Learning opening books in partially observable games: Using random seeds in Phantom Go

Tristan CAZENAVE; Jialin LIU; Fabien TEYTAUD; Olivier TEYTAUD

doi:10.1109/CIG.2016.7860389

Learning opening books in partially observable games: Using random seeds in Phantom Go

Tristan CAZENAVE, Jialin LIU, Fabien TEYTAUD, Olivier TEYTAUD

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Referred Conference Paper › peer-review

4 Citations (Scopus)

Abstract

Many artificial intelligences (AIs) are randomized. One can be lucky or unlucky with the random seed; we quantify this effect and show that, maybe contrarily to intuition, this is far from being negligible. Then, we apply two different existing algorithms for selecting good seeds and good probability distributions over seeds. This mainly leads to learning an opening book. We apply this to Phantom Go, which, as all phantom games, is hard for opening book learning. We improve the winning rate from 50% to 70% in 5×5 against the same AI, and from approximately 0% to 40% in 5×5, 7×7 and 9×9 against a stronger (learning) opponent.

Original language	English
Title of host publication	Proceedings of the 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016
Publisher	IEEE Computer Society
Number of pages	7
ISBN (Electronic)	9781509018833
DOIs	https://doi.org/10.1109/CIG.2016.7860389
Publication status	Published - 2016
Externally published	Yes
Event	2016 IEEE Conference on Computational Intelligence and Games, CIG 2016 - Santorini, Greece Duration: 20 Sept 2016 → 23 Sept 2016

Publication series

Name	IEEE Conference on Computatonal Intelligence and Games, CIG
Volume	2016
ISSN (Print)	2325-4270
ISSN (Electronic)	2325-4289

Conference

Conference	2016 IEEE Conference on Computational Intelligence and Games, CIG 2016
Country/Territory	Greece
City	Santorini
Period	20/09/16 → 23/09/16

Bibliographical note

Publisher Copyright:
© 2016 IEEE.

Access to Document

10.1109/CIG.2016.7860389

Cite this

CAZENAVE, T., LIU, J., TEYTAUD, F., & TEYTAUD, O. (2016). Learning opening books in partially observable games: Using random seeds in Phantom Go. In Proceedings of the 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016 Article 7860389 (IEEE Conference on Computatonal Intelligence and Games, CIG; Vol. 2016). IEEE Computer Society. https://doi.org/10.1109/CIG.2016.7860389

@inproceedings{941d1c39b9b643aa8ef0e8f1efa91360,

title = "Learning opening books in partially observable games: Using random seeds in Phantom Go",

abstract = "Many artificial intelligences (AIs) are randomized. One can be lucky or unlucky with the random seed; we quantify this effect and show that, maybe contrarily to intuition, this is far from being negligible. Then, we apply two different existing algorithms for selecting good seeds and good probability distributions over seeds. This mainly leads to learning an opening book. We apply this to Phantom Go, which, as all phantom games, is hard for opening book learning. We improve the winning rate from 50% to 70% in 5×5 against the same AI, and from approximately 0% to 40% in 5×5, 7×7 and 9×9 against a stronger (learning) opponent.",

author = "Tristan CAZENAVE and Jialin LIU and Fabien TEYTAUD and Olivier TEYTAUD",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016 ; Conference date: 20-09-2016 Through 23-09-2016",

year = "2016",

doi = "10.1109/CIG.2016.7860389",

language = "English",

series = "IEEE Conference on Computatonal Intelligence and Games, CIG",

publisher = "IEEE Computer Society",

booktitle = "Proceedings of the 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016",

address = "United States",

}

CAZENAVE, T, LIU, J, TEYTAUD, F & TEYTAUD, O 2016, Learning opening books in partially observable games: Using random seeds in Phantom Go. in Proceedings of the 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016., 7860389, IEEE Conference on Computatonal Intelligence and Games, CIG, vol. 2016, IEEE Computer Society, 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016, Santorini, Greece, 20/09/16. https://doi.org/10.1109/CIG.2016.7860389

Learning opening books in partially observable games: Using random seeds in Phantom Go. / CAZENAVE, Tristan; LIU, Jialin; TEYTAUD, Fabien et al.
Proceedings of the 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016. IEEE Computer Society, 2016. 7860389 (IEEE Conference on Computatonal Intelligence and Games, CIG; Vol. 2016).

Research output: Book Chapters | Papers in Conference Proceedings › Conference paper (refereed) › Referred Conference Paper › peer-review

TY - GEN

T1 - Learning opening books in partially observable games: Using random seeds in Phantom Go

AU - CAZENAVE, Tristan

AU - LIU, Jialin

AU - TEYTAUD, Fabien

AU - TEYTAUD, Olivier

PY - 2016

Y1 - 2016

N2 - Many artificial intelligences (AIs) are randomized. One can be lucky or unlucky with the random seed; we quantify this effect and show that, maybe contrarily to intuition, this is far from being negligible. Then, we apply two different existing algorithms for selecting good seeds and good probability distributions over seeds. This mainly leads to learning an opening book. We apply this to Phantom Go, which, as all phantom games, is hard for opening book learning. We improve the winning rate from 50% to 70% in 5×5 against the same AI, and from approximately 0% to 40% in 5×5, 7×7 and 9×9 against a stronger (learning) opponent.

AB - Many artificial intelligences (AIs) are randomized. One can be lucky or unlucky with the random seed; we quantify this effect and show that, maybe contrarily to intuition, this is far from being negligible. Then, we apply two different existing algorithms for selecting good seeds and good probability distributions over seeds. This mainly leads to learning an opening book. We apply this to Phantom Go, which, as all phantom games, is hard for opening book learning. We improve the winning rate from 50% to 70% in 5×5 against the same AI, and from approximately 0% to 40% in 5×5, 7×7 and 9×9 against a stronger (learning) opponent.

UR - http://www.scopus.com/inward/record.url?scp=85015408729&partnerID=8YFLogxK

U2 - 10.1109/CIG.2016.7860389

DO - 10.1109/CIG.2016.7860389

M3 - Conference paper (refereed)

AN - SCOPUS:85015408729

T3 - IEEE Conference on Computatonal Intelligence and Games, CIG

BT - Proceedings of the 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016

PB - IEEE Computer Society

T2 - 2016 IEEE Conference on Computational Intelligence and Games, CIG 2016

Y2 - 20 September 2016 through 23 September 2016

ER -

Learning opening books in partially observable games: Using random seeds in Phantom Go

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this