来表达意义。本文旨在填补这一空白，系统评估MLLMs对艺术深层含义的理解。

Result: 在MuseBench基准上对28个SOTA MLLMs进行了零样本评估，最佳模型准确率仅为48.29%，远低于人类专家水平（87.18%），表明当前模型在艺术意图理解方面存在巨大差距。

Insight: 创新点在于构建了一个专注于艺术创作意图推理的综合性多模态基准，其问题通过结合专业视频评论、四阶段迭代生成流程（包括捷径过滤、对抗性干扰项和专家验证）来确保质量和深度，这为评估和理解MLLMs在复杂、开放式艺术分析任务上的能力提供了新工具和视角。

Abstract: Audiovisual arts encompass diverse creative disciplines, including cinema, visual arts, stage performance, and game design, where artistic meaning arises from deliberate combinations of visual, auditory, and narrative elements (e.g., fear amplified through claustrophobic framing, or grief conveyed through silence and lingering close-ups). True artistic understanding extends beyond recognizing what is depicted to reasoning about why it is expressed through particular creative choices. Despite the strong progress of multimodal large language models (MLLMs), this critical aspect of artistic understanding remains underexplored, as existing benchmarks largely measure perceptual recognition while overlooking reasoning about creative intent. To address this gap, we introduce Musebench, a comprehensive benchmark designed to evaluate MLLMs on nuanced artistic understanding. It comprises 4,016 questions spanning cinematic arts, static visual arts, stage performing arts, and game arts, distilled from over 10K candidate video essays that pair professional commentary with visual demonstration. To capture the open-ended nature of artistic analysis at scale, the benchmark combines single-select and variable-option multi-select questions. All questions are generated and refined through a four-phase iterative pipeline combining shortcut filtering, adversarial distractors, and expert validation. Comprehensive zero-shot evaluation of 28 state-of-the-art MLLMs reveals that even the best-performing model achieves only 48.29% accuracy, substantially below human expert performance of 87.18%, exposing a significant gap in current models’ creative domain expertise.

[158] Argus: Metric Panoramic 3D Reconstruction for Indoor Scenes cs.CVPDF

Xi Li, Linyuan Li, Yan Wu, Tong Rao, Kai Zhang

TL;DR: 本文提出了Realsee3D数据集和Argus网络，用于解决室内场景度量全景三维重建问题。Realsee3D是一个包含真实与合成场景的大规模全景RGB-D混合数据集，Argus是一个基于该数据集训练的前馈网络，旨在从稀疏无序的全景图像中实现精确的度量重建。

Details

Motivation: 由于缺乏大规模全景RGB-D训练数据，度量前馈三维重建方法在全景数据上尚未得到充分探索。本文旨在填补这一空白，并解决在稀疏无序采集设置下，不当的坐标系锚点选择会导致全局位姿漂移的问题。

Result: 在Realsee3D基准测试上，Argus在相机位姿估计、深度估计和点云重建方面均达到了最先进的度量性能。

Insight: 创新点包括：1) 引入一个学习的共视模块，用于选择几何最优的参考视图来锚定度量世界坐标系，以缓解全局位姿漂移；2) 将双向像素到世界的映射分解为可解释的子步骤，并施加每步监督和跨坐标联合约束，以增强多任务学习中预测分支间的几何一致性。

Abstract: Metric feed-forward 3D reconstruction for panoramic data remains under-explored due to the lack of large-scale panoramic RGB-D training data. We present Realsee3D, a hybrid dataset of 10K indoor scenes (1K real, 9K synthetic) with 299K panoramic viewpoints and precise metric annotations, and Argus, a feed-forward network trained on it for metric panoramic 3D reconstruction. In the sparse unordered capture setting of Realsee3D, a poorly chosen coordinate anchor can cause global pose drift. Argus addresses this with a learned covisibility module that selects the geometrically optimal reference view to anchor the metric world frame. To further improve multi-task learning, we decompose the bidirectional pixel-to-world mapping into interpretable sub-steps with per-step supervision and cross-coordinate joint constraints, reinforcing geometric consistency across prediction branches. On the Realsee3D benchmark, Argus achieves state-of-the-art metric performance in camera pose estimation, depth estimation, and point cloud reconstruction. Project page: https://argus-paper.realsee.ai.

[159] Walking in the Implicit: Interactive World Exploration via Neural Scene Representation cs.CVPDF

Zhiqi Li, Chengrui Dong, Zhenhua Du, Hangning Zhou, Cong Qiu

TL;DR: 本文提出了一种名为’Walking in the Implicit’的以场景为中心的交互式世界探索范式，其核心是将生成过程的展开变量从视频帧的隐变量转变为一种固定长度、可渲染的隐式状态，称为神经隐式场景（NIS）。基于此，作者构建了NeuWorld系统，它使用Transformer VAE从稀疏的带位姿帧中学习局部锚定的NIS，并使用扩散Transformer根据未来相机轨迹和几何感知的历史检索来演化NIS。该系统在公开的带位姿视图数据上从头训练，无需预训练视频主干或辅助3D重建器，即可实现强大的长时一致性。

Details

Motivation: 现有的交互式视频生成系统在相机控制的世界探索中，通过展开不断增长的隐视频帧序列，将状态转移与高频观测（即图像）合成过程耦合在一起，这导致了效率与一致性问题。本文旨在将这两个过程解耦，以提升交互生成的效率与一致性。

Result: NeuWorld在公开的带位姿视图数据上从头训练，无需依赖预训练视频模型或辅助3D重建器，实现了强大的长时一致性，并具有优越的推理效率。

Insight: 核心创新在于提出了’神经隐式场景（NIS）’这一紧凑、可渲染的隐式状态作为生成过程的展开变量，从而将交互生成分解为紧凑场景状态的随机转移和给定状态下的确定性位姿条件渲染。另一个关键点是复用VAE编码器作为统一的调节器，将相机、参考图像和历史线索映射到相同的NIS模态，避免了使用外部异构编码器，简化了系统设计。

Abstract: Interactive video generation systems for camera-controlled world exploration roll out growing sequences of latent video frames, entangling state transition with high-frequency observation synthesis. We propose Walking in the Implicit, a scene-centric paradigm that changes the rollout variable from frame latents to a fixed-length, renderable implicit state, termed Neural Implicit Scene (NIS). This factorizes interactive generation into stochastic transition of a compact scene state and deterministic pose-conditioned rendering given the sampled state. We instantiate this paradigm as NeuWorld: a transformer VAE learns locally anchored NIS from sparse posed frames, and a diffusion transformer evolves NIS conditioned on future camera trajectories and geometry-aware retrieved history. By reusing the VAE encoder as a unified conditioner, NeuWorld maps camera, reference-image, and history cues into the same NIS modality, avoiding external heterogeneous encoders. Trained from scratch on public posed-view data without pretrained video backbones or auxiliary 3D reconstructors, NeuWorld achieves strong long-horizon consistency with favorable inference efficiency.

[160] Illuminating Unified Multimodal Model for Free-form Interleaved Text-Image Generation cs.CVPDF

Chonghuinan Wang, Zhikai Chen, Chunwei Wang, Yecong Wan, Junwei Yang

TL;DR: 本文提出了ILLUME-X，一个用于自由形式交错文本-图像生成的高级统一多模态模型。它通过改进多模态数据效率和稳定训练过程，实现了高质量的交错生成。模型包含优化的训练数据管道、渐进式训练策略和新的评估方法ILScore。

Details

Motivation: 为了解决模型自主生成自由形式交错文本-图像序列的挑战，推动多模态智能进入新阶段，需要提升数据效率和训练稳定性。

Result: ILLUME-X在风格迁移、图像分解和故事叙述等多个交错文本-图像生成任务上超越了之前的统一模型，取得了更好的性能。

Insight: 创新点包括优化的交错数据管道、针对自由长度多模态序列的自适应目标渐进训练策略，以及客观全面的交错序列评估方法ILScore，这些共同提升了生成质量和任务适应性。

Abstract: The advancement of generative AI models capable of producing text and image marks a critical step forward in the realm of multimodal intelligence, particularly for tasks involving the interleaving of both modalities. To advance this intelligence to the next stage, it is crucial for models to autonomously generate free-form interleaved text-image sequences. In this paper, we introduce ILLUME-X, an advanced unified multimodal paradigm that enables high-quality, free-form interleaved text-image generation by improving multimodal data efficiency and stabilizing the multimodal training process. ILLUME-X comprises three key components: (i) an expanded training data pipeline optimized for interleaved text-image generation, (ii) a progressive training strategy with self-adaptive objectives for free-length multimodal token sequences, and (iii) an objective and comprehensive evaluation method ILScore for interleaved text-image sequences. Notably, our ILLUME-X outperforms previous unified models across multiple interleaved text-image generation tasks like style transfer, image decomposition and storytelling.

[161] Emergence of a Shared Canonical Object Frame from In-the-Wild Videos cs.CVPDF

Tom Fischer, Martin Sundermeyer, Adam Kortylewski, Eddy Ilg

TL;DR: 本文提出一种自监督方法，通过从野外拍摄的物体中心视频中学习，无需手动标注即可建立共享的规范物体坐标系。该方法利用运动恢复结构（SfM）提供的噪声相机姿态，通过一个共享的几何瓶颈（粗糙规范网格）学习密集对应关系，从而实现跨实例的物体姿态对齐。

Details

Motivation: 传统方法依赖手动标注来建立共享规范坐标系，这限制了类别和实例的多样性，成为扩展瓶颈。本文旨在通过自监督学习从野外视频中自动涌现出共享规范框架，以解决这一可扩展性问题。

Result: 在160,000个野外物体视频上进行自监督训练后，该方法在类别级姿态估计基准测试中达到了与依赖规范姿态监督方法相当的竞争性精度。

Insight: 创新点在于通过共享几何瓶颈（粗糙规范网格）结合多视图一致性和特征提取器的语义先验，使共享规范框架能够从噪声SfM几何中自动涌现，无需任何规范姿态标签或类别条件，这为无监督3D理解提供了新思路。

Abstract: Comparing object orientations and positions across different instances requires their poses to be expressed in a shared canonical frame. Establishing such frames has traditionally required manual annotation, creating a scaling bottleneck that limits category and instance diversity. We show that a shared canonical frame can instead emerge from self-supervised training on object-centric videos captured in the wild, using only noisy camera poses from Structure-from-Motion. Our key idea is to route all training sequences through a shared geometric bottleneck: a coarse canonical mesh that carries no category-specific detail. By learning dense correspondences from image pixels to this mesh, and estimating per-sequence alignments from noisy SfM geometry, a common canonical frame emerges from multi-view consistency and the semantic priors of the feature extractor, without any canonical pose labels or category conditioning. Trained in a self-supervised manner on 160,000 in-the-wild object videos, our method achieves competitive accuracy on category-level pose estimation benchmarks compared to methods that rely on canonical pose supervision. The code and checkpoint is available on https://github.com/Fischer-Tom/Emergent-Canonical-Frame/.

[162] CylindTrack: Depth-Aware Cylindrical Motion Modeling for Panoramic Multi-Object Tracking cs.CV | cs.RO | eess.IVPDF

Buyin Deng, Kai Luo, Lingxin Huang, Xinqi Liu, Fei Cheng

TL;DR: 本文提出CylindTrack，一种用于全景多目标跟踪的深度感知圆柱运动建模框架。该框架通过深度-时间轨迹建模提升深度信息的时序一致性，利用球面时空一致性学习增强深度表征的几何对齐，并设计拓扑感知圆柱运动模型在周期性全景域中进行连续角状态空间下的运动预测与关联。

Details

Motivation: 全景多目标跟踪面临周期性图像域破坏平面运动假设、IoU关联在0°/360°接缝处不可靠，以及大视场场景中物体多、尺度变化大、交互频繁导致在线关联对深度线索敏感等问题。

Result: 论文通过联合建模轨迹级深度一致性与全景拓扑结构，在具有挑战性的全景场景中提升了身份保持与轨迹连续性。

Insight: 创新点包括将深度从帧级线索提升为时序滤波的轨迹级状态，结合时序混合器与球面几何感知注意力增强深度表征的时空一致性，以及将水平运动提升至连续角状态空间进行接缝一致的运动预测与关联。

Abstract: Multi-Object Tracking (MOT) is a core capability for embodied perception, and panoramic cameras are attractive for embodied systems because their 360° field of view reduces blind spots and keeps surrounding targets observable for longer durations. However, panoramic MOT is not a straightforward extension of perspective MOT. In equirectangular panoramic videos, the horizontal image domain is periodic rather than Euclidean, which breaks planar motion assumptions and makes IoU-based association unreliable near the 0°/360° seam. Meanwhile, large-FoV scenes often contain more objects, stronger scale variation, and more frequent interactions, making online association particularly sensitive to unstable frame-wise depth cues. To address these issues, we propose CylindTrack, a depth-aware cylindrical tracking-by-detection framework for panoramic MOT. CylindTrack first introduces Depth-Temporal Trajectory Modeling (DTM), which promotes instance depth from an isolated frame-wise cue to a temporally filtered trajectory-level state. To improve the reliability of depth observations, we further develop Spherical Spatio-Temporal Consistency Learning (SSTC), which combines a Temporal Mixer and Spherical Geometry-aware Attention to enhance temporal coherence and panoramic geometric alignment in depth-aware representations. Finally, we design a Topology-Aware Cylindrical Motion Model (TCMM) that lifts horizontal motion into a continuous angular state space and performs seam-consistent motion prediction and association in the periodic panoramic domain. By jointly modeling trajectory-level depth consistency and panoramic topology, CylindTrack improves identity preservation and trajectory continuity in challenging panoramic scenes. The source code will be released at https://github.com/warriordby/CylindTrack.

[163] One Forward Beats Two: InnerZoom for Accurate and Efficient GUI Grounding cs.CVPDF

Chen Liu, Ling Chen, Hanzhang Zhou, Liangyu Chen, Chenglin Cai

TL;DR: 本文提出了一种名为InnerZoom的单次前向框架，用于解决基于MLLM的GUI定位任务中目标区域感知信息在解码过程中丢失的问题。该方法通过在解码层之间桥接和精炼跨层证据状态，实现了在不增加额外计算成本的情况下，达到甚至超越两阶段ZoomIn方法的定位精度。

Details

Motivation: 现有MLLM GUI定位方法将目标定位建模为自回归坐标生成，但诊断分析发现，目标区域感知信息在中间解码层出现，却未能在最终坐标预测中保留或转化。ZoomIn等两阶段方法通过外部裁剪和重新运行提高了精度，但增加了延迟和计算成本。

Result: 在六个GUI定位基准测试（OSWorld-G, UI-Vision, OSWorld-GR, MMBench-GUI等）上，InnerZoom-4B模型均取得了最先进的性能，平均超越之前最佳结果3.1个百分点。在4B参数设置下，其性能优于基线模型和两阶段ZoomIn方法，同时端到端延迟降低高达31.8%，计算量减少约29%。

Insight: 核心创新在于提出了跨层证据桥接机制，将单次前向传播中产生的目标相关线索转化为紧凑的证据状态，并在后续解码层中持续保留、精炼和注入该状态以指导坐标预测，从而在保持高精度的同时显著提升了效率。

Abstract: MLLM-based GUI grounding methods commonly formulate target localization as autoregressive coordinate generation, enabling models to leverage the strong instruction-following and semantic understanding capabilities of MLLMs. However, this formulation requires the model to retain region-level target evidence while decoding coordinate tokens with the spatial precision demanded by GUI clicking. Our diagnostic analysis reveals that target-region awareness emerges in intermediate decoder layers but is neither retained nor translated into the final coordinate prediction. Existing ZoomIn-style methods address this issue through an external crop-and-rerun pass, which improves localization but increases end-to-end latency and computational cost. To retain the accuracy benefits of two-pass zooming without this extra cost, we propose InnerZoom, a single-forward framework for cross-layer evidence bridging. InnerZoom transforms target-related cues from the original forward pass into a compact cross-layer evidence state, then preserves, refines, and reinjects this state throughout later decoding layers to guide coordinate prediction. Extensive experimental results suggest that InnerZoom-4B achieves state-of-the-art performance on all six GUI grounding benchmarks, obtaining 64.7 on OSWorld-G, 40.2 on UI-Vision, 73.1 on OSWorld-GR, and 87.6 on MMBench-GUI, surpassing the previous best results by 4.1, 3.2, 2.9, and 2.3 points, respectively. Under a controlled 4B setting, InnerZoom improves the same SFT+RL baseline by 5.3 points on average and outperforms two-pass ZoomIn by 1.3 points on average, while reducing end-to-end latency by up to 31.8% and TFLOPs by about 29%. Code and models will be publicly available.

[164] SciIR: A Large-scale Training Dataset and Benchmark for Scientific Image Reasoning Generation cs.CVPDF

Zhiyuan Ma, Zhengfeng Shi, Yuning An, Peize Li, Jiabao Wei

TL;DR: 本文提出了SciIR，一个用于科学图像推理生成的大规模训练数据集和基准测试。该工作基于皮尔士符号学三元论，将科学推理形式化为实体结构（图标）、科学过程（索引）和科学定律（符号）三个核心维度。作者构建了包含超过8万对高质量科学图文对的SciIR-82k数据集，并提出了包含原子检查表的SciIR-Bench评估基准。实验表明现有模型在科学推理上存在显著不足，而基于SciIR-82k微调的Qwen-Image-SciIR模型在基准测试上取得了显著提升。

Details

Motivation: 当前文本到图像（T2I）模型在生成逼真图像方面表现出色，但在生成需要严格语义对齐和逻辑推理的科学图像时仍存在困难。为了解决科学图像生成领域训练数据稀缺和评估标准缺失的问题，作者提出了SciIR资源。

Result: 实验揭示了当前模型在科学推理能力上的显著缺陷。通过在SciIR-82k数据集上进行微调，作者开发的Qwen-Image-SciIR模型在SciIR-Bench基准测试上取得了实质性改进，将最终得分从35%提升至43%。

Insight: 论文的主要创新点在于：1）基于皮尔士符号学理论，为科学图像生成任务形式化了一个包含三个维度的科学推理框架；2）构建了大规模、高质量、层次化组织的科学图像-文本对数据集SciIR-82k，并引入了科学推理思维链（Sci-RCoT）来显式建模底层视觉逻辑；3）提出了与理论框架对齐的评估基准SciIR-Bench，并采用原子检查表将结果导向的科学准确性转化为过程导向、可验证的细粒度问题，为评估提供了更可靠的方法。

Abstract: While Text-to-Image (T2I) models have shown remarkable success in generating photorealistic visual content, they still struggle with the rigorous semantic alignment and logical reasoning required for scientific imagery. Inspired by Peirce’s Semiotic Triad, we introduce Scientific Image Reasoning (SciIR), a comprehensive resource for training and evaluation of scientific image generation. We formalize scientific reasoning into three core dimensions: Entity Structure (Icon), Scientific Process (Index), and Scientific Law (Symbol). Specifically, to overcome the scarcity of training data in scientific image generation, we elaborately create SciIR-82k, a large-scale dataset containing over 80,000 high-quality scientific image-text pairs from cutting-edge publications. The dataset is hierarchically organized according to the semiotic dimensions and incorporates a Scientific Reasoning Chain-of-Thought (Sci-RCoT) to explicitly model underlying visual logic. For evaluation, we propose SciIR-Bench, which aligns with these three semiotic levels and employs an Atomic Checklist to convert the outcome-oriented scientific accuracy into process-oriented, verifiable, fine-grained questions. Our extensive experiments reveal significant deficiencies in current models’ scientific reasoning capabilities. Furthermore, by fine-tuning on the SciIR-82k dataset, we developed the Qwen-Image-SciIR model, which achieves a substantial improvement on the SciIR-Bench, increasing the final score from 35% to 43%, laying a solid foundation for future advances in scientific image generation.

[165] T2LDM++: A Self-Conditioned Representation Guided Diffusion Model for Realistic Text-to-LiDAR Scene Generation cs.CVPDF

Wentao Qu, Qi Zhang, Chenxu Wang, Guofeng Mei, Yongfei Liu

TL;DR: 本文提出了T2LDM++，一种用于真实文本到LiDAR场景生成的自我条件表示引导扩散模型。该模型通过自我条件表示引导（SCRG）缓解了训练先验不足的问题，并构建了可控的文本-LiDAR数据。此外，作者还创建了两个高质量文本-LiDAR基准数据集（>10万样本）并设计了方向位置先验以减少街道失真，同时模型支持多种条件生成模式。

Details

Motivation: 解决文本到LiDAR生成任务中因文本-LiDAR对稀缺导致的场景过度平滑和可控性有限的问题，旨在通过更好的先验建模和高质量数据构建来生成更真实、几何细节丰富的LiDAR场景。

Result: 在构建的高质量文本-LiDAR基准上，T2LDM++能够在无条件和有条件设置下生成具有丰富几何细节的真实LiDAR场景，其自我条件表示引导（SCRG）设计更有效且轻量，在推理时解耦以避免计算开销。

Insight: 创新点包括引入自我条件表示引导（SCRG）机制，通过重建引导使去噪网络学习几何感知表示，从而在DDPM中实现更准确的去噪；以及构建大规模文本-LiDAR基准和方向位置先验来提升场景保真度，并支持通过冻结去噪网络学习控制编码器来实现多条件生成。

Abstract: Recent progress in Text-to-Image generation benefits from large-scale Text-Image pairs. However, the scarcity of Text-LiDAR pairs often causes over-smoothed scenes and limited controllability. In this paper, we rethink the limitations of Text-LiDAR generation task, focusing on alleviating insufficient training priors and constructing controllable Text-LiDAR data. We propose a \textbf{T}ext-\textbf{to}-\textbf{L}iDAR \textbf{D}iffusion \textbf{M}odel for LiDAR scene generation, T2LDM++, with a Self-Conditioned Representation Guidance (SCRG). Specifically, to alleviate object over-smoothing, SCRG employs a Guidance Network (GN) to provide reconstruction-based soft supervision to the Denoising Network (DN). This enables DN to learn geometry-aware representations through reconstruction guidance, leading to more accurate denoising in DDPMs. Meanwhile, through analysis and design, SCRG exhibits more effective and lightweight, while decoupled in inference, avoiding computational overhead. Furthermore, we construct two high-quality Text-LiDAR benchmarks ($>$100K samples) using a generalized strategy of geometric annotations, along with a controllability metric. Moreover, a directional position prior is designed to mitigate street distortion, further improving scene fidelity. Additionally, T2LDM++ supports multiple conditions, including (Semantic, Box, BEV, Camera)-to-LiDAR, Sparse-to-Dense, and Dense-to-Sparse generation, by learning a control encoder via frozen DN. With effective prior modeling and high-quality Text-LiDAR benchmarks, T2LDM++ can generate realistic LiDAR scenes with rich geometric details in unconditional and conditional settings.

[166] Latent Noise Mask for Reducing Visual Redundancy in Multimodal Large Language Models cs.CVPDF

Kai Jiang, Ruishu Zhu, Siqi Huang, Hongyuan Zhang, Xuelong Li

TL;DR: 本文提出了一种名为LatEnt Noise maSk (Lens)的框架，旨在解决多模态大语言模型在细粒度视觉推理中，因图像令牌密集冗余而导致问题相关视觉线索被稀释的问题。该框架通过一个轻量级的Lens证据令牌来评分并保留与问题相关的视觉令牌，同时向低相关性令牌注入自适应潜在噪声以抑制干扰，从而在不改变模型主干或令牌序列的情况下，使用更干净的视觉证据提升模型推理能力。

Details

Motivation: 多模态大语言模型在细粒度视觉推理中经常失败，因为问题相关的视觉线索被密集且冗余的图像令牌所稀释。现有方法通常将思维链扩展到视觉或潜在空间，增加了中间推理状态，但忽视了冗余视觉令牌的负面影响。

Result: 在大多数视觉问答数据集上，Lens将基础MLLM的性能提升了2.4-6.4个百分点；在定位任务上提升了4.1-6.4个百分点。这些结果表明，与简单地扩展推理轨迹相比，使用更干净的问题相关视觉证据能更直接地使多模态推理受益。

Insight: 创新点在于提出了一个由问题条件驱动的视觉证据纯化框架，通过一个轻量级的控制令牌和噪声生成器，在潜在空间中自适应地抑制无关视觉令牌，从而直接净化推理所需的视觉线索。这种方法的核心洞察是，提升多模态推理的关键可能不在于增加复杂的推理步骤，而在于提供更干净、更相关的输入证据。

Abstract: Multimodal large language models (MLLMs) often fail in fine-grained visual reasoning, as question-relevant visual cues are diluted by dense and redundant image tokens. Recent multimodal reasoning methods usually extend chain-of-thought from language models into visual or latent spaces, seeking to add intermediate reasoning states while overlooking the negative impact of redundant visual tokens. We propose LatEnt Noise maSk (Lens), a question-conditioned visual evidence purification framework that empowers MLLMs to reason with cleaner visual cues in latent space. Lens introduces a lightweight Lens Evidence Token (LET) to score which visual tokens support the current question and preserve them during decoding. Guided by the LET scores, it injects adaptive latent noise into low-relevance tokens, softly suppressing distractors without changing the model backbone or token sequence. With only one temporary learnable control token and a lightweight noise generator, Lens adds minimal overhead while improving the base MLLM by 2.4-6.4 points on most VQA datasets and by 4.1-6.4 points on grounding tasks. These results show that multimodal reasoning can benefit more directly from cleaner question-relevant visual evidence than from simply extending the reasoning trace.

[167] HiRes: A Hierarchical Cascaded Method for Resistor Value Identification cs.CV | eess.IVPDF

Rama Y. AlHamidi, Aseel A. Mohamed, Mustafa A. Eltayeb, Osama Hasoneh, Mohammad Shaqfeh

TL;DR: 本文提出了一种名为HiRes的层次化级联方法，用于从无约束图像中识别电阻值。该方法结合了目标检测（YOLOv8n）、语义分割（UNet++ with EfficientNet-B2）以及沿电阻轴投影的结构化解码，实现了端到端的识别。实验表明，该方法在检测、分割和端到端识别精度上均表现优异，超越了现有基线方法。

Details

Motivation: 由于光照、方向、尺度和背景复杂度的变化，从无约束图像中准确识别电阻值仍然是一个具有挑战性的计算机视觉任务。本文旨在开发一个鲁棒的端到端识别系统来解决这一问题。

Result: 在多样化的真实世界图像实验中，HiRes实现了检测mAP50为0.9906，分割mIoU为0.8444，端到端识别准确率达到85.8%（95%置信区间：78.0-91.9%），超越了公开的经典基线CVResist，并且在作者提出的具有挑战性的测试集上优于最先进的多模态大语言模型（MLLMs）。

Insight: 论文的创新点在于将学习到的视觉表征（目标检测和语义分割）与结构化的几何推理（轴投影、几何滤波、间隙保持的色带分离）相结合，并利用E24电阻系列进行验证，从而构建了一个鲁棒、高效且可解释的识别框架。这为处理具有明确几何结构的物体识别任务提供了可借鉴的思路。

Abstract: Accurate identification of resistor values from unconstrained images remains a challenging computer vision task due to variations in lighting, orientation, scale, and background complexity. This paper presents HiRes, a hierarchical cascaded pipeline for end-to-end resistor value identification directly from full-frame images. The approach combines object detection (YOLOv8n), semantic segmentation (UNet++ with EfficientNet-B2), and structured geometric decoding via projection along the resistor axis. To improve robustness, we incorporate geometric filtering, gap-preserving band separation, and validation against the E24 resistor series. Experiments across diverse real-world images show that HiRes achieves a detection mAP50 of 0.9906, a segmentation mIoU of 0.8444, and an end-to-end identification accuracy of 85.8% (95% CI: 78.0-91.9%), outperforming the publicly available classical baseline, CVResist, which fails to generalize beyond controlled conditions. In addition, our architecture outperforms state-of-the-art MLLMs on our challenging test set, offering a lower cost, high efficiency, and an interpretable alternative method. These results demonstrate the effectiveness of integrating learned visual representations with structured reasoning for robust resistor interpretation. Code and dataset are available at https://github.com/HiRes491/HiRes.

