用时间逻辑解析视频问答，准确率提升24.5个百分点。
TLG: Temporal-Logic Grounding for Video Question Answering via Source-Annotation Reconstruction and Category-Targeted Reasoning
- 从公开标注重建动作时间线，解析问题为逻辑程序执行
- 在无标注时回退到强视觉语言模型，关键类别交由前沿推理模型处理
- 实验证明真实标注比更大模型更能提升性能，适合时序推理研究者
TimeLogic 挑战要求对视频进行形式化的时间逻辑推理，包含16种操作符（如before、after、until、since、always、co-occur、ordering等），以布尔或四选一选择题形式呈现。当前端到端视频语言模型（VLMs）在此任务上仅达到接近随机水平，因其将视频视为帧的集合，无法定位动作发生时间。本文提出TLG（Temporal-Logic Grounding），一个三层系统：(i) 从基准生成所用的公开源数据集标注中重构视频动作时间线，将每个问题解析为时间逻辑程序并确定性执行；(ii) 在无标注情况下回退至强大开源VLM；(iii) 将VLM表现最弱的问题类别路由至前沿推理模型。该方法将测试准确率从46.9%的基线提升至71.37%，绝对提升24.5个百分点，接近排行榜榜首。我们进行了详尽消融实验，包括三种基于模型的时间线重建变体，均表现低于整体式VLM，凸显时间定位是不可简化的瓶颈，并表明真实标注而非更大模型才是准确率提升的关键。
原文摘要 · Abstract (English)
The TimeLogic Challenge evaluates formal temporal-logic reasoning over video - 16 operators (before, after, until, since, always, co-occur, ordering, ...) in boolean and 4-way multiple-choice form. End-to-end video-language models (VLMs) hover near chance on this task because they treat video as a bag of frames and cannot localize when actions occur. We present TLG (Temporal-Logic Grounding), a three-tier system that (i) reconstructs each video's action timeline from the public source-dataset annotations the benchmark was generated from, parses every question into a temporal-logic program, and executes it deterministically; (ii) falls back to a strong open VLM where no annotation exists; and (iii) routes only the question categories where the VLM is empirically weakest to a frontier reasoning model. TLG raises test accuracy from a 46.9% VLM baseline to 71.37%, a +24.5 absolute gain, reaching within 3 points of the leaderboard top. We report extensive ablations, including three model-based timeline-reconstruction variants that all underperform a holistic VLM, isolating temporal grounding as the irreducible bottleneck and showing that real annotations - not larger models - drive accuracy.
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。