Authors

Shicheng Fang

Publications 2

Preprint Open access

MARCO: Multi-Round Agentic Reinforcement for Conditional Molecular Optimization

Shicheng Fang, Yuxin Wang, Zhuo Yang et al. · 2026

Molecular optimization is inherently iterative: a candidate is proposed, evaluated against several objectives, and revised while preserving a relationship to the source molecule. Most instruction-following models instead emit one edited molecule, forcing validity, property improvement, and similarity control into a sin …

Preprint Open access

ORPG: Reconciling Multiple Reward Objectives through Objective-wise Policy Gradients

Shicheng Fang, Yiwen Zhao, Wenbo Tian et al. · 2026

Multi-reward policy optimization requires a joint update that reflects both the learning signals and the intended relationships among objectives. We introduce Objective-wise Reconciled Policy Gradient (ORPG), which constructs a separate clipped policy objective for each reward and reconciles the resulting gradients int …

Co-authors

Yuxin Wang Zhuo Yang Xiaohu Xu Jiahao Lu Chuanyuan Tan Tong Zhu Yining Zheng Xipeng Qiu Yiwen Zhao Wenbo Tian