A Study of Plasticity Loss in On-Policy Deep Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Juliani, Arthur, Ash, Jordan T.
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917825169326080
author Juliani, Arthur
Ash, Jordan T.
author_facet Juliani, Arthur
Ash, Jordan T.
contents Continual learning with deep neural networks presents challenges distinct from both the fixed-dataset and convex continual learning regimes. One such challenge is plasticity loss, wherein a neural network trained in an online fashion displays a degraded ability to fit new tasks. This problem has been extensively studied in both supervised learning and off-policy reinforcement learning (RL), where a number of remedies have been proposed. Still, plasticity loss has received less attention in the on-policy deep RL setting. Here we perform an extensive set of experiments examining plasticity loss and a variety of mitigation methods in on-policy deep RL. We demonstrate that plasticity loss is pervasive under domain shift in this regime, and that a number of methods developed to resolve it in other settings fail, sometimes even performing worse than applying no intervention at all. In contrast, we find that a class of ``regenerative'' methods are able to consistently mitigate plasticity loss in a variety of contexts, including in gridworld tasks and more challenging environments like Montezuma's Revenge and ProcGen.
format Preprint
id arxiv_https___arxiv_org_abs_2405_19153
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle A Study of Plasticity Loss in On-Policy Deep Reinforcement Learning
Juliani, Arthur
Ash, Jordan T.
Machine Learning
Artificial Intelligence
Continual learning with deep neural networks presents challenges distinct from both the fixed-dataset and convex continual learning regimes. One such challenge is plasticity loss, wherein a neural network trained in an online fashion displays a degraded ability to fit new tasks. This problem has been extensively studied in both supervised learning and off-policy reinforcement learning (RL), where a number of remedies have been proposed. Still, plasticity loss has received less attention in the on-policy deep RL setting. Here we perform an extensive set of experiments examining plasticity loss and a variety of mitigation methods in on-policy deep RL. We demonstrate that plasticity loss is pervasive under domain shift in this regime, and that a number of methods developed to resolve it in other settings fail, sometimes even performing worse than applying no intervention at all. In contrast, we find that a class of ``regenerative'' methods are able to consistently mitigate plasticity loss in a variety of contexts, including in gridworld tasks and more challenging environments like Montezuma's Revenge and ProcGen.
title A Study of Plasticity Loss in On-Policy Deep Reinforcement Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2405.19153