Temporal Chain of Thought: Long-Video Understanding by Thinking in Frames

• URL: http://arxiv.org/abs/2507.02001v1
• Date: Tue, 01 Jul 2025 18:39:26 GMT
• Title: Temporal Chain of Thought: Long-Video Understanding by Thinking in Frames
• Authors: Anurag Arnab, Ahmet Iscen, Mathilde Caron, Alireza Fathi, Cordelia Schmid,
• Abstract summary: We present Temporal Chain of Thought, an inference strategy for video question-answering.<n>We use the VLM itself to iteratively identify and extract the most relevant frames from the video.<n>We demonstrate how leveraging more computation at inference-time to select the most relevant context leads to improvements in accuracy.
• Score: 70.93346841539626
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Despite recent advances in Vision-Language Models (VLMs), long-video understanding remains a challenging problem. Although state-of-the-art long-context VLMs can process around 1000 input frames, they still struggle to effectively leverage this sequence length, and succumb to irrelevant distractors within the context window. We present Temporal Chain of Thought, an inference strategy for video question-answering that curates the model's input context. We use the VLM itself to iteratively identify and extract the most relevant frames from the video, which are then used for answering. We demonstrate how leveraging more computation at inference-time to select the most relevant context leads to improvements in accuracy, in agreement with recent work on inference-time scaling of LLMs. Moreover, we achieve state-of-the-art results on 4 diverse video question-answering datasets, showing consistent improvements with 3 different VLMs. In particular, our method shines on longer videos which would not otherwise fit within the model's context window: On longer videos of more than 1 hour on LVBench, our approach using a context window of 32K outperforms the same VLM using standard inference with a 700K context window by 2.8 points.

Related papers

• Moment Sampling in Video LLMs for Long-Form Video QA [22.638644170177013]
  "moment sampling" is a model-agnostic approach that enables the model to select the most relevant frames according to the context of the question.<n>By focusing on the frames most pertinent to the given question, our method enhances long-form VideoQA performance in Video LLMs.
  arXiv  Detail & Related papers  (2025-06-18T03:23:56Z)
• Threading Keyframe with Narratives: MLLMs as Strong Long Video Comprehenders [62.58375366359421]
  Multimodal Large Language Models (MLLMs) for long video understanding remains a challenging problem.<n>Traditional uniform sampling leads to selection of irrelevant content.<n>Post-training MLLMs on thousands of frames imposes a substantial computational burden.<n>We propose threadings with narratives (Nar-KFC) to facilitate effective and efficient long video perception.
  arXiv  Detail & Related papers  (2025-05-30T03:04:28Z)
• Multimodal Long Video Modeling Based on Temporal Dynamic Context [13.979661295432964]
  We propose a dynamic long video encoding method utilizing the temporal relationship between frames, named Temporal Dynamic Context (TDC)<n>We segment the video into semantically consistent scenes based on inter-frame similarities, then encode each frame into tokens using visual-audio encoders.<n>To handle extremely long videos, we propose a training-free chain-of-thought strategy that progressively extracts answers from multiple video segments.
  arXiv  Detail & Related papers  (2025-04-14T17:34:06Z)
• BOLT: Boost Large Vision-Language Model Without Training for Long-form Video Understanding [51.49345400300556]
  Large video-language models (VLMs) have demonstrated promising progress in various video understanding tasks.<n>Traditional approaches, such as uniform frame sampling, often inevitably allocate resources to irrelevant content.<n>We introduce BOLT, a method to BOost Large VLMs without additional Training through a comprehensive study of frame selection strategies.
  arXiv  Detail & Related papers  (2025-03-27T13:18:40Z)
• LongVU: Spatiotemporal Adaptive Compression for Long Video-Language Understanding [65.46303012350207]
  LongVU is an adaptive compression mechanism that reduces the number of video tokens while preserving visual details of long videos. We leverage DINOv2 features to remove redundant frames that exhibit high similarity. We perform spatial token reduction across frames based on their temporal dependencies.
  arXiv  Detail & Related papers  (2024-10-22T21:21:37Z)
• Visual Context Window Extension: A New Perspective for Long Video Understanding [45.134271969594614]
  We tackle the challenge of long video understanding from the perspective of context windows. We propose to adapt LMMs for long video understanding tasks by extending the visual context window. Our method consistently improves the performance as the number of video frames increases.
  arXiv  Detail & Related papers  (2024-09-30T07:25:16Z)
• Long Context Transfer from Language to Vision [74.78422371545716]
  Video sequences offer valuable temporal information, but existing large multimodal models (LMMs) fall short in understanding extremely long videos. In this paper, we approach this problem from the perspective of the language model. By simply extrapolating the context length of the language backbone, we enable LMMs to comprehend orders of magnitude more visual tokens without any video training.
  arXiv  Detail & Related papers  (2024-06-24T17:58:06Z)
• Too Many Frames, Not All Useful: Efficient Strategies for Long-Form Video QA [40.21221568678641]
  Long-form videos that span across wide temporal intervals are highly information redundant.<n>All information necessary to generate a correct response can often be contained within a small subset of frames.<n>Recent literature explore use of large language models in LVQA benchmarks, achieving exceptional performance.
  arXiv  Detail & Related papers  (2024-06-13T17:59:16Z)
• Retrieval meets Long Context Large Language Models [59.431200671427064]
  Extending context window of large language models (LLMs) is getting popular recently. Retrieval-augmentation versus long context window, which one is better for downstream tasks? Can both methods be combined to get the best of both worlds? Our best model, retrieval-augmented Llama2-70B with 32K context window, outperforms GPT-3.5-turbo-16k and Davinci003 in terms of average score on nine long context tasks.
  arXiv  Detail & Related papers  (2023-10-04T17:59:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.