Conan：基于多尺度视觉证据的逐步学习以像侦探一样推理
Conan: Progressive Learning to Reason Like a Detective over Multi-Scale Visual Evidence
AI总结：
Conan通过多阶段渐进冷启动策略和AIR RLVR框架，实现证据基础的多步视频推理，超越基线模型，达到最先进的性能。
摘要：
视频推理需要跨帧的多步推理，仍然是多模态大语言模型（MLLMs）的主要挑战。虽然基于强化学习（RL）的方法增强了推理能力，但它们通常依赖于纯文本链，导致结论不切实际或虚假。相反，帧检索方法引入了视觉 grounding，但仍难以准确定位证据。为了解决这些限制，我们提出了Conan，一个用于证据基础的多步视频推理框架。Conan识别上下文和证据帧，跨帧线索推理，并自适应决定何时结论或进一步探索。为此，我们1）构建Conan-91K，一个大规模自动生成推理轨迹的数据集，包括帧识别、证据推理和动作决策；2）设计多阶段渐进冷启动策略，结合Identification-Reasoning-Action（AIR）RLVR训练框架，逐步激励多步视觉推理。在六个多步推理基准上的大量实验表明，Conan在准确性上平均超过基线Qwen2.5-VL-7B-Instruct超过10%，达到最先进的性能。此外，Conan有效地泛化到长视频理解任务，验证了其强大的可扩展性和鲁棒性。
英文摘要：
Video reasoning, which requires multi-step deduction across frames, remains a major challenge for multimodal large language models (MLLMs). While reinforcement learning (RL)-based methods enhance reasoning capabilities, they often rely on text-only chains that yield ungrounded or hallucinated conclusions. Conversely, frame-retrieval approaches introduce visual grounding, yet still struggle with inaccurate evidence localization. To address these limitations, we present Conan, a framework for evidence-grounded multi-step video reasoning. Conan identifies context and evidence frames, reasons over cross-frame clues, and adaptively decides when to conclude or explore further. To achieve this, we 1) construct Conan-91K, a large-scale dataset of automatically generated reasoning traces that include frame identification, evidence reasoning, and action decision, and 2) design a multi-stage progressive cold-start strategy combined with an Identification-Reasoning-Action (AIR) RLVR training framework to progressively incentivize multi-step visual reasoning. Extensive experiments on six multi-step reasoning benchmarks demonstrate that Conan surpasses the baseline Qwen2.5-VL-7B-Instruct by an average of over 10% in accuracy, achieving state-of-the-art performance. Furthermore, Conan generalizes effectively to long video understanding tasks, validating its strong scalability and robustness.
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。
