Asynchronous Methods for Deep Reinforcement Learning

02/04/2016 ∙ by Volodymyr Mnih, et al. ∙ 0

We propose a conceptually simple and lightweight framework for deep reinforcement learning that uses asynchronous gradient descent for optimization of deep neural network controllers. We present asynchronous variants of four standard reinforcement learning algorithms and show that parallel actor-learners have a stabilizing effect on training allowing all four methods to successfully train neural network controllers. The best performing method, an asynchronous variant of actor-critic, surpasses the current state-of-the-art on the Atari domain while training for half the time on a single multi-core CPU instead of a GPU. Furthermore, we show that asynchronous actor-critic succeeds on a wide variety of continuous motor control problems as well as on a new task of navigating random 3D mazes using a visual input.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 8

page 17

Code Repositories

async_deep_reinforce

Asynchronous Methods for Deep Reinforcement Learning


view repo

async-rl

Replicating "Asynchronous Methods for Deep Reinforcement Learning" (http://arxiv.org/abs/1602.01783)


view repo

async-deep-rl

A Tensorflow based implementation of "Asynchronous Methods for Deep Reinforcement Learning": https://arxiv.org/abs/1602.01783


view repo

deep_rl_acrobot

Using deep reinforcement learning (DDPG & A3C) to solve Acrobot


view repo

a3c-distributed_tensorflow

Distributed Tensorflow Implementation of Asynchronous Methods for Deep Reinforcement Learning


view repo
This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.