Uncertainty DMD: Restoring Diversity in Few-Step Autoregressive Video Distillation

• URL: http://arxiv.org/abs/2609.11265v1
• Date: Thu, 10 Sep 2026 08:59:47 GMT
• Title: Uncertainty DMD: Restoring Diversity in Few-Step Autoregressive Video Distillation
• Abstract summary: Few-step distillation improves the efficiency of autoregressive (AR) video generation, but often causes diversity collapse.<n>We analyze this degradation in Distribution Matching Distillation (DMD)-distilled AR video generators and find that, in the autoregressive setting, it takes the form of a structured uncertainty collapse.
• Score: 52.34982929108336
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Few-step distillation improves the efficiency of autoregressive (AR) video generation, but often causes diversity collapse: under the same prompt, different noise samples tend to produce highly similar videos with weakened motion dynamics. We analyze this degradation in Distribution Matching Distillation (DMD)-distilled AR video generators and find that, in the autoregressive setting, it takes the form of a structured uncertainty collapse: the mode-seeking bias of DMD maps different noise samples to nearly identical first chunks, and the deterministic AR cache then propagates this collapsed state to all subsequent chunks, turning a local loss of stochasticity at the rollout root into a global suppression of temporal variation. Based on this analysis, we propose Uncertainty DMD, a simple uncertainty-injection framework that restores stochasticity at two key stages of AR generation: a timestep perturbation for the first chunk to increase first-chunk diversity, and a stochastic cache-writing mechanism for later chunks to preserve uncertainty in autoregressive conditioning. The method requires no architectural changes and introduces only lightweight perturbation operations. The same perturbation mechanisms are used during both training and inference. Experiments show that Uncertainty DMD consistently improves diversity and motion dynamics while maintaining comparable per-sample visual quality.

Related papers

• AAD-1: Asymmetric Adversarial Distillation for One-Step Autoregressive Video Generation [66.55343616132005]
  AAD-1 is an Asymmetric Adversarial Distillation framework for one-step autoregressive image-to-video generation.<n>AAD-1 achieves state-of-the-art performance in one-step autoregressive video generation.
  arXiv  Detail & Related papers  (2026-06-02T17:55:30Z)
• Adaptive Video Distillation: Mitigating Oversaturation and Temporal Collapse in Few-Step Generation [43.89162138967428]
  We propose a novel distillation framework tailored specifically for video diffusion models.<n>Its core innovations include: (1) an adaptive regression loss that dynamically adjusts spatial supervision weights to prevent artifacts arising from excessive distribution shifts; (2) a temporal regularization loss that promotes smooth and physically plausible sampling trajectories; and (3) an inference-time frame strategy that reduces sampling overhead while preserving perceptual quality.<n>Experiments and ablation studies on the VBench and VBench2 benchmarks demonstrate that our method achieves stable few-step video synthesis, significantly enhancing perceptual fidelity and motion realism.
  arXiv  Detail & Related papers  (2026-03-23T11:54:33Z)
• Diversity-Preserved Distribution Matching Distillation for Fast Visual Synthesis [33.341277146439275]
  Distribution matching distillation (DMD) aligns a multi-step generator with its few-step counterpart to enable high-quality generation under low inference cost.<n>We propose a role-separated distillation framework that explicitly disentangles the roles of distilled steps.<n>We term this approach Diversity-Preserved DMD (DP-DMD), which, despite its simplicity, preserves sample diversity while maintaining visual quality on par with state-of-the-art methods in extensive text-to-image experiments.
  arXiv  Detail & Related papers  (2026-02-03T05:45:25Z)
• Improving Temporal Consistency and Fidelity at Inference-time in Perceptual Video Restoration by Zero-shot Image-based Diffusion Models [5.61537470581101]
  We address the challenge of improving temporal coherence in video restoration using zero-shot image-based diffusion models.<n>We propose two complementary inference-time strategies: Perceptual Straightening Guidance (PSG) and Ensemble Sampling (MPES)
  arXiv  Detail & Related papers  (2025-10-29T11:40:06Z)
• Large Scale Diffusion Distillation via Score-Regularized Continuous-Time Consistency [60.74505433956616]
  continuous-time consistency model (sCM) is theoretically principled and empirically powerful for accelerating academic-scale diffusion.<n>We first develop a parallelism-compatible FlashAttention-2 JVP kernel, enabling sCM training on models with over 10 billion parameters and high-dimensional video tasks.<n>We propose the score-regularized continuous-time consistency model (rCM), which incorporates score distillation as a long-skip regularizer.
  arXiv  Detail & Related papers  (2025-10-09T16:45:30Z)
• Restoration Score Distillation: From Corrupted Diffusion Pretraining to One-Step High-Quality Generation [82.39763984380625]
  We propose textitRestoration Score Distillation (RSD), a principled generalization of Denoising Score Distillation (DSD)<n>RSD accommodates a broader range of corruption types, such as blurred, incomplete, or low-resolution images.<n>It consistently surpasses its teacher model across diverse restoration tasks on both natural and scientific datasets.
  arXiv  Detail & Related papers  (2025-05-19T17:21:03Z)
• DiffusionAD: Norm-guided One-step Denoising Diffusion for Anomaly Detection [80.20339155618612]
  DiffusionAD is a novel anomaly detection pipeline comprising a reconstruction sub-network and a segmentation sub-network.<n>A rapid one-step denoising paradigm achieves hundreds of times acceleration while preserving comparable reconstruction quality.<n>Considering the diversity in the manifestation of anomalies, we propose a norm-guided paradigm to integrate the benefits of multiple noise scales.
  arXiv  Detail & Related papers  (2023-03-15T16:14:06Z)
• DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration [66.01846902242355]
  Blind face restoration usually synthesizes degraded low-quality data with a pre-defined degradation model for training. It is expensive and infeasible to include every type of degradation to cover real-world cases in the training data. We propose Robust Degradation Remover (DR2) to first transform the degraded image to a coarse but degradation-invariant prediction, then employ an enhancement module to restore the coarse prediction to a high-quality image.
  arXiv  Detail & Related papers  (2023-03-13T06:05:18Z)

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.