AI Paper Atlas

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

Alignment-Guided Score Matching for Text-to-Image Alignment in Diffusion Models

Jaa-Yeon Lee, Yeobin Hong, Taesung Kwon, Jong Chul Ye

ICML 2026 · Image generation › Text-to-image diffusion · approach: Inference-time T2I improvement · 0 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 AGSM soft tokensSD1.5 · COCO val5KImageReward (x100): 17.72 -> 34.50Improved
add AGSM soft tokensSD1.5 · COCO val5KFID: 24.59 -> 25.94Hurt
add AGSM soft tokensSDXL · COCO val5KImageReward (x100): 75.06 -> 84.22Improved
add AGSM soft tokensSD3 · COCO val5KImageReward (x100): 94.27 -> 103.3Improved
add AGSM soft tokensSD3 · COCO val5KFID: 31.59 -> 34.08Hurt
add AGSM soft tokensSD3 · GenEvalcounting accuracy: 0.56 -> 0.64Improved
use negative samples D- with separate soft tokens (psi+, psi-) vs positive only (ablation)AGSM · COCO val5KImageReward (x100): 94.79 -> 103.3Improved
use separate positive/negative tokens instead of shared tokens (ablation)AGSM · COCO val5KImageReward (x100): 47.33 -> 103.3Improved
use separate positive/negative tokens instead of shared tokens (ablation)AGSM · COCO val5KFID: 31.20 -> 34.08Hurt
sample with positive tokens only instead of negative tokens for unconditional (ablation)AGSM · COCO val5KImageReward (x100): 84.53 -> 103.3Improved
use Plackett-Luce multi-candidate loss instead of Bradley-Terry (ablation)AGSM · COCO val5KImageReward (x100): 29.67 -> 34.50Improved
use Plackett-Luce loss instead of Bradley-Terry (ablation)AGSM · COCO val5KFID: 24.76 -> 25.94Hurt

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