Newest papers

Ask About the Right Moment: Training-Free Object Identity Reasoning for Video-Language Models

Di Yang, Xiaochang Li, Muyang Liu, Yufei Wang, Yanhai Xiong · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31471 · cited by 0 works (OpenAlex) · open-access PDF

Arabic

سجلت العديد من الأسئلة المتعلقة بالفيديو على وجه الخصوص عن نفس السيناريو. أين ينتهي الكرة في وقت مبكر بعد أن تكون الكوبية مجهولة؟ ما هي القائمة؟ كم مرة تم تغطية السيناريو؟ نموذج اللغة الفوتوغرافية (Video-LMs) تعرف نفس السيناريوهات في السيناريوهات الفوتوغرافية، ولكن لا تبقيها منفصلة مع الوقت. في الفيديوهات الفوتوغرافية الحقيقية، يتم وضع الأدلة على ذلك على النماذج من 8B إلى 235B، وتتبعها بعد ذلك من قبل مرشح. \emph{Track، Then Bind} تطبق هذه المبادئ دون التدريب على أدوات التدريب الثلاثة. يطلب المرشح الفوتوغرافية الم

Machine translation (Meta M2M100); the English original follows.

Abstract

Many questions about a video concern one physical object. Where does the ball end up after the cups are shuffled? Which card is the queen? How often was the mug covered? Video-language models (Video-LMs) recognise such objects in single frames but do not keep them apart over time. On real shell-game videos, open models from 8B to 235B parameters, their thinking variants, a reinforcement-learned tool agent and three training-free tool agents all stay within ten points of chance. We separate semantic identity binding from temporal identity maintenance. An object's identity should be fixed at a moment when the evidence for it is visible, and carried afterwards by a tracker. \emph{Track, Then Bind} implements this principle without training. A promptable video tracker follows every candidate object and draws it with a track number. A frozen vision-language model answers a static question about the identifying evidence, which binds the target to a track. The answer is then derived from the tracked state in text, without asking the model to reinterpret the original video during read-out. A single configuration covers all benchmarks. With Qwen3-VL-32B it solves both VET-Bench tasks, where a fine-tuned trajectory model reaches 92\% and 90\% and the same VLM voting over 16 samples stays at 30--32\%. It answers 55.4\% of the hardest Perception Test cup games, where the best published model answers 41.5\%, and, using track-derived timelines, 48.6\% of TOC-Bench questions, against 37.1\% for the same model answering directly. Drawing the same tracks on the video and asking the question directly falls 12--19 points short of the agent on real cup games.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Ask price, Computer science, Identity (music), Artificial intelligence, Object (grammar), Perception

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange