Reinforcement Learning Reading Group

When: Mondays, 4-5pm

Where: ICICS Room 238 / Zoom (link in mail)

Our reading group consist in a person leading the paper presentation with the aim at fostering live discussions and critics on the ideas and methods proposed, with the goal to understand the why and the how a particular paper is interesting.

Use this spreadsheet to sign up as a presenter.

To subscribe to the mailing list for talk announcements and other updates, send a message to with the words subscribe rlrg-l in the body.

  • If the above command doesn't work, you should try again with an empty subject, and the following message body
                  subscribe rlrg-l YOUR-EMAIL-ADDRESS
  • If you are using gmail, try using the web version and select "plain text mode"
  • Try to remove the signature
  • If you want to unsubscribe, send an email to with the words unsubscribe rlrg-l YOUR-EMAIL-ADDRESS in the body.
  • For more info on how majordomo mailing system works, you can check this page.
  • If you are still having issues, contact Daniele ( and he will register you manually.

  • Both slides and directly presenting from the paper in hand is fine.
  • These are some template points we try to go through during the reading group (you don't need to understand and know everything, it's a *group* process):
    • Why did you choose this paper?
    • What can you tell us about the authors?
    • What is the general context for the ideas in this paper?
    • What is the high-level idea of the paper, in a few sentences?
    • Overview of method/algorithm/theory
    • What was difficult to understand?
    • Results: most important figures and tables
    • Discussion: impact/assumptions/evaluation
    • Future work: what should be done next?
  • A (non-complete) list of topics covered in this reading group is available here.

Current term

Date Presenter Paper or topic Link
Oct 16 Nick Open X-Embodiment: Robotic Learning Datasets and RT-X Models website
Oct 23 Jenny Eureka: Human-Level Reward Design via Coding Large Language Models website
Oct 30 Ruiyu METRA: Scalable Unsupervised RL with Metric-Aware Abstraction website
Nov 6 Niloofar Learning Modular Robot Control Policies paper
Nov 20 Setareh
Nov 27 Aaron
Dec 4 Shengran
Jan 8
Jan 15

Past presentations, January-June 2023

Date Presenter Paper or topic Link
Jan 17 Jenny Fine-Tuning Language Models from Human Preferences paper
Jan 24 Yuval Tassa (GUEST TALK) Predictive Sampling: Real-time Behaviour Synthesis with MuJoCo paper
Jan 31 Daniele Dreamer v3 - Mastering Diverse Domains through World Models website
Feb 7 Shengran Inner Monologue: Embodied Reasoning through Planning with Language Models website
Feb 14 Rui Discovering faster matrix multiplication algorithms with reinforcement learning paper
Feb 28 Niloofar PADL: Language-Directed Physics-Based Character Control paper
Mar 14 Aaron In-context Reinforcement Learning with Algorithm Distillation paper
Mar 21 Amit Bermano (GUEST TALK) On the Effects of Language Models in Visual Generation zoom recording
Apr 4 Yuni ALAN : Autonomously Exploring Robotic Agents in the Real World website
Apr 11 Nick Magnetic control of tokamak plasmas through deep reinforcement learning paper
Apr 18 Niloofar Offline Reinforcement Learning as One Big Sequence Modeling Problem website
Apr 25 Helen Deep Exploration via Randomized Value Functions paper
May 2 Yuni Learning and Adapting Agility Skills by Transferring Experience website
May 9 Ruiyu Learning Agile Soccer Skills for a Bipedal Robot with Deep Reinforcement Learning paper
May 16 Niloofar PMP: Learning to Physically Interact with Environments using Part-wise Motion Priors paper
May 23 Wilder Critic Sequential Monte Carlo paper
May 30 Daniele Trace and Pace: Controllable Pedestrian Animation via Guided Trajectory Diffusion website
June 6 Kyle Barkour: Benchmarking Animal-level Agility with Quadruped Robots paper
June 13 Ruiyu VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training website
June 20 Julian Voyager: An Open-Ended Embodied Agent with Large Language Models website
June 27 Shuyuan Making Better Decision By Directly Planning In Continuous Control paper

Past Presentations, 2021

Date Presenter Title Link
Jun 23 Daniele Decision Transformer: Reinforcement Learning via Sequence Modeling paper
Jun 18 Wilder Implicit Under-Parameterization Inhibits Data-Efficient Deep Reinforcement Learning paper
Jun 11 Tyler Learning a family of motor skills from a single motion clip paper
Jun 4 Yuni Learning Spring Mass Locomotion: Guiding Policies with a Reduced-Order Model paper/video
May 28 Zhaoming Learning quadrupedal locomotion over challenging terrain paper
May 21 Ben Local motion phases for learning multi-contact character movements paper
May 14 Daniele Data-Efficient Reinforcement Learning with Self-Predictive Representations paper
May 7 Nam Hee Geppetto: Enabling Semantic Design of Expressive Robot Behaviors paper
Apr 30 Michiel Embodied Intelligence via Learning and Evolution paper
Apr 16 Zhaoming AMP: Adversarial Motion Priors for Stylized Physics-Based Character Control paper
Apr 2 Matt On the Importance of Hyperparameter Optimization for Model-based Reinforcement Learning paper
Mar 26 Setareh What Matters for On-policy Deep Actor-critic Methods? A Large-scale Study paper
Mar 19 Tianxin Emergent Complexity and Zero-shot Transfer via Unsupervised Environment Design paper
Mar 5 Michiel Autonomous navigation of stratospheric balloons using reinforcement learning paper
Feb 26 Ben Attention Is All You Need paper
Feb 12 Daniele Deep Reinforcement Learning in Parameterized Action Space paper
Feb 5 Zhaoming Direct Policy Optimization using Deterministic Sampling and Collocation paper

Past Presentations, 2020

Date Presenter Paper or topic Link
Dec 18 Ben Discovering Symbolic Models from Deep Learning with Inductive Biases paper
Dec 11 Matthew Gamma-Models: Generative Temporal Difference Learning for Infinite-Horizon Prediction paper
Dec 4 Nam Hee Learning Cross-Domain Correspondence for Control with Dynamics Cycle-Consistency paper
Nov 27 Wilder Robust Asymmetric Learning in POMDPs paper
Nov 20 Daniele Decoupling Representation Learning from Reinforcement Learning paper
Nov 13 Michiel How Neural Networks Extrapolate: From Feedforward to Graph Neural Networks paper