AI Paper Atlas

Research map / Video / Video generation

Disen-Forcing: Disentangling Semantic Anchoring from Motion for Autoregressive Video Diffusion

Zhengyang Yu, Akio Hayakawa, Masato Ishii, Qingtao Yu, Xuesong Li, Takashi Shibuya, Yuki Mitsufuji, Jing Zhang

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 Disen-Forcing, a method for stabilizing long-horizon autoregressive video diffusion rollouts by disentangling semantic anchoring from motion, building on frame sink ideas.

Problem

Autoregressive video diffusion models degrade rapidly when the inference horizon exceeds the training horizon, due to compounding exposure bias in long rollouts.

Approach

Revisits the heuristic frame sink design (persistent KV cache of initial frames, inspired by attention sinks in streaming LLMs) and separates semantic anchoring from motion dynamics to stabilize long generation. Full details are not available from the abstract.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add disentangled context routing to Frame Sink baseline (ablation)Self-Forcing + Frame Sink · MovieGenBench 30sDynamic degree: 34.69 -> 52.56Improved
Add motion-disentangling regularization on top of routing (ablation)Self-Forcing + Frame Sink · MovieGenBench 30sDynamic degree: 52.56 -> 77.19Improved
Replace universal frame sink with Disen-ForcingSelf-Forcing · MovieGenBench 30sDynamic degree: 35.66 -> 73.12Improved
Use frame sink on Causal-ForcingCausal-Forcing · MovieGenBench 30sDynamic degree: 68.28 -> 17.75Hurt
Use frame sink at the 5s training horizonSelf-Forcing · VBench 5sDynamic degree: 65.00 -> 37.00Hurt
Replace identified anchor heads with random heads (ablation)Disen-Forcing · MovieGenBench 30sDynamic degree: 73.12 -> 51.16 (layer-matched), 34.69 (global)Hurt
Lower anchor threshold to p=70 (more anchors) (ablation)Disen-Forcing · MovieGenBench 30sDynamic degree: 77.19 -> 48.22Hurt
Raise anchor threshold to p=90 (fewer anchors) (ablation)Disen-Forcing · MovieGenBench 30simaging quality drift: 4.39 -> 4.77Hurt
Increase regularization weight from 0.002 to 0.005 (ablation)Disen-Forcing · MovieGenBench 30sDynamic degree and drift: 73.12 -> 75.97 dynamic; drift 4.03 -> 4.47Mixed
Remove VAE re-encoding of counterfactual context (ablation)Disen-Forcing · MovieGenBench 30simaging quality drift: 4.03 -> 4.85Hurt
Mask top-20% anchor heads at inference (ablation)Self-Forcing · MovieGenBenchconsistencyHurt
Apply regularization to four layers instead of layer 26 (ablation)Disen-Forcing · MovieGenBench 30sDynamic degree: 73.12 -> 78.38; drift 4.03 -> 4.83Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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.

Experimental setup

Builds on: frame sink attention sinks

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