My research is motivated by a key problem of AI which is making decisions in an interactive environment, besides perception and prediction. I am broadly interested in interactive machine learning (e.g. reinforcement learning and bandit) under the real-world constraints such as sample efficiency, safety, robustness and alignment.

Recently, I work on reinforcement learning methods in training foundation models (project: Amazon Bedrock) and foundation models for decision making.

Education and Experiences

Preprints and Publications


CS234: Reinforcement Learning, Teaching Assistant, Winter 2019-2020.

CS229: Machine Learning, Teaching Assistant, Spring 2020-2021.

Professional Service

Journal Reviewing: JMLR, IEEE TPAMI, MLJ, AIJ, Biometrika

Conference Reviewing: NeurIPS (2019 - 2021), ICLR (2019 - 2021, 2023), ICML(2020, 2021), AISTATS (2020 - 2022), UAI (2020), AAAI (2022, 2023)