Skip to content
DSML.KZ
Publications

Preprint · arXiv · March 21, 2017

Pseudorehearsal in value function approximation

Vladimir Marochko, Leonard Johard, Manuel Mazzara

Manuel Mazzara
Read on arXivPDFDOI 10.48550/arxiv.1703.07075

Abstract

Catastrophic forgetting is of special importance in reinforcement learning, as the data distribution is generally non-stationary over time. We study and compare several pseudorehearsal approaches for Q-learning with function approximation in a pole balancing task. We have found that pseudorehearsal seems to assist learning even in such very simple problems, given proper initialization of the rehearsal parameters.

Reinforcement Learning in RoboticsDomain Adaptation and Few-Shot LearningAI-based Problem Solving and Planningcatastrophic forgettingvalue function approximationQ-learning

More by Manuel Mazzara