Center for Human-Compatible AI

imitation

Clean PyTorch implementations of imitation and reward learning algorithms

overcooked_ai

A benchmark environment for fully cooperative human-AI performance.

adversarial-policies

Find best-response to a fixed policy in multi-agent RL

human_aware_rl

Code for "On the Utility of Learning about Humans for Human-AI Coordination"

evaluating-rewards

Library to compare and evaluate reward functions

overcooked-demo

Web application where humans can play Overcooked with AI agents.

seals

Benchmark environments for reward modelling and imitation learning algorithms.

eirli

An Empirical Investigation of Representation Learning for Imitation (EIRLI), NeurIPS'21

tensor-trust

A prompt injection game to collect data for robust ML research

ranking-challenge

Testing ranking algorithms to improve social cohesion

tensor-trust-data

Dataset for the Tensor Trust project

learning-from-human-preferences

Reproduction of OpenAI and DeepMind's "Deep Reinforcement Learning from Human Preferences"

overcooked-hAI-exp

Overcooked-AI Experiment Psiturk Demo (for MTurk experiments)

nn-clustering-pytorch

Checking the divisibility of neural networks, and investigating the nature of the pieces networks can be divided into.

recon-email

Script for automatically creating the reconnaissance email.

multiagent-competition

Code for the paper "Emergent Complexity via Multi-agent Competition"

reward-preprocessing

Preprocessing reward functions to make them more interpretable

assistance-games

Supporting code for Assistance Games as a Framework paper

reducing-exploitability

stable-baselines3

PyTorch version of Stable Baselines, improved implementations of reinforcement learning algorithms.

dmc2gym

OpenAI Gym wrapper for the DeepMind Control Suite

minerl

MineRL Competition for Sample Efficient Reinforcement Learning - Python Package

reward-function-interpretability

rl-baselines3-zoo

A training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.

sacred

Sacred is a tool to help you configure, organize, log and reproduce experiments developed at IDSIA.

katago-driver-bug-repro

Docker files to help reproduce bug described in https://forums.developer.nvidia.com/t/kernel-oops-null-pointer-dereference-when-closing-cuda-application-katago/211270/3

pytorch-summary

Model summary in PyTorch similar to `model.summary()` in Keras

ray

A fast and simple framework for building and running distributed applications. Ray is packaged with RLlib, a scalable reinforcement learning library, and Tune, a scalable hyperparameter tuning library.

sgf-viewer

A simple webpage that can visualize a sgf string encoded as a url fragment.

slack-diskbot

low disk space alerts posted to Slack