Benjamin Eysenbach

Assistant Professor of Computer Science at Princeton University.
Affiliated/Associated Faculty with the Princeton Program in Cognitive Science and the Princeton Language Initiative.

prof_pic.jpg

Room 416

35 Olden St

Princeton NJ 08544

eysenbach@princeton.edu

I design reinforcement learning (RL) algorithms: AI methods that learn how to make intelligent decisions from trial and error. I am especially interested in self-supervised methods, which enable agents to learn intelligent behaviors without labels or human supervision. Our group has developed some of the foremost algorithms and analysis for such self-supervised RL methods. Here are a few examples. I run the Princeton Reinforcement Learning Lab.

Bio: Before joining Princeton, I did by PhD in machine learning at CMU under Ruslan Salakhutdinov and Sergey Levine and supported by the NSF GFRP and the Hertz Fellowship. I spent a number of years at Google Brain/Research before and during my PhD. My undergraduate studies were in math at MIT.

Join us! (Fall 2024) I am recruiting 1 – 2 PhD students and 1 – 2 MS students to start in Fall 2025. I am also recruiting 1 postdoc and 1 predoc researcher to start in Spring 2025. Please read this page for more information and details on how to apply.

news

Sep 13, 2024 :fire: JaxGCRL: A new benchmark for goal-conditioned RL is blazing fast, allowing you to train at 1 million steps per minute on 1 GPU. Experiments run so fast that the algorithm design process becomes interactive. Tools like this not only make research much more accessible (e.g., you can now run a bunch of interesting experiments in a free Colab notebook before the 90 min timeout), but also will change how RL is taught (less fighting with dependencies, more experiments on complex tasks, less waiting for experiments to queue and finish); stay tuned for COS 435 this Spring!
Sep 12, 2024 Upcoming talks:
  • NYU GRAIL (Oct 2, 2024).
  • Facebook – Reasoning and Planning (Oct 8, 2024).
  • Colloquium at Queens College (Oct 21, 2024).
  • European Workshop on RL: Keynote (Oct 28, 2024).
Aug 13, 2024 Skills and directed exploration seem to emerge from contrastive RL! Check out the website for videos, code, and the full paper! Let by Grace Liu with Michael Tang.
Aug 9, 2024 In attempts to change perceptions about who does RL, we’ve put together a poster of Notable Women in RL!
Jul 1, 2024 Excited to share work that will be presented at ICML 2024!

selected publications

The aim is to highlight a small subset of the work done in the group, and to give a sense for the sorts of problems that we're working on. Please see Google Scholar for a complete and up-to-date list of publications.

2024

  1. cmd.png
    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
    In Forty-first International Conference on Machine Learning, 2024
  2. raj_augmentation.png
    Closing the Gap between TD Learning and Supervised Learning - A Generalisation Point of View
    Raj Ghugare, Geist Matthieu, Glen Berseth, and Benjamin Eysenbach
    In The Twelfth International Conference on Learning Representations, 2024
  3. td_cpc.png
    Contrastive Difference Predictive Coding
    Chongyi Zheng, Ruslan Salakhutdinov, and Benjamin Eysenbach
    In The Twelfth International Conference on Learning Representations, 2024
  4. stabilizing.gif
    Stabilizing Contrastive RL: Techniques for Robotic Goal Reaching from Offline Data
    Chongyi Zheng, Benjamin Eysenbach, Homer Walke, Patrick Yin, Kuan Fang, Ruslan Salakhutdinov, and Sergey Levine
    In The Twelfth International Conference on Learning Representations, 2024

2023

  1. cvl.png
    Contrastive value learning: Implicit models for simple offline rl
    Bogdan Mazoure, Benjamin Eysenbach, Ofir Nachum, and Jonathan Tompson
    In Conference on Robot Learning, 2023
  2. alm.png
    Simplifying Model-based RL: Learning Representations, Latent-space Models, and Policies with One Objective
    Raj Ghugare, Homanga Bharadhwaj, Benjamin Eysenbach, Sergey Levine, and Russ Salakhutdinov
    In International Conference on Learning Representations , 2023
  3. ac_connection.png
    A Connection between One-Step RL and Critic Regularization in Reinforcement Learning
    Benjamin Eysenbach, Matthieu Geist, Sergey Levine, and Ruslan Salakhutdinov
    In International Conference on Machine Learning, 2023
  4. laeo.png
    Contrastive Example-Based Control
    Kyle Beltran Hatch, Benjamin Eysenbach, Rafael Rafailov, Tianhe Yu, Ruslan Salakhutdinov, Sergey Levine, and Chelsea Finn
    In Learning for Dynamics and Control Conference, 2023
  5. thesis.png
    Probabilistic Reinforcement Learning: Using Data to Define Desired Outcomes, and Inferring How to Get There
    Benjamin Eysenbach
    PhD Thesis, Carnegie Mellon University, 2023

2022

  1. mnm_lower_bound.gif
    Mismatched No More: Joint Model-Policy Optimization for Model-Based RL
    Benjamin Eysenbach, Alexander Khazatsky, Sergey Levine, and Ruslan Salakhutdinov
    In Advances in Neural Information Processing Systems, 2022
  2. contrastive.gif
    Contrastive Learning as Goal-Conditioned Reinforcement Learning
    Benjamin Eysenbach, Tianjun Zhang, Ruslan Salakhutdinov, and Sergey Levine
    In Advances in Neural Information Processing Systems, 2022
  3. ocbc.gif
    Imitating Past Successes can be Very Suboptimal
    Benjamin Eysenbach, Soumith Udatha, Russ R Salakhutdinov, and Sergey Levine
    In Advances in Neural Information Processing Systems, 2022
  4. info_geometry.gif
    The Information Geometry of Unsupervised Reinforcement Learning
    Benjamin Eysenbach, Ruslan Salakhutdinov, and Sergey Levine
    In International Conference on Learning Representations, 2022
  5. maxent_robust.gif
    Maximum Entropy RL (Provably) Solves Some Robust RL Problems
    Benjamin Eysenbach, and Sergey Levine
    In International Conference on Learning Representations, 2022

2021

  1. rce.gif
    Replacing Rewards with Examples: Example-Based Policy Search via Recursive Classification
    Benjamin Eysenbach, Sergey Levine, and Ruslan Salakhutdinov
    Advances in Neural Information Processing Systems, 2021
  2. c_learning_sawyer.gif
    C-Learning: Learning to Achieve Goals via Recursive Classification
    Benjamin Eysenbach, Ruslan Salakhutdinov, and Sergey Levine
    In International Conference on Learning Representations, 2021
  3. rpc_teaser.gif
    Robust Predictable Control
    Benjamin Eysenbach, Ruslan Salakhutdinov, and Sergey Levine
    In Advances in Neural Information Processing Systems, 2021

2019

  1. sorb.png
    Search on the replay buffer: Bridging planning and reinforcement learning
    Benjamin Eysenbach, Ruslan Salakhutdinov, and Sergey Levine
    In Advances in Neural Information Processing Systems, 2019