AI Paper Atlas

Research map / Video / Video generation

DynamicsBoost: Dynamic Plausible Video Generation via Annotation-Free Continuation Preference Optimization

Jiaxing Li, Jiepeng Wang, Junyao Gao, Yang Liu, Eric Li, Bo An, Hao-Xiang Guo

CVPR 2026 · Video › Video generation · approach: Compositional text-to-video diffusion

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
compute DPO loss only on continuation regions (2)-(3), excluding shared prefix (ablation)DynamicsBoost · VBenchOverall Consistency: 22.15 -> 25.64Improved
restrict loss to Region (3) only instead of Regions (2)-(3) (ablation)DynamicsBoost · VBenchDynamic Degree: 44.92 -> 38.21Hurt
use randomized negative and positive continuation lengths instead of fixed N2=N (ablation)DynamicsBoost · VBenchOverall Consistency: 25.44 -> 25.64Improved
use pure T2V (N1=0) as negative samples (ablation)DynamicsBoost · VBenchOverall Consistency: 25.64 -> 25.31Hurt
fix N1=0, N2=N for continuation lengths (ablation)DynamicsBoost · VBenchDynamic Degree: 44.92 -> 34.32Hurt
add more reference frames in continuation (ablation)continuation model · VBench and VideoRewardAesthetic Quality: 36.10 -> 46.64Improved
add pure T2V (0 reference frames) as the lowest-ordered sample (ablation)continuation preference ordering · VBenchAesthetic Quality: 36.10 -> 44.83Hurt
add learnable continuation task prompt (ablation)continuation model (4/13) · VBenchAesthetic Quality: 40.21 -> 43.81Improved
add timestep masking (ablation)continuation model (4/13) · VBenchImaging Quality: 49.29 -> 57.98Improved
apply annotation-free continuation preference pairs vs Flow-DPO with VideoReward labelsFlow-DPO · VBenchDynamic Degree: 31.22 -> 44.92Improved
apply SFT cold start on small datasetpretrained model · VBenchMotion Smoothness: 98.81 -> 96.84Hurt
compare proposed method with Flow-DenseDPO on imaging qualityDynamicsBoost · VBenchImaging Quality: 67.91 -> 66.81Mixed

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