conferencePaper
2014 IEEE Conference on Computational Intelligence and Games
DOI 10.1109/cig.2014.6932861
IEEE
Guhe
Markus
Lascarides
Alex
The effectiveness of persuasion in The Settlers of Catan
August 2014
https://doi.org/10.1109%2Fcig.2014.6932861
attachment
Submitted Version
https://www.pure.ed.ac.uk/ws/files/19353900/CIG2014.pdf
2020-07-20 18:34:57
1
application/pdf
journalArticle
10
International Journal of Gaming and Computer-Mediated Simulations
DOI 10.4018/ijgcms.2018040103
2
Boda
Márton Attila
Avoiding Revenge Using Optimal Opponent Ranking Strategy in the Board Game Catan
April 2018
https://doi.org/10.4018%2Fijgcms.2018040103
Publisher: IGI Global
47–70
attachment
Full Text
https://sci-hub.se/downloads/2020-05-28/3d/boda2018.pdf#view=FitH
2020-07-20 18:22:11
1
application/pdf
conferencePaper
2014 IEEE Conference on Computational Intelligence and Games
DOI 10.1109/cig.2014.6932884
IEEE
Guhe
Markus
Lascarides
Alex
Game strategies for The Settlers of Catan
August 2014
https://doi.org/10.1109%2Fcig.2014.6932884
attachment
Submitted Version
https://www.pure.ed.ac.uk/ws/files/19351482/CIG2014_GS.pdf
2020-07-20 18:24:09
1
application/pdf
bookSection
Lecture Notes in Computer Science
Springer Berlin Heidelberg
Szita
István
Chaslot
Guillaume
Spronck
Pieter
Monte-Carlo Tree Search in Settlers of Catan
2010
https://doi.org/10.1007%2F978-3-642-12993-3_3
DOI: 10.1007/978-3-642-12993-3_3
21–32
attachment
Full Text
https://zero.sci-hub.se/5140/3f6b582d932254ee1b7d29e6e9683934/szita2010.pdf#view=FitH
2020-07-20 18:29:58
1
application/pdf
bookSection
Multi-Agent Systems
Springer International Publishing
Xenou
Konstantia
Chalkiadakis
Georgios
Afantenos
Stergos
Deep Reinforcement Learning in Strategic Board Game Environments
2019
https://doi.org/10.1007%2F978-3-030-14174-5_16
DOI: 10.1007/978-3-030-14174-5_16
233–248
attachment
Accepted Version
https://oatao.univ-toulouse.fr/22647/1/xenou_22647.pdf
2020-07-20 18:10:35
1
application/pdf
journalArticle
41
Journal of the Operational Research Society
DOI 10.1057/jors.1990.2
1
Maliphant
Sarah A.
Smith
David K.
Mini-Risk: Strategies for a Simplified Board Game
January 1990
https://doi.org/10.1057%2Fjors.1990.2
Publisher: Informa UK Limited
9–16
attachment
Full Text
https://zero.sci-hub.se/4681/0e142dbe029d345411eb5019cea0b10a/maliphant1990.pdf#view=FitH
2020-07-20 18:28:37
1
application/pdf
conferencePaper
Proceedings of the 2002 ACM symposium on Applied computing - SAC \textquotesingle02
DOI 10.1145/508791.508904
ACM Press
Neves
Atila
Brasāo
Osvaldo
Rosa
Agostinho
Learning the risk board game with classifier systems
2002
https://doi.org/10.1145%2F508791.508904
attachment
Full Text
https://dacemirror.sci-hub.se/proceedings-article/f9ce3c906d4e89b8aa3b90f15f0dfe20/neves2002.pdf#view=FitH
2020-07-20 18:26:40
1
application/pdf
journalArticle
70
Mathematics Magazine
DOI 10.1080/0025570x.1997.11996573
5
Tan
Bariş
Markov Chains and the RISK Board Game
December 1997
https://doi.org/10.1080%2F0025570x.1997.11996573
Publisher: Informa UK Limited
349–357
attachment
Full Text
https://twin.sci-hub.se/6853/3bdc3204e08f60618dca66f19b9cd1fc/markov-chains-and-the-risk-board-game-1997.pdf#view=FitH
2020-07-20 18:28:15
1
application/pdf
journalArticle
76
Mathematics Magazine
DOI 10.1080/0025570x.2003.11953165
2
Osborne
Jason A.
Markov Chains for the RISK Board Game Revisited
April 2003
https://doi.org/10.1080%2F0025570x.2003.11953165
Publisher: Informa UK Limited
129–135
attachment
Full Text
https://twin.sci-hub.se/6908/ad9e3c21b4a5edae31079e43ad12c8ce/osborne2003.pdf#view=FitH
2020-07-20 18:28:23
1
application/pdf
journalArticle
9
IEEE Trans. Evol. Computat.
DOI 10.1109/tevc.2005.856211
6
Vaccaro
J. M.
Guest
C. C.
Planning an Endgame Move Set for the Game RISK: A Comparison of Search Algorithms
December 2005
https://doi.org/10.1109%2Ftevc.2005.856211
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
641–652
attachment
Full Text
https://moscow.sci-hub.se/1819/0e61163e1173d174a5261879afc2c42d/vaccaro2005.pdf#view=FitH
2020-07-20 18:31:44
1
application/pdf
conferencePaper
2018 IEEE Conference on Computational Intelligence and Games (CIG)
DOI 10.1109/cig.2018.8490419
IEEE
Gedda
Magnus
Lagerkvist
Mikael Z.
