On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs)

Anas M. Albaghajati; Lahouari Ghouti

On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs)

Anas M. Albaghajati, Lahouari Ghouti

Department of Information and Computer Science

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Deep reinforcement learning (DRL) merges reinforcement (RL) and deep learning (DL). In DRL-based agents rely on high-dimensional imagery inputs to make accurate decisions. Such excessively high-dimensional inputs and sophisticated algorithms require very powerful computing resources and longer training times. To alleviate the need for powerful resources and reduce the training times, this paper proposes novel solutions to mitigate the curse-of-dimensionality without compromising the DRL agent performance. Using these solutions, the deep Q-network model (DQN) and its improved versions require less training times while achieving better performance.

Original language	English
Title of host publication	ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
Publisher	ESANN (i6doc.com)
Pages	185-190
Number of pages	6
ISBN (Electronic)	9782875870650
State	Published - 2019
Event	27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2019 - Bruges, Belgium Duration: 24 Apr 2019 → 26 Apr 2019

Publication series

Name	ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning

Conference

Conference	27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2019
Country/Territory	Belgium
City	Bruges
Period	24/04/19 → 26/04/19

Bibliographical note

Publisher Copyright:
© 2019 ESANN (i6doc.com). All rights reserved.

ASJC Scopus subject areas

Artificial Intelligence
Information Systems

Cite this

Albaghajati, A. M., & Ghouti, L. (2019). On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs). In ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (pp. 185-190). (ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning). ESANN (i6doc.com).

Albaghajati, Anas M. ; Ghouti, Lahouari. / On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs). ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. ESANN (i6doc.com), 2019. pp. 185-190 (ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning).

@inproceedings{66e486427d754f6582b3fc61e52102d2,

title = "On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs)",

abstract = "Deep reinforcement learning (DRL) merges reinforcement (RL) and deep learning (DL). In DRL-based agents rely on high-dimensional imagery inputs to make accurate decisions. Such excessively high-dimensional inputs and sophisticated algorithms require very powerful computing resources and longer training times. To alleviate the need for powerful resources and reduce the training times, this paper proposes novel solutions to mitigate the curse-of-dimensionality without compromising the DRL agent performance. Using these solutions, the deep Q-network model (DQN) and its improved versions require less training times while achieving better performance.",

author = "Albaghajati, {Anas M.} and Lahouari Ghouti",

note = "Publisher Copyright: {\textcopyright} 2019 ESANN (i6doc.com). All rights reserved.; 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2019 ; Conference date: 24-04-2019 Through 26-04-2019",

year = "2019",

language = "English",

series = "ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning",

publisher = "ESANN (i6doc.com)",

pages = "185--190",

booktitle = "ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning",

}

Albaghajati, AM & Ghouti, L 2019, On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs). in ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN (i6doc.com), pp. 185-190, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2019, Bruges, Belgium, 24/04/19.

On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs). / Albaghajati, Anas M.; Ghouti, Lahouari.
ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. ESANN (i6doc.com), 2019. p. 185-190 (ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs)

AU - Albaghajati, Anas M.

AU - Ghouti, Lahouari

PY - 2019

Y1 - 2019

N2 - Deep reinforcement learning (DRL) merges reinforcement (RL) and deep learning (DL). In DRL-based agents rely on high-dimensional imagery inputs to make accurate decisions. Such excessively high-dimensional inputs and sophisticated algorithms require very powerful computing resources and longer training times. To alleviate the need for powerful resources and reduce the training times, this paper proposes novel solutions to mitigate the curse-of-dimensionality without compromising the DRL agent performance. Using these solutions, the deep Q-network model (DQN) and its improved versions require less training times while achieving better performance.

AB - Deep reinforcement learning (DRL) merges reinforcement (RL) and deep learning (DL). In DRL-based agents rely on high-dimensional imagery inputs to make accurate decisions. Such excessively high-dimensional inputs and sophisticated algorithms require very powerful computing resources and longer training times. To alleviate the need for powerful resources and reduce the training times, this paper proposes novel solutions to mitigate the curse-of-dimensionality without compromising the DRL agent performance. Using these solutions, the deep Q-network model (DQN) and its improved versions require less training times while achieving better performance.

UR - http://www.scopus.com/inward/record.url?scp=85071335936&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85071335936

T3 - ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning

SP - 185

EP - 190

BT - ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning

PB - ESANN (i6doc.com)

T2 - 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2019

Y2 - 24 April 2019 through 26 April 2019

ER -

Albaghajati AM, Ghouti L. On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs). In ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. ESANN (i6doc.com). 2019. p. 185-190. (ESANN 2019 - Proceedings, 27th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning).

On the speedup of deep reinforcement learning deep Q-networks (RL-DQNs)

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus subject areas

Fingerprint

Cite this