AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Entropy-Adaptive Diffusion Policy Optimization with Dynamic Step Alignment

RenYe Yan, Jikang Cheng, Yaozhong Gan, Shikun Sun, You Wu, Yunfan Yang, Liang Ling, Jinlong Lin, Yeshuang Zhu, Jie Zhou, Jinchao Zhang, Junliang Xing and 2 more

ICCV 2025 · Reinforcement learning › RL from feedback · approach: RL fine-tuning of generative models · 6 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 dynamic step alignment (early stop of denoising)DDPO · Training to aesthetic score 7training time (hours): 11.98h -> 5.52hImproved
add dynamic step alignment (early stop of denoising)DPOK · Training to aesthetic score 7training time (hours): 14.76h -> 7.16hImproved
add dynamic step alignment (early stop of denoising)DDPO · Training to aesthetic score 7total denoising steps: 1.843e6 -> 0.242e6Improved
add dynamic step alignment (early stop of denoising)DPOK · Training to aesthetic score 7total denoising steps: 2.015e6 -> 0.294e6Improved
use AdaEnt instead of DDPOSDv1.4 aesthetic fine-tuning · Unseen metrics, Inception Score meanInception Score: 22.95 -> 24.34Improved
use AdaEnt instead of TDPOSDv1.4 aesthetic fine-tuning · Unseen metrics, ImageReward meanImageReward: 0.8453 -> 1.1212Improved
scale entropy up to same order as KL (mu_e = 1.0) (ablation)AdaEnt · Reward learning on Simple-animalsaesthetic rewardHurt
add entropy bonus at moderate scaling factors (ablation)AdaEnt · Reward learning on Simple-animalsaesthetic rewardImproved
use adaptive gamma predictor instead of fixed entropy/KL weights (ablation)AdaEnt · Denoising trajectoryentropy-KL balanceImproved
use AdaEnt on SD-XL instead of DDPO, DPOK, D3POSD-XL aesthetic fine-tuning · SD-XLaesthetic rewardImproved
use AdaEnt instead of TDPO on SDv1.5SDv1.5 aesthetic fine-tuning · Unseen metrics, Inception ScoreInception Score: 2.40 -> 18.99Improved

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