Butler
Martin
Monte Carlo Methods for the Game Kingdomino
August 2018
https://doi.org/10.1109%2Fcig.2018.8490419
attachment
Submitted Version
https://arxiv.org/pdf/1807.04458
2020-07-20 18:29:37
1
application/pdf
journalArticle
88
Mathematics Magazine
DOI 10.4169/math.mag.88.5.323
5
Cox
Christopher
Silva
Jessica De
Deorsey
Philip
Kenter
Franklin H. J.
Retter
Troy
Tobin
Josh
How to Make the Perfect Fireworks Display: Two Strategies forHanabi
December 2015
https://doi.org/10.4169%2Fmath.mag.88.5.323
Publisher: Informa UK Limited
323–336
attachment
Full Text
https://moscow.sci-hub.se/5019/aae1c968ecb4576818556c669a20e535/christophercox2015.pdf#view=FitH
2020-07-20 18:26:02
1
application/pdf
conferencePaper
2017 IEEE Congress on Evolutionary Computation (CEC)
DOI 10.1109/cec.2017.7969465
IEEE
Walton-Rivers
Joseph
Williams
Piers R.
Bartle
Richard
Perez-Liebana
Diego
Lucas
Simon M.
Evaluating and modelling Hanabi-playing agents
June 2017
https://doi.org/10.1109%2Fcec.2017.7969465
attachment
Accepted Version
https://repository.essex.ac.uk/20341/1/1704.07069v1.pdf
2020-07-20 18:16:01
1
application/pdf
journalArticle
280
Artificial Intelligence
DOI 10.1016/j.artint.2019.103216
Bard
Nolan
Foerster
Jakob N.
Chandar
Sarath
Burch
Neil
Lanctot
Marc
Song
H. Francis
Parisotto
Emilio
Dumoulin
Vincent
Moitra
Subhodeep
Hughes
Edward
Dunning
Iain
Mourad
Shibl
Larochelle
Hugo
Bellemare
Marc G.
Bowling
Michael
The Hanabi challenge: A new frontier for AI research
March 2020
https://doi.org/10.1016%2Fj.artint.2019.103216
Publisher: Elsevier BV
103216
attachment
Full Text
https://pdf.sciencedirectassets.com/271585/1-s2.0-S0004370219X00120/1-s2.0-S0004370219300116/main.pdf?X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHoaCXVzLWVhc3QtMSJGMEQCIH1Rh0U0FaCQrRRyUSiT86nBmiJQoMaX1VTfIk3CSCBuAiBVtUw9mFSXO4P2IR2s2Cqs5kI4CwPLTkYQLjfgja598Sq0AwgzEAMaDDA1OTAwMzU0Njg2NSIMyb025En2S0rT1ZZeKpEDm01bDtAwa5C41YjSxTlisMeVuPdMt%2Fpp3ln9ZE3BEsM0TrGG0EvzEx3c85DHMJoG3oK32WnIGWwPieFnqEZGQJ5kP%2BEtnPQMUzxY7WXMipZQ%2Fmkao3oL%2Fu%2BbKsSHJ%2BAolhMh0G%2F1YvpVoepC6rGy8rku6DXcS0XWgvUzcoLJlcPRRsF5pGQp6xFOR1RW2pR6oq9UoJWEImFDw6X2g2MgPaFMRXMo71FNVX7Zix2oWGhaDRS0hMEerYFvmj1Lv0rmQbAo2h0lvvTZkGWrcauFFEjILxJFadqwcK4Xfe%2BfFDR2H61VZ7B792qzjSC8vCsAToK8BSVepuCnVpDM04cKPnrsiqtR2WMuCMYlS2w%2BLRCk03EjXQvU8ZR8J63MmEPpbJ7pS6JnI%2B0nCcCYtvb7yqcWTzmHjPN6ssNUpPX1ajjjlZyaFm5ntqyCxL0tYar6ra3TkmF2Lbesk%2F1wLTuqnLhsIEmzB01wzw%2BEd%2BxW7PZnDHKM%2FeBPaeg1Z2VaL%2BuwDGWMhBnk8O2sUafc2qT%2BRzMwuLHX%2BAU67AH1wVDVhIqhENwF1N9Pv%2BLhicrQXiHJ79HC1JGf9ulqBM9sLRnFjdyxRYUm6O%2F9RPSV6OTVARQGNQpBBqJUN5%2BCfvHvl%2BVgfEa3fTERLxkX1QBTsTGKAzKee1BjQylRYNnTLhUm0CV56l3jCBMG9LwpodFZgoMHURoqbkWDV9HkAS9W6LnIvcy4L7e76c3qrpB%2B1XoEzqX%2BpQ3S4lRCVR5NcTLnUmxww4%2Br8nzca5CibEM9l033knReVVZnjt2JZVzZVW6zXpO60OH%2F6W2N1qYvsY3oOLT7kX5w7GNCi4qrNANQLGrPXDDqLOuI2A%3D%3D&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20200720T183502Z&X-Amz-SignedHeaders=host&X-Amz-Expires=300&X-Amz-Credential=ASIAQ3PHCVTY6ZAVFJHH%2F20200720%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Signature=15beeda55689b6fd5e9d2d05b5ac322d5b98fdc5d4f70683cb530821a98eef3c&hash=9fe7e987a2d4bc206328d68796fc691838e0a0ff5373120293f9b5d82fefded1&host=68042c943591013ac2b2430a89b270f6af2c76d8dfd086a07176afe7c76c2c61&pii=S0004370219300116&tid=spdf-54a6ffb6-7f42-4d75-84e1-5f15f936d894&sid=76772d953af62548ae38ab517857dea18c60gxrqb&type=client
2020-07-20 18:35:12
1
application/pdf
conferencePaper
2019 IEEE Conference on Games (CoG)
DOI 10.1109/cig.2019.8848008
IEEE
Walton-Rivers
Joseph
Williams
Piers R.
