{"slug": "temporal-backtracking-search-for-test-time-generative-video-reasoning", "title": "Temporal Backtracking Search for Test-time Generative Video Reasoning", "summary": "Researchers introduce Temporal Backtracking Search (TBS), a test-time scaling method for generative video reasoning that searches over the temporal axis rather than denoising steps. TBS transforms video generation into an iterative generate-verify-restart loop, achieving 22.7% success in out-of-distribution settings where one-shot generation collapses to 0.7%. The method demonstrates that video models' local reasoning competence far exceeds single-shot rollouts.", "body_md": "arXiv:2606.13861v1 Announce Type: new\nAbstract: While test-time scaling has revolutionized reasoning in large language models, generative video reasoning remains bottlenecked by a single-shot paradigm. We demonstrate that searching over denoising steps cannot rescue logically flawed rollouts because spatial trajectories commit early in the diffusion process. Root-level Best-of-N (BoN) sampling is similarly inefficient: reasoning errors cluster early in the temporal axis, and resampling blindly discards verified upstream progress. To unlock effective test-time scaling for video models, we introduce Temporal Backtracking Search (TBS), which shifts the search space to the temporal axis. TBS transforms video generation into an iterative generate-verify-restart loop via three core mechanisms: (1) variable-K conditioning to resume generation from arbitrary clean prefixes; (2) temporal process verification to localize failures and extract valid restart anchors; and (3) prefix-based search to reallocate compute toward extending correct trajectories rather than root resampling. Across algorithmic, navigation, and robotics domains, TBS Pareto-dominates matched-budget BoN. In a strict out-of-distribution setting where one-shot generation collapses (0.7% for BoN), TBS achieves 22.7%, with every solved episode stemming from a restarted branch. Ultimately, TBS reveals that the local reasoning competence of video models far exceeds what single-shot rollouts indicate, providing a scalable test-time framework to unlock it.", "url": "https://wpnews.pro/news/temporal-backtracking-search-for-test-time-generative-video-reasoning", "canonical_source": "https://arxiv.org/abs/2606.13861", "published_at": "2026-06-15 04:00:00+00:00", "updated_at": "2026-06-15 04:13:20.811864+00:00", "lang": "en", "topics": ["artificial-intelligence", "generative-ai", "computer-vision", "ai-research"], "entities": ["Temporal Backtracking Search", "Best-of-N", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/temporal-backtracking-search-for-test-time-generative-video-reasoning", "markdown": "https://wpnews.pro/news/temporal-backtracking-search-for-test-time-generative-video-reasoning.md", "text": "https://wpnews.pro/news/temporal-backtracking-search-for-test-time-generative-video-reasoning.txt", "jsonld": "https://wpnews.pro/news/temporal-backtracking-search-for-test-time-generative-video-reasoning.jsonld"}}