Xinyi Ni

Xinyi Ni

Biography

I am currently a Machine Learning Scientist in industry, with continued interest in academic research. If you have exciting ideas or potential collaboration opportunities, feel free to reach out.

I completed my Ph.D. in ECE at UC Davis in Dec 2024, under the supervision of Prof. Lifeng Lai. My research focuses on risk-sensitive reinforcement learning (RL), enhancing algorithmic efficiency, sample complexity, and robustness.I also explored risk-sensitive RL in diverse settings, including reward-free frameworks and with human feedback (RLHF). During my internship at Microsoft, I applied large language model (LLM) reasoning and causal analysis to innovate on AI-driven game testing techniques.

I received my Bachelor’s degree in Information Engineering from Zhejiang University, China, in 2019. I collaborated with Prof. Jiangtao Huangfu on integrating deep learning (DL) into medical image diagnosis and autonomous driving.

Publications

  • Xinyi Ni and Lifeng Lai. “Risk-Sensitive Reinforcement Learning with Coherent Risk Measures.” Ph.D Thesis. 2025. University of California, Davis. Paper
  • Xinyi Ni and Lifeng Lai. “Risk-Sensitive Reinforcement Learning with $\phi$-Divergence-Risk.” IEEE Transaction on Information Theory. 2025. Paper
  • Xinyi Ni, Guanlin Liu and Lifeng Lai. “Risk-Sensitive Reward-Free Reinforcement Learning with CVaR.” International Conference on Machine Learning (ICML). 2024. Paper
  • Xinyi Ni and Lifeng Lai. “Robust Risk-Sensitive Reinforcement Learning with Conditional Value-at-Risk.” IEEE Information Theory Workshop (ITW) 2024. Paper
  • Xinyi Ni and Lifeng Lai. “Policy Gradient Based Entropic-VaR Optimization in Risk-Sensitive Reinforcement Learning.” Allerton Conference on Communication, Control, and Computing. IEEE, 2022. Paper
  • Xinyi Ni and Lifeng Lai. “Risk-sensitive reinforcement learning via Entropic-VaR optimization.” Asilomar Conference on Signals, Systems, and Computers. IEEE, 2022. Paper
  • Xinyi Ni and Lifeng Lai. “EVaR Optimization for Risk-Sensitive Reinforcement Learning.” UC Davis, 2021. Paper

Academic Activities

Reviewer

  • International Conference on Machine Learning (ICML)
  • IEEE Transactions on Information Theory
  • IEEE Transactions on Mobile Computing
  • IEEE Control Systems Letters