PhD student in computer science at Berkeley
I'm a second year PhD student in Computer Science at UC Berkeley, advised by Pieter Abbeel.
I work on finding computational solutions to achieve general machine intelligence.
At the moment, I'm interested in building neural networks for reinforcement learning.
- Hybrid Discriminative-Generative Training via Contrastive Learning. Hao Liu,
Pieter Abbeel. arXiv, 2020 [pdf]
- Improving Policy Gradient via Parameterized Reward. Hao Liu,
Pieter Abbeel. Tech report, 2019 [pdf]
- Taming MAML: Efficient Unbiased Gradient-based Meta-reinforcement learning. Hao Liu, Richard Socher, and Caiming Xiong. In International Conference on Machine Learning(ICML), 2019 [pdf]
- Competitive experience replay. Hao Liu, Alexander Trot, Richard Socher, and
Caiming Xiong. In International Conference on Learning Representations(ICLR), 2019
- Variational Inference with Tail-adaptive f-Divergence. Dilin Wang, Hao Liu, and Qiang Liu. In Advances in
Neural Information Processing Systems(NeurIPS), pages 5741--5751, Dec 2018
[pdf] Oral Presentation
- Action-dependent Control Variates for Policy Optimization via Stein Identity.
Hao Liu, Yihao Feng, Yi Mao, Dengyong Zhou, Jian Peng, and Qiang Liu. In International Conference
on Learning Representations(ICLR), 2018 [pdf]
In Deep Reinforcement Learning Workshop at NeurIPS, 2017 Oral Presentation
- 2018: B.E. from UESTC Advisor: Zenglin Xu
- 2018: Intern at Salesforce Advisors: Richard Socher, Caiming Xiong
- 2017: Intern at UT Austin Advisor: Qiang Liu
- Reviewer for JMLR, NeurIPS, ICML, ACL
- UC Berkeley EECS Department Fellowship
- ICLR, ICML, NeurIPS Conference Travel Grant
- Outstanding Undergraduate Thesis Award
- First Prize in The Chinese Mathematics Competitions