Bartle
Richard
The 2018 Hanabi competition
August 2019
https://doi.org/10.1109%2Fcig.2019.8848008
attachment
Accepted Version
https://repository.essex.ac.uk/26898/2/hanabi.pdf
2020-07-20 18:34:35
1
application/pdf
conferencePaper
2019 IEEE Conference on Games (CoG)
DOI 10.1109/cig.2019.8847944
IEEE
Canaan
Rodrigo
Togelius
Julian
Nealen
Andy
Menzel
Stefan
Diverse Agents for Ad-Hoc Cooperation in Hanabi
August 2019
https://doi.org/10.1109%2Fcig.2019.8847944
attachment
Submitted Version
https://arxiv.org/pdf/1907.03840
2020-07-20 18:11:10
1
application/pdf
journalArticle
45
Mathematics Magazine
DOI 10.1080/0025570x.1972.11976187
1
Ash
Robert B.
Bishop
Richard L.
Monopoly as a Markov Process
January 1972
https://doi.org/10.1080%2F0025570x.1972.11976187
Publisher: Informa UK Limited
26–29
attachment
Submitted Version
https://www.math.uiuc.edu/%7Ebishop/monopoly.pdf
2020-07-20 18:37:15
1
application/pdf
journalArticle
4
IEEE Trans. Comput. Intell. AI Games
DOI 10.1109/tciaig.2012.2204883
4
Cowling
Peter I.
Ward
Colin D.
Powley
Edward J.
Ensemble Determinization in Monte Carlo Tree Search for the Imperfect Information Card Game Magic: The Gathering
December 2012
https://doi.org/10.1109%2Ftciaig.2012.2204883
Publisher: Institute of Electrical and Electronics Engineers (IEEE)
241–257
attachment
Accepted Version
https://eprints.whiterose.ac.uk/75050/1/EnsDetMagic.pdf
2020-07-20 18:14:45
1
application/pdf
journalArticle
31
The College Mathematics Journal
DOI 10.1080/07468342.2000.11974103
1
Bosch
Robert A.
Optimal Card-Collecting Strategies for Magic: The Gathering
January 2000
https://doi.org/10.1080%2F07468342.2000.11974103
Publisher: Informa UK Limited
15–21
attachment
Full Text
https://zero.sci-hub.se/6795/ba844bedd2d417e4393d7af19bb3dd47/bosch2000.pdf#view=FitH
2020-07-20 18:37:22
1
application/pdf
conferencePaper
2009 IEEE Symposium on Computational Intelligence and Games
DOI 10.1109/cig.2009.5286501
IEEE
Ward
C. D.
Cowling
P. I.
Monte Carlo search applied to card selection in Magic: The Gathering
September 2009
https://doi.org/10.1109%2Fcig.2009.5286501
attachment
Full Text
https://dacemirror.sci-hub.se/proceedings-article/dfcfc3f5502682650ac71b68af8f9b19/ward2009.pdf#view=FitH
2020-07-20 18:29:50
1
application/pdf
bookSection
Lecture Notes in Computer Science
Springer Berlin Heidelberg
Demaine
Erik D.
Demaine
Martin L.
Uehara
Ryuhei
Uno
Takeaki
Uno
Yushi
UNO Is Hard, Even for a Single Player
2010
https://doi.org/10.1007%2F978-3-642-13122-6_15
DOI: 10.1007/978-3-642-13122-6_15
133–144
attachment
Submitted Version
https://dspace.mit.edu/bitstream/1721.1/62147/1/Demaine_UNO%20is.pdf
2020-07-20 18:36:18
1
application/pdf
journalArticle
Information Processing Letters
DOI 10.1016/j.ipl.2020.105995
Mishiba
Shohei
Takenaga
Yasuhiko
QUIXO is EXPTIME-complete
July 2020
https://doi.org/10.1016%2Fj.ipl.2020.105995
Publisher: Elsevier BV
105995
attachment
Full Text
https://pdf.sciencedirectassets.com/271527/AIP/1-s2.0-S002001902030082X/main.pdf?X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHoaCXVzLWVhc3QtMSJGMEQCIFtkSBpgjQCGD2t9HhDGKXByeuMLxjq5SpZiHiVJRtD2AiBGNfyOHc5rhR9YOWBJqfm4Q4sk9A7DiAYQK4bE21l10yq0AwgyEAMaDDA1OTAwMzU0Njg2NSIMTbV17TBAQ%2BDOw4NqKpEDFYUi3wRhC%2Baj5%2FaTwaaOsbSSQ1WVXW9J%2FkDFGuUgFScfYqdG0aRaazztSFianGDgj1FEpVC%2FwLMP8LEWFghexDo2fLhZpoaNA5v8DQIvrvb839ZJGlCB9HEcbyeStsLWWrl8pM1lYBckbsmir72eSqxkPqyFfdxni2pG4HcCVuJHe6pPJwoGPGoTndv1mCghHzuk9rvPiegQ9iaKu947uL9xnhB1c7TzMUf2EGPeKuB2jm4F5duW8V3IzqQjPf3tMPSRNn