Pseudorehearsal in actor-critic agents with neural network function approximation

TL;DR

This paper investigates how pseudorehearsal techniques can mitigate catastrophic forgetting in actor-critic reinforcement learning agents using neural networks, demonstrating improved learning stability in a pole balancing task.

Contribution

It introduces the application of pseudorehearsal to actor-critic agents with neural networks and compares different approaches to reduce forgetting.

Findings

Pseudorehearsal helps decrease catastrophic forgetting.

It improves learning stability in reinforcement learning.

Different pseudorehearsal methods vary in effectiveness.

Abstract

Catastrophic forgetting has a significant negative impact in reinforcement learning. The purpose of this study is to investigate how pseudorehearsal can change performance of an actor-critic agent with neural-network function approximation. We tested agent in a pole balancing task and compared different pseudorehearsal approaches. We have found that pseudorehearsal can assist learning and decrease forgetting.