Robust Risk-Sensitive Reinforcement Learning from Corrupted Human Feedback
Xinyi Ni and Lifeng Lai · Submitted to IEEE Journal on Selected Areas in Information Theory (JSAIT), 2026 · arXiv:2609.38938
Publications
Research on risk-sensitive reinforcement learning, efficient exploration, and learning from human feedback.
Submitted manuscript
Peer-reviewed papers
Provably Efficient Risk-Sensitive Reinforcement Learning with Human Feedback
Xinyi Ni and Lifeng Lai · IEEE International Symposium on Information Theory (ISIT), 2026
Risk-Sensitive Reinforcement Learning with ϕ-Divergence-Risk
Xinyi Ni and Lifeng Lai · IEEE Transactions on Information Theory, 2025
Risk-Sensitive Reward-Free Reinforcement Learning with CVaR
Xinyi Ni, Guanlin Liu, and Lifeng Lai · International Conference on Machine Learning (ICML), 2024
Robust Risk-Sensitive Reinforcement Learning with Conditional Value-at-Risk
Xinyi Ni and Lifeng Lai · IEEE Information Theory Workshop (ITW), 2024
Policy Gradient Based Entropic-VaR Optimization in Risk-Sensitive Reinforcement Learning
Xinyi Ni and Lifeng Lai · Allerton Conference on Communication, Control, and Computing, 2022
Risk-sensitive reinforcement learning via Entropic-VaR optimization
Xinyi Ni and Lifeng Lai · Asilomar Conference on Signals, Systems, and Computers, 2022
Dissertation and earlier work
Risk-Sensitive Reinforcement Learning with Coherent Risk Measures
Ph.D. dissertation · University of California, Davis
EVaR Optimization for Risk-Sensitive Reinforcement Learning
Xinyi Ni and Lifeng Lai · UC Davis, 2021