[168] Few-Shot Domain Incremental Learning via Continual Vision-Language Consolidation cs.CV | cs.AI | cs.LGPDF

Naeem Paeedeh, Mahardhika Pratama, Wolfgang Mayer, Mukesh Prasad, Weiping Ding

TL;DR: 本文提出了一种名为持续视觉-语言巩固（CVLC）的新算法，以解决少样本领域增量学习（FSDIL）问题。该方法通过基础域中的潜在空间保留和双融合投影（DCP）技术，结合视觉和语言原型，在数据稀缺的情况下高效适应新领域。实验表明，CVLC在多个基准测试中显著优于现有方法。

Details

Motivation: 现有领域增量学习（DIL）方法需要大量数据来适应新领域，在数据稀缺时容易过拟合。本文旨在解决这一极端数据短缺问题，提出了少样本领域增量学习（FSDIL）这一相对未探索的挑战。

Result: CVLC在多个基准问题上进行了测试，与现有技术相比，性能提升高达16%，显示出显著优势。

Insight: 创新点包括：通过潜在空间保留和DCP作为参数高效微调方法，结合视觉原型校准和LLM生成的语言原型，以及共享与领域特定组件结构来融合通用知识和领域细节。从客观角度看，该方法有效整合了视觉-语言多模态信息，提升了少样本场景下的领域适应能力。

Abstract: Existing domain-incremental learning (DIL) strategies call for massive amounts of data to adapt to new domains and suffer from the overfitting problem in the case of data scarcity. This paper puts forward a relatively uncharted problem, namely, few-shot domain incremental learning (FSDIL), taking into account the problem of extreme data shortages in the realm of DIL. A novel algorithm, namely Continual Vision-Language Consolidation (CVLC), is proposed to address the FSDIL problem, where the key idea lies in the concept of latent space reservation in the base domain coupled with dual coalescent projection (DCP) as a parameter-efficient fine-tuning method. First, the vision prototype is calibrated while multiple templates and synonyms are generated via LLMs to induce the language prototype. The vision and language prototypes are fused. Adaptation to never-ending arrivals of new domains is done by the DCP technique, fine-tuned in such a way to prepare the model to unseen domains via latent-space reservations committed in the base domain. CVLC is structured under shared and domain-specific components to combine general knowledge and domain-specific details. The advantage of our approach is demonstrated through a range of benchmark problems and comparisons with prior arts, in which CVLC outperforms them by up to a 16% gap. Our codes are shared publicly in https://github.com/Naeem-Paeedeh/CVLC .

[169] From Accuracy to Visual Dependence: Auditing and Filtering Modality Collapse in Traffic VideoQA cs.CVPDF

Sena Korkut, María Alejandra Bravo Sarmiento, Sanghwan Kim, Zeynep Akata

TL;DR: 本文研究了交通事故事频问答（VideoQA）中的模态崩溃问题，发现即使移除视频输入，多个开源视觉语言模型在基准测试中仍能保持甚至提升准确率，表明模型可能依赖文本捷径而非视觉证据。作者提出了数据集级别的诊断指标（Blind Gap和Visual Gain）以及实例级别的Shortcut Score，用于量化和过滤易产生捷径的问题，从而提高视觉依赖性。

Details

Motivation: 解决VideoQA中高基准准确率可能掩盖模型未真正利用视觉证据的问题，特别是在交通安全关键领域，确保模型依赖场景特定的视觉信息而非文本捷径进行推理至关重要。

Result: 在MM-AU等四个公共基准上的审计显示，移除视频输入后模型准确率有时更高，增加视频帧数反而降低性能；通过提出的Shortcut Score过滤问题子集，减少了捷径偏差并改善了视觉基础性能。

Insight: 创新点包括引入数据集级诊断指标量化视觉依赖性，以及提出无需训练的实例级Shortcut Score来连续过滤捷径问题；客观分析认为，该方法强调了在安全关键VideoQA中评估视觉基础质量而非仅追求高准确率的重要性。

Abstract: High benchmark accuracy does not guarantee genuine use of visual evidence. We study this problem in traffic accident Video Question Answering (VideoQA), where correct answers should depend on scene-specific visual evidence but may instead be inferred from textual shortcuts. Through an audit of four public benchmarks, we find that several recent open-weight Vision-Language Models (VLMs) perform competitively, and sometimes better, without video input. On the MM-AU benchmark, removing video consistently improves accuracy, and adding more frames further degrades performance. To quantify visual dependence, we introduce two dataset-level diagnostics: Blind Gap, measuring above-chance text-only performance, and Visual Gain, measuring the marginal benefit of adding video. We further propose an instance-level Shortcut Score that combines text-only confidence with visual necessity signals, enabling continuous, training-free filtering of shortcut-prone questions. The resulting subsets reduce shortcut bias and improve visual grounding. Our findings reveal large differences in grounding quality across benchmarks and show that visually grounded evaluation, not just high accuracy, is essential in safety-critical VideoQA.

[170] Your Data Manifold is Secretly a Reward Model: Shell-LCC for Text-to-Video Generation cs.CV | cs.LGPDF

Shihao Zhang, Yuguang Yan, Junzhe Zhang, Wei Zhao, Bohan Wang

TL;DR: 本文提出Shell-LCC方法，通过显式建模高质量SFT数据的流形结构，将其作为隐式的奖励模型，为文本到视频生成提供密集、可微且几乎无成本的奖励信号，以提升视频质量，特别是减少低频失真。

Details

Motivation: 现有T2V扩散模型依赖外部奖励信号（如奖励模型或DPO）进行对齐，但计算开销大、需要人工标注，且对局部细节改善有限。

Result: 实验表明，该方法提高了视频的真实感，增强了高频细节，减少了过度平滑伪影并缓解了运动模糊。

Insight: 创新点在于将数据流形本身视为奖励模型，并提出Shell-LCC来建模流形表面，避免均值回归问题，从而保留高频细节。

Abstract: Recent text-to-video (T2V) diffusion models rely heavily on auxiliary reward signals (e.g., via reward models or DPO) to align generated content with human aesthetics and improve realism. These signals, however, incur substantial computational overhead, require costly human annotations, and often yield limited improvement in fine-grained local details. In this paper, we argue that your data manifold is secretly a reward model. By explicitly modeling the manifold structure of high-quality Supervised Fine-Tuning (SFT) data and encouraging video latents to lie on this manifold, we derive dense, differentiable, and nearly cost-free reward signals that significantly improve video quality, particularly in mitigating low-level distortions. Our modeling builds upon Local Coordinate Coding (LCC), which captures the skeleton' of the manifold. However, directly applying LCC suffers from mean regression, pulling latents toward the geometric mean and losing high-frequency details. We therefore extend it to Shell Local Coordinate Coding (Shell-LCC), which models the manifold surface’ as an isotropic shell to align with the true high-density region. Experiments demonstrate that our approach improves realism, enhances high-frequency details, reduces over-smoothing artifacts, and alleviates motion blur.

Kun Li, Shengxi Gui, Francesco Nex, Michael Ying Yang

TL;DR: 本文提出了一种名为S4ECA的新型参数高效调优框架，用于解决遥感图像指代分割任务中现有方法计算成本高、泛化能力弱以及跨模态对齐不足的问题。该框架通过双编码器适配器架构，利用可学习查询提取语义语言代理，并结合语言引导的多尺度空间选择机制，动态增强相关视觉上下文，实现高效精确的跨模态对齐。

Details

Motivation: 现有遥感图像指代分割模型主要依赖全参数微调，计算开销大且可能损害预训练模型的泛化能力；而现有的参数高效调优方法多为单模态优化，难以处理遥感图像与自然语言之间的复杂跨模态依赖和领域差异。

Result: 在RRSIS-D和RefSegRS数据集上，仅更新主干网络2.4%的参数，即达到了最先进的性能，在复杂航空场景中表现出优越的效率和精度。

Insight: 创新点包括：设计双编码器适配器架构，分别通过可学习查询提取语义语言代理实现早期接地，以及通过语言引导的多尺度空间选择机制动态聚焦相关视觉区域；从客观角度看，该工作将参数高效调优与跨模态对齐相结合，为遥感多模态任务提供了一种轻量且有效的适配方案。

Abstract: Referring Remote Sensing Image Segmentation (RRSIS) seeks to localize and segment the target object or region specified by a natural language expression in a remote sensing image. While existing RRSIS models have benefited from large-scale foundation models, they predominantly rely on full fine-tuning. These approaches are computationally intensive and may weaken the generalization ability of pre-trained models, as extensive fine-tuning on significantly smaller downstream datasets can distort the well-structured feature representations learned during large-scale pre-training. Although Parameter-Efficient Tuning (PET) offers a potential alternative, existing PET frameworks primarily focus on single-modal optimization, failing to capture the complex cross-modal dependencies required for multimodal reasoning, while simultaneously struggling to bridge the substantial domain gap between natural scenes and aerial imagery. To address these limitations, we propose a novel framework, Semantic-driven Scale and Spatial Selection for Efficient Cross-modal Alignment (S4ECA), which enables effective and efficient cross-modal interaction through parameter-efficient adaptation. Specifically, we design a dual-encoder adapter architecture. The textual adapter employs learnable queries to distill highly semantic language proxies from word-level embeddings, facilitating early grounding. Simultaneously, the visual adapter refines hierarchical feature representations through a multi-scale dense extractor, followed by a language-guided scale and spatial selection mechanism that dynamically emphasizes relevant visual contexts, ensuring precise cross-modal alignment. By updating only 2.4% of the backbone parameters, our proposed model achieves state-of-the-art performance on the RRSIS-D and RefSegRS datasets, demonstrating superior efficiency and precision in complex aerial scenarios.

[172] VisReflect: Latent Visual Reflection for Fine-Grained Perception in Long Visual Context cs.CVPDF

Xiaoqian Shen, Mohamed Elhoseiny

TL;DR: 本文提出VisReflect框架，旨在解决大视觉语言模型在处理高分辨率图像和长视频时细粒度感知能力不足的问题。该方法通过在潜在空间中生成连续的视觉反射来强调与问题相关的视觉特征，从而在单次前向传播中引导注意力，避免了传统方法依赖离散标记和额外计算开销的缺点。

Details

Motivation: 现有方法在处理长视觉上下文时，视觉注意力沉没现象严重，导致无关标记吸收过多注意力，而基于裁剪和重编码的方法依赖不可靠的离散标记定位且计算成本高。

Result: 在BLINK、V*、HRBench-4K/8K等高分辨率图像基准和MVBench、VideoMME、MLVU等视频理解基准上，该方法相比强基线分别提升4.1%和1.8%，且相比基于缩放的方法，在达到相当性能的同时将视频推理时间减少约44%。

Insight: 创新点在于提出潜在视觉反射机制，将中间预测表示为连续潜在特征而非离散标记，从而在单次前向传播中实现细粒度感知，避免了额外的计算开销和定位误差。

Abstract: Large Vision Language Models (LVLMs) have achieved remarkable success on vision-language tasks, yet fine-grained perception over high-resolution images and long-context videos remains challenging. As the number of visual tokens increases, the visual attention sink phenomenon becomes increasingly severe, causing irrelevant tokens to absorb a disproportionate amount of attention mass. Recent approaches attempt to mitigate this issue by explicitly predicting bounding boxes or temporal spans and re-encoding the cropped visual regions. Such methods depend on unreliable numeric localization in the discrete token space and incur significant computational overhead due to additional forward passes. In this work, we propose VisReflect, a simple yet effective framework that improves fine-grained perception in long visual contexts through latent visual reflection. Instead of decoding intermediate predictions into discrete tokens, the model generates continuous visual reflection that represents question-relevant visual features in the latent space. These reflections selectively emphasize salient regions or frames, guiding attention towards relevant visual tokens within a single forward pass. We conduct comprehensive evaluations on challenging high-resolution image benchmarks, including BLINK, V*, and HRBench-4K/8K, as well as video understanding benchmarks such as MVBench, VideoMME, and MLVU. Our method consistently improves over strong baselines, achieving gains of 4.1% on image benchmarks and 1.8% on video benchmarks. Compared with zooming-based methods, our model achieves comparable performance while reducing inference time by roughly 44% on video understanding.

[173] The Surprising Effectiveness of Video Diffusion Models for Hand Motion Reconstruction cs.CVPDF

Yuxi Wang, Chengkai Jin, Yufei Liu, Wenqi Ouyang, Tianyi Wei

TL;DR: 本文提出ViDiHand方法，利用预训练视频扩散模型的表示能力进行4D双手运动重建，通过手部覆盖渲染目标调整模型特征，无需检测器或测试时优化，在多个数据集上显著超越现有方法。

Details

Motivation: 现有方法存在局限性：基于图像的方法依赖检测器，在严重遮挡下失效；基于视频的方法仅从稀缺的手部姿态标注中学习，难以建模运动动态、遮挡推理和手物交互。视频生成模型在互联网规模训练中隐含获得了这些能力，因此被用作基础。

Result: 在ARCTIC、HOT3D和HOI4D数据集上，ViDiHand大幅优于先前方法，确立了视频扩散模型作为手部运动重建的强大新基础，并有望为具身AI提供可扩展的野外数据收集途径。

Insight: 创新点在于利用预训练视频扩散模型的隐式世界先验（如运动动态和遮挡推理）进行手部运动重建，通过手部覆盖渲染目标专门化特征，同时保持先验知识，避免了传统检测和优化流程，提升了鲁棒性和效率。

Abstract: 4D hand motion reconstruction from egocentric video is bottlenecked by clear limitations of existing methods: image-based pipelines depend on a detector that fails under heavy occlusion, while video-based methods rely on temporal modules learned only from scarce hand-pose annotations, a narrow signal insufficient to model motion dynamics, occlusion reasoning, and hand-object interaction. These capabilities, however, are exactly what video generative models must implicitly acquire when trained to synthesize coherent video at internet scale. Motivated by this, we present ViDiHand, which leverages the representations of a pretrained video diffusion model to reconstruct 4D two-hand pose. We adapt it via a hand-overlay rendering objective that specializes its features for hands while preserving its world priors. A decoder then recovers metric-scale pose from the adapted features. The whole pipeline operates directly on full frames–no detector, no infiller, and no test-time optimization. On ARCTIC, HOT3D, and HOI4D, ViDiHand substantially outperforms prior methods, establishing video diffusion models as a powerful new foundation for hand motion reconstruction and a promising route to scalable in-the-wild data collection for embodied AI. Project page: https://vidihand.github.io.

[174] TRACE: A Concept Bottleneck Model for Longitudinal 3D Glioblastoma Response Assessment cs.CV | cs.LGPDF

Alia Tarek, Hamsa Saberr, Hamza Elghonemy, Youssef Afify, Tamer Basha

TL;DR: 本文提出了TRACE，一种基于概念瓶颈模型的可解释性方法，用于纵向3D胶质母细胞瘤反应评估。该方法通过共享3D视觉编码器处理多模态MRI扫描，预测有临床意义的肿瘤测量值作为根概念，并通过确定性规则计算下游RANO衍生概念，将反应评估构建为结构化概念推理而非直接的图像到标签预测。

Details

Motivation: 现有深度学习方法直接从影像特征预测反应标签，限制了临床检查、验证和修正。因此，作者旨在开发一种与RANO 2.0标准对齐的可解释模型，以增强胶质母细胞瘤纵向评估的透明度和临床一致性。

Result: 在LUMIERE数据集上进行5折患者级交叉验证，TRACE在4类宏F1上达到0.4769，在二元进展与非进展宏F1上达到0.7085。其性能优于概念瓶颈基线，并与已发表的非可解释深度学习方法相当。

Insight: 创新点在于将反应评估构建为结构化概念推理，通过预测临床可解释的肿瘤测量值作为中间概念，并整合确定性规则和扫描间隔信息。这为纵向医学影像分析提供了一种透明且与临床标准对齐的建模方向。

Abstract: Longitudinal glioblastoma response assessment requires comparing subtle tumor changes across MRI time points using structured clinical criteria such as RANO. However, most deep learning methods predict response labels directly from imaging features, which limits clinical inspection, verification, and correction. We introduce TRACE, a RANO 2.0-aligned concept bottleneck model for interpretable 4-class glioblastoma response classification on longitudinal 3D MRI. TRACE processes paired baseline and follow-up multimodal MRI scans with a shared 3D vision encoder, predicts clinically meaningful tumor measurements as root concepts, computes downstream RANO-derived concepts through deterministic rules, and incorporates scan interval and new-lesion information as passthrough concepts. This design frames response assessment as structured concept reasoning rather than direct image-to-label prediction. Using 5-fold patient-wise cross-validation on the LUMIERE dataset, TRACE achieves a 4-class macro F1 of 0.4769 and a binary progression-versus-non-progression macro F1 of 0.7085. It improves over a concept bottleneck baseline and remains within the range of published non-interpretable deep learning approaches. Ablation studies show that the expert RANO graph and intervention-consistency training are important for performance, while intervention experiments demonstrate that correcting concepts can improve downstream predictions. These results suggest that structured concept bottlenecks offer a transparent and clinically aligned direction for longitudinal glioblastoma response assessment, while highlighting the need for larger protocol-aligned datasets and external validation.

[175] A Point Cloud Transformer for Remote Monitoring and Automated Assessment of Physical Rehabilitation Exercises cs.CVPDF

Kazi Rafat, Md. Ismail Hossain, M M Lutfe Elahi, Sifat Momen, Fuad Rahman

TL;DR: 本文提出了一种基于Transformer的点云框架，用于通过分析RGBD数据采集的关节位置来提取特征并评估康复训练质量。该方法采用基于曲线的点云特征聚合技术增强点云信息，并利用轴向自注意力机制识别关键关节及其作用，以帮助用户更好地执行训练。该系统在Kimore、UI-PRMD和IRDS三个基准数据集上表现出色，具有模型小、推理快、泛化能力强等实用优势。

Details

Motivation: 康复训练对恢复患者身体功能至关重要，但传统方法成本高、需专家监督且难以普及。现有自动评估方法存在特征提取不相关、缺乏实用性、预处理昂贵或忽略关键特征等问题，因此需要一种高效、实用的自动化评估方案。

Result: 在Kimore、UI-PRMD和IRDS三个康复训练基准数据集上，所提出的引导系统超越了现有方法，达到了先进水平（SOTA），同时因其模型尺寸小、推理速度快以及对类似训练中特定关节的泛化能力而具有实际应用价值。

Insight: 创新点包括将Transformer架构应用于点云数据以提取康复训练特征，并引入基于曲线的点云特征聚合技术来增强信息表示；客观分析认为，轴向自注意力机制能有效识别关键关节角色，这为基于关节位置的运动分析提供了可借鉴的轻量化建模思路。

Abstract: Rehabilitation exercises are essential in restoring lost physical functions of patients suffering from various diseases (e.g., Parkinson’s, back pain). Carrying out these rehabilitation exercises, often prescribed by health experts, is costly, unavailable, and requires expert supervision. The availability of RGBD images and movement/position data of joints along with expert annotation of exercise data has prompted the use of automatic assessment of the quality of rehabilitation exercises, which is cost-effective and can be carried out at home. However, existing approaches do not extract relevant features, lack practical application, require expensive pre-processing, or overlook crucial features. This study proposes a transformer-based framework for point clouds to extract features and assess rehabilitation exercises by analyzing joint positions collected through RGBD data. We adapt and utilize a curve-based point-cloud feature aggregation technique to augment point-cloud information that aids model output. The transformer architecture also uses axial self-attention, recognizing important joints and their roles to assist users in performing the exercise better. The guided system outperforms existing approaches and is also practically relevant due to its small size, fast inference, and generalization on specific joints in similar exercises. We conduct our experiments on three crucial baseline datasets for rehabilitation exercises: Kimore, UI-PRMD, and IRDS.

[176] BrainJanus: A Unified Model for Understanding and Generation across Brain, Vision, and Language cs.CV | cs.LGPDF

Haitao Wu, Qirui Zhang, Zhouheng Yao, Shangquan Sun, Qihao Zheng

TL;DR: BrainJanus是首个在单一框架内统一大脑、视觉和语言建模的模型。它通过一个统一的大脑分词器将连续的神经动态量化为离散的token，并在共享的Omni空间中与视觉和语言表征对齐。基于此，模型采用一个全能的（All-in-One）自回归架构，利用下一个token预测来实现图像/文本到大脑的编码以及大脑到图像/文本的解码等任意到任意的无缝生成。

Details

Motivation: 现有方法主要将大脑编码和解码视为孤立任务，严重依赖单模态对齐和外部先验，而忽视了大脑作为多模态整合系统的内在本质。本文旨在解决这些局限，建模外部感官刺激与内部神经活动之间的双向对应关系。

Result: 大量实验表明，BrainJanus在多个基准测试中取得了卓越的性能。此外，该框架展现出零样本泛化能力，并保留了可解释的生物地形图。

Insight: 核心创新在于提出了首个统一的大脑-视觉-语言模型框架，通过共享表征空间和自回归生成架构，实现了任意模态间的双向映射。其统一的大脑分词器和全能的自回归架构设计，为通用大脑建模范式提供了新思路，并强调了多模态整合的重要性。

Abstract: Modeling the bidirectional correspondence between external sensory stimuli and internal neural activity has emerged as a critical frontier in neuroscience. However, existing approaches predominantly treat brain encoding and decoding as isolated tasks, relying heavily on unimodal alignment and external priors while overlooking the brain’s intrinsic nature as a multimodal integration system. To address these limitations, we propose BrainJanus, the first unified brain model that integrates brain, vision, and language within a single framework. Specifically, we introduce a Unified Brain Tokenizer to quantize continuous neural dynamics into discrete tokens aligned with visual and linguistic representations in a shared Omni space. Building on this, we utilize an All-in-One autoregressive architecture that leverages next-token prediction to enable seamless any-to-any generation, which encompasses image-to-brain and text-to-brain encoding, and brain-to-image and brain-to-text decoding. Extensive experiments demonstrate that BrainJanus achieves superior performance across diverse benchmarks. Furthermore, our framework exhibits zero-shot generalization and preserves interpretable biological topography, highlighting its potential as a general-purpose brain modeling paradigm. The code is available at \href{https://github.com/HaitaoWuTJU/BrainJanus}{GitHub}.

[177] MUSE: Unlocking Timestep as Native Task Steering for One-Step Dense Prediction cs.CVPDF

Shuo Zhou, Zhaoxin Li, Xiujuan Chai

TL;DR: 本文提出了一种名为MUSE的参数无关单模型多任务学习方法，用于密集预测任务。该方法通过重新利用一步扩散模型中固有的正弦时间步嵌入作为内生任务引导信号，实现了在单个模型中同时处理多个密集预测任务（如单目深度估计和法线估计），而无需引入额外的参数化适配器或专家模块。

Details

Motivation: 现有基于扩散模型的多任务学习方法通常需要引入参数繁重的适配器、专家模块或可学习的任务令牌，导致计算冗余。本文旨在探索一种更简洁高效的多任务学习范式，通过挖掘一步扩散模型的内在机制来避免额外的参数开销。

Result: 在10个数据集上的广泛实验表明，MUSE在单目深度估计和法线估计任务上取得了极具竞争力的性能。该方法在U-Net和DiT架构上均有效，证明了其泛化能力。

Insight: 核心创新点在于揭示了时间步嵌入可作为原生任务引导信号的内在机制，并通过流形解耦理论进行解释。这为利用现有生成基础设施构建通用视觉模型提供了一条简洁高效的路径，无需修改模型参数即可实现多任务学习。

Abstract: Monocular dense prediction has recently seen remarkable success by repurposing pre-trained diffusion models. This opens a promising yet challenging avenue for more efficient multi-task learning paradigm. However, existing multi-task diffusion methods often introduce parameter-heavy adapters, experts, or learnable task tokens, leading to computational redundancy. In this paper, we reveal an inherent mechanism within one-step diffusion models: the native, fixed sinusoidal timestep embedding can be repurposed as an endogenous task steering signal. Based on this discovery, we propose Multi-task Unified eStimation via timestep Embedding (MUSE), a parameter-free, single-model multi-tasking approach for dense prediction. We interpret this mechanism via Manifold Decoupling, where discrete, fixed timestep values deterministically steer the generation process towards decoupled, task-specific manifolds in the latent space. Extensive experiments across 10 datasets demonstrate that MUSE achieves highly competitive performance on both monocular depth and normal estimation, and its efficacy generalizes across U-Net and DiT architectures. Our work offers a concise and efficient path toward generalist vision models by simply unlocking the latent potential of existing generation infrastructure.

[178] FastPano3D: Feed-Forward Indoor Panoramic 3D Reconstruction from a Single Image cs.CVPDF

Jianqiang Li, Liumei Zhang, Wenjia Guo, Tianlong Feng, Yongzhi Liao

TL;DR: 本文提出了FastPano3D，一个端到端的框架，能够从单张全景图像直接生成可渲染的3D高斯表示，用于室内场景的快速三维重建。该方法通过轻量级特征编码器、自适应高斯采样和点云引导的细化策略，无需测试时优化，即可在数秒内完成高保真重建，推理速度比之前的SOTA方法快达156倍。

Details

Motivation: 为了解决从最少输入（单张全景图）快速重建详细3D室内场景的挑战，并权衡渲染质量、推理效率与数据依赖性。全景图像固有的等距柱状投影畸变和非均匀特征分布，使得直接前馈式高斯生成尤为困难。

Result: 在推理速度上，比之前最先进的方法（如Pano2Room）快达156倍，且参数量仅为其一半。广泛的实验表明，其渲染质量可与基于NeRF和3DGS的重建方法相媲美，为快速单视图3D场景推理设立了新基准。

Insight: 创新点在于提出了一个无需多视角监督或逐场景优化、直接从单张全景图前馈生成3D高斯的端到端框架。其轻量级编码、自适应采样和点云引导的细化策略，有效解决了全景图畸变和非均匀特征带来的挑战，实现了效率与质量的平衡。

Abstract: Recent advances in 3D scene reconstruction have highlighted the intricate trade-offs among rendering quality, inference efficiency, and data dependency. To address the challenge of rapidly reconstructing detailed 3D indoor scenes from minimal input, we introduce FastPano3D, an end-to-end framework that directly generates renderable 3D Gaussian representations from a single panoramic image. Unlike perspective-based methods, panoramic images inherently suffer from equirectangular projection distortions and spatially non-uniform feature distributions, making direct feed-forward Gaussian generation particularly challenging. In contrast to existing Gaussian Splatting based methods that rely on multi-view supervision or per-scene optimization, FastPano3D employs a lightweight feature encoder, adaptive Gaussian sampling, and a point-cloud-guided refinement strategy to achieve efficient and accurate scene generation without any test-time optimization. Our approach reconstructs high-fidelity 3D scenes within seconds, achieving up to 156 times faster inference than prior state-of-the-art methods such as Pano2Room, while using only half the parameters. Extensive experiments demonstrate that FastPano3D delivers rendering quality comparable to NeRF- and 3DGS-based reconstructions, establishing a new benchmark for rapid, single-view 3D scene inference.

