← 研踪 PaperTrailarXiv:2609.30981 · v2
RevisitQA: Short Videos Are Not Easy under Dense Revisits
首次提交 2026-09-25 16:27
版本更新 2026-11-19 13:42
雷达观测 2026-11-20 19:58
摘要
Video question answering benchmarks have increasingly focused on longer recordings, yet short videos can also be challenging when scenes are revisited frequently. As a camera repeatedly leaves and returns to the same workspace, objects may change out of view, making earlier observations outdated or misleading. Understanding such videos requires models to track evolving states across repeated visits rather than simply retrieve past observations. To study this challenge, we introduce RevisitQA, a benchmark containing 2,577 questions across 622 episodes with a median duration of 44.8 seconds, drawn from five real-world egocentric domains and two simulated environments. RevisitQA covers six question types and evaluates models at specific query times, requiring them to answer from the observed history or abstain when it is insufficient. Across 15 offline and seven streaming video LLMs, the median model correctly answers only 55.5% of answerable questions and abstains on just 30.3% of unanswerable ones. Removing earlier frames further reveals that historical observations help answer questions but can also make models less likely to abstain when answers remain unresolved. These findings highlight dense episodic revisits as an important and unresolved challenge for video question answering. Code is available at https://github.com/siruzhong/RevisitQA.
作者备注:Renamed from STORM-Bench to RevisitQA. This revision updates the benchmark questions, evaluation metrics, and model experiments
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-19 13:42 · 最新观测版本
RevisitQA: Short Videos Are Not Easy under Dense Revisits
Video question answering benchmarks have increasingly focused on longer recordings, yet short videos can also be challenging when scenes are revisited frequently. As a camera repeatedly leaves and returns to the same workspace, objects may change out of view, making earlier observations outdated or misleading. Understanding such videos requires models to track evolving states across repeated visits rather than simply retrieve past observations. To study this challenge, we introduce RevisitQA, a benchmark containing 2,577 questions across 622 episodes with a median duration of 44.8 seconds, drawn from five real-world egocentric domains and two simulated environments. RevisitQA covers six question types and evaluates models at specific query times, requiring them to answer from the observed history or abstain when it is insufficient. Across 15 offline and seven streaming video LLMs, the median model correctly answers only 55.5% of answerable questions and abstains on just 30.3% of unanswerable ones. Removing earlier frames further reveals that historical observations help answer questions but can also make models less likely to abstain when answers remain unresolved. These findings highlight dense episodic revisits as an important and unresolved challenge for video question answering. Code is available at https://github.com/siruzhong/RevisitQA.
查看此版本解读 →查看此版本原文 ↗