8Ztv1qlO8vUhpXTsyI5dH%2BURZTqOVp0fVn4En6CRNrkv05g%2B1rxq6b6gQmlfUeAIPaTwUfI2glYGtZKNkvlkYrZKoWvHkv9XzLd3%2FKiuaeKxM9nk4hZjJqtcWwaD3Gp9yr63IUPqUZW5BI2YJHNW%2B9SIbRzBmubE0b01LVFubW9rJo3hPtgKRHPpIEIm0j%2FjoszFdpyL4chFaML0HxrCmQeh7HkvJBMvERUM%2F882V%2FBm2zHRuKsPNpLUj%2BJ%2BDh6%2FQaE0pmoIOfTbNN8xIwnaLX%2BAU67AFOxpNyX5biC5h3HLfyBGY1KrsDmnyo3bcOIqAwepos4Dw%2BlQ8II9AjVQwzEJGvd8LQ9sYaZzntH7rnuZG15wBizwUDkD2G37c91hT%2BG9hKPKHkW9jZp6XijVMHYWhd34TF6iW%2BQSM5bMzKdQaXzoilRts%2B5DaLCeYk%2Fzc2FFcSMOT3pXBXWHr%2F16lr5Sp8Gh9FS9HnwI2O8pxy6E1lqGM0wP%2FwaWBT3HgdR2tvjQzn%2BcjDDKHABtqj3oo6janO3IKOPaFzYHFqiL0DS8Pet5gUVYynm9m37o5M3%2B6y7YoaXnoq1o1goiX6Zv0S5A%3D%3D&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20200720T183148Z&X-Amz-SignedHeaders=host&X-Amz-Expires=300&X-Amz-Credential=ASIAQ3PHCVTYWL2N32C3%2F20200720%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Signature=99c8222a858071c0b65fa6dac4ea0e1b548c1a4eb23ccac06cf864c7013a7593&hash=d2c7df4c2396ff204caa66ac0553e18f1b2712399c07b6e674199859ccc1b7f9&host=68042c943591013ac2b2430a89b270f6af2c76d8dfd086a07176afe7c76c2c61&pii=S002001902030082X&tid=spdf-742da070-9da3-43f1-861e-f733776129ef&sid=76772d953af62548ae38ab517857dea18c60gxrqb&type=client
2020-07-20 18:31:53
1
application/pdf
bookSection
Case-Based Reasoning Research and Development
Springer International Publishing
Woolford
Michael
Watson
Ian
SCOUT: A Case-Based Reasoning Agent for Playing Race for the Galaxy
2017
https://doi.org/10.1007%2F978-3-319-61030-6_27
DOI: 10.1007/978-3-319-61030-6_27
390–402
attachment
Woolford and Watson - 2017 - SCOUT A Case-Based Reasoning Agent for Playing Ra.pdf
application/pdf
journalArticle
85
Mathematics Magazine
DOI 10.4169/math.mag.85.2.083
2
Coleman
Ben
Hartshorn
Kevin
Game, Set, Math
April 2012
https://doi.org/10.4169%2Fmath.mag.85.2.083
Publisher: Informa UK Limited
83–96
attachment
Full Text
https://dacemirror.sci-hub.se/journal-article/768dabc67f6adcaa34a4c087b56b4283/game-set-math-2012.pdf#view=FitH
2020-07-20 18:24:32
1
application/pdf
journalArticle
125
The American Mathematical Monthly
DOI 10.1080/00029890.2018.1412661
3
Glass
Darren
The Joy of SET
February 2018
https://doi.org/10.1080%2F00029890.2018.1412661
Publisher: Informa UK Limited
284–288
attachment
Full Text
https://twin.sci-hub.se/6684/b949dbda2e3438aae344825abb7d0ff3/glass2018.pdf#view=FitH
2020-07-20 18:35:34
1
application/pdf
bookSection
Communications in Computer and Information Science
Springer Berlin Heidelberg
Lazarusli
Irene A.
Lukas
Samuel
Widjaja
Patrick
Implementation of Artificial Intelligence with 3 Different Characters of AI Player on “Monopoly Deal” Computer Game
2015
https://doi.org/10.1007%2F978-3-662-46742-8_11
DOI: 10.1007/978-3-662-46742-8_11
119–127
bookSection
Computers and Games
Springer Berlin Heidelberg
Pawlewicz
Jakub
Nearly Optimal Computer Play in Multi-player Yahtzee
2011
https://doi.org/10.1007%2F978-3-642-17928-0_23
DOI: 10.1007/978-3-642-17928-0_23
250–262
conferencePaper
2007 IEEE Symposium on Computational Intelligence and Games
DOI 10.1109/cig.2007.368089
IEEE
Glenn
James R.
Computer Strategies for Solitaire Yahtzee
2007
https://doi.org/10.1109%2Fcig.2007.368089
attachment
Submitted Version
http://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=707B7E09A9652A1E4F2AB4BF608C410A?doi=10.1.1.111.1724&rep=rep1&type=pdf
2020-07-20 18:09:04
1
application/pdf
journalArticle
18
Expert Systems
DOI 10.1111/1468-0394.00160
2
Maynard
Ken
Moss
Patrick
Whitehead
Marcus
Narayanan
S.
