Author

Shilong Zhang

View on Semantic Scholar

1 paper indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#machine learning Sep 2025

Advantage Weighted Matching: Aligning RL with Pretraining in Diffusion Models

This work establishes a novel theoretical analysis: DDPO is an implicit form of score/flow matching with noisy targets, which increases variance and slows convergence, and introduces Advantage Weighted Matching (AWM), a policy-gradient method for diffusion.

Shuchen Xue, Chongjian Ge, Shilong Zhang et al. · arXiv.org · 52 citations · ⚡11