Xing Chen

Make things simple and effective, not only in reinforcement learning.

I received my PhD in Artificial Intelligence from Jilin University, where I worked on deep reinforcement learning — off-policy analysis of policy optimization, exploration for continuous control, and RLHF — and spent a research visit at NTU's CCDS. I led a project on RL for quantitative trading at Sapient Intelligence. I now lead an ML team at Ragentile Intelligence, focusing on LLM post-training and infrastructure optimization.

Latest Updates

2026.09 Blog Halving the Activated Experts of a MoE Without Training.
2026.08 Paper “SRJudge: Empowering Large Language Models with Selective Reasoning for Fine-Grained Knowledge” (IJCAI, 2026).
2026.08 Career Joined Ragentile Intelligence as ML Researcher, Team Leader.
2026.04 Paper “Why Attend to Everything? Focus is the Key” (arXiv:2604.03260, 2026).
2026.03 Paper “A Tighter Bound for Reward Learning in Reinforcement Learning from Human Feedback” (Transactions on Machine Learning Research, 2026).
2026.03 Paper “Thin Keys, Full Values: Reducing KV Cache via Low-Dimensional Attention Selection” (arXiv:2603.04427, 2026).