[179] Residual-Guided Expert Specialization for Incomplete Multimodal Learning cs.CV | cs.AIPDF

Seunghun Baek, Jihwan Park, Jaeyoon Sim, Minjae Jeong, Hoseok Lee

TL;DR: 本文提出了一种名为MARS（Missingness-Aware Residual-guided Specialization）的混合专家框架，用于解决推理时模态缺失的不完整多模态学习（IML）问题。该方法通过对比完整与不完整输入产生的任务表示差异，生成一个特权残差信号来指导专家专门化，并设计了特征路由器在部署时模仿此行为，同时引入差异感知噪声正则化来增强鲁棒性。

Details

Motivation: 现实世界的预测系统在推理时经常面临模态缺失，导致不完整模态学习成为一个实际挑战。现有方法旨在学习对缺失输入鲁棒的表征，但由于证据缺失，不完整模态的表征不可避免地会偏离其完整模态的对应部分。

Result: 在多模态分类（CASIA-SURF、CREMA-D、UPMC Food-101）和分割（MCubeS）任务上的缺失场景实验中，MARS始终超越基线方法，同时保持高效性并可扩展到不同的骨干网络和任务。

Insight: 创新点在于明确利用模态缺失导致的表征偏差，通过残差信号引导专家专门化，并设计了双路由器架构（残差路由器和特征路由器）以及差异感知噪声正则化，以在训练和测试条件不一致时保持性能，这为处理模态缺失问题提供了新的模型设计思路。

Abstract: As real-world prediction systems often face missing modalities at inference, incomplete multimodal learning (IML) remains a practical challenge. While prior methods aim to learn representations robust to missing inputs, representations from incomplete modalities inevitably deviate from their full-modality counterparts due to missing evidence. To explicitly leverage these deviations, we propose MARS (Missingness-Aware Residual-guided Specialization), a mixture-of-experts framework that guides expert specialization based on how representations are reshaped by missingness. By contrasting task representations derived from incomplete inputs with their complete counterparts during training, we derive a privileged residual signal that captures this representational gap. The residual signal guides a residual router to assign samples to experts specialized for the corresponding deviation patterns. In parallel, a feature router learns to imitate this routing behavior using only incomplete inputs, enabling deployment without access to full modalities. To mitigate this train-test router gap, we develop a discrepancy-aware noise regularization that adaptively perturbs the residual router’s decisions when the feature router deviates, enhancing expert robustness under imperfect imitation. Experiments on multimodal classification (CASIA-SURF, CREMA-D, UPMC Food-101) and segmentation (MCubeS) under missing scenarios show that MARS consistently surpasses baselines while remaining efficient and extensible to diverse backbones and tasks.

[180] Set-Inclusive Uncertainty Modeling for Robust Brain Tumor Segmentation cs.CV | cs.AI | cs.LGPDF

Seunghun Baek, Jihwan Park, Jaeyoon Sim, Hoseok Lee, Seungjoo Lee

TL;DR: 本文提出了一种用于脑肿瘤分割的概率表示框架，通过将多模态MRI特征建模为高斯分布，其中均值捕获任务信息、方差度量因模态缺失引起的不确定性。该方法引入了一种集合包容策略，利用模态子集的层次结构并强制执行排序约束，以保持不确定性关系的一致性。在BraTS 2018和2020数据集上的实验表明，该方法在多种模态缺失场景下优于基线模型。

Details

Motivation: 解决多模态MRI脑肿瘤分割中，推理时难以获取全部模态导致信息丢失和不确定性增加的问题。现有方法未建模这种不确定性，将不完整证据编码为看似合理但不可靠的确定性表示。

Result: 在BraTS 2018和2020基准测试中，该方法在多种模态缺失场景下均优于基线模型，表现出优越的性能。

Insight: 创新点在于将表示建模为高斯分布以分离任务信息和不确定性，并引入集合包容策略利用模态子集层次结构来约束不确定性关系，为处理模态缺失问题提供了概率化且结构化的解决方案。

Abstract: Multimodal MRI is essential for accurate brain tumor segmentation. However, acquiring all modalities at inference is often challenging in practice, which causes intrinsic uncertainty due to unavoidable information loss. Without modeling this uncertainty, existing methods encode incomplete evidence into deterministic representations that appear plausible but lack reliability. In this regime, we propose a probabilistic representation framework that models representations as Gaussian distributions, where their mean captures task information and their variance measures uncertainty from missing evidence. To make variance reflect information deficiency, we regularize the mean from each partial configuration toward its full-modality counterpart, while scaling the variance with the discrepancy between their aligned means. We further introduce a set-inclusive strategy that exploits the hierarchical structure of modality subsets and enforces an ordering constraint to maintain their consistent uncertainty relationships. Extensive experiments on BraTS 2018 and 2020 demonstrate that our approach offers superior performance over baselines across diverse missing-modality scenarios. Code and model checkpoint are available at https://github.com/atlas-sky/SIUM.

[181] SADL: What to Ignore? A Benchmark for Subject-Aware Distractor Localization cs.CVPDF

Cao-Tri Nguyen, Nguyen-Khoa Luong, Vinh-Tiep Nguyen, Minh-Triet Tran

TL;DR: 本文提出了主题感知干扰物定位任务，并构建了首个真实世界基准数据集SADL，包含1000张照片中的1800个案例和14,617个标注候选对象。通过评估七种视觉语言模型在干扰物分类和排除过滤流程中的表现，发现模型能有效识别干扰物但存在过度排除问题。

Details

Motivation: 现有显著性模型和开放词汇检测器缺乏主题感知能力，无法根据用户意图自适应识别干扰物，且上下文无关的物体移除可能破坏场景语义连贯性。

Result: 在SADL基准测试中，视觉语言模型展现出较强的干扰物识别能力，但在排除过滤阶段会系统性地过度排除真实干扰物，揭示了多模态系统主题条件推理的关键瓶颈。

Insight: 创新点在于首次形式化定义了主题感知干扰物定位任务，并构建了包含硬负例的基准数据集；客观分析表明，模型在干扰物排除校准方面存在系统性缺陷，为多模态推理提供了诊断工具。

Abstract: Photographs frequently contain \emph{visual distractors} besides foregrounds and backgrounds of the intended subject, competing for attention and weakening composition. While modern editing tools streamline object removal, identifying which objects to remove remains a mostly manual process. Existing saliency models and open-vocabulary detectors operate without subject awareness, failing to adapt to shifting user intent. Furthermore, context-agnostic removal may disrupt the scene’s semantic coherence (e.g., keep the person but remove the chair they are sitting on). To address these limitations, we formalize the task of subject-aware distractor localization, which identifies distractors while retaining compositionally essential objects. This paper introduces \textsc{SADL}, the first real-world benchmark for this task, comprising 1,800 subject-aware cases across 1,000 photographs to enable systematic evaluation and facilitate future research. In total, there are 14,617 annotated candidates, including a robust set of 1,938 hard negatives to stress-test exclusion calibration. We evaluate seven proprietary and open-weight Vision-Language Models (VLMs) on a sequential pipeline of distractor classification followed by exclusion filtering, structured around five inclusion factors and three contextual exclusion rules. Our analysis reveals that VLMs are highly capable of identifying distractors, but then over-apply exclusion, which systematically suppresses true distractors at scale. By exposing this critical bottleneck, \textsc{SADL} provides a foundational diagnostic tool to advance subject-conditioned reasoning in multimodal systems.

[182] OmniCoT: A Benchmark for Global and Multi-Step Panoramic Reasoning cs.CVPDF

Haocong He, Chenfei Liao, Zichen Wen, Zihao Dongfang, Xu Zheng

TL;DR: 本文提出了OmniCoT全景空间推理基准，旨在评估和提升多模态大语言模型在全景图像中的全局多步推理能力。该基准包含用于评估的OmniCoT-B数据集、用于量化仿真与现实差距的OmniCoT-Real数据集，以及用于训练并带有结构化逐步思维链标注的OmniCoT-T数据集。作者还提出了一种两阶段训练策略，以强化模型在全景空间中的几何一致性推理。

Details

Motivation: 现有全景图像基准大多关注依赖局部线索或简单单步/少步推理的查询，未能充分利用全景图像支持复杂全局多步推理的根本优势。本文旨在填补这一空白，重新校准全景空间推理的难度，以更好地与全景图像的内在能力对齐。

Result: 论文提出了一个包含评估集（OmniCoT-B，6.7K数据）和真实世界标注子集（OmniCoT-Real，1K数据）的基准套件，用于衡量答案准确性和推理质量，并量化仿真与现实差距。基于训练集（OmniCoT-T，14.3K数据）和两阶段训练策略（SFT与GRPO）开发的模型旨在提升全局360度空间一致性推理能力。

Insight: 主要创新点在于构建了首个专注于全景图像全局多步推理的基准，并引入了带有结构化思维链标注的训练数据以及针对几何复杂全景空间定制的两阶段训练策略（SFT锚定证据，GRPO惩罚几何不一致路径），以巩固模型对全景空间的全局一致性理解。

Abstract: Multimodal Large Language Models (MLLMs) have demonstrated promising spatial reasoning capabilities, while these abilities remain underexplored in the emerging visual modality of panoramic imagery. The full 360°$\times$180° field of view of panoramas essentially supports complex global multi-step reasoning, which is also the fundamental advantage of panoramas in applications such as embodied intelligence. However, existing panoramic benchmarks largely focus on simplistic queries that rely on local cues or single-/few-step reasoning, thereby ignoring the fundamental advantage of panoramas and failing to fully exploit their potential. To address this gap, we introduce OmniCoT, a panoramic spatial reasoning suite designed to enable MLLMs to use global evidence and perform multi-step inference across viewpoints. It includes OmniCoT-B (6.7K data) for evaluation, which measures both answer accuracy and reasoning quality, OmniCoT-Real (1K data) as a manually annotated real-world subset to quantify the Sim-to-Real gap. For training, OmniCoT-T (14.3K data) is purpose-built with structured stepwise Chain-of-Thought annotations that explicitly link intermediate reasoning steps to panoramic evidence. Based on OmniCoT-T, we introduce OmniCoT-R1 and adopt a two-stage training strategy tailored to the geometrically complex panoramic space, where Supervised Fine-tuning (SFT) anchors reasoning to panoramic evidence (e.g., bearings, proximity) and GRPO penalizes geometrically incoherent paths to consolidate global 360° spatial consistency. Through OmniCoT, we aim to recalibrate the difficulty of panoramic spatial reasoning to better align with the intrinsic capabilities of panoramic imagery, thereby fostering meaningful progress in this research area.

[183] OWMDrive: Causality-Aware End-to-End Autonomous Driving via 4D Occupancy World Model cs.CVPDF

Junjie Cheng, Ruiqi Song, Ye Wu, Nanxing Zeng, Ximiao Li

TL;DR: OWMDrive是一个基于占用世界模型的生成式端到端自动驾驶框架，通过4D占用预测来建模交通交互中的时空因果动态，以指导基于扩散模型的规划器生成更前瞻和鲁棒的驾驶轨迹。

Details

Motivation: 现有基于学习的方法通常从静态场景表示做决策，缺乏对未来状态的显式推演和交通交互中时序因果动态的建模，导致在高不确定性场景（如遮挡和意外事件）下规划不稳定或过于保守。

Result: 大量实验表明，OWMDrive显著提高了规划的可靠性和安全性，尤其是在具有挑战性和部分可观测的驾驶场景中。

Insight: 创新点在于将4D占用世界模型作为条件先验来引导扩散规划，通过显式建模未来场景演化和时空因果依赖，实现更前瞻的轨迹生成；该方法将生成式世界模型与端到端规划相结合，提升了系统对复杂动态环境的适应能力。

Abstract: Autonomous driving systems are steadily moving toward end-to-end paradigms to mitigate the limited adaptability of rule-based pipelines in complex traffic environments. However, most existing learning-based methods still make decisions from static representations of the current scene, without explicit future rollouts or modeling of the temporal causal dynamics in traffic interactions. This limitation often results in unstable or overly conservative planning under high-uncertainty conditions, such as occlusions and unexpected events. To overcome these challenges, we introduce OWMDrive, a generative end-to-end driving framework built upon an Occupancy World Model for multi-step 3D occupancy forecasting, which serves as a conditional prior to guide diffusion-based planning. Conditioned on both current observations and predicted future states, the planner iteratively refines trajectory candidates to generate a reinforced driving trajectory. By explicitly modeling scene evolution over future horizons, OWMDrive captures key spatiotemporal causal dependencies, which leads to more foresighted and robust trajectory generation. Extensive experiments demonstrate that OWMDrive significantly improves planning reliability and safety, especially in challenging and partially observable driving scenarios.

[184] SA-Homo: Scale Adaptive Homography Estimation for Scale Variation Scenarios cs.CVPDF

Shangxuan Xie, Haifeng Wu, Yuhang Wang, Huarong Jia, Wen Li

TL;DR: 本文提出了一种名为SA-Homo的新型尺度自适应单应性估计框架，旨在解决图像对之间存在显著尺度差异时的鲁棒对齐问题。该方法采用分层尺度对齐策略，先通过包含多尺度线性注意力级联和跨尺度相似性矩阵块的模块进行初始粗对齐，再使用轻量级迭代模块进行局部精修。

Details

Motivation: 现有深度学习单应性估计方法在图像对尺度差异显著时性能严重下降，因为它们依赖的有限位移假设和局部特征一致性在大尺度差异下不再成立。本文旨在设计一个能适应广泛尺度差异比的鲁棒对齐框架。

Result: 在专门为尺度变化挑战构建的高分辨率多模态卫星基准数据集HMSA上进行的广泛实验表明，SA-Homo即使在8倍尺度差异下也能保持高精度，在传统尺度相似场景和具有挑战性的尺度变化场景中均优于最先进的方法。

Insight: 创新点在于分层尺度对齐策略，以及用于初始对齐的尺度感知差异桥接模块（SDBM），该模块结合了多尺度线性注意力级联（MLAC）和全局跨尺度相似性矩阵块（CSMB）以捕获长程依赖和鲁棒的跨尺度相关性。此外，构建了专门的HMSA数据集来推动该方向研究。

Abstract: Homography estimation, as one of the fundamental problems in computer vision, remains challenged by scale variation scenarios where image pairs potentially exhibit significant scale discrepancies. Existing deep learning frameworks frequently suffer from a significant performance degradation in such cases, as they rely on limited displacement assumptions and local feature consistency that might not hold under large scale gaps. In this paper, we propose SA-Homo, a novel scale-adaptive homography estimation framework designed to achieve robust alignment across a wide range of scale discrepancy ratios. We adopt a hierarchical scale alignment strategy that transitions from the global perspective with a heavy module to a local perspective with a light module. Specifically, we introduce the Scale-aware Discrepancy Bridging Module (SDBM) for initial alignment, which utilizes a Multi-scale Linear Attention Cascade (MLAC) to capture long-range dependencies and mitigate feature inconsistencies, along with a global Cross-scale Similarity Matrix Block (CSMB) for scale robust correlation representation. Once the initial scale gap is bridged, a lightweight Iterative Homography Estimation Refinement Module (IHERM) progressively polishes the result using local correlations. To facilitate this research, we contribute the HMSA dataset, a high-resolution, multi-modal satellite benchmark specifically tailored for scale-variant challenges. Extensive experiments demonstrate that SA-Homo maintains high precision even under 8$\times$ scale discrepancies, outperforming state-of-the-art methods in both conventional scale-similar scenarios and challenging scale variation scenarios. Code and collected datasets are available at https://github.com/shangxuanx330/SA_Homo

[185] Cross-Resolution Semantic Transfer for Robust Text-to-Image Retrieval in Low-Resolution Surveillance cs.CVPDF

Wenjie Qian, Bin Yang, Xiao Wang, Wenke Huang, Ling Mei

TL;DR: 本文提出了一种名为跨分辨率语义迁移（CRST）的CLIP风格框架，用于解决真实监控场景中文本到图像行人重识别（TIPR）因分辨率差异导致的性能下降问题。该框架通过分辨率条件推理、文本引导细化和CR-RDA模块，分别抑制退化视觉证据、恢复判别性线索以及稳定混合分辨率下的检索排序。在多个基准数据集上的实验表明，CRST显著提升了超低分辨率下的检索性能，同时保持了高分辨率检索的准确性。

Details

Motivation: 现有文本到图像行人重识别方法大多忽略了真实监控场景中普遍存在的分辨率差异问题，这导致了两种耦合的失效模式：证据可靠性崩溃（ERC）和排序分布漂移（RDD），严重影响了检索的鲁棒性。

Result: 在CUHK-PEDES、ICFG-PEDES和RSTPReid数据集上的实验表明，CRST方法将超低分辨率下的Rank-1和mAP指标平均提升了5.7%和5.3%，并且在稳定混合分辨率检索的同时，没有牺牲高分辨率检索的准确率。

Insight: 论文的创新点在于明确识别并建模了跨分辨率TIPR中的两种失效模式，并提出了一个统一的框架来协同处理证据可靠性和排序稳定性问题。其核心洞察是将高分辨率的邻域几何结构迁移到低分辨率检索中，以实现语义信息的跨分辨率鲁棒传递，这为处理视觉-语言任务中的域差异问题提供了新思路。

Abstract: Text-to-image person re-identification (TIPR) retrieves target persons using natural language descriptions. However, existing methods largely overlook resolution variance in real-world surveillance. They characterize cross-resolution TIPR through two coupled failure modes: Evidence Reliability Collapse (ERC), where degraded visual tokens become unreliable for grounding fine-grained text, and Ranking Distribution Drift (RDD), where mixed-resolution galleries distort similarity neighborhoods and destabilize retrieval rankings. To address this challenge, we propose Cross-Resolution Semantic Transfer (CRST), a CLIP-style framework with three modules: resolution-conditioned reasoning, text-guided refinement and CR-RDA. Resolution-conditioned reasoning estimates token reliability to suppress corrupted evidence. Text-guided refinement injects semantic priors to recover discriminative cues. CR-RDA transfers HR neighborhood geometry to stabilize LR ranking under mixed resolutions. Experiments on CUHK-PEDES, ICFG-PEDES, and RSTPReid show that CRST improves ultra-low-resolution Rank-1 and mAP on average by 5.7% and 5.3%, while stabilizing mixed-resolution retrieval without sacrificing high-resolution accuracy.The code will be made publicly available.

Mengzhu Ding, Xin Song, Xiaoke Ding, Hongwei Ding, Xuecong Liu

TL;DR: RBE-Flow是一个用于跨模态图像配准的新框架，它将密集光流估计重新表述为在学习的特征流形上的闭环循环贝叶斯估计问题。该方法通过将特征度量的非线性优化与概率状态更新深度耦合，建立了一个鲁棒的自校正机制，在多个具有挑战性的基准测试中实现了最先进的性能。

Details

Motivation: 跨模态图像配准由于严重的非线性辐射差异和几何畸变而具有根本性挑战，现有的确定性匹配方法缺乏不确定性感知，难以处理高度非凸的优化问题，并常在模糊区域累积误差。

Result: 在具有挑战性的OSdataset、WHU-OPT-SAR和RoadScene基准测试上的大量实验表明，RBE-Flow始终实现了最先进的性能，显著优于现有方法，尤其是在严格的亚像素标准下。

Insight: 核心创新在于将密集跨模态流估计重新表述为循环贝叶斯估计问题，并引入了循环流形优化（RMO）块和不确定性自适应概率更新（UAPU）模块，形成了一个具有自校正能力的闭环系统。此外，为防止方差崩溃而设计的几何感知修正负对数似然损失也是一种有效的训练策略。

Abstract: Cross-modal image registration is essential for multi-sensor perception but remains fundamentally challenging due to severe non-linear radiometric discrepancies and geometric distortions. Existing deterministic matching methods lack uncertainty awareness, struggling to navigate the resulting highly non-convex optimization landscape and frequently accumulating errors in ambiguous regions. In this paper, we propose RBE-Flow, a novel framework that reformulates dense cross-modal flow estimation as a closed-loop recurrent Bayesian estimation problem on learned feature manifolds. Diverging from standard feed-forward regression, RBE-Flow establishes a robust self-correcting mechanism by deeply coupling feature-metric non-linear optimization with probabilistic state updates. Specifically, a Recurrent Manifold Optimization (RMO) block iteratively generates flow observations and their associated uncertainties, which are then optimally assimilated into the prior state via an Uncertainty-Adaptive Probabilistic Update (UAPU) using deterministic sigma-point projection. Crucially, the resulting calibrated posterior covariance is fed back to adaptively regularize the damping of subsequent optimization steps, allowing the system to modulate its convergence based on predictive confidence. To ensure stable probabilistic training, we introduce a hybrid supervision scheme featuring a geometry-aware rectified NLL loss that structurally prevents variance collapse. Extensive experiments on challenging OSdataset, WHU-OPT-SAR, and RoadScene benchmarks demonstrate that RBE-Flow consistently achieves state-of-the-art performance, outperforming existing methods by a significant margin, particularly under strict sub-pixel criteria. Project page: https://github.com/NEU-Liuxuecong/RBE-Flow

[187] PS-MOT: Cultivating Instance Awareness from Point Seeds for Multi-Object Tracking cs.CV | cs.RO | eess.IVPDF

Kai Luo, Fei Teng, Mengfei Duan, Wanjun Jia, Xu Wang

TL;DR: 本文提出了一种名为PS-MOT（点监督多目标跟踪）的新范式，旨在用成本更低的点标注替代传统的边界框监督。为应对点监督带来的空间模糊性和身份漂移挑战，作者提出了PS-Track框架，该框架在数据、模型和损失三个层面进行设计，包括时序反馈提示、点激发的波束注意力模块和不确定性引导的高斯学习。实验在多个跟踪数据集上验证了其有效性，并达到了点监督跟踪的新SOTA水平。

Details

Motivation: 传统多目标跟踪依赖边界框标注，成本高昂。本文旨在探索一种更经济的点监督方法，但其面临空间模糊和身份漂移的挑战，因此需要设计新的方法来弥补点标注在几何结构和尺度信息上的缺失。

Result: 在DanceTrack、EmboTrack、SportsMOT和JRDB等多个多目标跟踪基准数据集上进行了实验。结果表明，PS-Track方法为点监督跟踪提供了一个可行且有效的替代方案，并在点监督跟踪任务上建立了新的最先进（SOTA）性能。

Insight: 创新点在于从数据、模型和损失三个层次构建了一个从点到实例的层次化处理流程。具体包括：利用时序反馈和运动先验生成一致性伪标签的数据策略；通过语义关联激活高频分量以“幻觉”出物体边界的注意力模块；以及将伪标签建模为概率分布以动态校准监督强度的损失函数设计。这为弱监督学习，特别是点标注在复杂视觉任务中的应用提供了新思路。

Abstract: We introduce Point-supervised Multi-Object Tracking (PS-MOT) as a cost-effective alternative to traditional bounding box supervision, shifting the focus from spatial fitting to topological center-driven representation. However, PS-MOT faces challenges, e.g., spatial ambiguity and identity drift due to the lack of explicit geometric structure and scale constraints. To address these, we propose PS-Track, a hierarchical pipeline transitioning from points to instances across data, model, and loss levels. At the data level, we introduce Temporal-Feedback Prompting (TFP) to evolve points into temporally consistent pseudo-labels using negative spatial cues and motion priors. At the model level, we design the Point-Excited Wavelet Attention (PEWA) module, which leverages semantic correlations to activate high-frequency components, ``hallucinating’’ object boundaries. At the loss level, Uncertainty-Guided Gaussian Learning (UGL) models pseudo-labels as probabilistic distributions, dynamically calibrating supervision intensity. Experiments on DanceTrack, EmboTrack, SportsMOT, and JRDB demonstrate that PS-Track provides a feasible and effective point-supervised alternative across diverse tracking scenarios, establishing a new state-of-the-art for point-supervised tracking. The source code is available at https://github.com/xifen523/PS-MOT.

[188] PGE-SAM: Prompt-Guided Feature Enhancement for Interactive Segmentation under Degradation cs.CVPDF

Tuan-Duc Nguyen, Anh-Tuan Mai, Duc-Trong Le

TL;DR: 该论文提出了PGE-SAM框架，旨在解决SAM模型在图像存在噪声、模糊和压缩等退化情况下性能显著下降的问题。该方法通过提示引导的特征增强，利用用户提示和先验掩码预测来指导特征恢复，并引入多尺度特征交互和前景重建损失来恢复细节。同时，论文构建了一个包含多种退化类型的医学图像交互分割基准DM-Seg。实验表明，PGE-SAM在医学和自然图像上均实现了SOTA的鲁棒性，且参数量增加较少。

Details

Motivation: SAM模型在真实世界图像退化（如噪声、模糊、压缩）下性能严重下降，现有方法全局恢复特征而未聚焦于分割相关区域，且忽略了SAM的迭代优化机制，导致在交互式分割场景中表现不佳。

Result: 在提出的DM-Seg基准（涵盖多种成像模态和退化严重程度）以及自然图像上进行广泛实验，PGE-SAM在多个退化级别上均达到了SOTA的鲁棒性，同时保持了在干净图像上的泛化能力，且增加的参数量少于先前方法的五分之一。

Insight: 创新点在于显式利用用户提示和先验掩码通过提示引导生成器进行空间引导的特征恢复，并结合多尺度特征交互与前景重建损失来恢复退化丢失的细节。客观来看，其将交互分割与特征增强紧密结合，并构建了专门的医学图像退化分割基准，为退化场景下的模型鲁棒性研究提供了新思路。

Abstract: Segment Anything Model (SAM) has revolutionized promptable image segmentation with strong zero-shot generalization. However, its performance degrades substantially under real-world imaging artifacts such as noise, blur, and compression. Existing methods restore features globally without focusing on segmentation-relevant regions and neglect SAM’s iterative refinement mechanism, leading to suboptimal performance in interactive settings. We propose Prompt-Guided Feature Enhancement SAM (PGE-SAM), a framework that explicitly leverages user prompts and prior mask predictions to spatially guide the feature restoration process toward regions of interest through a Prompt Guidance Generator. To recover fine-grained details lost under degradation, we introduce Multi-Scale Features Interaction to incorporate low-level encoder features, along with a Foreground Reconstruction Loss that restricts feature-level supervision to the segmentation target. Furthermore, we present DM-Seg, a benchmark for interactive segmentation on degraded medical images, spanning multiple imaging modalities with both general and modality-specific degradations at varying severity levels. Extensive experiments demonstrate that PGE-SAM achieves SOTA robustness on both medical and natural image domains across multiple degradation levels, while maintaining generalization to clean images and adding less than one-fifth of the parameters of prior methods.

[189] On the Faithfulness of Post-Hoc Concept Bottleneck Models cs.CV | cs.AIPDF

Laines Schmalwasser, Jan Blunk, Niklas Penzel, Julia Niebling, Joachim Denzler

TL;DR: 本文分析了后验概念瓶颈模型（post-hoc CBMs）的忠实性问题，指出仅依赖目标任务精度评估模型会掩盖概念语义是否真实有意义的问题。作者识别了两种导致概念表示不忠实的失败案例：辅助数据带来的协变量偏移，以及视觉语言模型生成的概念标签中的系统性噪声。为此，论文提出了新的度量标准，将概念忠实性与预测精度解耦，并在真实和合成基准测试中验证了这些度量能有效检测标准精度评估无法发现的非忠实行为。

Details

Motivation: 后验概念瓶颈模型旨在通过将潜在特征映射到可解释的概念空间来弥合深度学习表示与人类理解之间的鸿沟，但当前主要依赖目标任务精度作为成功标准，这无法保证所学概念具有真实的语义意义，可能仅是预测性伪影。

Result: 在真实世界和合成基准测试上的实证结果表明，新提出的度量标准能够识别出标准基于精度的评估所无法检测的非忠实行为，验证了其有效性。

Insight: 创新点在于直接分析学习到的投影本身，而非仅看预测精度，从而形式化了两种失败模式（协变量偏移和系统性标签噪声），并引入了将概念忠实性与预测精度解耦的新评估指标，为可解释性研究提供了更可靠的评估框架。

Abstract: Human decision-making interprets the world through high-level concepts, such as recognizing a bird by its belly color. To bridge the gap between opaque deep learning representations and human understanding, Post-Hoc Concept Bottleneck Models (post-hoc CBMs) project latent features onto interpretable concept spaces using auxiliary datasets or vision-language models. However, relying on target task accuracy as the primary measure of post-hoc CBM success obscures whether the learned concepts are semantically meaningful or merely predictive artifacts. For example, random concept projections can achieve competitive accuracy despite being semantically meaningless. In this work, we analyze the learned projections directly and identify two failure cases: First, for concept projections learned from auxiliary data, covariate shifts can lead to unfaithful concept representations for the target task. In particular, we provide an upper bound on the error introduced by this shift. Second, systematic label noise in surrogate concept labels generated by vision-language models leads to unfaithful projections. After formalizing these failure modes, we introduce novel metrics that decouple concept faithfulness from predictive accuracy. Our empirical results across real-world and synthetic benchmarks confirm that these metrics identify unfaithful behaviors that standard accuracy-based evaluation fails to detect.