Garay
Matt
Brannon
Nathan
Kantamneni
Raj Gopal
Kustra
Todd
Modeling expert problem solving in a game of chance: a Yahtzeec case study
May 2001
https://doi.org/10.1111%2F1468-0394.00160
Publisher: Wiley
88–98
attachment
Full Text
https://cyber.sci-hub.se/MTAuMTExMS8xNDY4LTAzOTQuMDAxNjA=/maynard2001.pdf#view=FitH
2020-07-20 18:29:00
1
application/pdf
bookSection
Computers and Games
Springer International Publishing
Oka
Kazuto
Matsuzaki
Kiminori
Systematic Selection of N-Tuple Networks for 2048
2016
https://doi.org/10.1007%2F978-3-319-50935-8_8
DOI: 10.1007/978-3-319-50935-8_8
81–92
attachment
Full Text
https://sci-hub.se/downloads/2020-05-25/5f/oka2016.pdf#view=FitH
2020-07-20 18:32:30
1
application/pdf
conferencePaper
2016 Conference on Technologies and Applications of Artificial Intelligence (TAAI)
DOI 10.1109/taai.2016.7880154
IEEE
Matsuzaki
Kiminori
Systematic selection of N-tuple networks with consideration of interinfluence for game 2048
November 2016
https://doi.org/10.1109%2Ftaai.2016.7880154
attachment
Full Text
https://twin.sci-hub.se/6299/d9bbecbbec212dab7fe6e6a67213b1cb/matsuzaki2016.pdf#view=FitH
2020-07-20 18:32:39
1
application/pdf
conferencePaper
2014 IEEE Conference on Computational Intelligence and Games
DOI 10.1109/cig.2014.6932920
IEEE
Rodgers
Philip
Levine
John
An investigation into 2048 AI strategies
August 2014
https://doi.org/10.1109%2Fcig.2014.6932920
attachment
Full Text
https://zero.sci-hub.se/3377/2e196ce6e3cb06a636bf1ffdee8f5b6f/rodgers2014.pdf#view=FitH
2020-07-20 18:21:23
1
application/pdf
journalArticle
arxiv:2006.04635
Anthony
Thomas
Eccles
Tom
Tacchetti
Andrea
Kramár
János
Gemp
Ian
Hudson
Thomas C.
Porcel
Nicolas
Lanctot
Marc
Pérolat
Julien
Everett
Richard
Singh
Satinder
Graepel
Thore
Bachrach
Yoram
Learning to Play No-Press Diplomacy with Best Response Policy Iteration
2020
http://arxiv.org/abs/2006.04635v2
attachment
Full Text
https://arxiv.org/pdf/2006.04635v2.pdf
2020-07-20 18:27:18
1
application/pdf
journalArticle
arxiv:1909.02128
Paquette
Philip
Lu
Yuchen
Bocco
Steven
Smith
Max O.
Ortiz-Gagne
Satya
Kummerfeld
Jonathan K.
Singh
Satinder
Pineau
Joelle
Courville
Aaron
No Press Diplomacy: Modeling Multi-Agent Gameplay
2019
http://arxiv.org/abs/1909.02128v2
attachment
Full Text
https://arxiv.org/pdf/1909.02128v2.pdf
2020-07-20 18:31:04
1
application/pdf
journalArticle
arxiv:1902.06996
Tan
Hao Hao
Agent Madoff: A Heuristic-Based Negotiation Agent For The Diplomacy Strategy Game
2019
http://arxiv.org/abs/1902.06996v1
attachment
Full Text
https://arxiv.org/pdf/1902.06996v1.pdf
2020-07-20 18:21:06
1
application/pdf
journalArticle
arxiv:1807.04458
Gedda
Magnus
Lagerkvist
Mikael Z.
