Awesome Sample-Efficient Reinforcement Learning

This repository is a paper list of sample-efficient reinforcement learning where agents are expected to learn policies from limited interaction data. It's still being updated.

Please cite this repo if you find it helpful.

@techreport{Yu_A_Survey_on_2021,
author = {Yu, Tao},
month = {12},
title = {{A Survey on Sample-Efficient Reinforcement Learning}},
year = {2021}
}

Model-Free Methods

Proceeding	Method	Title	Resource
NeurIPS 2019	DER	When to use parametric models in reinforcement learning?	[pdf] [code]
AAAI 2021	SAC-AE	Improving Sample Efficiency in Model-Free Reinforcement Learning from Images	[pdf] [project]
ICML 2020	CURL	CURL: Contrastive Unsupervised Representations for Reinforcement Learning	[pdf] [code]
NeurIPS 2020	PI-SAC	Predictive Information Accelerates Learning in RL	[pdf]
NeurIPS 2020	SLAC	Stochastic Latent Actor-Critic: Deep Reinforcement Learning with a Latent Variable Model	[pdf]
NeurIPS 2020	RAD	Reinforcement Learning with Augmented Data	[pdf] [code]
ICLR 2021	DrQ	Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels	[pdf] [code]
ICLR 2021	SPR	Data-Efficient Reinforcement Learning with Self-Predictive Representations	[pdf] [code]
NeurIPS 2021	PlayVirtual	PlayVirtual: Augmenting Cycle-Consistent Virtual Trajectories for Reinforcement Learning	[pdf] [code]

Model-Based Methods

Proceeding	Method	Title	Resource
ICML 2019	PlaNet	Learning Latent Dynamics for Planning from Pixels	[pdf] [code]
ICLR 2020	SimPLe	Model-Based Reinforcement Learning for Atari	[pdf] [code]
ICLR 2020	Dreamer	Dream to Control: Learning Behaviors by Latent Imagination	[pdf] [code]
ICLR 2021	Dreamer V2	Mastering Atari with Discrete World Models	[pdf] [code]
NeurIPS 2021	EfficientZero	Mastering Atari Games with Limited Data	[pdf] [code]