AI Paper Atlas

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

DenseGRPO: From Sparse to Dense Reward for Flow Matching Model Alignment

Haoyou Deng, Keyu Yan, Chaojie Mao, Xiang Wang, Yu Liu, Changxin Gao, Nong Sang

ICLR 2026 · Image generation › Text-to-image diffusion · approach: Preference optimization for diffusion · 23 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 ODE-based step-wise dense reward (DenseGRPO) over Flow-GRPOFlow-GRPO · Human Preference AlignmentPickScore: 23.31 -> 24.64Improved
add DenseGRPO over Flow-GRPOFlow-GRPO · Compositional Image GenerationGenEval: 0.95 -> 0.97Improved
add DenseGRPO over Flow-GRPOFlow-GRPO · Visual Text RenderingOCR accuracy: 0.92 -> 0.95Improved
use ODE-based dense reward instead of latent-similarity (CoCA) dense rewardFlow-GRPO · Human Preference AlignmentPickScore: 23.63 -> 24.64Improved
add latent-similarity step reward (CoCA) to Flow-GRPOFlow-GRPO · Human Preference AlignmentPickScore: 23.31 -> 23.63Improved
use dense reward R_{t-1} per step instead of sparse terminal reward (ablation)Flow-GRPO · Human Preference AlignmentPickScoreImproved
use calibrated timestep-specific noise level instead of uniform a=0.7 (ablation)DenseGRPO · Human Preference AlignmentPickScoreImproved
use uniform a=0.7 noise with dense reward instead of Flow-GRPO (ablation)DenseGRPO · Human Preference AlignmentPickScoreImproved
increase ODE denoising steps from n=1 to n=t (ablation)DenseGRPO · Human Preference AlignmentPickScoreImproved
use single-step ODE (n=1) for dense reward instead of Flow-GRPO sparse reward (ablation)DenseGRPO · Human Preference AlignmentPickScoreHurt
increase ODE steps n from 1 to 2 to t (ablation)DenseGRPO · training costGPU hours for 20 steps: 11, 13, and 19 GPU hoursHurt
apply DenseGRPO on FLUX.1-devFlow-GRPO · FLUX.1-devPickScoreImproved

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