gkswamy98

fast_irl

Contains implementation of the FILTER algorithm for exponentially faster inverse reinforcement learning.

pillbox

Contains implementation of AdVIL, AdRIL, and DAeQuIL algorithms from the ICML '21 Paper Of Moments and Matching.

dotfiles

some of my configs

causal_il

Contains implementation of the DoubIL and ResiduIL algorithms from the ICML '22 paper Causal Imitation Learning under Temporally Correlated Noise.

sequence_model_il

Contains sequence-model implementations of on and off-policy imitation learning algorithms for problems with unobserved contexts.

valuedice

Fork of ValueDICE code that supports discrete action spaces, pybullet, and is truly off-policy.

meta-rl-bci

meta learning + maxent deeprl for shared autonomy from eeg signals

replay_est

Contains implementation of the replay estimation algorithm from "Minimax Optimal Online Imitation Learning via Replay Estimation."

adversarial_rl

CS 294-131 Project, reimplementing https://arxiv.org/pdf/1702.02284.pdf.

ViZDoom

Doom-based AI Research Platform for Reinforcement Learning from Raw Visual Information. :godmode:

icl

294-149-coding-proj

Code for https://www.overleaf.com/project/5bb426f847a81409e6f4fd86

294-149_final_proj

CS 294-149 Final Project: Information-Theoretic Selection of Classifiers

autosort

Human assisted few-shot object sorting

baselines

OpenAI Baselines: high-quality implementations of reinforcement learning algorithms

causil

D4RL

A collection of reference environments for offline reinforcement learning

DQfD

An implementation of Deep Q-Learning from Demonstrations (DQfD) for playing Atari 2600 video games

filter

gym-minigrid

Minimalistic gridworld package for OpenAI Gym

il_envs

jaco_learning

Control, planning, and learning system for human-robot interaction with a JACO2 7DOF robotic arm.

mjrl

Reinforcement learning algorithms for MuJoCo tasks

mmil

Website for ICML'21 paper.

mu4e-dashboard

A dashboard for mu4e (mu for emacs)

replay

sequil

serializable

Utilities for creating serializable classes

softlearning

Softlearning is a reinforcement learning framework for training maximum entropy policies in continuous domains. Includes the official implementation of the Soft Actor-Critic algorithm.

valentinp.github.com

My public page.