MAVEN：一种多阶段代理标注管道用于视频推理任务
MAVEN: A Multi-stage Agentic Annotation Pipeline for Video Reasoning Tasks
AI总结：
本文提出MAVEN，一种多阶段代理标注管道，通过链式推理轨迹生成多任务训练数据，用于视频事件推理任务，核心方法是多尺度时空事件描述，支持代理驱动的领域适应，通过分层细化循环改进数据质量，并在多个数据集上验证了其有效性。
摘要：
训练视频事件推理的视觉语言模型（VLMs）需要高质量的结构化标注，这些标注不仅要描述发生了什么，还要捕捉何时、何地、为何以及后果。我们提出了MAVEN（多阶段代理视频事件标注），一种多阶段代理管道，通过链式推理（CoT）轨迹将原始视频转换为多任务训练数据，围绕指定的事件焦点组织。在核心部分，MAVEN从三个互补的标题级别合成多尺度时空事件描述（MSTED），该显式中间体是下游问答生成的唯一输入，适用于多种任务格式。关键的是，MAVEN支持代理驱动的领域适应：给定新的视频数据集和目标问题示例，代理可以重新设计所有提示，而无需手动重新工程。分层细化循环进一步将注释错误分类到分类学中，追溯根本原因到起始管道阶段，并应用有针对性的编辑，重写提示或修改管道结构本身，迭代改进数据质量。我们应用MAVEN标注超过5,300个交通视频，并在生成的数据上微调Cosmos-Reason2-8B。在私人CCTV评估集上，微调优于Gemini 2.5 Pro和3.1 Flash，包括在零样本情况下MCQ准确率提高了38.8个百分点。在AccidentBench上，仅使用CCTV训练提升了Cosmos-Reason2的MCQ分数10.7分，并在没有dashcam视频的情况下与Gemini 2.5 Pro持平；添加代理适应的dashcam注释缩小了与Gemini 3.1 Flash的差距，RL后训练将总体性能推过了Gemini基线。对仓库监控和公共安全视频的定性结果进一步表明，代理工作流能够轻松适应新领域。
英文摘要：
Training Vision Language Models (VLMs) for video event reasoning requires high-quality structured annotations capturing not only what happened, but when, where, why, and with what consequence, at a scale manual labelling cannot support. We present MAVEN (Multi-stage Agentic Video Event aNnotation), a multi-stage agentic pipeline that turns raw videos into multi-task training data with Chain-of-Thought (CoT) reasoning traces, organized around a designated Event of Focus. At its core, MAVEN synthesizes a Multi-Scale Spatio-Temporal Event Description (MSTED) from three complementary caption levels; this explicit intermediate serves as the sole input to downstream Q&A generation across multiple task formats. Crucially, MAVEN supports agent-driven domain adaptation: given a new video dataset and target question examples, the agent redesigns all prompts top-down without manual re-engineering. A hierarchical refinement loop further classifies annotation errors against a taxonomy, traces root causes to the originating pipeline stage, and applies targeted edits that rewrite prompts or modify the pipeline structure itself, iteratively improving data quality. We apply MAVEN to label over 5,300 traffic videos and fine-tune Cosmos-Reason2-8B on the resulting data. On a private CCTV evaluation set, fine-tuning surpasses both Gemini 2.5 Pro and 3.1 Flash, including a $+38.8$-point gain in MCQ accuracy over zero-shot. On AccidentBench, CCTV-only training lifts Cosmos-Reason2 by $+10.7$ MCQ points and matches Gemini 2.5 Pro despite seeing no dashcam videos; adding agent-adapted dashcam annotations narrows the gap to Gemini 3.1 Flash, and RL post-training pushes overall performance past both Gemini baselines. Qualitative results on warehouse surveillance and public safety videos further show the agentic workflow readily adapts the pipeline to new domains.
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。
