AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

ViPO: Visual Preference Optimization at Scale

Ming Li, Jie Wu, Justin Cui, Xiaojie Li, Rui Wang, Chen Chen

ICLR 2026 · Image generation › Text-to-image diffusion · approach: Preference optimization for diffusion · 1 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
add polynomial confidence term (alpha=8) to Diffusion-DPODiffusion-DPO SD1.5 · GenEval (Pick-a-Pic V2 training)overall: 43.00 -> 49.87Improved
add polynomial confidence term to Diffusion-DPODiffusion-DPO SDXL · GenEval (Pick-a-Pic V2 training)overall: 58.02 -> 60.34Improved
add polynomial confidence term to Diffusion-DPODiffusion-DPO SD1.5 · Pick-a-Pic V2 testImageReward: 0.297 -> 0.679Improved
add polynomial confidence term to Diffusion-DPODiffusion-DPO SD1.5 · Pick-a-Pic V2 testHPSv2.1: 26.12 -> 28.30Improved
add polynomial confidence term vs Diffusion-KTODiffusion-KTO SD1.5 · Pick-a-Pic V2 testPickScore: 21.06 -> 21.48Improved
use negative alpha on over-simple shuffled-loser dataset (ablation)Poly-DPO SD1.5 · synthetic over-simple preference datasetHPSv2.1 (figure)Improved
vary alpha on high-quality ViPO-Image-1M (ablation)Poly-DPO SD1.5 (SFT-initialized) · ViPO-Image-1MHPSv2.1 (figure)No clear effect
add SFT then Poly-DPO on ViPO-Image-1MSD1.5 · GenEvaloverall: 0.42 -> 0.54Improved
add Poly-DPO after SFT on ViPO-Image-1MSD3.5-Medium SFT · DPG-Benchoverall: 84.24 -> 87.71Improved
add Poly-DPO after SFT on ViPO-Image-1MFLUX.1-dev SFT · CVTG-2Kword accuracy: 0.2126 -> 0.6859Improved
add Poly-DPO after SFT on ViPO-Image-1MSD3.5-Medium SFT · GPT-4o human qualityaccuracy: 77.50 -> 85.25Improved
add SFT stage before Poly-DPO (ablation)Poly-DPO SD1.5 · SD1.5 SFT ablationImageReward: 0.6391 -> 1.0161Improved

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.

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