Graying the black box: Understanding DQNs

Tom Zahavy, Nir Ben Zrihem, Shie Mannor

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In recent years there is a growing interest in using deep representations for reinforcement learning. In this paper, we present a methodology and tools to analyze Deep Q-networks (DQNs) in a non-blind matter. Using our tools we reveal that the features learned by DQNs aggregate the state space in a hierarchical fashion, explaining its success. Moreover we are able to understand and describe the policies learned by DQNs for three different Atari2600 games and suggest ways to interpret, debug and optimize deep neural networks in reinforcement learning.

Original languageEnglish
Title of host publication33rd International Conference on Machine Learning, ICML 2016
EditorsKilian Q. Weinberger, Maria Florina Balcan
Pages2809-2822
Number of pages14
ISBN (Electronic)9781510829008
StatePublished - 2016
Event33rd International Conference on Machine Learning, ICML 2016 - New York City, United States
Duration: 19 Jun 201624 Jun 2016

Publication series

Name33rd International Conference on Machine Learning, ICML 2016
Volume4

Conference

Conference33rd International Conference on Machine Learning, ICML 2016
Country/TerritoryUnited States
CityNew York City
Period19/06/1624/06/16

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Software
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Graying the black box: Understanding DQNs'. Together they form a unique fingerprint.

Cite this