Lifelong Reinforcement Learning

Lifelong reinforcement learning (LRL) focuses on developing agents capable of continuously learning and adapting to new tasks throughout their operational lifetime, without catastrophic forgetting of previously acquired skills. Current research emphasizes improving sample efficiency and robustness to environmental changes through techniques like selective experience replay, parameter-free optimization methods, and biologically-inspired architectures incorporating neuromodulation or modularity. These advancements are crucial for deploying reinforcement learning agents in real-world scenarios characterized by non-stationary environments and limited training data, impacting fields such as robotics, autonomous driving, and medical imaging.

Papers

July 28, 2022

Latent Properties of Lifelong Learning Systems
Corban Rivera, Chace Ashcraft, Alexander New, James Schmidt, Gautam Vallabha
Latent Feature Lifelong Reinforcement Learning Interpretable Surrogate Model Lifelong Learning System

July 12, 2022

Reactive Exploration to Cope with Non-Stationarity in Lifelong Reinforcement Learning
Christian Steinparz, Thomas Schmied, Fabian Paischer, Marius-Constantin Dinu, Vihang Patil, Angela Bitto-Nemling, Hamid Eghbal-zadeh, Sepp Hochreiter
Policy Gradient Domain Shift Non Stationarity Lifelong Reinforcement Learning

July 1, 2022

Modular Lifelong Reinforcement Learning via Neural Composition
Jorge A. Mendez, Harm van Seijen, Eric Eaton
Lifelong Reinforcement Learning Neural Module Compositional Reinforcement Learning Neural Composition Compositional Reasoning Task

June 1, 2022

Provably Efficient Lifelong Reinforcement Learning with Linear Function Approximation
Sanae Amani, Lin F. Yang, Ching-An Cheng
Sublinear Regret Linear Function Approximation Lifelong Reinforcement Learning Contextual Markov Decision Process

May 22, 2022

A Dirichlet Process Mixture of Robust Task Models for Scalable Lifelong Reinforcement Learning
Zhi Wang, Chunlin Chen, Daoyi Dong
Reinforcement Learning Lifelong Reinforcement Learning Task Distribution Dirichlet Process Robust Task

March 14, 2022

L2Explorer: A Lifelong Reinforcement Learning Assessment Environment
Erik C. Johnson, Eric Q. Nguyen, Blake Schreurs, Chigozie S. Ewulum, Chace Ashcraft, Neil M. Fendley, Megan M. Baker, Alexander New, Gautam K. Vallabha
Reinforcement Learning Lifelong Learning Continual Reinforcement Learning Continual Learning Algorithm Lifelong Reinforcement Learning Reinforcement Learning Solution

December 30, 2021

Constructing a Good Behavior Basis for Transfer using Generalized Policy Updates
Safa Alver, Doina Precup
Reward Function Formality Transfer Unseen Task Lifelong Reinforcement Learning

November 18, 2021

Lifelong Reinforcement Learning with Temporal Logic Formulas and Reward Machines
Xuejing Zheng, Chao Yu, Chen Chen, Jianye Hao, Hankz Hankui Zhuo
Temporal Logic Continuous Learning Lifelong Reinforcement Learning Reward Machine Temporal Logic Formula

Lifelong Reinforcement Learning

Papers

Latent Properties of Lifelong Learning Systems

Reactive Exploration to Cope with Non-Stationarity in Lifelong Reinforcement Learning

Modular Lifelong Reinforcement Learning via Neural Composition

Provably Efficient Lifelong Reinforcement Learning with Linear Function Approximation

A Dirichlet Process Mixture of Robust Task Models for Scalable Lifelong Reinforcement Learning

L2Explorer: A Lifelong Reinforcement Learning Assessment Environment

Constructing a Good Behavior Basis for Transfer using Generalized Policy Updates

Lifelong Reinforcement Learning with Temporal Logic Formulas and Reward Machines