Massively Parallel Methods for Deep Reinforcement Learning
Abstract
We present the first massively distributed architecture for deep reinforcement learning. This architecture uses four main components: parallel actors that generate new behaviour; parallel learners that are trained from stored experience; a distributed neural network to represent the value function or behaviour policy; and a distributed store of experience. We used our architecture to implement the Deep Q-Network algorithm (DQN). Our distributed algorithm was applied to 49 games from Atari 2600 games from the Arcade Learning Environment, using identical hyperparameters. Our performance surpassed non-distributed DQN in 41 of the 49 games and also reduced the wall-time required to achieve these results by an order of magnitude on most games.
Keywords
Cite
@article{arxiv.1507.04296,
title = {Massively Parallel Methods for Deep Reinforcement Learning},
author = {Arun Nair and Praveen Srinivasan and Sam Blackwell and Cagdas Alcicek and Rory Fearon and Alessandro De Maria and Vedavyas Panneershelvam and Mustafa Suleyman and Charles Beattie and Stig Petersen and Shane Legg and Volodymyr Mnih and Koray Kavukcuoglu and David Silver},
journal= {arXiv preprint arXiv:1507.04296},
year = {2015}
}
Comments
Presented at the Deep Learning Workshop, International Conference on Machine Learning, Lille, France, 2015