视频过程推理
Process-of-Thought Reasoning for Videos
AI总结：
视频过程推理框架通过结构化推理步骤提升视频理解的准确性与可解释性。
摘要：
视频理解不仅需要识别视觉内容，还需要对长时间和嘈杂的观测进行时间相关、多步骤推理。我们提出了视频过程推理（PoT）框架，该框架通过将视频推理结构化为一系列轻量且可验证的步骤，使推理过程变得明确。PoT结合（i）时间证据选择、（ii）逐步状态更新和（iii）受约束的答案合成，使模型能够逐步完善假设，同时保持对视频证据的可追溯性。该框架设计为模型无关，可以插入到现有的视觉-语言骨干网络中，支持闭书推理和借助外部工具的证据增强推理。我们进一步引入了统一的PoT轨迹表示，使中间决策与时间片段对齐，从而提高对干扰项的鲁棒性，并减少幻觉解释。在标准视频推理任务上的大量实验表明，PoT在事实正确性和时间定位方面表现一致，同时为诊断和下游应用提供了可解释的推理轨迹。
英文摘要：
Video understanding requires not only recognizing visual content but also performing temporally grounded, multi-step reasoning over long and noisy observations. We propose Process-of-Thought (PoT) Reasoning for Videos, a framework that makes the reasoning process explicit by structuring video inference into a sequence of lightweight, verifiable steps. PoT interleaves (i) temporal evidence selection, (ii) step-wise state updates, and (iii) constrained answer synthesis, enabling the model to progressively refine hypotheses while maintaining traceability to video evidence. The framework is designed to be model-agnostic and can be plugged into existing vision-language backbones, supporting both closed-book reasoning and evidence-augmented reasoning with external tools. We further introduce a unified representation for PoT traces that aligns intermediate decisions with temporal segments, which improves robustness to distractors and reduces hallucinated explanations. Extensive experiments on standard video reasoning tasks demonstrate that PoT consistently improves factual correctness and temporal grounding, while providing interpretable reasoning traces for diagnosis and downstream use.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
