AI Paper Atlas

Research map / Video / Video generation

Mode Seeking meets Mean Seeking for Fast Long Video Generation

Shengqu Cai, Weili Nie, Chao Liu, Julius Berner, Lvmin Zhang, Nanye Ma, Hansheng Chen, Maneesh Agrawala, Leonidas Guibas, Gordon Wetzstein, Arash Vahdat

ICML 2026 · Video › Video generation · approach: Autoregressive long video generation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use decoupled DDT dual heads instead of a single velocity head (ablation)Ours · Ablation (30s)quality: 0.5298 -> 0.6359Improved
add sliding-window teacher DMD alignment (ablation)Ours · Ablation (30s)quality: 0.6075 -> 0.6359Improved
add long-video SFT flow-matching anchor (ablation)Ours · Ablation (30s)consistency: 0.9579 -> 0.9615Improved
add long-video SFT flow-matching anchor (ablation)Ours · Ablation (30s)motion: 0.9690 -> 0.9685Mixed
use mode-seeking teacher alignment instead of long-context SFTOurs vs Long-context SFT · 30-second VBench-Longaesthetic quality: 0.4973 -> 0.5735Improved
train on mixed-length videos instead of long videos onlyMixed-length SFT vs Long-context SFT · 30-second VBench-Longaesthetic quality: 0.4973 -> 0.5467Improved
use ours instead of InfinityRoPEOurs · 30-second VBench-Longdynamic degree: 0.7188 -> 0.9453Improved
use ours instead of rollout baselines at 60 secondsOurs vs LongLive · 60-second stress testdynamic degree: 0.8203 -> 0.9297Improved
use ours vs CausVid for subject consistencyOurs vs CausVid · 30-second VBench-Longsubject consistency: 0.9736 vs 0.9682Hurt
use ours instead of other methods in user studyOurs · User study 30soverall quality: 3.68 vs 3.23 (InfinityRoPE)Improved

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.

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