[190] HASTE: A Framework for Training-Free, Dynamic, and Steerable Compression of Pre-Trained Convolutional Neural Networks cs.CVPDF

Lukas Meiner, Jens Mehnert, Alexandru Paul Condurache

TL;DR: 本文提出了HASTE框架，一种无需训练、动态且可调控的预训练卷积神经网络压缩方法。该框架通过局部敏感哈希在推理时识别并合并特征图中的冗余通道，从而降低计算成本。实验表明，HASTE能在不重新训练的情况下显著减少FLOPs，同时保持较高的准确率。

Details

Motivation: 解决在资源受限设备上部署大型预训练CNN时，现有动态执行方法通常需要专门训练或微调，且依赖数据访问的问题。

Result: 在CIFAR-10和ImageNet数据集上的多种架构实验中，ResNet34在CIFAR-10上实现了46.2%的FLOPs减少，仅损失1.25%的准确率，无需任何重新训练。

Insight: 创新点在于利用局部敏感哈希实现无需训练的通道合并，将通道合并方案与视觉Transformer中的token合并任务概念上联系起来，为预训练CNN的运行时可调控压缩提供了有效解决方案。

Abstract: Deploying large convolutional neural networks (CNNs) on resource-constrained devices is challenging due to their high computational cost. While dynamic execution methods are promising, existing approaches for CNNs typically require specialized training or fine-tuning, limiting their effectiveness when applied to pre-trained models and requiring data access. To address this gap, we propose HASTE (Hashing for Tractable Efficiency), a plug-and-play convolution module that enables training-free, dynamic compression of large pre-trained CNNs. At inference time, HASTE uses locality-sensitive hashing to identify and merge redundant channels of latent feature maps on a patch-wise basis. This process simultaneously compresses the depth of both input features and their corresponding filters, resulting in computationally cheaper convolutions. We conduct extensive experiments on CIFAR-10 and ImageNet across a range of architectures, demonstrating a 46.2% FLOPs reduction in a ResNet34 on CIFAR-10 with only a 1.25% drop in accuracy, without any retraining. We support our claims by comprehensive ablation studies to validate our core design choices, an analysis of the method’s properties and limitations, and a discussion that connects our channel merging scheme to the conceptually related task of token merging in Vision Transformers. Our results demonstrate that HASTE provides an effective solution for steerable compression of pre-trained CNNs at runtime, opening new possibilities for the deployment of efficient deep learning methods.

[191] 3D Scene-Adaptive Trajectory-Controllable Human Image Animation with Camera Movement cs.CVPDF

Deyin Liu, Jicheng Xu, Lin Yuanbo Wu, Xiaowei Zhao, Xiatian Zhu

TL;DR: 本文提出了一种场景自适应的人体图像动画框架，能够在重建的3D环境中同时控制人体运动和相机轨迹以生成视频。该方法首先开发了一种地面自适应的3D运动重定向方法，使用户能友好地控制运动轨迹并自动适应地面高度和方向的变化；其次设计了一种视角自适应的潜在融合机制，通过场景可见性掩码将点云几何先验注入生成过程，为相机控制下的视角变化提供精确引导。

Details

Motivation: 现有基于扩散/流模型的动画方法虽已开始使用3D建模条件，但在自然场景中合成相机视角变化下的轨迹可控人体运动仍面临挑战。本文旨在解决在变化的相机视角下，在自然场景中合成轨迹可控人体运动的难题。

Result: 在两个标准人体图像动画基准数据集上的实验表明，该方法在相关视频生成指标上相比现有技术（SOTA）取得了显著提升。

Insight: 创新点包括：1）地面自适应的3D运动重定向，实现了对复杂地形（如坡度、方向）的自动适应和用户友好控制；2）视角自适应的潜在融合机制，利用点云几何先验和场景可见性掩码，精确引导相机视角变化下的生成过程。从客观角度看，该方法将3D场景重建与生成模型紧密结合，为可控、场景感知的人体动画提供了系统性的解决方案。

Abstract: Human image animation, which aims to generate a video of a reference subject following a provided action sequence, has received increasing research interest. With the development of diffusion-based/flow-based video foundation models, existing animation works have began to upgrade the guidance information from 2D skeleton/pose to 3D modeling conditions. Despite achieving reasonable results, these approaches face challenges in synthesizing trajectory-controllable human motion within natural scene under changed camera views. In this work, we present a scene-adaptive human image animation framework that controls both human motion and camera trajectories within a reconstructed 3D environment for video generation. To achieve this, we first develop a ground-adaptive 3D motion retargeting approach to enable user-friendly motion trajectory control adapting to the changes of elevations of ground and orientations automatically. Then we design a viewpoint-adaptive latent fusion mechanism to inject point-cloud geometric priors through scene-visibility masking into the generative process, providing precise guidance of viewpoint changes under camera control. Experiments on two standard human image animation benchmark datasets demonstrate remarkable improvements of our method over the state of the arts in related video generation metics. Project page: https://robinhood256100.github.io/web-disp

[192] Orca: The World is in Your Mind cs.CVPDF

Yihao Wang, Yuheng Ji, Mingyu Cao, Yanqing Shen, Runze Xiao

TL;DR: Orca是一个通用的世界基础模型，它通过多模态世界信号学习统一的世界潜在空间，并通过多模态读出接口进行暴露。该模型以’下一状态预测’为核心，结合无意识学习（从连续视频中学习密集自然状态转换）和有意识学习（通过语言描述的事件和VQA监督建模稀疏有意义的状态转换）两种互补范式进行训练。在包含12.5万小时视频和1.6亿事件标注的大规模数据集上预训练后，Orca在文本生成、图像预测和具身动作生成三个下游任务上表现出色，超越了同等规模的专用基线模型。

Details

Motivation: 为了解决现有模型通常孤立优化下一词、下一帧或下一动作预测的局限性，论文旨在构建一个能够统一理解和预测世界状态转换的通用世界基础模型。

Result: 在文本生成、图像预测和具身动作生成三个代表性下游任务上，仅训练轻量级模态特定解码器（主干网络冻结）的Orca超越了同等规模的专用基线模型，验证了更强的世界潜在空间能实现更强的下游读出能力。

Insight: 创新点在于提出了以’下一状态预测’为核心的世界建模统一框架，以及结合无意识学习和有意识学习的互补训练范式。这为构建能够统一理解、预测和作用于世界的通用人工智能模型提供了一个有前景的路径。

Abstract: We introduce Orca, an initial instantiation of a general world foundation model. Orca learns a unified world latent space from multimodal world signals and exposes it through multimodal readout interfaces. Rather than optimizing isolated next-token, next-frame, or next-action prediction, we are centered on Next-State-Prediction modeling, offering a unified state-transition modeling route toward understanding, predicting, and acting upon the world. Orca learns through two complementary paradigms: unconscious learning captures dense natural state transitions from continuous videos, and conscious learning models sparse meaningful state transitions by language-described events and VQA supervision. For pre-training, we construct a large-scale world-learning inventory data, including 125K hours of video data and 160M event annotations. After pre-training, Orca learns a unified world latent space. To examine whether the learned latent supports downstream, we evaluate it by three representative downstream readouts: text generation, image prediction, and embodied action generation. Orca’s backbone is frozen, and only the lightweight modality-specific decoders are trainable. Experiments show the scalability of the proposed paradigm and verify that stronger world latent enables stronger downstream readouts. Orca outperforms similar-sized specialized baselines. These results show that Orca, as a general world foundation model, presents a promising approach to understanding, predicting, and acting upon the world. Finally, we discuss the current limitations, aiming to provide useful insights and inspiration for the community.

[193] EcoVideo: Entropy-Orchestrated Video Generation Paradigm in Cloud-Edge Dynamics cs.CVPDF

Jiayu Chen, Hengyi Zhang, Maoliang Li, Minyu Li, Zihao Zheng

TL;DR: 本文提出EcoVideo，一种基于熵编排的视频生成框架，用于在云边动态环境中优化扩散Transformer视频生成的延迟问题。该框架通过训练无关的自注意力熵估计帧间信息密度，选择关键帧由云端大模型去噪，边缘轻量模型通过运动感知插值和细化重建其余帧，并动态调整关键帧预算和边缘细化深度以适应实时带宽和计算资源，从而优化端到端延迟。

Details

Motivation: 现有DiT视频生成因迭代全帧去噪而延迟高，且先前的云边方法多依赖静态的步骤间解耦，无法利用帧间相似性或适应系统动态变化，因此需要一种能动态解耦帧间处理并适应资源约束的框架。

Result: 在代表性DiT视频生成器上的实验表明，EcoVideo在质量-效率权衡上有所改进，在低带宽、计算受限的边缘设置下实现了高达2.9倍的端到端加速。

Insight: 创新点包括：利用自注意力熵作为训练无关的帧信息密度估计器进行动态关键帧选择；云边协同的稀疏关键帧去噪与运动感知插值重建；以及基于实时带宽和计算可用性的自适应资源分配策略，以优化延迟。

Abstract: DiT video generation is latency-intensive due to iterative full-frame denoising, while prior cloud-edge methods largely rely on static inter-step decoupling and cannot leverage inter-frame similarity or adapt to system dynamics. We propose EcoVideo, an entropy-orchestrated framework for dynamic inter-frame decoupling: early-stage self-attention entropy provides a training-free estimate of frame-wise information density for frame selection; a cloud large model denoises sparse high-entropy keyframes; and an edge lightweight model reconstructs the remaining frames via motion-aware interpolation with refinement for temporal stability. EcoVideo further adapts the keyframe budget and edge refinement depth to real-time bandwidth and compute availability, optimizing end-to-end latency under constraints. Experiments on representative DiT video generators show improved quality–efficiency trade-offs and up to 2.9x end-to-end speedup in low-bandwidth, compute-limited edge settings. Code is available at https://github.com/IF-LAB-PKU/EcoVideo.

[194] Beyond 2D Matching: A Unified Single-Stage Framework for Geometry-Aware Cross-View Object Geo-Localization cs.CV | cs.AIPDF

Liyao Wang, Ruipu Wu, Haojun Xu, Lei Shi, Linjiang Huang

TL;DR: 本文提出了一种超越传统2D匹配的单阶段几何感知跨视图目标地理定位框架GAGeo，并构建了一个包含22万对地面-卫星和无人机-卫星图像的大规模高精度数据集。该框架基于π^3等变3D基础模型，通过整合视觉特征、参考提示和可学习任务令牌，在单次前向传播中联合预测边界框、分割掩码和相机姿态，并利用对比损失以卫星视图为通用锚点隐式对齐地面和无人机表示，实现零样本地面到无人机定位。

Details

Motivation: 现有跨视图目标地理定位方法过度依赖2D外观匹配，且受限于缺乏几何元数据、多样化提示和标准视场图像的数据集，因此需要解决这些交织的挑战。

Result: 大量实验表明，该方法在多个基准上显著优于现有最先进方法，并在未见场景和新型跨视图设置中展现出卓越的泛化能力。

Insight: 创新点包括构建大规模多模态提示数据集、提出基于3D先验的单阶段联合预测框架，以及利用卫星视图作为通用锚点的对比损失设计，实现了无需三元组训练数据的零样本跨模态定位。

Abstract: Cross-view object geo-localization (CVOGL) aims to locate a target object from a query view (e.g., ground or drone) within a geo-tagged reference image (e.g., satellite). Existing approaches heavily rely on 2D appearance matching and are constrained by limited datasets lacking geometric metadata, diverse prompts, and standard field-of-view imagery. To address these intertwined challenges, we first introduce \dataset, a large-scale, high-fidelity building dataset comprising over 220,000 ground-satellite and drone-satellite pairs. It provides multi-modal prompts (points, boxes, masks) and camera poses to enable flexible target referring and explicit spatial modeling. Furthermore, we propose a novel single-stage Geometry-Aware Geo-localization framework (GAGeo), built upon the permutation-equivariant 3D foundation model $π^3$. By seamlessly integrating visual features, referring prompts, and learnable task tokens, our model adapts the inherited 3D prior to jointly predict bounding boxes, segmentation masks, and camera poses in a single forward pass. Additionally, we introduce a contrastive loss that utilizes the satellite view as a universal anchor, implicitly aligning ground and drone representations to enable zero-shot ground-to-drone localization without requiring triplet training data. Extensive experiments demonstrate that our approach significantly outperforms state-of-the-art methods, exhibiting exceptional generalization ability in unseen scenes and novel cross-view setups.

[195] Learning from Reliable Latent Prompts for Visual Recognition with Missing Modalities cs.CVPDF

Taixi Chen, Nancy Guo

TL;DR: 该论文提出了一种名为’从可靠潜在提示学习’的新范式，用于解决多模态视觉识别中模态缺失的问题。该方法将可学习的提示建模为与输入无关的稳定潜在锚点，而非从可能损坏的实例级特征动态生成，从而在模态严重缺失的情况下仍能提供鲁棒的指导和跨模态知识补偿。

Details

Motivation: 现实场景中模态缺失输入普遍存在，导致为完整模态数据优化的模型性能急剧下降。现有基于实例级特征生成动态提示的方法，其可靠性会随着缺失率升高而降低，限制了模型性能。

Result: 在三个基准数据集上的实验结果表明，该方法在广泛的模态缺失场景下（包括高达90%的极端缺失率）均取得了最先进的性能。

Insight: 核心创新在于假设并利用可学习的潜在提示本身封装了稳定的、模态内在的先验知识，这些知识与损坏的输入解耦。这提供了一种新的、输入无关的提示学习范式，通过稳定的潜在锚点实现鲁棒的跨模态知识补偿，为解决模态缺失问题提供了更稳健的方案。

Abstract: Large-scale multimodal models (LMMs) have achieved superior performance in visual recognition by synergizing information across diverse, massive-scale paired modalities. In real-world scenarios, however, missing-modality inputs are ubiquitous, causing models optimized for modality-complete data to exhibit precipitous performance degradation. Existing research has introduced prompt learning to mitigate this issue, typically by generating dynamic prompts from instance-level features, regardless of whether the input modalities are complete or partially absent. However, such input-conditioned strategies are hindered by the escalating unreliability of instance-level features; as higher missing rates increase the proportion of incomplete modalities, the resulting instability in prompt learning limits the model’s performance. To address this limitation, we hypothesize that learnable latent prompts themselves encapsulate stable, modality-intrinsic priors that are decoupled from corrupted inputs. Consequently, we propose a novel paradigm: Learning from Reliable Latent Prompts. Unlike prior methods, we model input-agnostic learnable prompts as stable latent anchors that enable robust guidance and effective cross-modal knowledge compensation, even under extreme missing rates (e.g., 90%). Empirical results across three benchmark datasets demonstrate that our “learn-from-latent-prompts” approach achieves state-of-the-art performance across a wide range of missing-modality scenarios. Extensive experiments further confirm the effectiveness of this paradigm in providing a robust solution to the missing-modality problem.

[196] Towards in-the-wild Egocentric 3D Hand-Object Pose Estimation cs.CVPDF

Siddhant Bansal, Zhifan Zhu, Shashank Tripathi, Jiahe Zhao, Michael J. Black

TL;DR: 本文提出了一个名为EPIC-Contact的野外第一人称视角数据集，包含密集的3D手-物体接触对应关系和网格姿态，并提出了一个名为HOPformer的端到端Transformer模型，用于联合预测双手和物体的姿态。该模型在ARCTIC数据集上取得了82.4%的成功率，在EPIC-Contact数据集上显著提升了性能。

Details

Motivation: 解决从野外第一人称RGB视频中估计精确3D手-物体姿态的挑战，这些挑战主要源于严重遮挡和接触模糊。现有基于学习的方法泛化能力不足且受限于监督数据的稀缺。

Result: 在实验室3D数据集ARCTIC上，HOPformer达到82.4%的成功率，比当前SOTA提升6.2个百分点。在新提出的EPIC-Contact数据集上，成功率和接触偏差分别提升近一倍和减少75%。

Insight: 创新点包括引入大规模野外第一人称接触数据集EPIC-Contact，以及提出HOPformer这一端到端Transformer架构，其交叉注意力解码器利用手部先验来增强物体特征，从而实现鲁棒的联合姿态估计。

Abstract: Estimating accurate 3D hand-object pose from in-the-wild egocentric RGB remains challenging due to severe occlusions and ambiguous contact. Existing learning-based methods often struggle to generalise to in-the-wild scenes and are limited by the scarcity of supervision. We address these issues with two contributions. First, we introduce EPIC-Contact, an in-the-wild egocentric dataset of 2.3K clips (62.3K frames) with dense, bijective 3D hand-object contact correspondences and posed meshes. Second, we propose HOPformer, an end-to-end transformer that jointly predicts bi-manual hand and object pose in a single forward pass. A cross-attention decoder conditions object features on hand priors, producing robust pose estimation. We test HOPformer on the in-lab 3D dataset, ARCTIC, as well as our newly introduced EPIC-Contact dataset. HOPformer reaches 82.4% success rate on ARCTIC (+6.2 pts over current SOTA). On EPIC-Contact, it nearly doubles the success rate while reducing contact deviation by 75%. EPIC-Contact, HOPformer code and checkpoints are released: https://sid2697.github.io/epic-contact.

[197] UnfoldArt: Zero-Shot Recovery of Full Articulated 3D Objects from Text or Image cs.CVPDF

Mohamed el amine boudjoghra, Ivan Laptev, Angela Dai

TL;DR: 本文提出UnfoldArt，一种基于辩论驱动智能体方法的零样本全关节3D物体重建技术，能够从文本或图像输入中恢复物体的关节结构、隐藏几何和内部结构。该方法通过高层和低层智能体进行两轮结构化辩论，结合视频生成先验，联合推断关节参数并重建完整的3D关节物体。

Details

Motivation: 关节3D物体在具身AI、机器人和虚拟现实中至关重要，但现有方法因缺乏监督数据或先验知识，难以从稀疏观测中可靠恢复关节结构、隐藏几何和内部结构。

Result: 该方法能够生成高保真几何、内部结构以及运动一致的状态，超越了直接观测表面的限制，实现了对完整3D关节物体的零样本重建。

Insight: 创新点在于首次提出辩论驱动的智能体方法，结合高层语义推理与低层参数估计，并利用视频生成先验暴露遮挡几何，实现了关节结构与完整几何的联合推断。

Abstract: Articulated 3D objects are essential for interactive environments in embodied AI, robotics, and virtual reality, but reconstructing their structure and motion from sparse observations remains challenging. Existing approaches remain largely constrained by lack of supervised data or lack the priors needed to reliably recover articulation, hidden geometry, and internal object structure. We present the first debate-driven agentic approach to articulated 3D object reconstruction from text or image inputs that both grounds articulation reasoning in concrete motion and exposes the occluded geometry revealed under articulation. High-level agents reason about object semantics and motion using knowledge from vision-language and video models, while low-level agents estimate articulation parameters and interaction points; together, they engage in a two-round structured debate that first exploits global–local disagreement and then grounds the agents in freely generated video. The same video prior, conditioned on the agreed articulation, then drives each part through its motion to expose occluded interiors and geometry that cannot be inferred from a single static view. By combining agentic reasoning with a video generative prior, our approach jointly infers articulation and reconstructs complete 3D articulated objects, producing high-fidelity geometry, internal structure, and motion-consistent states beyond directly observed surfaces.

[198] Goku: A Million-Scale Universal Dataset and Benchmark for Instruction-Based Video Editing cs.CVPDF

Sen Liang, Cong Wang, Zhentao Yu, Fengbin Guan, Zhengguang Zhou

TL;DR: 本文提出了Goku，一个包含200万高质量指令对齐视频编辑对的大规模通用数据集，首次将任务边界从基础外观编辑扩展到多任务和结构操作（如精确控制主体运动）。为了解决这些复杂任务中的数据合成挑战，作者设计了一个高效的数据合成流程，将复杂编辑分解为可控子问题，并引入了渐进式过滤系统以确保数据可靠性。此外，作者探索了在Goku上的最优网络结构，提出了Goku-Edit模型，该模型利用MLLM作为文本编码器，并采用解耦的双分支设计：专用掩码分支处理结构控制，主分支专注于外观渲染。同时，作者还提出了一个全面的视频编辑基准Goku-Bench，包含1000个人工验证的测试用例和7个新颖的编辑特定指标。在Goku-Bench上评估，Goku-Edit在指令遵循方面比其他开源模型提升了高达8%。

Details

Motivation: 现有基于指令的视频编辑数据集通常专注于单任务外观编辑，无法满足现实世界场景中复杂的创意需求。为了弥补这一差距，作者旨在创建一个大规模、多任务的指令视频编辑数据集和基准。

Result: 在提出的Goku-Bench基准上评估，Goku-Edit模型在指令遵循方面比其他开源模型提升了高达8%。

Insight: 主要创新点包括：1) 构建了首个大规模、覆盖多任务和结构操作的指令视频编辑数据集Goku；2) 设计了高效的数据合成流程和渐进式过滤系统以应对复杂编辑任务的数据挑战；3) 提出了Goku-Edit模型，采用基于MLLM的文本编码器和解耦的双分支架构，分别处理结构控制和外观渲染；4) 建立了包含新颖评估指标的综合性基准Goku-Bench。

Abstract: Existing instruction-based video editing datasets commonly focus on single-task appearance editing, failing to meet the complex creative demands of real-world scenarios. To bridge this gap, we present Goku, a large-scale dataset featuring 2 million high-quality, instruction-aligned video editing pairs, which is the first to extend task boundaries from basic appearance editing to multi-task and structural manipulations(e.g., precise control of subject movement). To tackle the data synthesis challenges inherent in these complex tasks, we design an efficient data synthesis pipeline that decomposes complex edits into controllable sub-problems and introduce a progressive filtering system for data reliability throughout the whole process. Furthermore, we explore the optimal network structures on Goku, and propose Goku-Edit. To deeply comprehend complex editing instructions, Goku-Edit leverages an MLLM as its text encoder and adopts a decoupled dual-branch design: a dedicated mask branch handles structural control, freeing the main branch for appearance rendering. A comprehensive video editing benchmark, Goku-Bench, is also proposed with 1,000 human-verified test cases and 7 novel editing-specific metrics. Evaluated on Goku-Bench, Goku-Edit obtains up to +8% improvement on other open-source models in terms of instruction following.

[199] Reweighting Framewise Attention in Video Transformers for Facial Expression Understanding cs.CVPDF

Seongro Yoon, Donghyeon Cho, Jinsun Park, François Brémond

TL;DR: 本文提出了一种名为MiRA（Marginal-induced Attention Redistribution）的即插即用框架，用于增强视频Transformer模型在面部表情理解任务中对细微面部动态的时空选择性。该方法通过分析自注意力图，估计帧级重要性并重新分配注意力，无需引入额外可训练参数。实验在面部表情识别基准上验证了其对强ViT基线的持续改进。

Details

Motivation: 现有基于ViT的视频模型虽然通过大规模自监督预训练表现出色，但其注意力机制往往侧重于主导的全局运动和粗糙的时间动态，限制了其对细粒度面部变化的敏感性。本文旨在解决这一局限性，以更好地理解无约束条件下视频中微妙且局部化的面部动态。

Result: 在具有挑战性的面部表情识别基准测试中，MiRA框架在强ViT基线上取得了持续的性能提升，表明其有效性。

Insight: 主要创新点在于提出了一个无需额外参数的注意力重分配框架MiRA，它通过帧级置信度和帧内集中度统计量来估计帧重要性，并精确或高效地（通过flashLite模式）将注意力重新导向时空局部化的面部线索，从而增强模型对细微动态的捕捉能力。

Abstract: Understanding facial expressions in videos requires modeling subtle and localized facial dynamics under unconstrained conditions. Although recent Vision Transformer(ViT)-based video models have shown strong performance through large-scale self-supervised pretraining, their attention mechanisms often emphasize dominant global motions and coarse temporal dynamics, limiting sensitivity to fine-grained facial variations. To address this limitation, we propose MiRA (Marginal-induced Attention Redistribution), a plug-in frame-marginal attention redistribution framework for ViT backbones that enhances spatio-temporal selectivity toward subtle facial dynamics without introducing additional trainable parameters. MiRA derives frame-level confidence and intra-frame concentration statistics from self-attention maps to estimate frame-wise marginal importance and redistribute attention toward spatiotemporally localized facial cues. We first introduce a principled \textit{exact mode} based on post-softmax attention redistribution. To further improve efficiency, we propose \textit{flashLite mode}, a lightweight pre-softmax approximation that integrates frame-marginal redistribution into FlashAttention kernels while preserving the effectiveness of the exact formulation. Experimental results on challenging Facial Expression Recognition(FER) benchmarks demonstrate consistent improvements over strong ViT baselines.

