AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Seeing What Matters: Visual Preference Policy Optimization for Visual Generation

Ziqi Ni, Yuanzhi Liang, Rui Li, Yi Zhou, Haibin Huang, Chi Zhang, Xuelong Li

CVPR 2026 · Reinforcement learning › RL from feedback · approach: Reward modeling versus direct optimization · 5 citations (OpenAlex)

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
replace scalar GRPO advantage with perceptual allocation-map advantages (DINO)DanceGRPO on FLUX.1-dev · HPD testHPSv2.1: 0.3203 -> 0.3321Improved
add ViPO to DanceGRPOFLUX.1-dev · HPD testImageReward (OOD): 1.0392 -> 1.1883Improved
DanceGRPO training aloneFLUX.1-dev · HPD testImageReward / PickScore: 1.1495 -> 1.0392 (ImageReward); 22.7038 -> 22.5962 (PickScore)Hurt
use SAM backbone instead of DINO in PSM (ablation)ViPO · HPD testHPSv2.1: 0.3321 -> 0.3219Hurt
use ResNet backbone in PSM instead of DINO (ablation)ViPO · HPD testPickScore: 22.8305 -> 22.8492Improved
apply ViPO to videoWan2.1-T2V-14B DanceGRPO · VidProM / VBenchVBench Total: 80.84 -> 81.70Improved
add ViPO to DanceGRPO on videoWan2.1 · VBenchQuality: 69.68 -> 72.59Improved
use semantically structured allocation map instead of uniform all-ones map (ablation)ViPO · HPD testHPSv2.1: 0.3043 -> 0.3321Improved
apply allocation map to advantage rather than reward (ablation)ViPO · HPD testImageReward: 1.0058 -> 1.1883Improved
variance-weighted rather than average aggregation of components (ablation)ViPO · HPD testImageReward: 1.1318 -> 1.1883Improved
vary retained PCA components K from 1 to 5 (ablation)ViPO · HPD testHPSv2.1/PickScore/ImageReward: K=3: 0.3321/22.8305/1.1883; K=4 HPS 0.3337Mixed
add Gaussian smoothing sigma=1 vs none (ablation)ViPO · HPD testImageReward: 1.1618 -> 1.1883Improved

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