Jiajun Fan

ICLR 2025

Online Reward-Weighted Fine-Tuning of Flow Matching with Wasserstein Regularization

In brief

What the paper does.

We introduce ORW-CFM-W2, a self-evolving RLHF framework enabling flow matching models to continuously optimize through online reward feedback without relying on human-collected datasets. We derive a tractable Wasserstein-2 bound providing the first theoretical guarantee for collapse-free policy evolution.

Published in
International Conference on Learning Representations 2025 (ICLR 2025)
Paper
openreview.net/forum?id=2IoFFexvuw
Project page
jiajunfan.com/projects/orw-cfm-w2/

Cite

Cite the paper.

@inproceedings{fan2025online,
  title={Online Reward-Weighted Fine-Tuning of Flow Matching
  with Wasserstein Regularization},
  author={Jiajun Fan and Shuaike Shen and Chaoran Cheng
  and Yuxin Chen and Chumeng Liang and Ge Liu},
  booktitle={The Thirteenth International Conference on
  Learning Representations},
  year={2025},
  url={https://openreview.net/forum?id=2IoFFexvuw}
}