[200] Open-Vocabulary and Referring Segmentation for 3D Gaussians Using 2D Detectors cs.CVPDF

Jameel Hassan, Yasiru Ranasinghe, Vishal Patel

TL;DR: 本文提出了GaussDet方法，旨在为3D高斯泼溅（3DGS）场景重建技术赋予开放词汇和指代表达式的语义理解能力。该方法通过利用2D开放词汇目标检测器，为每个3D高斯点学习实例特征，并通过多视图检测结果的语义投票聚合，生成鲁棒的视图聚合语义标签分布（VASD），从而避免了直接蒸馏CLIP特征的限制。

Details

Motivation: 现有方法通过将CLIP特征蒸馏到3DGS场景中来实现语言驱动的理解，但存在实例分组机制需要预定义数量或受噪声影响的问题，且依赖CLIP限制了语义理解仅为简单名词短语，无法处理复杂的空间推理和指代表达式。

Result: 在开放词汇分割（LeRF-OVS, ScanNet）和指代表达式定位（Ref-LeRF）两个关键任务上的广泛评估表明，GaussDet相比现有方法取得了持续改进，特别是在严格的零样本设置下，指代表达式定位任务实现了16.7% mIoU的显著提升。

Insight: 创新点在于利用离散的2D开放词汇检测器替代密集的CLIP特征蒸馏，通过多视图语义投票聚合（VASD）来正则化实例分组噪声，从而实现了从简单语言查询到复杂指代表达式定位的零样本扩展，增强了3D场景的语义理解能力。

Abstract: 3D Gaussian Splatting (3DGS) has emerged at the forefront of 3D scene reconstruction. Extending 3DGS with language-driven, open-vocabulary understanding has gained significant attention for real-world applications such as embodied AI. Recent methods achieve this by learning an instance feature attribute and assigning semantics by distilling high-dimensional Contrastive Language-Image Pretraining (CLIP) features directly into the scene representation. However, the instance grouping mechanisms of these methods either require a predefined number of instances or suffer from noise in their bottom-up grouping strategies. Furthermore, the reliance on CLIP restricts semantic understanding to simple noun phrases, preventing complex spatial reasoning and referential expression grounding. We present GaussDet, a method that circumvents the need for dense CLIP features by leveraging discrete, open-vocabulary 2D object detectors with referring expression capabilities. We learn instance features for individual Gaussians to decompose the scene into 3D instance groups. By rendering these groups and aggregating semantic votes from multi-view 2D detections, we generate a robust View-Aggregated Semantic Label Distribution (VASD) for each 3D instance. This view-aggregation strategy acts as a strong regularizer, attenuating spurious labels caused by low-quality instance grouping. Our approach enables a straightforward, zero-shot extension from simple language queries to complex referential grounding. Extensive evaluations across two key tasks – open-vocabulary segmentation (LeRF-OVS, ScanNet) and referring expression grounding (Ref-LeRF) – demonstrate that GaussDet achieves consistent improvements over existing methods. Most notably, we achieve a substantial 16.7% mIoU improvement in referential grounding within a strict zero-shot setting.

cs.AI [Back]

[201] GPTNT: Benchmarking Real-Time Collaboration Between Multimodal Agents on Keep Talking And Nobody Explodes cs.AI | cs.CLPDF

Amit Parekh, Sabrina McCallum, Kareem Al-Hasan, Malvina Nikandrou, Alessandro Suglia

TL;DR: 本文提出了GPTNT基准测试，基于合作视频游戏《Keep Talking and Nobody Explodes》，用于评估多模态智能体在实时协作任务中的表现。该基准模拟了时间压力、信息不对称和不完美通信等真实协作条件，要求两个智能体（一个能操作炸弹但无拆弹手册，另一个有手册但无法操作）通过高效沟通协作拆弹。

Details

Motivation: 现有基准通常孤立地研究协作所需的条件，而真实协作往往同时涉及时间压力、信息不对称和不完美通信。本文旨在填补这一空白，创建一个能综合评估多模态智能体实时协作能力的基准。

Result: 测试表明，当前最先进的闭源和开源模型在GPTNT基准上均无法在实时条件下成功拆除任何炸弹，而人类玩家可以轻松达到这一标准。控制实验揭示了模型在状态跟踪、时间压力下的高效行动、模糊处理及错误恢复等方面存在关键弱点。

Insight: 创新点在于构建了一个能分离协作能力与记忆解决方案的基准，通过可替换指令手册或合作伙伴来评估模型的即时推理能力。该基准基于真实游戏，受益于程序生成和活跃的模组社区，可随模型进步而持续演化，避免一次性被解决而失效。

Abstract: Multimodal models are increasingly deployed to solve tasks collaboratively with humans or other artificial agents. Existing benchmarks show that these models possess many of the required component capabilities, but the conditions that coincide in collaboration, including time pressure, information asymmetry, and imperfect communication, are usually studied in isolation. We introduce GPTNT, a benchmark built on the cooperative video game Keep Talking and Nobody Explodes, in which two agents must coordinate to defuse procedurally generated bomb puzzles against a live countdown. One agent can see and manipulate the bomb but does not have the defusal instructions; the other has the instructions but cannot see or manipulate the bomb. Neither agent can succeed alone: success requires effective and efficient communication. Unlike turn-based proxies, GPTNT requires agents to act asynchronously and communicate in real time. GPTNT is designed to separate collaboration from reliance on memorized solutions: the instruction manual, the partner, or both can be withheld to isolate what a model derives in the moment from what it already knows. We show that GPTNT poses a substantial challenge for state-of-the-art systems: none of the closed- or open-source models we test defuses a single bomb in real time, a bar that human players clear. Through controlled experiments, we identify critical weaknesses in state tracking, efficient action under time pressure, ambiguity handling, and error recovery. We release GPTNT as a benchmark for collaborative performance that current evaluations leave unmeasured. Because it runs on the real game, GPTNT benefits from procedural generation and inherits a living modding community, allowing the benchmark to evolve as models improve rather than being solved once and retired.

[202] Low-cost concept-based localized explanations: How far can we get with training-free approaches? cs.AI | cs.CL | cs.CVPDF

Darian Fernández-Gutiérrez, Rafael Bello, Marilyn Bello, Natalia Díaz-Rodríguez

TL;DR: 本文评估了中等规模多模态大语言模型在零样本条件下执行局部化概念命名的能力，提出了可复现的零样本概念命名评估协议，包括封闭集类别约束提示和基于嵌入相似度的开放集策略。实验表明，四种MLLM模型在多个数据集上达到62%-88%的对象级精确匹配准确率，证明了无需训练的局部区域概念标注潜力。

Details

Motivation: 概念可解释人工智能缺乏细粒度概念标注数据，限制了验证能力，研究旨在探索训练无关方法在局部概念命名任务中的极限表现。

Result: 在严格零样本条件下，四种7B-32B参数的MLLM模型在对象级概念命名任务中达到62%-88%的精确匹配准确率，展示了跨数据集的一致性性能趋势。

Insight: 创新点在于设计了可复现的零样本评估框架，通过封闭集提示和嵌入相似度策略分别应对中等和大规模标签空间，为低成本C-XAI研究提供了标准化工具。

Abstract: Concept-based Explainable AI (C-XAI) seeks human-understandable explanations grounded in semantic concepts, yet validation is limited by the scarcity of fine-grained concept annotations. We evaluate whether mid-scale Multimodal Large Language Models (MLLMs) can perform localized concept naming under strict zero-shot conditions by assigning labels to bounding-box regions at both object and part levels. We propose a reproducible zero-shot evaluation protocol for Concept Naming (CoNa) with (i) closed-set, category-constrained prompting for moderate vocabularies and (ii) Open-CoNa, an embedding-similarity-based strategy for large label spaces. Experiments with four MLLMs (7B-32B) show consistent performance trends across datasets, reaching 62%-88% object-level exact-match accuracy, highlighting the potential of training-free concept annotation from localized regions. We discuss limitations and failure modes and release a reproducible framework to support future low-cost C-XAI research.

[203] Evidence-Informed LLM Beliefs for Continual Scientific Discovery cs.AI | cs.CL | cs.LGPDF

Dhruv Agarwal, Reece Adamson, Andrew McCallum, Peter Clark, Ashish Sabharwal

TL;DR: 本文针对LLM在开放式科学发现中使用的’贝叶斯惊奇’指标提出改进，指出其静态特性与人类持续推理的非平稳性不匹配，进而提出证据驱动的LLM信念更新机制，通过结合检索增强生成技术更新先验信念，并改进搜索策略以过滤虚假奖励并最大化多样性，在五个发现领域平均提升非平稳惊奇积累30.62%。

Details

Motivation: 现有方法（如AutoDiscovery）将贝叶斯惊奇视为静态量，但人类科学发现中的惊奇感是随经验演变的非平稳过程，这阻碍了LLM实现持续科学发现。

Result: 在五个发现领域的实验中，改进后的方法比原始搜索程序平均增加30.62%的非平稳惊奇积累，并识别出37.5%的静态惊奇值为虚假奖励。

Insight: 创新点在于提出证据驱动的非平稳信念更新框架，结合嵌入检索增强生成实现先验更新，并通过信念更新过滤与多样性最大化双重机制优化搜索过程，为持续科学发现提供了动态奖励评估范式。

Abstract: Open-ended scientific discovery with large language models (LLMs) increasingly operates as a long-horizon loop of hypothesis search and verification, where a reward signal guides which hypotheses to test next. A notable recent example is AutoDiscovery, which uses “Bayesian surprise” - the belief shift an LLM undergoes after observing evidence for a hypothesis - as both a discovery metric and a reward for search. We first observe that AutoDiscovery treats surprisal as a static quantity, while surprisal in human reasoning is non-stationary - it is defined relative to beliefs that evolve with experience, a prerequisite for continual scientific discovery. We address this mismatch with evidence-informed LLM beliefs: priors updated with evidence from previous hypotheses to compute non-stationary surprisal for new hypotheses. We compare in-context belief-updating mechanisms and find that embedding-based retrieval-augmented generation over prior discoveries best anticipates eventual posteriors, identifying 37.5% of static surprisals as spurious. We then modify search to avoid these spurious rewards and prioritize hypotheses that remain surprising under non-stationary beliefs. Concretely, we introduce two complementary changes to the original search procedure: belief-update filtering and diversity maximization. Across five discovery domains, our method increases accumulated non-stationary surprisal by 30.62% on average compared to the original search procedure, demonstrating that continual scientific discovery with LLMs requires not only better belief measurement but also search procedures that avoid redundancy and encourage diversity.

[204] PolicyGuard: A Dialogue-Grounded Sub-Agent Verifier for Policy Adherence in LLM Agents cs.AI | cs.CLPDF

Seongjae Kang, Taehyung Yu, Sung Ju Hwang

TL;DR: 本文提出了POLICYGUARD，一个用于确保LLM智能体在对话中遵守公司政策的子代理验证器。它通过利用完整的对话上下文、对政策和对话进行自我推理，并为智能体的下一轮对话提供可操作的反馈，来解决多轮、依赖对话内容的政策遵从性问题。

Details

Motivation: 现有方法将政策遵从性视为安全保障问题，通过外部检查来阻止不合规的代理动作，但这低估了真实工作流程的复杂性。真实流程涉及多轮对话、需要明确的用户确认和前提条件读取，并且依赖于整个对话内容而非单个参数值。

Result: 在tau^2-BENCH airline基准测试中，针对GPT-5.4、Claude Sonnet 4.6和Gemini 2.5 Pro三个模型供应商，POLICYGUARD将PASS4分数分别提升了+12.0、+6.0和+12.0个百分点。逐调用分析表明，它在保持更高政策违规召回率的同时，其拦截频率仅为参数级防护器的一半左右。

Insight: 论文的创新点在于将政策遵从性视为一个需要完整对话上下文和内部推理的广泛问题，而非简单的参数过滤。POLICYGUARD作为一个子代理验证器，其核心设计是共享代理的对话视图、进行上下文推理并提供具体修复指导，这为构建更可靠、更符合业务流程的LLM智能体提供了新的架构思路。

Abstract: LLM agents handle user requests on behalf of organizations through tool calls and must follow the company policies stated in their system prompts. Prior work approaches this as a safeguarding problem – external checks that block non-compliant agent actions. We argue that policy adherence is a broader problem: real workflows unfold across many turns, require explicit user confirmation and prerequisite reads, and hinge on the content of the dialogue rather than on any single argument value. Meeting this bar requires (i) full conversation context, (ii) self-reasoning over the policy and the current dialogue, and (iii) conversation-specific remediation that guides the agent’s next turn – three capabilities that prior safeguard work has often underestimated. We introduce POLICYGUARD, a sub-agent verifier that shares the agent’s view of the dialogue, reasons over the policy in context, and provides actionable feedback for the agent’s next turn. On tau^2-BENCH airline across three vendors (GPT-5.4, Claude Sonnet 4.6, Gemini 2.5 Pro) with four trials per setting, POLICYGUARD improves PASS4 by +12.0 / +6.0 / +12.0 pp. Per-call analyses show POLICYGUARD achieves higher policy-violation recall while blocking roughly half as often as argument-level guards.

[205] The Complexity Ceiling Benchmark: A Multi-Domain Evaluation of Sequential Reasoning Under Depth Scaling cs.AI | cs.CLPDF

Shubh Chapra, Dhruv Kumar, Murari Mandal, Yash Sinha

TL;DR: 论文提出了复杂度天花板基准（CCB），用于评估语言模型在序列推理任务中随着所需步骤增加而性能衰减的模式。该基准通过固定任务语义内容、仅改变深度N（5到50步），在三个结构不同的领域（空间状态跟踪、抽象符号指针操作、传递关系推理）上测试了五个前沿和开源LLM。研究发现模型性能呈现几何级逐步衰减，且不同领域的性能天花板差异显著：前两个领域最强模型在N=50时保持pd>0.92，而第三个领域所有模型在N=5时崩溃，最佳模型的50%成功率仅约4.7步。

Details

Motivation: 为了解决现有基准在评估语言模型长序列推理能力方面的不足，特别是缺乏对推理深度扩展下性能衰减模式的系统性研究，作者旨在量化模型在不同结构任务中随步骤增加而出现的推理崩溃现象。

Result: 在CCB基准的6000次试验中，所有测试模型均表现出几何级逐步衰减模式，其中传递关系推理领域性能崩溃最严重（最佳模型50%成功率仅4.7步）。提出的跟踪指标（TFBC）显示14.5%的正确答案通过错误中间推理得出，强制详细状态跟踪未能提升性能天花板（McNemar p=1.000）。

Insight: 创新点在于设计了控制语义内容、仅变化深度的多领域基准，揭示了模型推理的几何衰减规律和领域特异性天花板；提出的k*指标（首次推理分歧步骤）比参数量更能预测领域内准确率，为长视野推理能力提供了可解释的量化框架。

Abstract: We introduce the Complexity Ceiling Benchmark (CCB), a controlled evaluation of how language-model reasoning decays as the number of required sequential steps grows. CCB fixes the semantic content of a task and varies only its depth N in {5,…,50} across three structurally distinct regimes: grounded spatial state-tracking, abstract symbolic pointer manipulation, and transitive relational inference. Across 6,000 trials over five frontier and open-weight LLMs we find a consistent pattern of geometric per-step decay with widely separated domain ceilings: on the first two regimes the strongest models retain pd>0.92 across N=50; on the third every model collapses by N=5, with the best model’s 50%-success horizon at H0.5~4.7 steps despite pd=0.863. A trace-level metric (TFBC) shows that 14.5% of correct answers across the benchmark are reached via incorrect intermediate reasoning. Forced verbose state-tracking does not move the ceiling (McNemar p=1.000), and the mean step at which reasoning first diverges, k*, predicts within-domain accuracy better than parameter count. CCB and the geometric decay model together reduce a model’s long-horizon reasoning profile to one interpretable number per task family.

[206] Mixture of Debaters: Learn to Debate at Architectural Level in Multi-Agent Reasoning cs.AI | cs.CL | cs.MA | cs.MMPDF

Dayong Liang, Kaisong Gong, Yi Cai, Changmeng Zheng, Xiao-Yong Wei

TL;DR: 本文提出了一种名为Mixture of Debaters（MoD）的统一框架，旨在解决现有多智能体辩论框架的两个关键局限：静态架构和计算开销大。MoD利用混合专家（Mixture-of-Experts）范式，在单个模型内实现动态自辩论，通过双路由、动量切换和统一自辩论三个关键技术，在降低延迟和令牌消耗的同时提升推理性能。

Details

Motivation: 现有基于多智能体的辩论框架存在两个主要问题：一是其架构是静态的，智能体角色和协作模式在设计时即固定，缺乏灵活性；二是需要实例化多个模型副本，导致巨大的计算开销。MoD旨在通过一个统一的、动态的架构来解决这些问题。

Result: 在多模态基准测试上的广泛实验表明，MoD在性能上超越了单模型基线和传统的多智能体系统，实现了更高的准确率，同时延迟降低了3.7倍，令牌消耗减少了87%。

Insight: 核心创新点在于将混合专家范式与辩论推理相结合，提出了三个关键技术：1）双路由机制，将角色分配与流程解耦，动态决定何时辩论与何时综合；2）动量切换，利用局部上下文平滑令牌级路由，减少专家切换的抖动；3）统一自辩论，将不同的辩论角色封装到轻量级专家模块中，在保持行为多样性的同时消除了智能体间通信的开销。这为在单个模型内实现高效、动态的多角色推理提供了新思路。

Abstract: Existing multi-agent debate frameworks suffer from two critical limitations: they rely on static architectures where agent roles and coordination patterns are fixed at design time, and they require instantiating multiple model copies, incurring substantial computational overhead. We propose Mixture of Debaters (MoD), a unified framework that enables dynamic self-debate within a single model by leveraging the Mixture-of-Experts paradigm. We address three key challenges in adapting MoE for dialectical reasoning: (1) dual-routing that decouples role allocation from process flow, dynamically determining when to debate versus when to synthesize; (2) momentum switching that smooths token-level routing with local context, reducing expert-switch jitter; and (3) unified self-debate that encapsulates diverse debating personas into lightweight expert modules, eliminating inter-agent communication while preserving behavioral diversity. Extensive experiments on multimodal benchmarks demonstrate that MoD outperforms both single-model baselines and conventional multi-agent systems, achieving superior accuracy with 3.7x lower latency and 87% reduction in token consumption.The source code can be accessed at https://github.com/YongLD/MoD.

[207] UCOB: Learning to Utilize and Evolve Agentic Skills via Credit-Aware On-Policy Bidirectional Self-Distillation cs.AI | cs.CLPDF

Songjun Tu, Chengdong Xu, Qichao Zhang, Yiwen Ma, Yaocheng Zhang

TL;DR: 本文提出了UCOB框架，用于通过信用感知的在线策略双向自蒸馏来学习和进化智能体技能。该方法将技能条件提示和无技能提示视为同一模型的两个在线策略上下文视图，通过比较它们在相同任务和锚定状态下的预期回报，将更高回报的视图作为局部教师，从而内部化有用的技能行为并纠正误导性技能使用。

Details

Motivation: 现有技能记忆方法通过重用过去的经验作为文本指导来改进智能体强化学习，但检索到的技能并非总是可靠的，它们可能在某个状态下有帮助，却在另一个状态下误导同一策略。这使得常见的特权教师假设（即技能条件提示可被视为无技能提示的固定教师）变得脆弱。

Result: 在ALFWorld、WebShop和Search-QA等智能体任务上的实验表明，UCOB在多个模型规模上均优于无技能RL、技能记忆基线方法和自蒸馏方法，在ALFWorld和WebShop上分别比SOTA基线提升了23.5和18.0个点。

Insight: 核心创新在于提出了信用感知的在线策略双向自蒸馏机制，通过动态比较技能条件与无技能策略在同一状态下的预期回报来分配信用，从而更可靠地利用和进化技能记忆，并指导技能检索与反思自训练，避免了固定教师假设的脆弱性。

Abstract: Skill memories can improve agentic reinforcement learning by reusing past experience as textual guidance, but retrieved skills are not oracular: they may help in one state while misleading the same policy in another. This makes the common privileged-teacher assumption fragile, namely that a skill-conditioned prompt can be treated as a fixed teacher for the no-skill prompt. We introduce UCOB, a framework for learning to utilize and evolve agentic skills via credit-aware on-policy bidirectional self-distillation. UCOB treats skill-conditioned and no-skill prompts as two on-policy context views of the same model, compares their return-to-go within the same task and anchor state, and uses the higher-return view as the local teacher. This local credit signal internalizes useful skill-conditioned behavior, corrects misleading skill usage, and guides task/state skill memory updates, utility-aware retrieval, and reflection self-training. Experiments on agentic tasks, including ALFWorld, WebShop, and Search-QA, show that UCOB outperforms skill-free RL, skill-memory baselines, and self-distillation methods across model scales, with up to 23.5 and 18.0 point gains over SOTA baselines on ALFWorld and WebShop. Ablations and analyses further validate its core mechanisms and efficiency.

[208] GUICrafter: Weakly-Supervised GUI Agent Leveraging Massive Unannotated Screenshots cs.AI | cs.CL | cs.CVPDF

Sunqi Fan, Lingshan Chen, Runqi Yin, Qingle Liu, Yongming Rao

TL;DR: GUICrafter是一个弱监督的图形用户界面（GUI）智能体，旨在解决GUI智能体领域数据稀缺且标注成本高昂的问题。它通过一个两阶段的课程学习框架，首先利用大规模未标注的屏幕截图和网页学习视觉定位，然后使用少量高质量数据通过强化学习进行校准，从而大幅减少对人类标注的依赖。

Details

Motivation: 当前GUI智能体面临数据难以从互联网直接获取、大规模收集成本高的问题，导致其跨设备泛化能力差和对细粒度GUI元素的视觉定位能力有限。本研究的动机是探索如何利用大量未标注数据来应对这一数据挑战。

Result: 实验表明，GUICrafter仅使用UI-TARS系统0.1%的数据量，就达到了与之相当甚至更优的性能。在相同标注数据量下，它超越了所有先前的方法，如GUI-R1。

Insight: 论文的核心创新点在于提出了一个弱监督的课程学习框架，通过利用GUI交互中固有的丰富上下文信号从未标注数据中学习视觉定位，从而显著降低了对昂贵人工标注的依赖。从客观角度看，这种利用大规模未标注数据预训练、再用少量标注数据微调的模式，为数据稀缺的GUI智能体领域提供了一种高效且可扩展的解决方案。

Abstract: Data, as the fundamental substrate of modern intelligence, has greatly driven the development of current foundation models. Naturally, researchers aim to extend this paradigm to the domain of GUI agents, hoping to build strong GUI agents through a similar paradigm. However, GUI agent data cannot be directly harvested from the internet, making it costly and difficult to collect at scale. As a result, current GUI agents suffer from poor cross-device generalization and limited visual grounding ability for fine-grained GUI elements. As an attempt to address data challenge in GUI agents, we propose GUICrafter, a weakly-supervised GUI agent leveraging massive unannotated screenshots to substantially reduce the reliance on expensive human annotations. GUICrafter explores a curriculum learning framework for training GUI agents through two progressive stages. First, the model learns visual grounding from large-scale unannotated screenshots and webpages, leveraging the rich contextual signals inherent in GUI interactions without human annotations. Then, in Stage 2, we leverage a small amount of high-quality data to calibrate the model via reinforcement learning. Experiments show that GUICrafter achieves competitive, or even superior, performance to advanced systems like UI-TARS while using only 0.1% of its data. Furthermore, under the same amount of annotated data, GUICrafter surpasses all previous methods such as GUI-R1. Code, data, and models are available at https://github.com/fansunqi/GUICrafter.

[209] Does Verbose Chain-of-Thought Really Help? In-Distribution Evidence that Content, Not Length, Matters cs.AI | cs.CLPDF

Wenlong Wang, Fergal Reid

TL;DR: 本文通过两种实验方法探究了思维链（CoT）提示中冗长性对LLM推理能力的影响。研究发现，额外的推理步骤本身（即长度）并不直接提升准确性，关键在于这些步骤所携带的语义内容（如验证和检查内容）。

Details

Motivation: 研究动机是厘清思维链提示提升大语言模型推理能力的根源：究竟是中间步骤的语义内容有帮助，还是仅仅因为生成了更多token从而获得了额外计算时间？

Result: 在分布内实验中，对25个模型的分析显示，额外token基本不改变准确性。在受控干预实验中，冗长轨迹确实能小幅提升准确性（通常在1-4个百分点），但效果取决于冗长文本的质量而非长度，且在最大程度数字遮盖时效果被放大。

Insight: 核心洞见是：影响推理性能的关键是额外token所执行的功能（即它们携带的推理和验证内容），而非其数量。这既不能完全用纯前向计算解释，也不能用纯语义内容解释，为理解CoT机制提供了新视角。

Abstract: Chain-of-thought (CoT) prompting improves LLM reasoning, but the source is contested: do the intermediate steps help because they carry useful semantic content, or because conditioning on more tokens buys extra computation before the model commits to an answer? We bring two lines of evidence to bear. First, in distribution: we repeatedly sample each model on the same question and pair a shorter with a longer of its own natural generations that follow the same reasoning plan, so nothing is rewritten and both traces are genuinely in-distribution. Across 25 models the extra tokens leave accuracy essentially unchanged for every independently-trained reasoner, and a blind analysis of the surplus tokens shows that what gain exists elsewhere tracks validation- and checking-content, not verbosity per se. Second, as a controlled intervention, we ask whether two traces expressing the same semantic content (the same facts, operations, and intermediate values, verified through directed acyclic graph equivalence) produce different outcomes when one is more verbose, using a dual-validator design across four targets and eight benchmarks with number-redacted completion and stratified bootstrap confidence intervals. Verbose traces do improve accuracy (25 of 32 benchmark-target cells are positive under at least one validator), but the effects are modest (typically 1-4 points) and depend on the quality of the verbose prose, not merely its length. Under maximum numerical redaction the effect is amplified (median 3.24x across four arithmetic benchmarks), and length-matched non-reasoning filler recovers none of it. Both lines converge: what matters is what the extra tokens do (the reasoning and validation content they carry), not how many there are, a picture neither a pure forward-pass-compute nor a pure semantic-content account fully explains.

[210] Self-Evolving World Models for LLM Agent Planning cs.AI | cs.CLPDF

Xuan Zhang, Wenxuan Zhang, See-Kiong Ng, Yang Deng

TL;DR: 本文提出了WorldEvolver，一个自演化的世界模型框架，旨在通过部署时上下文修订来提升LLM智能体的长期规划能力。该框架包含情景记忆、语义记忆和选择性前瞻三个模块，在不改变智能体或模型参数的情况下，利用真实交互经验改进世界模型的预测可靠性。

Details

Motivation: 现有世界模型为LLM智能体提供行动后果的前瞻预测，但其预测可能不可靠，若被忽略、误用甚至可能损害下游决策性能。本文旨在解决世界模型预测的不可靠性问题，以增强智能体的规划能力。

Result: 在ALFWorld和ScienceWorld基准上，WorldEvolver在Word2World任务上取得了最高的预测准确率，并在AgentBoard评估的下游智能体成功率上超越了其他世界模型基线，表明其在三个骨干模型上均能同时提升预测保真度和规划性能。

Insight: 创新点在于提出了一个冻结模型参数、仅通过部署时上下文修订来使世界模型自我演化的框架，其核心是结合了基于检索的模拟（情景记忆）、从预测-观测不匹配中提取启发式规则（语义记忆）以及对低置信度预测进行过滤（选择性前瞻）的机制。

Abstract: World models offer a principled way to equip long-horizon LLM agents with foresight: predictions of action consequences before execution. However, unreliable foresight can be ignored, misused, or even degrade downstream decision-making. In this paper, we introduce WorldEvolver, a self-evolving world model framework that revises its deployment-time context while keeping the downstream agent and all model parameters frozen. WorldEvolver integrates three modules: (i) Episodic Memory, which exploits real action transitions through retrieval-based simulation; (ii) Semantic Memory, which extracts persistent heuristic rules from prediction-observation mismatches; and (iii) Selective Foresight, which filters low-confidence predictions before integrating them into agent reasoning context. We evaluate WorldEvolver on ALFWorld and ScienceWorld, measuring world model prediction accuracy on Word2World and downstream agent success rate on AgentBoard. Extensive experiments show that WorldEvolver achieves the highest prediction accuracy across three backbones and leads other world model baselines on downstream agent success rate, demonstrating that test-time memory revision enhances both predictive fidelity and planning performance.

cs.CY [Back]

[211] LLM-Ideoplasticity: Measuring Ideological Plasticity in the Political Behavior of LLMs as a Context-Conditioned Distribution cs.CY | cs.AI | cs.CLPDF

Adib Sakhawat, Syed Rifat Raiyan, Tahsin Islam, Takia Farhin, Hasan Mahmud

TL;DR: 该论文提出LLM-Ideoplasticity框架，通过实证研究表明大型语言模型的政治意识形态并非固定点，而是依赖于上下文的条件分布。研究评估了九个当前主流LLM，使用基于VAA-CHES投影模型的统一测量框架，在六个上下文轴上映射到三个验证维度（lrgen、lrecon、galtan）。研究发现模型对上下文高度敏感，但整体意识形态分布范围狭窄，仅占欧洲主要政党分布范围的三分之一。

Details

Motivation: 解决现有研究将LLM政治立场简化为单一固定点的问题，强调意识形态应被视为随上下文变化的分布，以更准确刻画LLM在政治行为上的复杂性。

Result: 在VAA-CHES基准上的定量结果显示：说服性框架和低资源语言分别使坐标偏移达0.57和0.52单位，思维链推理会加剧释义不稳定性；多特质多方法（MTMM）分析支持结论，模型群体的意识形态分布范围仅为欧洲主要政党的三分之一。

Insight: 创新点在于将LLM政治意识形态建模为条件概率分布，并开发了系统测量框架；客观来看，该研究揭示了上下文对LLM政治输出的显著影响，为评估模型偏见提供了新方法论，强调需用‘形状’而非‘点’来表征模型行为。

Abstract: We argue, with systematic empirical evidence, that a large language model’s political ideology is not a fixed point, but a conditional distribution $\mathbb{P}($position$\mid$context$)$ over a real political space. We evaluate nine current LLMs using a unified measurement framework anchored by VAA-CHES projection models, which map responses onto three validated dimensions (lrgen, lrecon, galtan) across six contextual axes. Our findings reveal high sensitivity to context: persuasive framing and under-represented languages displace coordinates by up to 0.57 and 0.52 units, respectively, while chain-of-thought reasoning often amplifies rather than dampens paraphrase instability. Despite this local plasticity, the model cohort occupies a remarkably narrow Overton envelope overall, occupying roughly one-third the spread of major European parties. Supported by a multi-trait multi-method (MTMM) analysis, we conclude that a single point cannot summarize LLM political behavior; it must be characterized as a shape. Our code and data are publicly available at https://github.com/sakhadib/LLM-Ideoplasticity.

