arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.14294cs.CVcs.AIcs.LGcs.RO
收藏

在扩散噪声中寻求物理规律

Seeking Physics in Diffusion Noise

Brown University(布朗大学);University of Edinburgh(爱丁堡大学);MIT(麻省理工学院)

Chujun Tang, Lei Zhong, Fangqiang Ding

AI总结:

研究通过分析预训练扩散变换器的中间去噪表示,发现物理合理与不合理视频在中层特征空间部分可分离,提出渐进轨迹选择策略提升物理一致性并降低推理成本。

摘要:

视频扩散模型是否编码了预测物理合理性的信号?我们探测预训练扩散变换器(DiT)的中间去噪表示,发现物理合理和不合理视频在不同噪声水平下的中层特征空间部分可分离。这种分离性无法完全归因于视觉质量和生成器身份,表明冻结的DiT特征中存在可恢复的物理相关线索。基于此观察,我们引入渐进轨迹选择,一种推理时策略,通过使用轻量级物理验证器在几个中间检查点评分并修剪低分候选轨迹。在PhyGenBench上的广泛实验表明,我们的方法在提高物理一致性的同时减少推理成本,实现与Best-of-K采样相当的结果,但使用显著更少的去噪步骤。

英文摘要:

Do video diffusion models encode signals predictive of physical plausibility? We probe intermediate denoising representations of pretrained Diffusion Transformers (DiTs) and find that physically plausible and implausible videos are partially separable in mid-layer feature space, even at high noise levels. Within-source and perceptual-quality controls suggest that this signal is not fully explained by generator identity or generic visual quality. We distill the signal into a lightweight, backbone-specific physics verifier trained on frozen features and use it in two complementary inference-time mechanisms under a fixed multi-trajectory budget: progressive trajectory selection, which scores trajectories at intermediate checkpoints and prunes weak candidates early, and reward-gradient guidance, which steers surviving trajectories by backpropagating through only the first few DiT blocks. Experiments on PhyGenBench and Physics-IQ across CogVideoX-2B/5B and Wan 2.1-14B show that progressive selection matches verifier-based Best-of-4 on CogVideoX-2B while reducing wall-clock inference time by 37%, whereas reward-gradient guidance substantially improves physical consistency on CogVideoX-5B, all without fine-tuning the video generator.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