Butler
Martin
Monte Carlo Methods for the Game Kingdomino
2018
http://arxiv.org/abs/1807.04458v2
attachment
Full Text
https://arxiv.org/pdf/1807.04458v2.pdf
2020-07-20 18:29:18
1
application/pdf
journalArticle
arxiv:1909.02849
Nguyen
Viet-Ha
Perrot
Kevin
Vallet
Mathieu
NP-completeness of the game Kingdomino
2019
http://arxiv.org/abs/1909.02849v3
attachment
Full Text
https://arxiv.org/pdf/1909.02849v3.pdf
2020-07-20 18:31:12
1
application/pdf
journalArticle
arxiv:1912.02318
Lerer
Adam
Hu
Hengyuan
Foerster
Jakob
Brown
Noam
Improving Policies via Search in Cooperative Partially Observable Games
2019
http://arxiv.org/abs/1912.02318v1
attachment
Full Text
https://arxiv.org/pdf/1912.02318v1.pdf
2020-07-20 18:26:28
1
application/pdf
journalArticle
arxiv:1603.01911
Baffier
Jean-Francois
Chiu
Man-Kwun
Diez
Yago
Korman
Matias
Mitsou
Valia
Renssen
André van
Roeloffzen
Marcel
Uno
Yushi
Hanabi is NP-hard, Even for Cheaters who Look at Their Cards
2016
http://arxiv.org/abs/1603.01911v3
attachment
Full Text
https://arxiv.org/pdf/1603.01911v3.pdf
2020-07-20 18:25:31
1
application/pdf
journalArticle
arxiv:2004.13710
Canaan
Rodrigo
Gao
Xianbo
Togelius
Julian
Nealen
Andy
Menzel
Stefan
Generating and Adapting to Diverse Ad-Hoc Cooperation Agents in Hanabi
2020
http://arxiv.org/abs/2004.13710v2
attachment
Full Text
https://arxiv.org/pdf/2004.13710v2.pdf
2020-07-20 18:25:19
1
application/pdf
journalArticle
arxiv:2004.13291
Canaan
Rodrigo
Gao
Xianbo
Chung
Youjin
Togelius
Julian
Nealen
Andy
Menzel
Stefan
Evaluating the Rainbow DQN Agent in Hanabi with Unseen Partners
2020
http://arxiv.org/abs/2004.13291v1
attachment
Full Text
https://arxiv.org/pdf/2004.13291v1.pdf
2020-07-20 18:22:45
1
application/pdf
journalArticle
arxiv:2003.05119
Biderman
Stella
Magic: the Gathering is as Hard as Arithmetic
2020
http://arxiv.org/abs/2003.05119v1
attachment
Full Text
https://arxiv.org/pdf/2003.05119v1.pdf
2020-07-20 18:27:42
1
application/pdf
journalArticle
arxiv:1904.09828
Churchill
Alex
Biderman
Stella
Herrick
Austin
Magic: The Gathering is Turing Complete
2019
http://arxiv.org/abs/1904.09828v2
attachment
Full Text
https://arxiv.org/pdf/1904.09828v2.pdf
2020-07-20 18:27:51
1
application/pdf
journalArticle
arxiv:1810.03744
Zilio
Felipe
Prates
Marcelo
Neural Networks Models for Analyzing Magic: the Gathering Cards
2018
http://arxiv.org/abs/1810.03744v1
attachment
Full Text
https://arxiv.org/pdf/1810.03744v1.pdf
2020-07-20 18:30:42
1
application/pdf
conferencePaper
Proceedings of the 2020 4th International Conference on Intelligent Systems, Metaheuristics & Swarm Intelligence
DOI 10.1145/3396474.3396492
ACM
Grichshenko
Alexandr
Araújo
Luiz Jonatã Pires de
Gimaeva
Susanna
Brown
Joseph Alexander
Using Tabu Search Algorithm for Map Generation in the Terra Mystica Tabletop Game
March 2020
https://doi.org/10.1145%2F3396474.3396492
attachment
Submitted Version
https://arxiv.org/pdf/2006.02716
2020-07-20 18:36:31
1
application/pdf
journalArticle
arxiv:1009.1031
Migdał
Piotr
A mathematical model of the Mafia game
2010
http://arxiv.org/abs/1009.1031v3
attachment
Full Text
https://arxiv.org/pdf/1009.1031v3.pdf
2020-07-20 18:20:44
1
application/pdf
journalArticle
arxiv:1003.2851
Demaine
Erik D.
Demaine
Martin L.
Harvey
Nicholas J. A.
Uehara
Ryuhei
Uno
Takeaki
Uno
Yushi
The complexity of UNO
2010
http://arxiv.org/abs/1003.2851v3
attachment
Full Text
https://arxiv.org/pdf/1003.2851v3.pdf
2020-07-20 18:34:43
1
application/pdf
journalArticle
arxiv:1603.00928
Almanza
Matteo
Leucci
Stefano
Panconesi
Alessandro
Trainyard is NP-Hard
2016
http://arxiv.org/abs/1603.00928v1
attachment
Full Text
https://arxiv.org/pdf/1603.00928v1.pdf
2020-07-20 18:36:08
1
application/pdf
journalArticle
arxiv:1505.04274
Langerman
Stefan
Uno
Yushi
Threes!, Fives, 1024!, and 2048 are Hard
2015
http://arxiv.org/abs/1505.04274v1
attachment
Full Text
https://arxiv.org/pdf/1505.04274v1.pdf
2020-07-20 18:35:46
1
application/pdf
journalArticle
arxiv:1804.07396
Eppstein
David
Making Change in 2048
2018
http://arxiv.org/abs/1804.07396v1
attachment
Full Text
https://arxiv.org/pdf/1804.07396v1.pdf
2020-07-20 18:28:01
1
application/pdf
journalArticle
arxiv:1804.07393
Das
Madhuparna
Paul
Goutam
Analysis of the Game "2048" and its Generalization in Higher Dimensions
2018
http://arxiv.org/abs/1804.07393v2
attachment
Full Text
https://arxiv.org/pdf/1804.07393v2.pdf
2020-07-20 18:21:31
1
application/pdf
journalArticle
arxiv:1606.07374
Yeh
Kun-Hao
Wu
I.-Chen
Hsueh
Chu-Hsuan
Chang
Chia-Chuan
Liang
Chao-Chin
Chiang
Han
Multi-Stage Temporal Difference Learning for 2048-like Games
2016
http://arxiv.org/abs/1606.07374v2
attachment
Full Text
https://arxiv.org/pdf/1606.07374v2.pdf
2020-07-20 18:30:19
1
application/pdf
journalArticle
arxiv:1408.6315
Mehta
Rahul
2048 is (PSPACE) Hard, but Sometimes Easy
2014
http://arxiv.org/abs/1408.6315v1
attachment
Full Text
https://arxiv.org/pdf/1408.6315v1.pdf
2020-07-20 18:20:36
1
application/pdf
computerProgram
Settlers of Catan bot trained using reinforcement learning