[212] Situation Perception: A Necessary Primitive to Artificial Superintelligence cs.CY | cs.AI | cs.CL | cs.ETPDF

Ziqin Yuan, Jaymari Chua

TL;DR: 本文认为当前大型语言模型虽具备强大的模式识别能力，但缺乏真正的通用智能，并提出通往人工超智能的关键在于发展‘情境感知’能力——即构建、修正和基于潜在时间跨度的可能世界内部模拟进行行动的能力。

Details

Motivation: 论文的动机是指出当前大型语言模型仅是统计引擎，缺乏类似人类婴儿从零开始逐步发现物体恒存性、因果关系等基础认知能力，因此需要发展‘情境感知’这一核心能力以实现人工超智能。

Result: 摘要未提及具体的定量实验结果或基准测试，而是提出了衡量机器能否模拟未来、追求自主目标并可能评判其创造者的适当测试框架。

Insight: 论文的创新点在于将‘情境感知’定义为人工超智能的必要基础，并分解其三个核心组件：抽象预测、长期压缩记忆和基于目标的主动学习，为超越当前语言模型局限提供了新的研究方向。

Abstract: Current large language models are extraordinary statistical engines. They compress vast amounts of text into useful patterns and can explain science, write code, imitate reasoning, and participate in philosophical conversation. Yet pattern mastery is not the same as general intelligence. A human infant begins with little explicit knowledge, but gradually discovers object permanence, cause and effect, other minds, bodily agency, and the persistence of the physical world. We make an argument that the path to artificial superintelligence (ASI) depends on a missing capacity we call \emph{situation perception}: the ability to construct, revise, and act within internal simulations of possible worlds across latent time. \emph{ perception} requires at least three core components: abstract prediction, long-term compressed memory, and active learning guided by objectives. In this work, we analyse why modern large language models remain incomplete, and propose the appropriate tests for measuring progress and consequences of machines that can simulate futures, pursue self-directed goals, and possibly judge their own creators.

cs.MM [Back]

[213] A Good Talk Does not Look Like a Summary, It Teaches You! Measuring Takeaways from Paper-to-Video Talks cs.MM | cs.CLPDF

Ishani Mondal, Aparna Garimella, Ananya Sai, Pannaga Shivaswamy, Jordan Boyd-Graber

TL;DR: 本文提出了EffectivePresentationScorer框架，用于评估科学论文演示视频的教学质量，而不仅仅是视觉质量或内容覆盖。研究发现，现有论文转视频系统生成的视频能提及正确主题并遵循论文结构，但未能有效解释先决概念或阐明方法原理。

Details

Motivation: 现有自动生成科学论文视频的评估指标主要关注视觉质量或关键点是否出现，而未能评估视频是否真正帮助观众理解论文思想。

Result: 将EffectivePresentationScorer应用于现有论文转视频生成系统后发现，生成的视频在提及正确主题和遵循论文结构方面表现良好，但在解释先决概念和阐明方法原理方面存在不足。

Insight: 创新点在于从教学有效性（如清晰解释主要思想、引入必要背景概念、将技术细节与论文主要贡献联系起来）而非单纯的内容存在性来评估科学演示视频，揭示了现有评估指标的盲区。

Abstract: Automatically generated videos from scientific papers are increasingly used for education and research dissemination. However, existing evaluation metrics mainly measure visual quality or whether key points from the paper appear in the video without assessing whether the video actually helps viewers understand the ideas. We introduce EffectivePresentationScorer, a framework for evaluating the instructional quality of scientific presentation videos. It checks whether a video explains the main ideas clearly, introduces needed background concepts, and connects technical details to the main contribution of the paper. When we apply EffectivePresentationScorer to the existing paper-to-video generation systems, we find that generated videos mention the correct topics and follow the structure of the paper but fail to explain prerequisite concepts or clarify why the method works. These failures are often ignored by existing video evaluation metrics, which focus on content presence rather than explanatory quality.

cs.IR [Back]

Anisha Saha, Vaibhav Rathore, Abhisek Tiwari, Akash Ghosh, Sai Ruthvik Edara

TL;DR: 该论文提出了一个名为$M^3QAFrame$的多模态多跨度医学问答框架，旨在解决现有系统在真实场景中处理同时包含文本和视觉内容的医学文档时的不足。该框架利用视觉线索，从文本和图像的不同跨度中生成包含文本答案和相关图像的综合答案。

Details

Motivation: 现有医学问答系统主要处理纯文本，而现实中的医学文档常包含图文内容，需要结合图像来提供更全面的答案。因此，论文旨在填补这一空白，开发一个能同时处理文本和视觉信息的多模态多跨度问答系统。

Result: 论文在自建的$M^3 QuestionIng$数据集上进行了广泛实验，结果表明，该方法在多种评估指标上均持续优于现有方法。

Insight: 创新点包括提出首个多模态多跨度医学问答框架，通过结合视觉信息增强答案的全面性；同时，构建了包含用户意图和查询类型标注的数据集，以提升对查询和上下文的理解能力。

Abstract: The growing adoption of AI in healthcare, particularly in preventive care, highlights the critical need for accessibility and precision in Medical Question Answering (MedQA). In recent years, significant efforts have been made to develop multi-span medical question-answering systems, where the answer to a query may span multiple sections or paragraphs of a source document. However, existing systems fall short of aligning with real-world scenarios, where source documents often include both textual and visual content, requiring answers to incorporate images for better comprehension. To address this gap, we propose $M^3QAFrame$, a multi-modal, multi-span medical question-answering framework that leverages visual cues to enhance the generation of comprehensive answers drawn from diverse textual and visual spans. The model takes the context, query, and images as input and outputs an answer containing both textual answers and relevant images. The text and image embeddings are processed using a transformer-based architecture to determine the sentence and image relevance. We curate a multi-modal, multi-span medical question-answering ($M^3 QuestionIng$) dataset containing queries, medical contexts, associated medical images, and extractive answers. Additionally, each query-answer pair is labeled with user intent and query type to enhance query and context comprehension. Extensive experiments show that our approach consistently outperforms existing methods across various evaluation metrics.

[215] PIXELRAG: Web Screenshots Beat Text for Retrieval-Augmented Generation cs.IR | cs.AI | cs.CL | cs.CV | cs.LGPDF

Yichuan Wang, Zhifei Li, Zirui Wang, Paul Teiletche, Lesheng Jin

TL;DR: PixelRAG是一种新的检索增强生成方法，它直接以网页的视觉形式（即截图）表示网站，并在像素空间中进行检索和阅读，避免了复杂的文本解析流程。该方法基于视觉嵌入模型构建，通过微调处理截图数据，将检索到的截图直接输入视觉语言模型，无需中间文本转换。实验表明，PixelRAG在多个基准测试中优于无检索和基于文本的RAG基线，同时通过图像压缩实现了效率提升。

Details

Motivation: 当前基于检索增强的LLM主要依赖文本检索，但网页本质上是视觉化的，现有系统通过解析HTML线性化文本时会丢弃布局、视觉结构和格式信息，导致信息损失。PixelRAG旨在解决这一问题，直接在网页的原生视觉形式上进行操作，以更完整地利用网页信息。

Result: PixelRAG在多个基准测试中表现优异：在文本中心任务（如NQ和SimpleQA）上超越基线；在多模态开放域QA（如MMSearch）、噪声新闻语料库（如LiveVQA）和智能体基准（如MoNaCo）上实现显著提升，准确率最高比基于文本的基线提高18.1%。此外，通过图像压缩，在较低分辨率下可减少高达3倍的token成本，同时保持准确性。

Insight: 论文的创新点在于首次提出完全在像素空间中操作的全端到端RAG架构，消除了文本抽象，直接利用网页截图进行检索和生成。从客观角度看，这挑战了传统RAG依赖文本表示的必要性，展示了视觉形式在提升性能和效率方面的潜力，特别是通过图像压缩优化token使用，为多模态检索提供了新思路。

Abstract: Augmenting large language models (LLMs) with retrieved web text has become a dominant paradigm, yet the web is not natively textual: existing systems depend on complex parsing pipelines that linearize HTML and discard layout, visual structure, and formatting. We introduce PixelRAG, a new retrieval-augmented method that represents websites in their native visual form and performs retrieval and reading entirely in pixel space, enabling an end-to-end architecture that eliminates text abstraction. PixelRAG is, to our knowledge, the first pipeline to operate over a full Wikipedia corpus in this form, scaling to a datastore of 30 million screenshot images with an efficient visual retrieval index. Built on an existing visual embedding model (i.e., Qwen3-VL-Embedding), PixelRAG further fine-tunes this model on screenshot data with carefully curated contrastive training data. Retrieved screenshots are then fed directly as pixel inputs to a VLM, without intermediate text conversion. PixelRAG consistently outperforms both no-retrieval and text-based RAG baselines, most surprisingly on widely studied text-centric tasks such as NQ and SimpleQA. It also achieves strong gains on multimodal open-domain QA (e.g., MMSearch), benchmarks over noisy news corpora (e.g., LiveVQA), and agentic benchmarks (e.g., MoNaCo), improving accuracy by up to 18.1% over text-based baselines. Finally, pixel representations enable a new efficiency lever for RAG through image compression, achieving up to 3x token cost reduction at lower resolutions while maintaining accuracy. Our results challenge the necessity of text representations in web retrieval, suggesting that web RAG can operate directly in the web’s native visual form while improving both performance and efficiency.

[216] How Do LLMs Cite? A Mechanistic Interpretation of Attribution in Retrieval-Augmented Generation cs.IR | cs.AI | cs.CLPDF

Ian van Dort, Maria Heuss

TL;DR: 本文首次从机制上解释了大语言模型在检索增强生成中如何决定是否添加内联引用，通过激活修补方法在PopQA数据集上发现引用机制是一个分布式的’归因集成’，而非单一组件，并证明通过调控关键注意力头和MLP层可显著修复引用错误。

Details

Motivation: 旨在验证RAG中内联引用的忠实性，即模型是否真正依据外部文档生成答案，这是一个尚未验证的关键假设。

Result: 在PopQA数据集上，通过调控关键组件修复了超过90%的漏引和69%的虚假引用，且不影响答案准确性；在HotpotQA基准上效果有限但趋势一致，表明机制非数据集特定。

Insight: 揭示了引用机制是分布式、多阶段的’归因集成’，而非局部化组件；同时指出内联引用可能造成虚假的安全感，因为模型表面推理与内部计算路径存在潜在脱节。

Abstract: Retrieval-Augmented Generation (RAG) aims to enhance the trustworthiness of Large Language Models (LLMs) by grounding their outputs in external documents, often using inline citations for verifiability. However, the faithfulness of these citations – whether the model genuinely uses a source to generate an answer – remains a critical, unverified assumption. This paper offers the first mechanistic account of how a large language model decides whether to attach an inline citation while answering a factoid question. Using the Llama-3.1-8B-Instruct model in a controlled experimental environment based on the PopQA dataset, we employ an activation patching approach. We map the underlying mechanism responsible for citation, discovering that it is not a single, localized component but a distributed, multi-stage “attributional ensemble” of attention heads and MLP layers. We show that amplifying or attenuating only those critical heads and MLPs repairs over 90% of missed citations and eliminates 69% of spurious ones on PopQA without harming answer accuracy. Although gains on the multi-document HotpotQA benchmark are modest, the same component set still moves citation rates in the intended direction, indicating that the underlying mechanism is not dataset-specific. The results reveal a potential disconnect between the model’s apparent reasoning and its internal computational pathway, suggesting that inline citations can create a false sense of security.

[217] LEDGER: Scaling Agentic Document Editing with Dependency-aware Graph Retrieval cs.IR | cs.AI | cs.CLPDF

Mike Hang Wang, Utkarsh Garg, Reza Davari, Huitian Jiao, Hao Cheng

TL;DR: LEDGER是一个用于智能文档编辑的上下文工程框架，通过构建轻量级依赖图来显式建模文档结构（包括层次组织、显式引用、隐式依赖和语义关系），并利用图引导检索为每次编辑选择必要上下文，从而在避免全文档处理的同时保持一致性。

Details

Motivation: 解决智能文档编辑中的新型上下文工程挑战，即在对长篇结构化文档进行局部编辑时，需要高效应用编辑而不破坏交叉引用或语义一致性。

Result: 在包含1.9k个测试用例、涵盖多种文档类型和长度的基准测试中，LEDGER将六个最先进模型的一致性从56%提升至76%，同时减少了token使用量；在低推理成本下，LEDGER匹配了高推理成本基线的性能。

Insight: 核心创新在于通过显式的依赖图表示来建模文档结构，并利用图引导检索实现高效的上下文选择；这表明在智能文档编辑中，显式的依赖表示可以部分替代昂贵的内部推理过程。

Abstract: We introduce LEDGER to tackle the novel context engineering challenge of agentic document editing, where localized edits to long, structured documents must be applied efficiently without breaking cross-references or semantic consistency. LEDGER constructs a lightweight dependency graph that explicitly models document structure, including hierarchical organization, explicit references, implicit dependencies, and semantic relationships. For each edit, graph-guided retrieval selects only the necessary context, avoiding full-document processing while preserving consistency. We evaluate LEDGER on a curated benchmark of 1.9k test cases with various document types and lengths, spanning six state-of-the-art models: LEDGER improves consistency from 56% to 76% across all six models and test scenarios while reducing token usage. Notably, LEDGER with low reasoning effort matches baseline performance at high reasoning effort using fewer tokens, showing that explicit dependency representations can partially substitute for expensive internal reasoning in agentic document editing.

Yini Huang, Wenlong Zhang

TL;DR: 本文提出了一种名为UniCA（统一交叉注意力编码器）的多模态检索模型，旨在解决现有方法依赖文本编码器自注意力进行隐式融合、缺乏显式跨模态语义对齐的问题。该模型引入了双向交叉注意力块和正相似性损失，并在WebQA基准测试中取得了优于基线模型的性能。

Details

Motivation: 现实应用中视觉-文本数据日益增长，但现有多模态检索框架仅通过文本编码器的自注意力进行隐式融合，限制了跨模态语义的显式对齐，因此需要更高效的融合机制。

Result: 在WebQA基准测试的混合任务和图文任务上，UniCA均优于基线模型，其中混合任务的Recall@5提升了4.09%，Recall@10提升了3.28%，MRR@1提升了3.96%，达到了新的SOTA水平。

Insight: 创新点包括：1）双向交叉注意力块在拼接前实现视觉与文本token的主动语义交换；2）正相似性损失直接优化查询与正样本嵌入的绝对语义邻近度；3）构建轻量级数据集UMR-S10以降低计算成本。这些设计增强了跨模态对齐的显性与效率。

Abstract: Multi-modal retrieval has become increasingly critical for handling the growing volume of integrated visual-textual data in real-world applications, but existing frameworks rely on implicit fusion via text encoder self-attention, limiting explicit cross-modal semantic alignment. To address this gap, this paper proposes UniCA (Unified Cross-Attention Encoder), a multi-modal retrieval model with four key innovations: 1) a bi-directional cross-attention (Bi-CA) block that enables active semantic exchange between visual and textual tokens prior to concatenation, capturing inter-modal correlations more efficiently. 2) a Positive Similarity Loss that optimizes absolute semantic proximity between query and positive candidate embeddings. 3) a streamlined dataset UMR-S10 (Universal Multimodal Retrieval Sample 10%) to reduce computational costs while retaining semantic diversity and task representativeness. 4) an experimental validation on the WebQA benchmark demonstrates that UniCA outperforms the baseline model across Hybrid and Image-Text tasks, achieving improvements of up to 4.09% in Recall@5, 3.28% in Recall@10, and 3.96% in MRR@1 for the hybrid task. UniCA provides an efficient and robust solution for multi-modal retrieval, lowering deployment barriers through its lightweight dataset and enhanced fusion mechanism.

[219] Multimodal Graph RAG for Long-range Visually Rich Document Understanding cs.IR | cs.CVPDF

Yi-Cheng Wang, Chu-Song Chen

TL;DR: 本文提出了一种基于多模态图结构的检索增强生成方法，用于解决长文档视觉问答任务中全局理解不足的问题。该方法通过构建多模态知识图谱来总结文档的全局知识，并引入了一个新的文档级视觉问答基准DLVQA进行评估。实验表明，该方法在多个多跳问答基准和DLVQA上优于现有的多模态检索增强生成或基于知识图谱的方法。

Details

Motivation: 现有多模态大语言模型受限于上下文窗口长度，难以理解长文档；而多模态检索增强生成方法虽然能检索相关页面，但在需要全局理解的视觉问答任务上仍有不足。

Result: 在多个多跳问答/视觉问答基准和提出的DLVQA基准上，该方法超越了现有的多模态检索增强生成或基于知识图谱的方法，取得了更好的性能。

Insight: 创新点在于将多模态知识图谱引入检索增强生成框架，以捕获文档的全局视觉和文本信息；同时创建了DLVQA基准，为文档级视觉问答提供了直接评估数据集，弥补了以往依赖间接指标评估的不足。

Abstract: Multimodal large language models (MLLMs) are widely applied to visual document understanding. However, comprehending long documents remains an issue by the limited context window. Though recent multimodal retrieval-augmented generation (MMRAG) can address this challenge by retrieving relevant pages. It still struggles with the visual question answering (VQA) requiring holistic comprehension of a document. To cope with this, knowledge graph (KG) that summarizes global knowledge of a document can provide an effective solution. However, most existing LLM-based KG construction methods handle only the language modality, leaving the automatic creation of multimodal KGs (MMKGs) for visually rich documents largely unexplored. In this paper, we introduce a multimodal graph-based RAG approach to tackle this problem. Existing LLM-based KG methods evaluate the QA performance relying on indirect evidence such as comprehensiveness, diversity, empowerment, and so on. The lack of annotated datasets for comprehensive document-level VQA poses a significant challenge to effective model evaluation. To overcome this limitation, we also introduce a new benchmark, DLVQA (document-level VQA), which provides reference summaries and corresponding supporting facts for global document-level questions. Experimental results show that our approach outperforms existing MMRAG or KG-based approaches on multi-hop QA/VQA benchmarks and DLVQA.

cs.SD [Back]

[220] LoRA-Tuned Large Language Models for Dementia Detection via Multi-View Speech-Derived Features cs.SD | cs.AI | cs.CL | cs.LGPDF

Jonghyeon Park, Olivier Jiyoun Jung, Myungwoo Oh

TL;DR: 该论文提出了一种基于低秩适应（LoRA）微调的大语言模型（LLM），用于通过多视角语音特征进行痴呆症早期检测。该方法将自动语音识别（ASR）转录文本（含停顿标记）、话语级主题线索、时间流畅性统计和音位序列这四种互补的语音衍生信号统一编码到提示中，使单个LLM能够进行结构化多视角推理，而无需特定模态编码器或后期融合。

Details

Motivation: 早期痴呆检测需要及时干预，而反映认知障碍的自发性语音提供了一种非侵入性筛查方式。传统方法通常只关注单一表征维度（如声学描述符、停顿建模、ASR转录文本或多模态融合），限制了跨异构认知症状的整合推理能力。

Result: 在ADReSSo基准测试上，最佳模型取得了90.14%的F1分数。消融实验证实了每个视角的互补贡献。

Insight: 创新点在于提出了一种统一的提示工程方法，将多种互补的语音衍生特征（包括语言内容、话语结构、时间动态和音位信息）整合到单个LLM中进行端到端的多视角结构化推理，避免了复杂的多模态编码器或融合策略，并通过LoRA实现高效微调。

Abstract: Early detection of dementia enables timely intervention, and reflecting cognitive impairment, spontaneous speech offers a non-invasive screening modality. Conventional approaches often focus on a single representational dimension – such as acoustic descriptors, pause modeling, automatic speech recognition (ASR) transcripts, or multimodal fusion – limiting integrative reasoning across heterogeneous cognitive symptoms. We propose a low-rank adaptation (LoRA)-tuned large language model (LLM) that performs structured multi-view reasoning over four complementary speech-derived signals: ASR transcripts with pause markers, discourse-level topic cues, temporal fluency statistics, and phonological sequences. These cues are encoded within a unified prompt, enabling a single LLM to learn a coherent decision function without modality-specific encoders or late-stage fusion. On ADReSSo, our best model achieves an F1-score of 90.14%, and ablation confirms the complementary contribution of each view.

cs.RO [Back]

Carmen Ng, Gjergji Kasneci

TL;DR: 本文提出了一种基于梯度的审计框架，用于多语言评估LLM在文化偏好梯度下的道德权衡行为。该框架通过对称控制的场景（如护理、教育和服务领域），将‘拯救谁’的困境转化为‘优先帮助谁’的决策，并审计了四个LLM在四个国家-语言对中的表现。研究发现，LLM存在持续且文化不对称的梯度追踪失败，提示工程难以可靠纠正，因此多语言、多元化的审计应作为LLM-机器人部署前的关键环节。

Details

Motivation: 随着LLM治理的社交机器人越来越多地决定现实世界中谁优先获得帮助，而优先规范因文化（如年龄、地位、群体规模）而异，缺乏多元校准可能导致不平等。现有LLM道德审计以英语为中心，很少测试具身环境，因此在LLM-机器人部署加剧的背景下，多元校准成为一个紧迫的诊断缺口。

Result: 在四个LLM上进行了57,600次决策的审计，并以国家特定的MME偏好梯度为基准。顺序一致性测试显示，模型在区分文化背景方面存在持续失败：西方语言决策的质量校准强度几乎是中文和日语的两倍；多数优先的权衡中的高确定性常常抹杀跨文化梯度；对基于年龄和地位规范的敏感性不足可能边缘化少数群体。提示效果不均，仅对比示例能带来一致增益。

Insight: 创新点在于引入了基于梯度的多语言审计框架，将经典道德机器实验转化为具身的‘优先帮助’困境，并提出了一个治理类型学来映射梯度分化、方向倾向和审议中的漏洞。客观分析表明，模型内在因素（而非仅提示工程）是更稳健的改进杠杆，强调了在部署前进行多语言、多元化审计的必要性。

Abstract: LLM-governed social robots increasingly decide who receives real-world assistance first. As prioritization norms vary across cultures by age, status, and group size, failure to calibrate pluralistically can scale into unequal access. Yet LLM moral audits remain English-centered, rarely test embodied contexts, leaving pluralistic calibration as an urgent diagnostic gap amid intensifying LLM-robot deployment. We introduce a gradient-based audit framework for multilingual evaluation of LLM moral trade-off behavior against cultural preference gradients. Grounded in nine cross-domain social robotics reviews (>8,000 papers), we derive symmetry-controlled scenarios across care, education, and services, translating the Moral Machine Experiment’s “whom to spare” into “whom to assist first” dilemmas with preserved identity trade-offs (many vs. few; young vs. old; higher vs. lower status). We audit four LLMs across four country-language pairs in four prompting regimes (57,600 decisions), benchmarked against country-specific MME preference gradients. Ordinal concordance tests whether models differentiate cultural contexts; a governance typology maps vulnerabilities in gradient differentiation, directional tendency, and deliberation. We find persistent, culturally asymmetric gradient tracking failures that prompting alone cannot reliably correct: quality calibration is nearly twice as strong for Western-language decisions as for Chinese and Japanese; high determinism in majority-first trade-offs often erases cross-cultural gradients; partial sensitivity to age- and status-based norms risks sidelining minorities. Prompting effects are uneven; only contrastive exemplars yield consistent gains, while reasoning-only prompts can worsen tracking. Our results motivate multilingual, pluralistic audits as an LLM-robot pre-deployment gate and suggest model factors are a more robust lever than prompting alone.

[222] When Stopping Fails: Rethinking Minimal Risk Conditions through Human-Interactive Autonomous Driving for Safe Transportation Systems cs.RO | cs.AI | cs.CV | cs.HCPDF

Yash Tandon, Giovanni Tapia Lopez, Marcus Blennemann, Mohan Trivedi, Ross Greer

TL;DR: 本文分析了自动驾驶车辆（AV）在现实部署中因采用最小风险条件（MRC）行为（如停车）而导致的安全问题，指出单纯停车策略无法确保与人类交通系统的安全融合，并提出了通过增强人机交互能力来改进AV安全框架的方向。

Details

Motivation: 当前自动驾驶安全框架主要围绕碰撞避免和MRC行为设计，但实际部署表明，停车等被动策略会阻碍交通、干扰应急响应并造成可访问性挑战，因此需要重新思考如何实现安全的人机交互集成。

Result: 论文未提供具体定量结果，但基于公开记录的事件分析，对AV停车行为与人-AV交互失败进行了分类，并指出了现有安全范式的关键缺陷。

Insight: 创新点在于系统性地将AV故障行为归类为感知、规划和控制限制，并强调需引入人机交互感知、语言基础规划、远程引导等能力，以推动AV从被动回退策略转向主动协作的交互式自主性。

Abstract: Autonomous vehicles (AVs) are increasingly deployed in urban environments, yet their safety frameworks remain primarily designed around collision avoidance and minimal risk condition (MRC) behaviors such as slowing or stopping when uncertainty arises. Although effective in reducing immediate crash risk, real-world deployments indicate that stopping alone does not guarantee safe integration into human-governed roadway systems. Incidents reported by municipalities and public records show that AV fallback behaviors can obstruct traffic, interfere with emergency response operations, and create accessibility challenges for passengers and pedestrians. This paper presents an analysis of publicly documented incidents involving AV stopping behavior and human-AV interaction failures. We categorize these incidents according to limitations in perception, planning, and control within current AV architectures. Using this taxonomy, we identify key gaps in existing safety paradigms, particularly the lack of mechanisms for interpreting human authority, responding to multimodal instructions, and adapting to dynamic, socially regulated traffic conditions. We then review emerging research directions that support human-interactive perception, language-grounded and accessibility-aware planning, and assisted control through remote guidance and teleoperation. The analysis highlights the need to augment current AV safety frameworks with capabilities that enable cooperative interaction with human agents and infrastructure. These findings suggest that reliable urban deployment of AVs requires moving beyond passive fallback strategies toward human-interactive autonomy.

[223] FalconTrack: Photorealistic Auto-Labeled Perception and Physics-Aware Vision-Based Aerial Tracking cs.RO | cs.AI | cs.CVPDF

Yan Miao, Karteek Gandiboyina, Noah Giles, Hideki Okamoto, Bardh Hoxha

TL;DR: 本文提出了FalconTrack，一个用于视觉空中跟踪的统一感知与跟踪框架。该框架利用基于高斯泼溅的光真实感可编辑模拟器自动生成标注数据，并通过结合多头感知模块与物理感知的扩展卡尔曼滤波器进行零样本模拟到现实的迁移。

Details

Motivation: 解决在GPS拒止环境中视觉空中跟踪的可靠感知问题，传统光真实感数据集依赖繁重的人工标注且耗时，需要一种自动化标注和高效模拟到现实迁移的方法。

Result: 在零样本模拟到现实迁移中，感知模型在三个几何多样物体和两种环境上达到96-100%的分类准确率，优于两个基线。在真实硬件闭环视觉跟踪中，系统以约25Hz运行，在F1-tenth和门跟踪任务中实现100%成功率，而基线在快速出视野场景下成功率降至60%。

Insight: 创新点包括基于高斯泼溅模拟器的自动化标注流程，以及结合多头感知与类条件动力学先验的物理感知跟踪框架，实现了高效的零样本模拟到现实迁移和鲁棒的实时跟踪性能。

Abstract: Vision-based aerial tracking is critical in GPS-denied environments. Reliable perception for tracking depends on large-scale labeled data, yet most photorealistic datasets rely on heavy manual annotation and are time-consuming to produce. We present FalconTrack, a unified perception-and-tracking framework that (i) leverages a photorealistic editable simulator for automated label generation and (ii) combines multi-head perception with physics-aware tracking for zero-shot sim-to-real transfer. FalconTrack provides an automated labeling pipeline in a Gaussian Splatting simulator that isolates target Gaussians from short object videos and composites them with randomized backgrounds to generate RGB, mask, class, and 6-DoF pose labels, producing about 10k labeled images in under 20 minutes. Using this dataset, we train a multi-head perception module with staged learning and reprojection consistency, and fuse its outputs with class-conditioned dynamics priors in an EKF for tracking. Our perception model outperforms two baselines and reaches 96-100% class accuracy in zero-shot sim-to-real transfer on three geometrically diverse objects and two environments, while maintaining consistent performance in unseen simulated and real scenes. In real hardware closed-loop visual tracking, the onboard system runs at about 25 Hz and achieves 100% success in sim-to-real F1-tenth and gate tracking in five trajectories across two environments, while a mask-centered vision baseline drops to 60% success on F1-tenth during fast out-of-view scenarios.

