Highly scalable 2D JAX physics engine.

Reinforcement learning on general 2D physics environments in JAX. ICLR 2025 Oral.

Pytorch implementation of Stable Opponent Shaping (https://openreview.net/pdf?id=SyGjjsC5tQ).

hanabi_learning_environment is a research platform for Hanabi experiments.

A list of Hanabi strategies

Code release for Learning with Opponent-Learning Awareness and variations.

Exploring the Input Switch Affine Networks model

From Pixels to Torques: Policy Learning using Deep Dynamical Convolutional Neural Networks (DDCNN)