https://jonzia.github.io/Catan/
MATLAB
conferencePaper
34
Proceedings of the Annual Meeting of the Cognitive Science Society
Guhe
Markus
Lascarides
Alex
Trading in a multiplayer board game: Towards an analysis of non-cooperative dialogue
2012
https://escholarship.org/uc/item/9zt506xx
Issue: 34
attachment
Guhe and Lascarides - 2012 - Trading in a multiplayer board game Towards an an.pdf
application/pdf
journalArticle
POMCP with Human Preferencesin Settlers of Catan
https://www.aaai.org/ocs/index.php/AIIDE/AIIDE18/paper/viewFile/18091/17217
attachment
POMCP with Human Preferencesin Settlers of Catan.pdf
application/pdf
blogPost
The impact of loaded dice in Catan
https://izbicki.me/blog/how-to-cheat-at-settlers-of-catan-by-loading-the-dice-and-prove-it-with-p-values.html
journalArticle
Monte Carlo Tree Search in a Modern Board Game Framework
https://project.dke.maastrichtuniversity.nl/games/files/bsc/Roelofs_Bsc-paper.pdf
attachment
Full Text
https://project.dke.maastrichtuniversity.nl/games/files/bsc/Roelofs_Bsc-paper.pdf
2020-07-20 18:47:19
1
application/pdf
conferencePaper
Pfeiffer
Michael
Reinforcement Learning of Strategies for Settlers of Catan
2004
https://www.researchgate.net/publication/228728063_Reinforcement_learning_of_strategies_for_Settlers_of_Catan
attachment
Pfeiffer - 2004 - Reinforcement Learning of Strategies for Settlers .pdf
application/pdf
presentation
Michael Wolf
An Intelligent Artificial Player for the Game of Risk
20/04/2005
http://www.ke.tu-darmstadt.de/lehre/archiv/ss04/oberseminar/folien/Wolf_Michael-Slides.pdf
attachment
An Intelligent Artificial Player for the Game of R.pdf
application/pdf
journalArticle
RISKy Business: An In-Depth Look at the Game RISK
https://scholar.rose-hulman.edu/rhumj/vol3/iss2/3
attachment
RISKy Business An In-Depth Look at the Game RISK.pdf
application/pdf
journalArticle
RISK Board Game ‐ Battle Outcome Analysis
http://www.c4i.gr/xgeorgio/docs/RISK-board-game%20_rev-3.pdf
attachment
Full Text
https://www.c4i.gr/xgeorgio/docs/RISK-board-game%20_rev-3.pdf
2020-07-20 18:54:23
1
application/pdf
thesis
Blekinge Institute of Technology, School of Engineering, Department of Systems and Software Engineering.
Olsson
Fredrik
A multi-agent system for playing the board game risk
Risk is a game in which traditional Artificial-Intelligence methods such as for example iterative deepening and Alpha-Beta pruning can not successfully be applied due to the size of the search space. Distributed problem solving in the form of a multi-agent system might be the solution. This needs to be tested before it is possible to tell if a multi-agent system will be successful at playing Risk or not. In this thesis the development of a multi-agent system that plays Risk is explained. The system places an agent in every country on the board and uses a central agent for organizing communication. An auction mechanism is used for negotiation. The experiments show that a multi-agent solution indeed is a prosperous approach when developing a computer based player for the board game Risk.
2005
http://urn.kb.se/resolve?urn=urn:nbn:se:bth-3781
51
Independent thesis Advanced level (degree of Master (One Year))
attachment
Full Text
http://bth.diva-portal.org/smash/get/diva2:831093/FULLTEXT01
2021-07-24 08:26:48
1
application/pdf
attachment
Full Text
https://www.diva-portal.org/smash/get/diva2:831093/FULLTEXT01.pdf
2021-07-24 08:28:25
3
blogPost
State Representation and Polyomino Placement for the Game Patchwork
https://zayenz.se/blog/post/patchwork-modref2019-paper/
journalArticle
arXiv:2001.04233 [cs]
Lagerkvist
Mikael Zayenz
Computer Science - Artificial Intelligence
State Representation and Polyomino Placement for the Game Patchwork
Modern board games are a rich source of entertainment for many people, but also contain interesting and challenging structures for game playing research and implementing game playing agents. This paper studies the game Patchwork, a two player strategy game using polyomino tile drafting and placement. The core polyomino placement mechanic is implemented in a constraint model using regular constraints, extending and improving the model in (Lagerkvist, Pesant, 2008) with: explicit rotation handling; optional placements; and new constraints for resource usage. Crucial for implementing good game playing agents is to have great heuristics for guiding the search when faced with large branching factors. This paper divides placing tiles into two parts: a policy used for placing parts and an evaluation used to select among different placements. Policies are designed based on classical packing literature as well as common standard constraint programming heuristics. For evaluation, global propagation guided regret is introduced, choosing placements based on not ruling out later placements. Extensive evaluations are performed, showing the importance of using a good evaluation and that the proposed global propagation guided regret is a very effective guide.
