Towards Continual Reinforcement Learning: A Review and Perspectives

Authors: Khimya Khetarpal, Matthew Riemer, Irina Rish, Doina Precup

JAIR 2022 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this article, we aim to provide a literature review of different formulations and approaches to continual reinforcement learning (RL), also known as lifelong or non-stationary RL. We begin by discussing our perspective on why RL is a natural fit for studying continual learning. We then provide a taxonomy of different continual RL formulations by mathematically characterizing two key properties of non-stationarity, namely, the scope and driver non-stationarity. This offers a unified view of various formulations. Next, we review and present a taxonomy of continual RL approaches.
Researcher Affiliation Collaboration Khimya Khetarpal EMAIL Mila, Mc Gill University Matthew Riemer EMAIL Mila, Universit e de Montr eal, IBM Research Irina Rish EMAIL Mila, Universit e de Montr eal Doina Precup EMAIL Mila, Mc Gill University, Deep Mind
Pseudocode No The paper is a literature review and taxonomy of continual RL. It describes various approaches conceptually and mathematically, but does not present any algorithms or pseudocode in a structured format.
Open Source Code No The paper is a literature review and taxonomy. It does not describe any specific new methodology for which open-source code would be provided. There are no statements regarding the release of code for this work.
Open Datasets No The paper is a review and does not present experimental results using specific datasets. It discusses various benchmarks such as 'Open AI Gym', 'Arcade Learning Environment (ALE)', 'Jelly Bean World', 'Causal World', and 'Meta-world' in the context of evaluating continual RL agents, but it does not use them for experiments within this work.
Dataset Splits No This paper is a literature review and taxonomy of continual reinforcement learning. It does not describe any experiments that would require specifying training, validation, or test dataset splits.
Hardware Specification No This paper is a literature review and taxonomy of continual reinforcement learning. It does not describe any experiments that would require specific hardware for execution.
Software Dependencies No This paper is a literature review and taxonomy of continual reinforcement learning. It does not describe any experiments or implementations that would require specific software dependencies with version numbers.
Experiment Setup No This paper is a literature review and taxonomy of continual reinforcement learning. It does not describe any experiments, training procedures, or hyperparameter settings in its content.