长视频时间定位本质是搜索问题，而非识别难题。
Natural-Language Temporal Grounding in Hour-Long Videos is a Search Problem: A Benchmark and Empirical Decomposition
- 将长视频定位视为搜索任务，提出新基准测试框架
- 95%失败源于搜索能力不足，非事件识别能力缺陷
- 检索+定位混合模型性能提升6.7倍，适合长视频应用
时间定位——为自然语言查询返回视频中的时间段区间 $[t_s, t_e]$——是长视频的语言接口，但以往研究多集中于短视频；小时级自然语言定位的动态机制仍待探索。我们认为，在小时尺度下，核心瓶颈是搜索而非识别：视频大模型的限制不在于定位附近事件，而在于面对自然语言查询时，从长视频中搜寻相关片段的能力。为验证此观点，我们发布 ExtremeWhenBench，首个公开的小时级定位基准（2,273个查询，194段视频，平均时长75.7分钟，最长达9小时），采用开放形式查询分布。所有开放视频大模型表现崩溃，而帧级检索基线表现更优；故障分类显示85%的失败源于搜索问题；采用‘检索-再定位’混合架构可实现6.7倍性能提升，类比开放域问答中的‘检索-阅读’范式。
原文摘要 · Abstract (English)
Temporal grounding--returning the interval $[t_s, t_e]$ for a natural-language query over a video--is the language interface to long-form video, yet has been studied on short videos; the dynamics of hour-scale natural-language grounding remain underexplored. We take the position that at hour-scale, the binding constraint is search, not recognition: Video-LLMs are bottlenecked not by localizing a nearby event, but--given a natural-language query--by searching for the relevant region of a long video. To test this, we release ExtremeWhenBench, the first open hour-scale grounding benchmark (2,273 queries over 194 videos, mean 75.7 min, max 9 hr) with an open-form query distribution. Every open Video-LLM collapses while a frame-level retrieval baseline outperforms them; a failure taxonomy attributes 85% of failures to search; and a retrieve-then-ground hybrid recovers 6.7x over the monolithic Video-LLM--mirroring retrieve-then-read in open-domain QA.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。