Che Wang

PhD student in Computer Science.

E-mail: chewang[at]nyu[dot]edu

Links: [GitHub][Google Scholar]

Curriculum Vitae

About me

Hi! I am Che Wang, a PhD student at New York University, advised by Professor Keith Ross. My current research focuses on improving sample efficiency and achieving a better understanding of deep reinforcement learning from a fundamental level. Recently, I have been especially interested in studying bias reduction and representation learning techniques for deep reinforcement learning.

I have been working as a teaching assistant for machine learning and reinforcement learning class. I often lead workshops to help students set up environments and work with hpc clusters. I also design research projects that help students learn about the entire research workflow, from coding to research paper writing.

Before becoming a PhD student, I was an undergraduate student at New York University Shanghai, as a member of the inaugural class of 2017. In addition to deep learning, I have also been interested in other topics such as game design, robotics and data visualization and have created a number of related projects during my undergraduate study.

In 2022, I very recently received the Pearl Brownstein Doctoral Research Award.

Publications

• VRL3: A Data-Driven Framework for Visual Deep Reinforcement Learning. Che Wang, Xufang Luo, Keith Ross, Dongsheng Li. NeurIPS 2022.
• Reinforcement Learning with Automated Auxiliary Loss Search. Tairan He, Yuge Zhang, Kan Ren, Minghuan Liu, Che Wang, Weinan Zhang, Yuqing Yang, Dongsheng Li. NeurIPS 2022.
• On the Convergence of the Monte Carlo Exploring Starts Algorithm for Reinforcement Learning. Che Wang, Shuhan Yuan, Kai Shao, Keith Ross. ICLR 2022.
• Randomized Ensembled Double Q-Learning: Learning Fast Without a Model. Xinyue Chen*, Che Wang*, Zijian Zhou*, Keith Ross. ICLR 2021.
• BAIL: Best-Action Imitation Learning for Batch Deep Reinforcement Learning. Xinyue Chen, Zijian Zhou, Zheng Wang, Che Wang, Yanqiu Wu, Keith Ross. NeurIPS 2020.
• Striving for Simplicity and Performance in Off-Policy DRL: Output Normalization and Non-Uniform Sampling. Che Wang*, Yanqiu Wu*, Quan Vuong, Keith Ross. ICML 2020.
• Portfolio Online Evolution in StarCraft. Che Wang, Pan Chen, Yuanda Li, Christoffer Holmgard, Julian Togelius. Oral presentation at AIIDE 2016.

Work in progress

• Uncertainty Dampening: Correcting the Q-Landscape in Offline Reinforcement Learning. Che Wang, Keith Ross.
• Aggressive Q-Learning with Ensembles: Achieving Both High Sample Efficiency and High Asymptotic Performance. Yanqiu Wu, Xinyue Chen, Che Wang, Yiming Zhang, Zijian Zhou, Keith Ross
• On the Convergence of Monte Carlo UCB for Random-Length Episodic MDPs. Z. Dong, Che Wang, K. Ross. Under review.

Teaching

• In 2019 I have taken a collegiate teaching course led by Professor Jace Hargis at NYU Shanghai.
• Since 2018 I have been working for the ML and RL class, I found that step-by-step video tutorials can be very helpful. I also tried design experimental full-scale research projects to help students learn the research workflow.
• From 2016-2018 I have worked as a mentor and instructor at BiggerLab. A few student teams that I worked with went to participate in the 2018 MIT Battlecode contest. My experience shows that even high school students can learn AI concepts very quickly when the class is exciting enough.
• In 2016 I was a peer tutor for the Intro to CS course and hosted a coding competition. The top prizes were drones generously sponsored by Professor Olivier Marin.

Service

• Serving as a reviewer for ICML, ICLR and NeurIPS.

Hobbies

• I found teaching to be very interesting and constantly seek to improve my skills.
• I have been a dancer since 2013, I was one of the co-founders of the Xiaolong Shakers dance club at NYU Shanghai.
• I am a fan of science fiction and video games.