AI Paper Atlas

Research map / Video / Video generation

Beyond Distribution Matching: Self-Supervised Representation Forcing for Few-Step Video Generation

Junyi Chen, Haiyu Zhang, ZHOUJIE FU, Tengfei Wang, Chunchao Guo

NeurIPS 2026 · Video › Video generation · approach: Efficient video diffusion models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a self-supervised representation forcing objective added to Distribution Matching Distillation to improve few-step video diffusion generation.

Problem

Distribution Matching Distillation (DMD) for few-step video diffusion models optimizes only in output space and gives no signal shaping the student's internal representations.

Approach

Adds a self-supervised representation-level training signal to DMD so that the student's internal features are explicitly shaped, rather than relying solely on output-space distribution matching; applicable to both bidirectional and causal student architectures.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add self-supervised representation loss (EMA teacher, heterogeneous noise) to DMD (ablation)DMD on HunyuanVideo-1.5 · 3D re-rendering probe, 120 static-scene promptsPSNR: 20.64 -> 21.24Improved
add representation loss to DMD (ablation)DMD on HunyuanVideo-1.5 · VBench Temporal FlickeringTemporal Flickering: 96.39 -> 96.86Improved
add representation loss to DMD (ablation)DMD on HunyuanVideo-1.5 · in-scene text rendering (qualitative)text legibilityImproved
add representation loss to causal Self-Forcing-style distillationCausal-Forcing student, Wan2.1-1.3B · VBench TotalVBench Total score: 84.04 -> 84.15Improved
add representation loss to bidirectional DMDWan2.1-1.3B bidirectional student · VisionReward (100 prompts)VisionReward score: 5.310 -> 7.060 (vs CausVid)Improved
set masking ratio max to 0.4 (bidirectional) or 0.2 (causal) instead of other values in 0-1.0 (ablation)SF2 masking ratio rho_max · 400-prompt motion-rich validation setVBench-style scoreImproved
use space-only mask sampling instead of agnostic per-token masking (ablation)SF2 causal student · 400-prompt motion-rich validation setDynamic Degree vs Background ConsistencyMixed
add SF2 to WorldPlay distillation pipelineWorldPlay · interactive camera/action followingtranslation, rotation and action accuracyImproved
add SF2 branch to DMD trainingDMD training cycle · per-cycle wall-clock% of cycle: SF2 branch 6.2 of 100.0 percent of cycleNo clear effect
do not reset projector learning rate multiplier after warmup (ablation)SF2 projector warmup · training stabilitytraining stabilityHurt
do not stop gradient on teacher path (ablation)SF2 EMA teacher · training stabilityEMA update stabilityHurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Surprising findings

Experimental setup

Baselines: Distribution Matching Distillation

Builds on: Distribution Matching Distillation

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy