Sparse Reward Environment

Sparse reward environments, characterized by infrequent or delayed feedback signals, pose a significant challenge for reinforcement learning (RL) agents. Current research focuses on improving exploration strategies through intrinsic rewards (e.g., novelty-based methods, information-theoretic approaches), hierarchical RL architectures to decompose complex tasks, and leveraging auxiliary models like vision-language models to provide richer reward signals. These advancements aim to enhance sample efficiency and enable RL agents to solve complex, real-world problems where dense reward functions are impractical or impossible to design, impacting fields like robotics and human-computer interaction.

Papers

August 22, 2022

Some Supervision Required: Incorporating Oracle Policies in Reinforcement Learning via Epistemic Uncertainty Metrics
Jun Jet Tai, Jordan K. Terry, Mauro S. Innocente, James Brusey, Nadjim Horri
Reinforcement Learning Reinforcement Learning Benchmark Sparse Reward Environment Test Oracle Actor Critic Reinforcement Learning Epistemic Uncertainty Quantification Uncertainty Aware Exploration

June 28, 2022

Dext-Gen: Dexterous Grasping in Sparse Reward Environments with Full Orientation Control
Martin Schuck, Jan Brüdigam, Alexandre Capone, Stefan Sosnowski, Sandra Hirche
Reinforcement Learning Robotic Grasping Robotic Hand Sparse Reward Environment Dexterous Grasping Orientation Control

June 24, 2022

Phasic Self-Imitative Reduction for Sparse-Reward Goal-Conditioned Reinforcement Learning
Yunfei Li, Tian Gao, Jiaqi Yang, Huazhe Xu, Yi Wu
Sparse Reward Sparse Reward Environment Effective Reinforcement Learning

June 12, 2022

Dealing with Sparse Rewards in Continuous Control Robotics via Heavy-Tailed Policies
Souradip Chakraborty, Amrit Singh Bedi, Alec Koppel, Pratap Tokekar, Dinesh Manocha
Sparse Reward Sparse Reward Environment Continuous Control Task Heavy Tailed Policy

April 20, 2022

Understanding and Preventing Capacity Loss in Reinforcement Learning
Clare Lyle, Mark Rowland, Will Dabney
Reinforcement Learning Human Understanding Sparse Reward Environment Capacity Loss Deep Reinforcement Learning Agent

April 2, 2022

Safe Reinforcement Learning via Shielding under Partial Observability
Steven Carr, Nils Jansen, Sebastian Junges, Ufuk Topcu
Reinforcement Learning Deep Reinforcement Learning Safe Reinforcement Learning Partial Observability Sparse Reward Environment Safe Exploration

January 11, 2022

STIR$^2$: Reward Relabelling for combined Reinforcement and Imitation Learning on sparse-reward tasks
Jesus Bujalance Martin, Fabien Moutarde
Imitation Learning Feature Imitation Reward Report Sparse Reward Environment Policy Algorithm

December 2, 2021

Hindsight Task Relabelling: Experience Replay for Sparse Reward Meta-RL
Charles Packer, Pieter Abbeel, Joseph E. Gonzalez
Reinforcement Learning Sparse Reward Meta Reinforcement Learning Experience Replay Sparse Reward Environment Proxy Reward Hindsight Goal Relabeling

November 2, 2021

Discovering and Exploiting Sparse Rewards in a Learned Behavior Space
Giuseppe Paolo, Miranda Coninx, Alban Laflaquière, Stephane Doncieux
Optimal Policy Scientific Discovery Sparse Reward Sparse Reward Environment Exploration Algorithm Behavior Space

Sparse Reward Environment

Papers

Some Supervision Required: Incorporating Oracle Policies in Reinforcement Learning via Epistemic Uncertainty Metrics

Dext-Gen: Dexterous Grasping in Sparse Reward Environments with Full Orientation Control

Phasic Self-Imitative Reduction for Sparse-Reward Goal-Conditioned Reinforcement Learning

Dealing with Sparse Rewards in Continuous Control Robotics via Heavy-Tailed Policies

Understanding and Preventing Capacity Loss in Reinforcement Learning

Safe Reinforcement Learning via Shielding under Partial Observability

STIR$^2$: Reward Relabelling for combined Reinforcement and Imitation Learning on sparse-reward tasks

Hindsight Task Relabelling: Experience Replay for Sparse Reward Meta-RL

Discovering and Exploiting Sparse Rewards in a Learned Behavior Space