Chongyi Zheng

prof_pic.jpg

I am a PhD student in Computer Science at Princeton University advised by Benjamin Eysenbach. I am interested in developing reinforcement learning (RL) algorithms that can leverage prior knowledge, reason over long horizons, and gather information in a unsupervised manner. To this end, I work on unsupervised RL, goal-conditioned RL, and representation learning for control. I completed my M.S. from Carnegie Mellon University advised by Ruslan Salakhutdinov and received my bachelor’s degree from Southeast University. I have had the great opportunity to collaborate with Sergey Levine.

publications

(*: equal contribution)

2026

  1. onestep_fb.svg
    Can We Really Learn One Representation to Optimize All Rewards?
    Chongyi Zheng*, Royina Karegoudra Jayanth*, and Benjamin Eysenbach
    arXiv preprint arXiv:XX, 2026
    PDF Code Website
  2. value_flows.png
    Value Flows
    Perry Dong*, Chongyi Zheng*, Chelsea Finn, Dorsa Sadigh, and Benjamin Eysenbach
    International Conference on Learning Representations, 2026
    PDF Code Website Thread
  3. infom.gif
    Intention-Conditioned Flow Occupancy Models
    Chongyi Zheng, Seohong Park, Sergey Levine, and Benjamin Eysenbach
    International Conference on Learning Representations, 2026
    PDF Code Website Thread

2025

  1. upskill.png
    UpSkill: Mutual Information Skill Learning for Structured Response Diversity in LLMs
    Devan Shah, Owen Yang, Daniel Yang, Chongyi Zheng, and Benjamin Eysenbach
    NeurIPS Workshop on Scaling Environments for Agents, 2025
    PDF
  2. cirl.png
    Consistent Zero-Shot Imitation with Contrastive Goal Inference
    Kathryn Wantlin, Chongyi Zheng, and Benjamin Eysenbach
    arXiv preprint arXiv:2510.17059, 2025
    PDF
  3. csf.jpg
    Can a MISL Fly? Analysis and Ingredients for Mutual Information Skill Learning
    Chongyi Zheng*, Jens Tuyls*, Joanne Peng, and Benjamin Eysenbach
    International Conference on Learning Representations (Oral Presentation < 1.8%), 2025
    PDF Code Website Thread

2024

  1. cmd.jpg
    Learning Temporal Distances: Contrastive Successor Features Can Provide a Metric Structure for Decision-Making
    Vivek Myers, Chongyi Zheng, Anca Dragan, Sergey Levine, and Benjamin Eysenbach
    International Conference on Machine Learning, 2024
    PDF Code
  2. gai.gif
    Generalized Animal Imitator: Agile Locomotion with Versatile Motion Prior
    Ruihan Yang*, Zhuoqun Chen*, Jianhan Ma*, Chongyi Zheng*, Yiyu Chen, Quan Nguyen, and Xiaolong Wang
    Conference on Robot Learning, 2024
    CoRL Deployable Workshop (Best Paper Award), 2023
    PDF Website
  3. td_infonce.gif
    Contrastive Difference Predictive Coding
    Chongyi Zheng, Ruslan Salakhutdinov, and Benjamin Eysenbach
    International Conference on Learning Representations, 2024
    PDF Code Website Thread
  4. stable_contrastive_rl.gif
    Stabilizing Contrastive RL: Techniques for Offline Goal Reaching
    Chongyi Zheng, Benjamin Eysenbach, Homer Walke, Patrick Yin, Kuan Fang, Ruslan Salakhutdinov, and Sergey Levine
    International Conference on Learning Representations (Spotlight Presentation < 5%), 2024
    PDF Code Website Thread

2023

  1. bridge_data_v2.png
    BridgeData V2: A Dataset for Robot Learning at Scale
    Homer Walke, Kevin Black, Abraham Lee, Moo Jin Kim, Max Du, Chongyi Zheng, Tony Zhao, Philippe Hansen-Estruch, Quan Vuong, Andre He, Vivek Myers, Kuan Fang, Chelsea Finn, and Sergey Levine
    In Conference on Robot Learning, 2023
    PDF Code Website

2021

  1. pasf.png
    Learning Domain Invariant Representations in Goal-conditioned Block MDPs
    Beining Han, Chongyi Zheng, Harris Chan, Keiran Paster, Michael Zhang, and Jimmy Ba
    Advances in Neural Information Processing Systems, 2021
    PDF Code

2020

  1. ndq_mmm.png
    Learning Nearly Decomposable Value Functions Via Communication Minimization
    Tonghan Wang*, Jianhao Wang*, Chongyi Zheng, and Chongjie Zhang
    In International Conference on Learning Representations, 2020
    PDF Code Website