PhyRubric：基于评分细则的 T2V 生成物理常识诊断评估框架

View PDF HTML (experimental)

Abstract:As text-to-video (T2V) models are increasingly positioned as world simulators, it is essential to evaluate whether their videos conform to physical commonsense. However, existing benchmarks rely on annotators who rate each video with a single score based on intuition, yielding ratings that are subjective and do not tell which physical principle is violated. We introduce PhyRubric, a rubric-grounded diagnostic framework for evaluating physical commonsense in T2V generation. For each caption, PhyRubric constructs a caption-specific rubric before any video is seen, decomposing the caption into semantic and physical items, and derives scores from item-level verdicts grounded in timestamped evidence. We instantiate this framework as PhyRubric-Bench, which spans 34 categories across five physical domains, and scale it up to PhyRubric-Train with 3,788 videos, on which we fine-tune the open-source PhyRubric-Judge. Empirical results across ten T2V models show that physical commonsense remains far from solved, and that PhyRubric-Judge correlates more closely with human judgments than existing open-source judges while matching the frontier proprietary Gemini 4 Argon. Further analysis provides actionable insights for T2V models.

Comments:	34 pages, 19 figures
Subjects:	Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Cite as:	arXiv:2611.13576 [cs.CV]
	(or arXiv:2611.13576v1 [cs.CV] for this version)
	https://doi.org/10.48550/arXiv.2611.13576 arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Huiyi Chen [view email]
[v1] Tue, 10 Nov 2026 04:50:38 UTC (8,069 KB)