arXiv：cs.LG· Hansen Jin Lillemark, Alex Rojas, Zachary Novack, Runqian Wang, Yilun Du, Yian Ma, Taylor Berg-Kirkpatrick, Rose Yu·· 16 小时前
Equilibrium Forcing：无需噪声条件训练的自适应视频生成框架
Equilibrium Forcing: Adaptive Video Generation Without Noise Conditioning
AI 导读
研究者提出 Equilibrium Forcing（EqF），一种不做噪声水平条件化的视频去噪生成框架，将去噪场学习与采样解耦，使推理阶段能以闭环方式根据样本反馈自适应调整。该框架在具有挑战性的自回归视频生成基准上提升了视频质量与一致性，性能超过标准的噪声水平条件化去噪视频方法。论文已被 NeurIPS 2026 接收。
正文
Abstract:Standard autoregressive video generation algorithms based on Diffusion and Flow Matching rely on rigid training objectives and static sampling schedules, limiting inference procedures from adapting to the data. We introduce Equilibrium Forcing (EqF), a simplified framework for video denoising generative models without noise level conditioning. EqF pioneers modular training- and inference-time designs for noise-unconditional generation that decouple learning the denoising field from sampling. This flexibility allows for inference-time algorithms that operate in a closed loop by adapting to feedback from the sample, improving video quality and consistency on challenging autoregressive video generation benchmarks. Extensive analysis elucidates exactly how removing the noise level conditioning enables EqF's data-dependent inference properties to surpass the performance of standard noise level-conditional denoising video methods.
| Comments: | Published at NeurIPS 2026. Project page: this https URL |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2608.14706 [cs.CV] |
| (or arXiv:2608.14706v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2608.14706 arXiv-issued DOI via DataCite |
Submission history
From: Hansen Lillemark [view email]
[v1]
Tue, 11 Aug 2026 00:02:32 UTC (9,033 KB)
[v2]
Wed, 14 Oct 2026 05:37:12 UTC (9,033 KB)
来源：arXiv：cs.LG · arxiv.org