Small batch deep reinforcement learning

Published: 21 Sept 2023, Last Modified: 02 Nov 2023NeurIPS 2023 posterEveryoneRevisionsBibTeX
Keywords: Reinforcement Learning, Deep Reinforcement Learning, Value based, Batch Size
TL;DR: We uncover the the surprising finding that decreasing the batch size improves the performance of many standard deep RL agents.
Abstract: In value-based deep reinforcement learning with replay memories, the batch size parameter specifies how many transitions to sample for each gradient update. Although critical to the learning process, this value is typically not adjusted when proposing new algorithms. In this work we present a broad empirical study that suggests reducing the batch size can result in a number of significant performance gains; this is surprising, as the general tendency when training neural networks is towards larger batch sizes for improved performance. We complement our experimental findings with a set of empirical analyses towards better understanding this phenomenon.
Supplementary Material: pdf
Submission Number: 14647