Publications

* Equal contribution · † Corresponding authors

2026

  1. ROSS: Relearning from Self-Generated Rollouts through Selective Supervision
    Zhiwei Zhang, Huayu Deng, Fei Zhao, Jiayan Fu, Bin Liang, Kam-Fai Wong†, and Mu Chuan†
    Technical Report , Sep 2026
  2. PACT: From Credit Assignment to Critic Alignment
    Jiayan Fu, Hang Xu, Yong Zhang, Zhaokai Luo, Yao Hu, Dongyan Zhao†, and Mu Chuan†
    Technical Report , Sep 2026
  3. ReAlloc: Hierarchical Optimization Budget Reallocation for LLM post-training
    Ye Mo, Chuan Zhou, Jiayan Fu, Junchi Li, Jiasheng Sun, Jiangwei Xie, Jingjing Liu, Jian Liu, Jiajun Bu, and Sheng Zhou†
    Submitted to ICLR 2027 , 2026