[224] Vision-Language-Action Models: Experimental Insights from a Real-World UR5 Platform cs.RO | cs.CVPDF

Mathilde Hochedel, Marc Lalonde

TL;DR: 本研究探讨了将最新的视觉-语言-动作（VLA）模型从受控的研究基准迁移到真实世界机器人平台（UR5e机械臂）的可行性，并建立了一个可复现的评估框架。研究发现，离线指标与真实系统闭环行为之间存在显著差距，这主要归因于动作语义、坐标系约定、模态间时序对齐等系统级因素，而非单纯的模型能力限制。

Details

Motivation: 解决VLA模型从仿真或受控基准向真实机器人平台部署时，其性能表现是否可靠、可复现的问题，并探究影响实际部署成功的关键因素。

Result: 在UR5e真实机器人平台上进行了系统实验，结果表明，尽管模型离线指标良好，但其闭环行为不稳定，揭示了仿真与真实世界之间的性能鸿沟。

Insight: 创新点在于将VLA机器人研究从以模型为中心的问题重构为系统级问题，强调数据-模型-控制整个管路的精确协调比单纯提升模型容量更为关键，并提供了包括数据采集、数据集工程、微调部署在内的完整可复现框架和实验观察。

Abstract: This project investigates whether recent Vision-Language-Action (VLA) models can be transferred from controlled research benchmarks to a real-world robotic platform, specifically a UR5e manipulator, in a reproducible and operationally meaningful manner. The work integrates real-robot data acquisition, dataset engineering (compatible with the RLDS format), and the fine-tuning and deployment of OpenVLA and OpenVLA-OFT models, with systematic validation of action representations and control interfaces. The project resulted in several foundational assets: (i) a complete real-robot data acquisition pipeline, (ii) a dataset conversion workflow aligned with RLDS standards, (iii) an initial fine-tuning and inference infrastructure for VLA models, and (iv) a structured set of experimental observations grounded in real-robot trials. These elements collectively establish a reproducible framework for evaluating learning-based manipulation systems beyond simulation. Empirically, the experiments reveal a consistent gap between promising offline indicators and unstable closed-loop behavior on the physical system: this gap cannot be attributed solely to model limitations, it is strongly influenced by action semantics, coordinate frame conventions, temporal alignment between modalities, image preprocessing consistency, and dataset coverage and quality. These observations lead to a key interpretation: the successful deployment of VLA systems in real-world settings depends less on incremental improvements in model capacity and more on precise control of the entire data-model-control pipeline. The project reframes VLA-based robotics from a primarily model-centric challenge to a system-level problem; it highlights the difficulty of running robust task execution on the real robot and provides a clear, experimentally grounded understanding of the conditions required for reliable deployment.

[225] Training Vision-Language-Action Models with Dense Embodied Chain-of-Thought Supervision cs.RO | cs.CVPDF

Haoyang Li, Guanlin Li, Youhe Feng, Chen Zhao, Zhuoran Wang

TL;DR: 这篇论文提出了ZR-0，一个拥有26亿参数、用于跨具身智能体迁移的端到端视觉-语言-动作模型。其核心创新是使用密集的具身思维链监督来对齐不同机器人平台间的表示。模型采用双流架构，结合了预训练的视觉语言模型进行推理和基于扩散Transformer的动作专家生成连续动作，并在推理时可跳过思维链生成。

Details

Motivation: 解决跨具身智能体（不同机器人平台）在视觉-语言-动作模型中迁移的挑战，因为低层状态和动作空间在不同平台间差异巨大。作者观察到，高层认知过程（如场景感知、任务规划）在不同具身体之间是共享的，并基于此进行模型设计。

Result: 在涵盖单臂、双臂和人形机器人具身体的三个模拟基准（LIBERO, RoboTwin 2.0, RoboCasa GR-1 Tabletop）以及xArm平台的真实世界实验中，ZR-0均表现出强大的性能。

Insight: 主要创新点在于提出了密集的具身思维链监督来对齐跨具身表示，以及双流架构设计（System 2用于生成结构化推理，System 1用于生成动作），该设计允许在推理时完全跳过思维链生成而不会损失性能，提高了效率。

Abstract: Cross-embodiment transfer in vision-language-action (VLA) models remains challenging because low-level state and action spaces differ fundamentally across robot platforms. We observe that the high-level cognitive process underlying manipulation, including scene perception, object identification, task planning, and sub-task decomposition, is largely shared across embodiments. Based on this observation, we present ZR-0, a 2.6 billion parameter end-to-end VLA model that uses dense Embodied Chain-of-Thought (ECoT) supervision to align cross-embodiment representations within the vision-language model (VLM). ZR-0 adopts a dual-stream architecture: a pre-trained VLM (System 2) generates structured ECoT reasoning during training, while a Diffusion Transformer-based action expert (System 1) produces continuous action chunks via flow matching. The two components are coupled through cross-attention, with an attention mask that restricts the action expert to input prompt features only, enabling ECoT generation to be entirely skipped at inference without any performance loss. ZR-0 is pre-trained on ProcCorpus-60M, a large-scale dataset comprising approximately 60 million frames (approximately 1,000 hours) from over 400K trajectories, with dense ECoT annotations covering 96.8% of all frames. We evaluate ZR-0 on three simulation benchmarks spanning single-arm (LIBERO), bimanual (RoboTwin 2.0), and humanoid (RoboCasa GR-1 Tabletop) embodiments, as well as real-world experiments on the xArm platform, demonstrating strong performance across all settings. Code and model checkpoints are available at https://github.com/RUCKBReasoning/ZR-0.

[226] GROW$^2$: Grounding Which and Where for Robot Tool Use cs.RO | cs.AI | cs.CVPDF

Yuhong Deng, Yuyao Liu, David Hsu

TL;DR: 本文提出了GROW$^2$模型，用于解决机器人工具使用中的开放世界可供性接地问题。该模型通过将接地过程分层为语义和几何两个层面，利用视觉语言模型进行常识推理以选择工具和识别相关部件，再通过视觉基础模型将这些部件精确地定位到3D区域中。

Details

Motivation: 为了解决机器人在缺乏专用工具时，需要创造性地使用开放类别物体作为工具（即开放世界可供性接地）的挑战，该研究旨在让机器人能够根据任务指令选择合适的物体并定位其具体作用区域。

Result: 在已建立的基准测试中，GROW$^2$在可供性预测基准上超越了最先进的基线方法。此外，它在开放类别物体上实现了零样本泛化，并且在模拟和真实世界的机器人工具使用实验中均优于基线方法。

Insight: 创新点在于利用物体部件作为自然抽象，将复杂的接地问题分解为语义和几何两个层次，从而避免了数据密集的端到端训练。该方法巧妙地结合了视觉语言模型的常识推理能力和视觉基础模型的几何感知能力，实现了高效且泛化性强的工具使用策略。

Abstract: Can the robot use a plate to cut a cake if no knife is available? Tool use greatly expands robot capabilities, but to use tools creatively beyond their intended functions, the robot faces the challenge of $\textit{open-world affordance grounding}$: select an open-category object to act as a tool and localize its specific region of action. To this end, we introduce GROW$^2$ (GROunding Which and Where), which leverages object parts as a natural abstraction to split the grounding process hierarchically into semantic and geometric levels, thus bypassing the need for data-heavy, end-to-end training. Semantically, GROW$^2$ harnesses the commonsense reasoning of Vision-Language Models (VLMs) to parse a natural-language task instruction, select a suitable object as the tool, and identify task-relevant parts on the tool and the target object. Geometrically, vision foundation models then ground the selected parts into precise 3D regions from a single RGB-D image. Experiments on established benchmarks show that GROW$^2$ outperforms state-of-the-art baselines on affordance prediction benchmarks. Further, it achieves zero-shot generalization over open-category objects and outperforms baselines in both simulated and real-world robot tool use experiments.

cs.LO [Back]

[227] LAMP: Lean-based Agentic framework with MCP and Proof Repair cs.LO | cs.AI | cs.CLPDF

Santhana Srinivasan R, Maithilee Patawar

TL;DR: 本文提出了LAMP框架，一个基于Lean 4的、利用MCP（模型上下文协议）和证明修复的多智能体框架，旨在解决大型语言模型在数学推理中生成证明不可靠且难以验证的问题。该框架通过一个特定领域（组合词论）的本体论在推理时提供显式的结构化领域知识，从而合成经过内核验证的Lean 4证明。

Details

Motivation: 动机在于解决大型语言模型生成的数学证明不可靠且难以验证的问题，同时克服交互式定理证明器（如Lean 4）因形式化知识库（如Mathlib）在某些专业领域（如组合词论）覆盖不足而导致的证明能力受限。

Result: 在涵盖组合词论所有八个模块和三个难度级别的90个定理测试套件中，LAMP为96.7%的定理合成了经过验证的证明，显著超过了无支架的基线方法和现有的专用证明器。消融实验表明，移除其工具接地的架构或规划器/构建器的分离都会导致性能下降约12个百分点。

Insight: 创新点包括：1) 首次在Lean 4中形式化了组合词论领域，提供了核心定义和基础引理；2) 提出了LAMP多智能体框架，通过领域特定本体论在推理时注入知识，而非微调证明器，从而实现了高可靠性的证明合成。其工具接地的架构和智能体分工设计被证明对性能至关重要。

Abstract: Large language models are increasingly capable of mathematical reasoning, but the proofs they generate are often unreliable and hard to verify. Interactive theorem provers such as Lean 4 address this by accepting only kernel-checked proofs; however, their reach is bounded by the formalized knowledge available. While Mathlib, a repository of formalized Lean 4 theorems that covers diverse mathematical areas, certain specialized areas remain underrepresented; notably, the domain of Combinatorics on Words (CoW). CoW studies sequences, exploring their properties such as periodicity, borders, conjugacy, and morphisms. As a result, specialized provers, trained on Mathlib-centered data, lack the lemmas to operate in CoW. We present two contributions. First, we introduce a Lean 4 formalization of CoW containing eight modules and \textbf{93} declarations of core definitions and foundational lemmas. Second, we present LAMP, a multi-agent framework that synthesizes kernel-verified Lean 4 proofs by providing explicit, structured domain knowledge at inference time through an ontology, rather than by fine-tuning a prover. LAMP coordinates a Planner, Builder, and Verifier with Model Context Protocol based access to a domain-specific CoW ontology. In a suite of 90 CoW theorems that span all eight modules and three difficulty levels, LAMP synthesizes verified proofs for 96.7% of theorems, substantially exceeding both an unscaffolded baseline and existing specialized provers. An ablation shows that removing LAMP’s tool-grounded architecture or its Planner/Builder separation each cost roughly 12 percentage points, even with the backbone model held fixed.

cs.LG [Back]

[228] An Agentic AI Pipeline for Appliance-Level Energy Anomaly Detection and LLM-Driven Recommendations cs.LG | cs.AI | cs.CLPDF

Dihia Falouz, Aida Douaibia, Amine Bechar, Youssef Elmir, Abbes Amira

TL;DR: 本文提出了一种面向办公楼电器级能耗异常检测的端到端智能代理流程，该流程结合了深度时间序列预测、变分异常检测和基于LLM的推理，以生成优先级排序且可操作的维护建议。系统采用混合奇异谱分析（SSA）和长短期记忆（LSTM）预测模型跟踪七种办公电器，并应用带注意力机制的电器级LSTM变分自编码器（VAE）来标记异常的日消耗事件。一个三阶段的LangChain流程（包含上下文代理、诊断代理和报告代理）利用动态检索生成结构化诊断和可读报告，并整合了操作员反馈的反射记忆层。

Details

Motivation: 办公楼电器级能耗监测会产生大量噪声警报，非专业的设施管理人员难以有效利用这些信息。本文旨在解决这一问题，通过自动化智能流程将原始数据转化为可执行的维护建议。

Result: 在一个包含持续/瞬态尖峰、意外关机和系统性事件等16种场景的基准测试中，动态检索在所有LLM后端上达到了与完整静态检索相当的性能，同时将每个事件的平均上下文来源从六个减少到三至六个。最佳后端在70分阈值下获得了90.4/100分，通过率为100%，一个完全本地的7B参数模型通过了全部16个场景。

Insight: 主要创新点在于构建了一个结合深度预测、变分异常检测和LLM推理的端到端代理式AI流程，并引入了基于事件特征的动态上下文检索机制（LangChain三阶段流程），在保证性能的同时显著减少了推理所需的上下文量。系统设计还包含了整合人类反馈的反射记忆层，增强了实用性和可迭代性。

Abstract: Appliance-level energy monitoring in office buildings produces noisy alerts that non-expert facility managers struggle to use. This paper proposes an end-to-end agentic pipeline that combines deep time-series forecasting, variational anomaly detection, and LLM-based reasoning to generate prioritized, actionable maintenance recommendations. The system tracks seven office appliances using a hybrid Singular Spectrum Analysis (SSA) and Long Short-Term Memory (LSTM) forecasting model, and applies a per-appliance LSTM Variational Autoencoder (VAE) with attention to flag abnormal daily consumption episodes. A three-stage LangChain pipeline begins with a Context Agent that always retrieves three core RAG sources (model reliability, hourly baseline, and expert knowledge) and conditionally adds up to three more (forecast context, anomaly history, global baseline) based on event characteristics, capped at eight reasoning steps. A Diagnosis Agent converts the evidence into a structured JSON diagnosis, and a Report Agent renders a human-readable narrative. A reflective memory layer incorporates operator feedback. The dashboard shows real-time 30-minute forecasts, intraday consumption, the previous day anomaly report, and a feedback form. We evaluate the forecasting model, anomaly detector with appliance-specific thresholds, and LLM reasoning on a 16-scenario benchmark including sustained and transient spikes, unexpected shutdowns, and systemic events, comparing five LLM backends under static vs. dynamic retrieval. Dynamic retrieval matches full static retrieval across all backends while cutting average context from six to three-six sources per event. The best backend scores 90.4/100 with a 100% pass rate at a 70-point threshold, and a fully local 7B-parameter model passes all 16 scenarios.

Yong Yi Bay, Kathleen A. Yearick

TL;DR: 本文研究了语言模型在测试时通过多次采样（test-time scaling）来回答难题的现象，指出虽然增加采样次数能提高覆盖率（至少一次尝试正确的概率），但实际部署系统必须选择一个答案，而选择能力存在上限。超过一定采样次数后，额外采样只会增加成本，甚至可能因模型对错误答案过于自信而降低性能。论文提出了’模态上限’和’相关性上限’的概念，并引入’有效样本数’作为确定最佳采样次数的指标。

Details

Motivation: 动机在于揭示当前语言模型通过增加测试时采样次数来提升性能的常见做法存在局限性：虽然覆盖率随采样增加而上升，但选择正确答案的能力会达到瓶颈，导致计算资源浪费和潜在性能下降。

Result: 论文通过理论分析和实验表明，在典型基准测试中，选择正确答案的’模态上限’在几十次采样内就已达到，而评估基准得分的’相关性上限’出现得更早，额外采样无益甚至有害。

Insight: 创新点在于提出了’模态上限’和’相关性上限’的概念，以及’有效样本数’这一实用指标，强调模型性能的瓶颈在于识别正确答案而非生成更多样本，为优化测试时采样策略提供了理论依据。

Abstract: People overthink; language models over-sample, and the extra effort can talk both into a worse answer. Reasoning systems answer a hard question by sampling it many times (test-time scaling), and the more they draw, the more often a correct answer turns up somewhere, so coverage, the fraction of problems with at least one correct try, climbs and appears to be progress. But a deployed system must return one answer, and choosing it, not knowing which try is right, is selection; selection is capped, and past a point extra samples only make the model surer of a confident mistake, even as every draw adds cost. The gap between climbing coverage and stalled selection, the identifiability gap, is the answer a model can produce but not pick. So the real question is not whether to sample but how far, and the answer is: not far. For picking an answer, the vote has already settled within a few dozen draws, the modal ceiling; for scoring a benchmark, sooner still, the correlation ceiling. Beyond that, extra draws cost compute and add nothing, and can even make the answer worse. This paper turns the cutoff into a single number, the effective number of samples, that any sampling run already reveals. The bottleneck is recognizing a right answer, not generating one.

[230] Do Models Read What They Write? Causal Registers in Scratchpad Reasoning cs.LG | cs.CLPDF

Benjamin Shih, John Winnicki, Eric Darve

TL;DR: 该论文研究了在思维链推理中模型是否真正读取并利用其写入的中间状态。通过设计受控的状态跟踪任务，实验发现Qwen2.5-Coder-7B等模型在训练写入中间状态后，能够基于编辑后的内部表示（而非可见文本）进行后续计算，表明模型确实将写入的状态作为计算的一部分使用。

Details

Motivation: 研究动机是验证过程监督中模型是否真正将思维链中的中间状态用于后续计算，以评估其对模型对齐的实际帮助。

Result: 在Qwen2.5-Coder-7B上，写入中间状态的模型在两种任务变体中对编辑后状态的下游影响预测准确率达到80%和91%，显著高于仅报告最终答案的对照组，且该因果使用模式在不同模型家族中可复现。

Insight: 创新点在于提出了更精确的思维链监督目标：不仅要使中间推理可读，更要训练模型将写入的状态作为其计算的一部分使用，这为模型对齐提供了新的评估视角。

Abstract: A central hope behind process supervision is that models can expose intermediate variables that matter for their later behavior. For this to help with alignment, a scratchpad must be tied to the computation: when the model writes a state, later steps should compute from that state. To test this requirement, we use a controlled state-tracking task with a known update rule, comparing models trained to report only the final state with models trained to write intermediate states before giving the final answer. At evaluation, we edit the internal representation of one written state while leaving the visible scratchpad text fixed. Because the transition rule is known, the edit has a single correct downstream consequence. In Qwen2.5-Coder-7B, the state-writing model predicts the next phase bit implied by the edited state on 80% and 91% of held-out examples across the two task variants, while pretrained and final-answer-only controls remain near baseline. Additional controls rule out generic next-token steering and copying another continuation: the prediction depends on both the edited state and the current move. The same causal-use pattern replicates across model families. Together, these results suggest a sharper goal for scratchpad oversight: not just to make intermediate reasoning legible, but to train written states that the model uses as part of its computation.

[231] A Diagnostic Framework and Multi-Evaluator Audit of Evaluator-Driven Preference Dynamics in Self-Adapting LLM Agents cs.LG | cs.CLPDF

Liu Zewen

TL;DR: 本文提出了一个诊断框架EPC（包括多模态偏好崩溃指数MPCI、评估器索引耦合矩阵和Jensen-Shannon散度JSD），用于检测和量化LLM评估器驱动的偏好动态中的不稳定性。通过对八个实验条件（共122次重复）的分析，发现评估器之间存在显著耦合差异，且GPT-4o从五月到六月的版本漂移甚至逆转了研究结论，突显了单次快照评估研究的不可靠性。

Details

Motivation: 解决专有LLM评估器的测量结果可能在短期内失效的问题，旨在提供一个可检测评估器偏好动态不稳定性的诊断工具，以增强评估研究的可靠性。

Result: 在八个实验条件下，耦合系数范围从0.00到1.18（平均变异系数约0.9），其中四个条件显示强耦合（如GPT-4o五月版），四个条件崩溃至近零（如GPT-4o六月版）。GPT-4o从五月到六月的版本漂移导致研究结论反转，自我评估则始终崩溃（97%为零，JSD=0.003）。

Insight: 创新点在于提出EPC框架来量化评估器耦合和不稳定性，核心洞察是评估器版本条件性不稳定的模式（而非单一耦合强度）使得单次快照研究不可靠，强调了动态监测的必要性。

Abstract: Measurements of proprietary LLM evaluators can become invalid within weeks – we document one case and provide the diagnostic framework to detect it. We introduce EPC – comprising the Multimodal Preference Collapse Index (MPCI), evaluator-indexed coupling matrix, and Jensen-Shannon divergence (JSD) – and apply it across eight experimental conditions (N=112 main + N=10 ablation = 122 unique repetitions, all reported). Coupling coefficients range from 0.00 to 1.18 across per-condition means (CV approx 0.9, n=8 conditions). Four conditions show strong coupling (N=36; GPT-4o May, GPT-4o-mini, Qwen3.7-plus, DashScope 30r); four collapse to near-zero (N=76; GPT-4o June, qwen-plus N=30, symmetric LR, DeepSeek self-eval). The May-to-June GPT-4o drift – an N=8 re-replication inverting the study’s conclusion – is the most informative measurement: a diagnostic instrument detecting its own instability demonstrates the fragility it was designed to measure. Self-evaluation (97% zero, JSD=0.003) consistently collapses, though floor effects are possible. Output-format confound analysis finds per-strategy aggregate rho=0.89 but per-instance rho=0.219 (p=0.093); PCI reported as preference-convergence metric. We release EPC with all data. The finding is not any single coupling magnitude but the pattern of version-conditional instability that makes single-snapshot evaluator studies unreliable.

[232] Can AI Draw Science? A Benchmark for Evaluating Scientific Figure Generation by Text-to-Image and Multimodal Models cs.LG | cs.CV | cs.GRPDF

Davie Chen

TL;DR: 该论文提出了SciDraw-Bench，一个专门用于评估文本到图像和多模态模型生成科学图表（如机制图、示意图、概念框架等）能力的基准测试。它包含32个结构化任务，覆盖8种图表类型和10个学科，并设计了一个包含文本保真度、语义正确性、结构质量和规范遵循性的四维评估协议。初步评估表明，领域专用系统SciDraw AI在所有维度和图表类型上都显著优于通用基线模型。

Details

Motivation: 现有的图像生成基准（如GenEval）主要评估自然图像的组合性、物体计数或真实感，缺乏对生成科学图表可用性的衡量标准，例如正确的文本标签、实体关系的忠实描绘、连贯的图表结构以及学科绘图规范的遵循。

Result: 在一个涵盖所有八种图表类型的试点评估中，领域专用系统SciDraw AI在文本保真度、语义正确性、结构质量和规范遵循性四个维度上，对每种图表类型都大幅优于代表性的通用文本到图像模型基线，其中在语义正确性和规范遵循性方面的差距最大。

Insight: 论文的核心创新在于首次构建了一个结构化、可机器检查的科学图表生成基准（SciDraw-Bench）及其多维评估协议，填补了该领域的评估空白。其提出的评估维度（特别是语义正确性和规范遵循性）为衡量生成内容的领域适用性而非仅仅是视觉质量提供了新视角。

Abstract: Text-to-image and multimodal generative models are increasingly used to produce scientific figures such as mechanism diagrams, experimental-design schematics, conceptual frameworks, and graphical abstracts. Yet existing image-generation benchmarks (e.g., GenEval, T2I-CompBench, DPG-Bench) evaluate natural images and measure compositionality, object counting, or photorealism. None of them measure what makes a generated scientific figure usable: correct and legible text labels, faithful depiction of entities and their relations, coherent diagrammatic structure, and adherence to disciplinary drawing conventions. We introduce SciDraw-Bench, a benchmark of 32 structured scientific-figure generation tasks spanning eight figure types and ten disciplines, where each task pairs a natural-language prompt with a machine-checkable specification of required labels, relations, components, conventions, and negative constraints. We propose a four-dimensional evaluation protocol: Text Fidelity (OCR-based label recall and character error rate), Semantic Correctness (vision-language-model judging against the specification), Structural Quality, and Convention Adherence, together with a meta-evaluation protocol and a preliminary inter-judge reliability analysis (human-rating validation is ongoing). We evaluate a domain-specific system, SciDraw AI, against representative general-purpose text-to-image models, and outline a code-to-figure baseline as a planned extension. In a pilot over all eight figure types, the domain-specific system substantially outperforms the general-purpose baselines on every dimension and figure type, with the largest gaps on semantic correctness and convention adherence; text fidelity remains the hardest dimension for all systems.

Chungpa Lee, Jihoon Kwon, Kyle Min, Jy-yong Sohn

TL;DR: 本文挑战了稀疏自编码器在跨模态嵌入空间中的隐含假设，即同一概念在不同模态下共享相同的特征方向。研究发现存在跨模态特征异质性现象，并提出了一种训练模态特定稀疏自编码器并事后对齐对应特征的方法，以提高重建保真度和跨模态检索性能。

Details

Motivation: 解决视觉语言模型联合嵌入空间中语义特征纠缠的问题，并质疑稀疏自编码器应用中关于跨模态特征方向一致的隐含假设。

Result: 所提方法在重建保真度、跨模态检索和概念引导任务上表现更优，但未明确提及具体基准测试或是否达到SOTA水平。

Insight: 揭示了跨模态特征异质性现象，并提出通过模态特定稀疏自编码器保持特征几何结构，再事后对齐对应特征，以解决模态分裂问题。

Abstract: Vision-language models map images and text into a joint embedding space. However, these embeddings often entangle multiple semantic features, which limits their interpretability and controllability. While sparse autoencoders have emerged as a useful tool for decomposing these embeddings into monosemantic features, their application to joint embedding spaces has largely relied on an implicit, untested assumption that semantically corresponding features share the same directions across modalities. In this paper, we challenge this assumption by identifying discrepancies in feature directions for the same concept across image and text modalities, a phenomenon we term cross-modal feature heterogeneity. We demonstrate that this heterogeneity is a key driver of the modality split, where a shared concept activates different latents depending on the modality. This finding further reveals why aligning latent activations alone is insufficient to resolve the underlying feature mismatch. Motivated by this observation, we propose an approach that trains modality-specific sparse autoencoders to preserve each modality’s feature geometry, and then aligns corresponding features post hoc. Our method improves reconstruction fidelity and enhances performance in cross-modal retrieval and concept steering.

