Tian Gao

I am a Ph.D. student in Computer Science at Stanford University, advised by Prof. Chelsea Finn and Prof. Dorsa Sadigh. Previously, I obtained my bachelor's degree from the Institute for Interdisciplinary Information Sciences (Yao Class) at Tsinghua University, advised by Prof. Yi Wu. I was honored to receive Yao Award (Gold Medal), the highest honor of our department. I was very fortunate to work with Prof. Yuke Zhu as a visiting researcher in UT Austin.

I am broadly interested in machine learning and robotics. My research goal is to develop efficient learning methods to build autonomous robots with robust and generalizable behaviors that help humans do a wide range of real-world tasks.

Email / Twitter / Google Scholar / Linkedin

Research

	Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone Tian Gao, Celine Tan, Catherine Glossop, Timothy Gao, Jiankai Sun, Kyle Stachowicz, Shirley Wu, Oier Mees, Dorsa Sadigh, Sergey Levine, Chelsea Finn, project page / arXiv SteerVLA uses VLM reasoning to steer a VLA policy for robust control in long-tail driving scenarios.
	Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone Max Sobol Mark, Tian Gao, Georgia Gabriela Sampaio, Mohan Kumar, Archit Sharma, Chelsea Finn, Aviral Kumar project page / arXiv Fine-tuning multiple policy classes with Actor-Critic RL.
	PRIME: Scaffolding Manipulation Tasks with Behavior Primitives for Data-efficient Imitation Learning Tian Gao, Soroush Nasiriany, Huihan Liu, Quantao Yang, Yuke Zhu IEEE Robotics and Automation Letters (RA-L), 2024 project page / arXiv A primitive-based data-efficient imitation learning framework that scaffolds manipulation tasks with behavior primitives, breaking down long human demonstrations into concise, simple behavior primitive sequences.
	Learning and Retrieval from Prior Data for Skill-based Imitation Learning Soroush Nasiriany, Tian Gao, Ajay Mandlekar, Yuke Zhu CoRL, 2022 project page / arXiv A skill-based imitation learning framework for tackling long-horizon robot manipulation tasks data-efficiently and robustly.
	Phasic Self-Imitative Reduction for Sparse-Reward Goal-Conditioned Reinforcement Learning Yunfei Li, Tian Gao, Jiaqi Yang, Huazhe Xu, Yi Wu ICML, 2022 project page / arXiv / video A phasic self-imitative reduction (PAIR) framework for tackling sparse-reward goal-conditioned RL problems effectively.

Website template from Jon Barron,