Generalized circle agent for geometry friends using deep reinforcement learning

Azmi Can Özgen; Mandana Fasounaki; Hazim Kemal Ekenel

doi:10.1109/SIU.2018.8404596

Generalized circle agent for geometry friends using deep reinforcement learning

Azmi Can Özgen, Mandana Fasounaki, Hazim Kemal Ekenel

Department of Computer Engineering

Istanbul Technical University

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

12 Citations (Scopus)

Abstract

Reinforcement learning began to perform at human-level success in game intelligence after deep learning revolution. Geometry Friends is a puzzle game, where we can benefit from deep learning and expect to have successful game playing agents. In the game, agents are collecting targets in two dimensional environment and they try to overcome obstacles in the way. In this paper, Q-learning approach is applied to this game and a generalized circle agent for different types of environment is implemented. Agent is trained by giving only screen pixels as input via a Convolutional Neural Network. Experimental results show that with the proposed method game completion rate and completion times are improved compared to random agent.

Original language	English
Title of host publication	26th IEEE Signal Processing and Communications Applications Conference, SIU 2018
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1-4
Number of pages	4
ISBN (Electronic)	9781538615010
DOIs	https://doi.org/10.1109/SIU.2018.8404596
Publication status	Published - 5 Jul 2018
Event	26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 - Izmir, Turkey Duration: 2 May 2018 → 5 May 2018

Publication series

Name	26th IEEE Signal Processing and Communications Applications Conference, SIU 2018

Conference

Conference	26th IEEE Signal Processing and Communications Applications Conference, SIU 2018
Country/Territory	Turkey
City	Izmir
Period	2/05/18 → 5/05/18

Bibliographical note

Publisher Copyright:
© 2018 IEEE.

Keywords

Convolutional Neural Networks
Game-playing AI
Q-learning
Reinforcement Learning

Access to Document

10.1109/SIU.2018.8404596

Cite this

Özgen, A. C., Fasounaki, M., & Ekenel, H. K. (2018). Generalized circle agent for geometry friends using deep reinforcement learning. In 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 (pp. 1-4). (26th IEEE Signal Processing and Communications Applications Conference, SIU 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SIU.2018.8404596

Özgen, Azmi Can ; Fasounaki, Mandana ; Ekenel, Hazim Kemal. / Generalized circle agent for geometry friends using deep reinforcement learning. 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 1-4 (26th IEEE Signal Processing and Communications Applications Conference, SIU 2018).

@inproceedings{901179f7d2524dc3b52574ea2292e702,

title = "Generalized circle agent for geometry friends using deep reinforcement learning",

abstract = "Reinforcement learning began to perform at human-level success in game intelligence after deep learning revolution. Geometry Friends is a puzzle game, where we can benefit from deep learning and expect to have successful game playing agents. In the game, agents are collecting targets in two dimensional environment and they try to overcome obstacles in the way. In this paper, Q-learning approach is applied to this game and a generalized circle agent for different types of environment is implemented. Agent is trained by giving only screen pixels as input via a Convolutional Neural Network. Experimental results show that with the proposed method game completion rate and completion times are improved compared to random agent.",

keywords = "Convolutional Neural Networks, Game-playing AI, Q-learning, Reinforcement Learning",

author = "{\"O}zgen, {Azmi Can} and Mandana Fasounaki and Ekenel, {Hazim Kemal}",

note = "Publisher Copyright: {\textcopyright} 2018 IEEE.; 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018 ; Conference date: 02-05-2018 Through 05-05-2018",

year = "2018",

month = jul,

day = "5",

doi = "10.1109/SIU.2018.8404596",

language = "English",

series = "26th IEEE Signal Processing and Communications Applications Conference, SIU 2018",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1--4",

booktitle = "26th IEEE Signal Processing and Communications Applications Conference, SIU 2018",

address = "United States",

}

Özgen, AC, Fasounaki, M & Ekenel, HK 2018, Generalized circle agent for geometry friends using deep reinforcement learning. in 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018. 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018, Institute of Electrical and Electronics Engineers Inc., pp. 1-4, 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018, Izmir, Turkey, 2/05/18. https://doi.org/10.1109/SIU.2018.8404596

Generalized circle agent for geometry friends using deep reinforcement learning. / Özgen, Azmi Can; Fasounaki, Mandana; Ekenel, Hazim Kemal.
26th IEEE Signal Processing and Communications Applications Conference, SIU 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 1-4 (26th IEEE Signal Processing and Communications Applications Conference, SIU 2018).

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

TY - GEN

T1 - Generalized circle agent for geometry friends using deep reinforcement learning

AU - Özgen, Azmi Can

AU - Fasounaki, Mandana

AU - Ekenel, Hazim Kemal

PY - 2018/7/5

Y1 - 2018/7/5

N2 - Reinforcement learning began to perform at human-level success in game intelligence after deep learning revolution. Geometry Friends is a puzzle game, where we can benefit from deep learning and expect to have successful game playing agents. In the game, agents are collecting targets in two dimensional environment and they try to overcome obstacles in the way. In this paper, Q-learning approach is applied to this game and a generalized circle agent for different types of environment is implemented. Agent is trained by giving only screen pixels as input via a Convolutional Neural Network. Experimental results show that with the proposed method game completion rate and completion times are improved compared to random agent.

AB - Reinforcement learning began to perform at human-level success in game intelligence after deep learning revolution. Geometry Friends is a puzzle game, where we can benefit from deep learning and expect to have successful game playing agents. In the game, agents are collecting targets in two dimensional environment and they try to overcome obstacles in the way. In this paper, Q-learning approach is applied to this game and a generalized circle agent for different types of environment is implemented. Agent is trained by giving only screen pixels as input via a Convolutional Neural Network. Experimental results show that with the proposed method game completion rate and completion times are improved compared to random agent.

KW - Convolutional Neural Networks

KW - Game-playing AI

KW - Q-learning

KW - Reinforcement Learning

UR - http://www.scopus.com/inward/record.url?scp=85050810609&partnerID=8YFLogxK

U2 - 10.1109/SIU.2018.8404596

DO - 10.1109/SIU.2018.8404596

M3 - Conference contribution

AN - SCOPUS:85050810609

T3 - 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018

SP - 1

EP - 4

BT - 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018

Y2 - 2 May 2018 through 5 May 2018

ER -

Özgen AC, Fasounaki M, Ekenel HK. Generalized circle agent for geometry friends using deep reinforcement learning. In 26th IEEE Signal Processing and Communications Applications Conference, SIU 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 1-4. (26th IEEE Signal Processing and Communications Applications Conference, SIU 2018). doi: 10.1109/SIU.2018.8404596

Generalized circle agent for geometry friends using deep reinforcement learning

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this