[234] Neural Subspace Reallocation: Continual Learning as Retrieval-Based Subspace Memory Management cs.LG | cs.AI | cs.CVPDF

Byeong Hoon Yoon

TL;DR: 本文提出神经子空间重分配（NSR）方法，将持续学习重新定义为参数子空间上的内存管理。该方法将低秩适应（LoRA）模块视为可压缩、可检索的内存单元，通过压缩、存储、检索和重分配的循环过程管理任务知识，在固定容量的骨干网络上实现高效的任务记忆与恢复。

Details

Motivation: 传统持续学习方法通常将每个任务的LoRA适配器视为一次性模块，导致任务切换时效率低下且容易遗忘。NSR旨在通过将LoRA模块作为可管理的记忆单元，解决在固定模型容量下高效复用历史知识、减少遗忘并加速循环任务恢复的问题。

Result: 在Split-CIFAR-100上，NSR将循环任务恢复时间减少了10倍；在异构的5-Datasets基准测试中，NSR取得了最高的准确率和最少的遗忘，其反向迁移（backward transfer）接近零的程度比无记忆启发式方法高约9倍。内存占用分析显示每个任务仅需0.29 MB参数存储。

Insight: 核心创新在于将持续学习视为基于检索的子空间内存管理，通过SVD压缩和嵌入相似性检索的简单机制驱动性能提升，而非依赖复杂的强化学习控制器。研究发现，记忆机制（压缩与相似性检索）本身比学习到的分配策略对固定容量下的持续学习性能更为关键。

Abstract: We introduce Neural Subspace Reallocation (NSR), which reframes continual learning as memory management over parameter subspaces. Instead of treating Low-Rank Adaptation (LoRA) modules as disposable per-task adapters, NSR manages them as compressible, retrievable memory units on a frozen backbone through a recurring cycle: (1) compress learned LoRAs via SVD, (2) reserve them in a TaskKnowledgeBank, (3) recall related past LoRAs by embedding similarity to warm-start new or returning tasks, and (4) reallocate the active subspace accordingly, with distillation protecting prior tasks. We prove that in cyclic environments any memoryless allocation policy incurs cumulative regret Omega(T(M-1)Delta_switch) relative to a history-aware policy backed by the Bank (Theorem 1). Empirically, on Split-CIFAR-100 the Bank reduces cyclic recovery time by 10x, exactly as predicted, and on the heterogeneous 5-Datasets benchmark NSR achieves the highest accuracy and the least forgetting, about 9x closer to zero backward transfer than the memoryless heuristics. Crucially, we run a controlled study that isolates which component matters: holding the Bank fixed and varying only the allocation rule, we find that a simple similarity-based retrieval rule matches or beats a learned reinforcement-learning controller (recovering recurring tasks in 0 vs 1.8 steps and reaching equal accuracy). Our central, honest finding is therefore that the memory mechanism – compression and similarity retrieval – rather than a learned allocation policy, drives continual-learning performance under fixed capacity. A memory-budget analysis confirms the compressed Bank stays small – 0.29 MB of parameter memory per task – so a top-K retention cap bounds the total footprint while preserving fast recovery for retained tasks.

[235] FlowAWR: Online Adaptive Flow Reinforcement via Advantage-Weighted Rectification cs.LG | cs.CVPDF

Zheming Fu, Ruizhe He, Wei Shang, Xiaoxiao Ma, Lei Wang

TL;DR: 本文提出了FlowAWR，一种用于连续生成流模型的在线自适应强化学习新范式。它通过将策略优化重新表述为对理论最优速度场的监督回归，避免了传统方法对随机微分方程采样器和无分类器引导的依赖，从而实现了更高效、更一致的训练与生成。

Details

Motivation: 现有基于密度近似的策略梯度方法在连续空间中对齐生成流模型时，依赖于随机SDE采样器来构建可处理的转移核，这导致了训练-推理不一致并需要无分类器引导。而如DiffusionNFT等隐式框架的启发式固定幅度校正，则阻碍了基于组内相对质量的优化强度。

Result: 在SD3.5-Medium上的比较评估中，FlowAWR实现了更好的对齐性能，并且相比DiffusionNFT获得了2到5倍的收敛加速（例如，在1.2k步达到24.12 PickScore，而DiffusionNFT在2.0k步达到23.82，FlowGRPO在>4k步达到23.50）。在多奖励约束下，FlowAWR保持了生成质量，满足结构规则的同时维持了稳定的域外性能。

Insight: 核心创新在于从KL约束的奖励最大化最优策略出发，推导出具有幅度感知、优势加权校正形式的最优速度场。这实现了无需SDE的优化和无需CFG的生成，为连续生成模型的强化学习对齐提供了一个理论驱动且更高效的监督回归框架。

Abstract: Aligning generative flow models on continuous spaces via online reinforcement learning is constrained by intractable trajectory likelihoods. Existing density-approximated policy gradient methods rely on stochastic SDE samplers to construct tractable transition kernels, which introduce training-inference inconsistencies and necessitates Classifier-Free Guidance (CFG). While implicit frameworks such as DiffusionNFT directly optimize forward-process velocity fields, its heuristic fixed-magnitude corrections prevent optimization strength from relative intra-group quality. We propose \textit{Flow Advantage-Weighted Rectification} (\textbf{FlowAWR}), a paradigm that recasts continuous generative policy optimization as supervised regression toward a theoretically optimal velocity field. Starting from the optimal policy of a KL-constrained reward maximization, FlowAWR derives the optimal velocity field that admits a magnitude-aware, advantage-weighted rectification form, yielding SDE-free optimization and CFG-free generation. In comparative evaluations on SD3.5-Medium, FlowAWR achieves improved alignment performance alongside a 2$\times$ to 5$\times$ convergence acceleration over DiffusionNFT (e.g., reaching a 24.12 PickScore in 1.2k steps, versus 23.82 in 2.0k steps for DiffusionNFT and 23.50 in $>$4k steps for FlowGRPO). Under multi-reward constraints, FlowAWR sustains generation quality, satisfying structural rules while maintaining stable out-of-domain performance.

[236] DreamForge-World 0.1 Preview: A Low-Compute Real-Time Controllable World Model cs.LG | cs.CVPDF

Daniyel Ayupov, Artur Markov-Tsoy

TL;DR: 本文介绍了DreamForge-World 0.1 Preview，这是一个用于实时交互式世界模拟的低计算量基础世界模型预览版。该系统基于LongLive 1自回归视频架构（源自Wan2.1-T2V-1.3B）进行改造，并引入了受Matrix-Game系列启发的残差动作通路。其核心目标是实现消费级GPU上的实时运行，支持键盘鼠标控制、多模态初始化等多种交互功能，在单张RTX 4090上能以480p分辨率达到约14-15 FPS。

Details

Motivation: 该研究的动机是探索与前沿大规模世界模拟器互补的路径，专注于低计算量适应、消费级GPU运行时以及广泛的交互能力覆盖，旨在为实时可控的世界模型预览提供一条实用且高性价比的途径。

Result: 模型在单张RTX 4090 GPU上，能以原生480p分辨率进行分钟级别的交互式推演，帧率达到14-15 FPS，且内存占用较低。论文未提及在特定基准测试上与SOTA模型的定量比较，而是强调其作为预览系统在低计算成本和实时交互性方面的实用演示。

Insight: 摘要宣称的创新点在于将现有开源视频主干网络（LongLive 1/Wan2.1）与受Matrix-Game启发的残差动作通路相结合，通过针对性的适应训练，构建了一个高性价比的实时可控世界模型预览系统。从客观角度看，其核心创新在于在低计算资源约束下，系统性地整合了多种实时交互功能（如键盘鼠标控制、多模态初始化、中途重新提示、双视图操作），为世界模型的实用化部署提供了一个可行的轻量级技术方案。

Abstract: We present DreamForge-World 0.1 Preview, a preview foundational world model for real-time interactive world simulation. The system adapts the LongLive 1 autoregressive video stack, itself derived from Wan2.1-T2V-1.3B, with a residual action pathway inspired by the Matrix-Game family. DreamForge-World 0.1 Preview focuses on a complementary axis to frontier-scale world simulators: low-compute adaptation, consumer-GPU runtime, and broad interactive capability coverage. It supports live keyboard and mouse control, multimodal initialization, mid-stream reprompting, dual-view operation, and minute-scale interactive rollouts at native 480p resolution, reaching up to 14 to 15 FPS FPS on a single RTX 4090 with a low memory footprint. By leveraging open video backbones and applying targeted adaptation runs, we build the preview system with high cost-efficiency. DF-World 0.1 Preview is not yet a memory-complete or frontier-quality world simulator, but demonstrates a practical low-compute route toward real-time controllable world-model previews on consumer GPUs.

eess.IV [Back]

[237] Anatomy-Grounded Synthetic Coronary Angiography for Geometry-Informed Multi-View Matching eess.IV | cs.CVPDF

In Kyu Lee, Sumin Seo, Jaesik Min

TL;DR: 本文提出了一种基于解剖学的合成冠状动脉造影数据生成框架，用于解决多视角血管造影图像匹配中缺乏真实标注数据的问题。该框架通过从3D冠状动脉CT血管造影（CCTA）体积生成高保真数字重建放射影像（DRR），并模拟C臂采集几何来获得密集且准确的3D到2D投影标签。在此基础上，作者设计了一个几何信息匹配模块（GIMM），将全局特征和解剖结构整合到对应关系学习中，以提升匹配性能。

Details

Motivation: 解决多视角冠状动脉造影图像对应匹配任务中，由于获取真实标注数据成本极高且难以扩展，导致鲁棒深度学习模型发展受阻的根本性数据瓶颈问题。

Result: 在提出的基于CT衍生的DRR数据集上进行了全面评估，结果表明该方法优于其他基线匹配模型。

Insight: 创新点在于提出了一个物理基础的数据生成框架，以零人工成本生成具有密集、高精度3D-2D投影标签的合成数据，从而绕过了真实数据标注的瓶颈；并设计了一个几何信息匹配模块，将解剖结构先验整合到特征匹配中，实现了无人工干预的评估。

Abstract: Accurate correspondence matching across multiple angiographic views is the prerequisite for 3D coronary reconstruction and interventional guidance. However, the development of robust deep learning models for this task has been stifled by a fundamental data bottleneck. Obtaining ground truth for matching tasks in angiography pairs is prohibitively expensive and hard to scale. To overcome this barrier, we introduce a physically-grounded data generation framework that synthesizes high-fidelity Digital Reconstructed Radiographs (DRRs) from 3D Coronary CT Angiography (CCTA) volumes. Our framework generates dense, highly accurate 3D-to-2D projection labels by simulating realistic C-arm acquisition geometry on patient anatomy at zero human cost. Leveraging this dense supervision, we propose a Geometry-Informed Matching Module (GIMM) that integrates global feature and anatomical structure into correspondence learning. Unlike real angiography where assessment relies on subjective human annotation, our dataset provides 2D correspondence labels with paired images, allowing human-free evaluation. We comprehensively evaluate our method on the proposed CT-derived DRR dataset and demonstrate improvements over other matching baseline models.

[238] BLUE: A Stale-Pixel Optical-Flow Compositor for Entropy-Efficient Surveillance Video Encoding eess.IV | cs.CVPDF

Shubham Baid, Akash James, Sahil Chachra, Nishant Sinha, Kunal Kislay

TL;DR: 本文提出了一种名为BLUE的预编码合成器，用于解决静态摄像头监控视频存储效率低下的问题。该方法通过维护一个持久的背景种子帧，在编码前将背景像素替换为种子像素，从而大幅减少背景重复编码的比特开销。实验表明，BLUE在保持前景质量的同时，显著降低了视频文件大小，并提供了选择性激活策略以平衡存储节省与感知质量。

Details

Motivation: 连续录制监控系统面临存储瓶颈，即使调整编码参数，静态背景仍占用大量比特重复编码。现有编解码器调优无法完全解决此问题，因此需要一种能利用场景静态特性的预处理方法。

Result: 在VIRAT Ground Surveillance Release 2.0数据集的308个子片段上测试，使用x264和x265编码器时，BLUE在CRF 28下平均减少文件大小34.6%（x264）和39.4%（x265）。前景区域PSNR在60.7%的片段中保持或提升（平均+0.36 dB），但全帧感知质量（VMAF）中位数下降6.75-8.59分。通过轻量级部署门控机制，可识别40%的片段实现近乎不可察觉的质量下降。

Insight: 创新点在于利用光流和种子帧合成技术，将静态背景编码为近乎零成本的SKIP宏块，从而显著提升编码效率。客观分析表明，该方法通过显式权衡存储节省与感知质量，并引入选择性激活策略，为静态场景监控编码提供了可借鉴的预处理框架。

Abstract: Continuous-recording surveillance systems face a storage problem that codec tuning alone cannot fully solve: even at aggressive CRF settings, a static-camera scene spends most of its bits re-encoding a background that has not changed. We present BLUE, a pre-encode compositor that exploits this structure by maintaining a persistent seed frame of the background and substituting background pixels with seed pixels before the encoder runs. The encoder then emits near-free SKIP macroblocks for the frozen background, while live pixels in foreground regions are carried unchanged at full quality. We evaluate BLUE on all 308 annotated short subclips from the VIRAT Ground Surveillance Release 2.0 dataset using a six-point CRF sweep with both x264 and x265. At CRF 28, BLUE reduces file size by a mean of 34.6% (x264) / 39.4% (x265) on 95.8% / 99.4% of clips respectively. Foreground-region PSNR, computed only over VIRAT object-annotation bounding boxes, is preserved or improved on 60.7% of clips (+0.36 dB mean, +5.48 dB maximum). Full-frame perceptual quality (VMAF) drops by a median of 6.75-8.59 points; we quantify and disclose this trade-off explicitly. A lightweight deployment gate measuring the compositor’s own VMAF on a 2-second prefix identifies the 40% of clips where even full-frame quality degradation is near-imperceptible (Delta VMAF <= -2.9), enabling a selective-activation strategy that retains both the storage benefit and acceptable perceptual fidelity.

[239] A Self-Supervised Learning Framework for Video Encoding Complexity Clustering eess.IV | cs.CVPDF

Krishna Srikar Durbha, Hassene Tmar, Ping-Hao Wu, Ioannis Katsavounidis, Alan C. Bovik

TL;DR: 本文提出了一种名为压缩回声对比学习（CECL）的自监督学习框架，用于根据视频的编码复杂度对视频进行聚类。该方法利用视频对压缩的响应（即压缩回声）作为监督信号，在预训练中捕获编码特性，以改进自适应视频流中编码设置的优化。

Details

Motivation: 自适应视频流在互联网视频传输中广泛应用，但确定每个视频的最佳编码设置是一个关键挑战，因为视频内容和特性差异很大。本文旨在通过自监督学习聚类视频编码复杂度，以解决这一优化问题。

Result: 实验表明，CECL在编码复杂度聚类下游任务中有效，其学习到的表示优于现有最先进的视觉编码器，并在固定比特率阶梯上实现了显著的比特率和质量节省。

Insight: 创新点在于引入压缩回声作为自监督信号来捕获视频编码特性，这是一种新颖的表示学习方法，可借鉴于视频处理领域，以无监督方式优化编码决策。

Abstract: Adaptive video streaming is a widely used technique for delivering video content over the internet. One of the key challenges is determining the optimal encoding settings for each video, which can vary significantly based on its content and characteristics. In this paper, we propose Compression Echo Contrastive Learning (CECL), a novel self-supervised learning framework for clustering videos based on their encoding complexity. Our method leverages the response of a video to compression - the Compression Echo - as a supervisory signal, allowing the model to capture underlying encoding characteristics during pretraining. We conduct extensive experiments to demonstrate the effectiveness of our learned representations for the downstream task of clustering videos by their encoding complexity. Our results show that CECL improves upon existing state-of-the-art visual encoders and delivers strong bitrate and quality savings against the fixed bitrate ladder.

cs.CR [Back]

[240] The Calibrated Deepfake Trust Score (CDTS): Competence-Coupled Trust Degradation Across Deepfake Detectors cs.CR | cs.CV | cs.LGPDF

Md Anas Biswas

TL;DR: 本文提出了一种名为校准深度伪造信任分数（CDTS）的框架，将深度伪造检测重新定义为一种经过校准的、可自我审计的信任工具。核心发现是检测器的校准质量与其判别能力之间存在强耦合关系：当检测器的判别能力下降时，其输出信任分数的校准度也会恶化。这一现象在多种检测器架构和数据集上得到验证，并可用于构建无需标签的校准风险监控以及改进基于置信度的路由策略。

Details

Motivation: 现代深度伪造检测器在内容审核、溯源和验证等实际应用中，其输出概率常被解读为信任度，因此其校准质量与原始准确率同等重要。论文旨在研究并量化影响这种信任工具可靠性的核心因素。

Result: 在32种实验配置中，检测器能力与校准度之间存在强负相关（汇总皮尔逊相关系数 r = -0.81）。这一结果在三个架构不同的检测器（两个卷积网络和一个CLIP视觉变换器）上均得到验证（r分别为-0.88, -0.83, -0.86）。基于能力估计的无标签监控可以标记未见生成器的校准风险，并且基于能力的路由策略在低到中覆盖率操作区域优于基于置信度的路由，降低了整体AURC。

Insight: 论文的核心创新点在于揭示了深度伪造检测器中“能力-校准耦合”这一普遍现象，并论证了“能力”是影响检测器可信度（包括校准公平性和解释忠实度）的统一驱动因素。提出的CDTS包装器机制，以及无需真实标签即可估计能力以监控校准风险的方法，为构建更可靠的深度伪造信任评估系统提供了新思路。

Abstract: Modern deepfake detectors are rarely consumed as bare classifiers. In moderation, provenance, and verification pipelines their output probability is read as a degree of trust, so its calibration matters as much as raw accuracy. We reframe deepfake detection as a calibrated, self-auditing trust instrument, the Calibrated Deepfake Trust Score (CDTS), and identify what governs its trustworthiness. Our central finding is a competence-calibration coupling: the calibration of the trust score degrades as the detector’s discriminative competence falls. We establish it across 32 configurations (pooled Pearson r = -0.81), demonstrate it within a single dataset, reinforce it by inducing low competence directly, and replicate it on a fourth held-out dataset the detectors never trained on. It holds across three architecturally distinct detectors, two convolutional networks and a CLIP vision transformer (r = -0.88, -0.83, -0.86). The result is also deployable: a single calibrator frozen on in-domain data fails on exactly the low-competence generators the coupling flags (its error tracks competence at r = -0.98), and competence is estimable without labels, so a label-free monitor flags calibration risk on unseen generators and routing source-batches on a reference-free competence estimate lowers overall AURC and improves the low-to-mid coverage operating region relative to confidence-based routing. The same competence factor also drives calibration inequity across demographic subgroups (distinct from accuracy inequity) and explanation faithfulness. We therefore argue that detector trustworthiness is organized by competence as a shared driver, that competence is the right quantity to estimate and condition on, and that trust scoring must be competence-aware. We offer the CDTS wrapper as the mechanism, and report openly where the unification is tight and where it is architecture-specific.

Table of Contents

cs.CL [Back]

[1] SEAD: Competence-Aware On-Policy Distillation via Entropy-Guided Supervision cs.CLPDF

[2] Developmental Trajectories of Situation Modeling and Mentalizing in Transformer Language Models cs.CLPDF

[3] Latent Bridges for Multi-Table Question Answering cs.CL | cs.AI | cs.DBPDF

[4] Memory-Managed Long-Context Attention: A Preliminary Study of Editable Request-Local Memory cs.CL | cs.LGPDF

[5] EVLA: An Electro-Aware Multimodal Assistant for Physically-Grounded Driving Reasoning and Control cs.CLPDF

[6] A3M: Adaptive, Adversarial and Multi-Objective Learning for Strategic Bidding in Repeated Auctions cs.CL | cs.LGPDF

[7] The strength of clinical evidence is recoverable from language model representations but not from their stated grades cs.CL | cs.AI | cs.IR | cs.LGPDF

[8] Masked Diffusion Decoding as $x$-Prediction Flow cs.CLPDF

[9] ThinkProbe: Beyond Accuracy – Structural Profiling of Open-Ended LLM Reasoning Traces via Non-Generative Thought Graphs cs.CLPDF

[10] Evolution Fine-Tuning: Learning to Discover Across 371 Optimization Tasks cs.CL | cs.LGPDF

[11] DistilledGemma: Balanced Efficiency-Accuracy for Person-Place Relation Extraction from Multilingual Historical Articles cs.CLPDF

[12] Can OCR-VLMs Read Devanagari? A Stress-Test Benchmark and Post-Correction Study cs.CL | cs.CVPDF

[13] Travel-Oriented Reasoning Large Language Model via Domain-Specific Knowledge Graphs cs.CL | cs.NEPDF

[14] MIThinker: A Plug-and-Play Policy-Optimized Thinker For Motivational Interviewing Counseling cs.CLPDF

[15] TriageRA-CCF: Source-Side Clinical Confidence and Coverage Signals for Adaptive Rank Budgeting in Medical LLMs cs.CLPDF

[16] EntroRouter: Learning Efficient Model Routing via Entropy Regulation cs.CLPDF

[17] To Reason or to Fabricate: Reasoning Without Shortcuts via Hint-Anchored Pairwise Aggregation cs.CL | cs.AIPDF

[18] The Verbose Context Problem in Medical Records cs.CL | cs.AIPDF

[19] Coverage-Driven KV Cache Eviction for Efficient and Improved Inference of LLM cs.CL | cs.AIPDF

[20] AURORA: Asymmetry and Update-Induced Rotation for Robust Hallucination Detection in Large Language Models cs.CLPDF

[21] Hybrid Retriever Evolution for Multimodal Document Reasoning Agents cs.CL | cs.AI | cs.LG | cs.MAPDF

[22] Two-Stage Prompt Optimization for Few-Shot Relation Extraction: From Reasoning-Guided Search to Gradient-Guided Refinement cs.CL | cs.AIPDF

[23] Resolution Thresholds in VLM Detection of Harmful ASCII Art Across Construction Modes and Languages cs.CLPDF

[24] How LLMs See Creativity: Zero-Shot Scoring of Visual Creativity with Interpretable Reasoning cs.CLPDF

[25] Can MLLMs Critique Like Humans? Evaluating Open-Ended Aesthetic Reasoning in Multimodal Large Language Models cs.CLPDF

[26] Why Struggle with Continuous Latents? Interpretable Discrete Latent Reasoning via Rendered Compression cs.CLPDF

[27] SEVA: Self-Evolving Verification Agent with Process Reward for Fact Attribution cs.CL | cs.AI | cs.LGPDF

[28] KbSD: Knowledge Boundary aware Self-Distillation for Behavioral Calibration in Agentic Search cs.CLPDF

[29] ARKD: Adaptive Reinforcement Learning-Guided Bidirectional KL Divergence Distillation for Text Generation cs.CL | cs.AIPDF

[30] Clinical Reasoning Graphs: Structured Evaluation of LLM Diagnostic Reasoning Reveals Competence Without Consistency cs.CL | cs.AI | q-bio.QMPDF

[31] Towards Physical Intuitions for Alignment Dynamics: A Case Study With Randomness Crystallization cs.CLPDF

[32] LatentRevise: Learning from Zero-Hit Reasoning cs.CLPDF

[33] Are We Measuring Strategy or Phrasing? The Gap Between Surface- and Approach-Level Diversity in LLM Math Reasoning cs.CLPDF

[34] Not-quite-human tastes: the stylized omnivorousness of LLM survey surrogates cs.CL | econ.GNPDF

[35] Efficient Retrieval-Augmented Generation via Token Co-occurrence Graphs cs.CL | cs.IRPDF

[36] CORTEX: High-Quality Cross-Domain Organization of Web-Scale Corpora through Ontological Corpus Graph cs.CLPDF

[37] Forewarned is Forearmed: When Non-Sequential Embedding Turns Into an Anomaly Detector cs.CL | cs.AI | cs.LG | eess.ASPDF

[38] DAIN: Dynamic Agent-Based Interaction Network for Efficient and Collaborative Multimodal Reasoning cs.CLPDF

[39] Before Thinking, Learn to Decide: Proactive Routing for Efficient Visual Reasoning cs.CLPDF

[40] Grounding LLM Reasoning under Incomplete Graph Evidence cs.CLPDF

[41] MOPD: Multi-Teacher On-Policy Distillation for Capability Integration in LLM Post-Training cs.CL | cs.LGPDF

[42] Regime-Aware Peer Specialization for Robust RAG under Heterogeneous Knowledge Conflicts cs.CLPDF

[43] Uncertainty-Aware Generation and Decision-Making Under Ambiguity cs.CL | cs.LGPDF

cs.CV [Back]

[44] RADIANT-PET: Reasoning-Augmented PET/CT Lesion Segmentation with Large Language Models and Reinforcement Learning cs.CV | cs.AI | cs.LGPDF

[45] SoccerNet 2026 Player-Centric Ball Action Spotting: Per-Player Attention with Agreement-Based Ensembling cs.CVPDF

[46] Transition-Aware best-of-N sampling for Longitudinal Chest X-ray Reports cs.CVPDF

[47] RadarTwin: Scene-Specific mmWave Radar Simulation and Learning for Mobile Indoor Perception cs.CV | cs.LGPDF

[48] GPU-Accelerated Inverse Structural Anastylosis from Block Collapse Dynamics cs.CV | cs.LGPDF

[49] CLOSER-VLN: Closed-Loop Self-Verified Retrieval-Augmented Reasoning for Aerial Vision-Language Navigation cs.CV | cs.AIPDF

[50] Meta-learning as a principle for human-like visual representations cs.CV | cs.LG | q-bio.NCPDF

[51] Vision-driven Preference Synthesis for Mitigating Hallucinations in VLMs cs.CV | cs.LGPDF

[52] AEGIS: A Semantic GAN and Evidential Learning Frameworkfor Robust Adversarial Detection in Vision Sensors cs.CV | cs.AIPDF

[53] CLEAR-MoE: Shared-Basis Expert Extraction from Frozen Vision Transformers via Calibration-Driven Layer Selection cs.CV | cs.DCPDF

[54] Detecting Clinical Hallucinations in LVLMs via Counterfactual Visual Grounding Uncertainty cs.CV | cs.CLPDF

[55] Digitizing Coaching Intelligence: An Agentic Framework for Holistic Athlete Profiling using VLM and RAG cs.CV | cs.AI | cs.MAPDF

[56] DataComp-VLM: Improved Open Datasets for Vision-Language Models cs.CV | cs.CL | cs.LGPDF

[57] SatSplat: Geometrically-Accurate Gaussian Splatting for Satellite Imagery cs.CVPDF

[58] IMU-HOI: A Symbiotic Framework for Coherent Human-Object Interaction and Motion Capture via Contact-Conscious Inertial Fusion cs.CVPDF

[59] Animation2Code: Evaluating Temporal Visual Reasoning in Video-to-Code Generation cs.CV | cs.AI | cs.CLPDF

[60] Obliviate: Erasing Concepts from Autoregressive Image Generation Models cs.CVPDF

[61] FedLAS: Feature-Modulated Bidirectional Label Smoothing for Neural Network Calibration cs.CV | cs.AI | cs.LGPDF

[62] BackTranslation2.0 – A Linguistically Motivated Metric to Assess Sign Language Production cs.CVPDF

[63] Predicting Metastatic Risk from Primary Tissue Architecture via Distance-Aware Spatial Modeling cs.CV | cs.AIPDF

[64] SATB-VR: Training Few-Step Video Restoration Diffusion Model using SNR-Aware Trajectory Blending cs.CVPDF

[65] Mitigating Batch Effects in Histopathology via Language-Mediated Robust Embedding Generation cs.CV | cs.CLPDF

[66] CCRC: A Change-Aware Captioning and Reasoning Chain for Image Change Captioning and Segmentation cs.CV | cs.AIPDF

[67] A Physics-Grounded Benchmark for Multi-Agent Dynamics in World Models cs.CV | cs.ROPDF

[68] X-Mind: Efficient Visual Chain-of-Thought via Predictive World Model for End-to-End Driving cs.CV | cs.AIPDF

[69] ViPSim: Collaborating Visual and Parameter Spaces for Consistent Long-Horizon Embodied World Models cs.CV | cs.ROPDF

[70] CoGS: Compositional Dynamic Human-Object Scenes Gaussian Splatting from Monocular Video cs.CVPDF

[71] DLGStream: Dynamic Language-embedded Guassian Splatting for Open-vocabulary Enabled Free-viewpoint Video Streaming cs.CVPDF

[72] Ground4D: Consistency-Aware 4D Reconstruction from Monocular Video cs.CVPDF

[73] Personalizing MLLMs via Reinforced Multimodal Reference Game cs.CVPDF

[74] On Test-Time Scaling for Vision-Language Models cs.CVPDF

[75] HKVLM: Faithful Reasoning Grounding by Binding Language Queries to a Frozen Detector cs.CVPDF

[76] ExACT: Exemplar-Driven Calibrated Refinement for Training-Free Visual Grounding in Remote Sensing Images cs.CVPDF

[77] Self-Evolving Agentic Image Restoration via Deliberate Planning and Intuitive Execution cs.CVPDF