AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Reinforcement Learning-driven Diffusion Models: Language, Vision, and Beyond

Zhiyuan Ma, Jiabao Wei, Tianyu Wang, Hanqing Wang, Jiadong Pan, Haowen Xu, Liangliang Zhao, Ruixun Liu, Yuhan Wang, Yuwei Zhang, Liang Feng, Junhao Xiao and 12 more

TPAMI 2026 · Reinforcement learning › RL from feedback · approach: RL fine-tuning of generative models · 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
apply Flow-GRPO RL alignmentStable Diffusion 3.5-M · GenEvalGenEval score: 0.63 -> 0.95Improved
apply Flow-GRPO RL alignmentStable Diffusion 3.5-M · OCR scoreOCR score: 0.59 -> 0.92Improved
apply DanceGRPOStable Diffusion v1.4 · HPS-v2.1HPS-v2.1: 0.239 -> 0.365Improved
use DanceGRPO instead of DDPOStable Diffusion v1.4 · HPS-v2.1HPS-v2.1: 0.297 -> 0.365Improved
apply DanceGRPO with joint rewardsStable Diffusion v1.4 · Pick-a-PicPick-a-Pic: 0.202 -> 0.215Improved
apply DanceGRPOFLUX.1-dev · Pick-a-PicPick-a-Pic: 0.224 -> 0.228No clear effect
apply DanceGRPOFLUX.1-dev · GenEvalGenEval score: 0.659 -> 0.687Improved
apply DPOKStable Diffusion v1.5 · DrawBench ImageRewardImageReward: 0.13 -> 0.58Improved
apply Qwen-Image-2.0-RL post-trainingQwen-Image-2.0-Base · T2I ArenaElo: 1115 -> 1193Improved
apply RL post-training (TraDo-8B-Thinking) to dLLMLLaDA-8B-Instruct · MATH500accuracy: 36.2 -> 87.4Improved
apply d1 RL to dLLMLLaDA-8B-Instruct · HumanEvalaccuracy: 49.4 -> 39.0Hurt
apply wd1 RL to dLLMLLaDA-8B-Instruct · Sudokuaccuracy: 11.7 -> 76.4Improved

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