2020-01-13
arXiv.org
http://arxiv.org/abs/2001.04233
2020-07-21 10:55:58
arXiv: 2001.04233
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2001.04233.pdf
2020-07-21 10:56:09
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2001.04233
2020-07-21 10:56:13
1
text/html
presentation
State Representation and Polyomino Placement for the Game Patchwork
https://zayenz.se/papers/Lagerkvist_ModRef_2019_Presentation.pdf
attachment
Full Text
https://zayenz.se/papers/Lagerkvist_ModRef_2019_Presentation.pdf
2020-07-21 10:56:59
1
application/pdf
journalArticle
arXiv:2001.04238 [cs]
Lagerkvist
Mikael Zayenz
Computer Science - Artificial Intelligence
Nmbr9 as a Constraint Programming Challenge
Modern board games are a rich source of interesting and new challenges for combinatorial problems. The game Nmbr9 is a solitaire style puzzle game using polyominoes. The rules of the game are simple to explain, but modelling the game effectively using constraint programming is hard. This abstract presents the game, contributes new generalized variants of the game suitable for benchmarking and testing, and describes a model for the presented variants. The question of the top possible score in the standard game is an open challenge.
2020-01-13
arXiv.org
http://arxiv.org/abs/2001.04238
2020-07-21 10:57:58
arXiv: 2001.04238
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2001.04238.pdf
2020-07-21 10:58:01
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2001.04238
2020-07-21 10:58:02
1
text/html
blogPost
Nmbr9 as a Constraint Programming Challenge
https://zayenz.se/blog/post/nmbr9-cp2019-abstract/
conferencePaper
DOI 10.1109/CIG.2019.8848097
Goodman
James
Re-determinizing MCTS in Hanabi
08 2019
1-8
attachment
Goodman - 2019 - Re-determinizing MCTS in Hanabi.pdf
application/pdf
conferencePaper
ISBN 978-1-5386-4359-4
2018 IEEE Conference on Computational Intelligence and Games (CIG)
DOI 10.1109/CIG.2018.8490449
Maastricht
IEEE
Canaan
Rodrigo
Shen
Haotian
Torrado
Ruben
Togelius
Julian
Nealen
Andy
Menzel
Stefan
Evolving Agents for the Hanabi 2018 CIG Competition
8/2018
DOI.org (Crossref)
https://ieeexplore.ieee.org/document/8490449/
2020-07-21 11:01:52
1-8
2018 IEEE Conference on Computational Intelligence and Games (CIG)
attachment
Submitted Version
https://arxiv.org/pdf/1809.09764
2020-07-21 11:01:56
1
application/pdf
bookSection
765
ISBN 978-3-319-67467-4 978-3-319-67468-1
BNAIC 2016: Artificial Intelligence
Cham
Springer International Publishing
Bosse
Tibor
Bredeweg
Bert
van den Bergh
Mark J. H.
Hommelberg
Anne
Kosters
Walter A.
Spieksma
Flora M.
Aspects of the Cooperative Card Game Hanabi
2017
DOI.org (Crossref)
http://link.springer.com/10.1007/978-3-319-67468-1_7
2020-07-21 11:02:26
Series Title: Communications in Computer and Information Science
DOI: 10.1007/978-3-319-67468-1_7
93-105
attachment
Full Text
https://twin.sci-hub.se/6548/49fca9bfed767f739defcd030c004bdb/vandenbergh2017.pdf#view=FitH
2020-07-21 11:02:31
1
application/pdf
bookSection
10664
ISBN 978-3-319-71648-0 978-3-319-71649-7
Advances in Computer Games
Cham
Springer International Publishing
Winands
Mark H.M.
van den Herik
H. Jaap
Kosters
Walter A.
Bouzy
Bruno
Playing Hanabi Near-Optimally
2017
DOI.org (Crossref)
http://link.springer.com/10.1007/978-3-319-71649-7_5
2020-07-21 11:02:53
Series Title: Lecture Notes in Computer Science
DOI: 10.1007/978-3-319-71649-7_5
51-62
conferencePaper
ISBN 978-1-5386-3233-8
2017 IEEE Conference on Computational Intelligence and Games (CIG)
DOI 10.1109/CIG.2017.8080417
New York, NY, USA
IEEE
Eger
Markus
Martens
Chris
Cordoba
Marcela Alfaro
An intentional AI for hanabi
8/2017
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/8080417/
2020-07-21 11:03:36
68-75
2017 IEEE Conference on Computational Intelligence and Games (CIG)
attachment
Full Text
https://zero.sci-hub.se/6752/bcf6e994ee7503ab821bd67848727b05/eger2017.pdf#view=FitH
2020-07-21 11:03:40
1
application/pdf
conferencePaper
Osawa
Hirotaka
Solving Hanabi: Estimating Hands by Opponent's Actions in Cooperative Game with Incomplete Information
A unique behavior of humans is modifying one’s unobservable behavior based on the reaction of others for cooperation. We used a card game called Hanabi as an evaluation task of imitating human reflective intelligence with artificial intelligence. Hanabi is a cooperative card game with incomplete information. A player cooperates with an opponent in building several card sets constructed with the same color and ordered numbers. However, like a blind man's bluff, each player sees the cards of all other players except his/her own. Also, communication between players is restricted to information about the same numbers and colors, and the player is required to read his/his opponent's intention with the opponent's hand, estimate his/her cards with incomplete information, and play one of them for building a set. We compared human play with several simulated strategies. The results indicate that the strategy with feedbacks from simulated opponent's viewpoints achieves more score than other strategies.
2015
https://aaai.org/ocs/index.php/WS/AAAIW15/paper/view/10167
AAAI Workshops
attachment
Osawa - 2015 - Solving Hanabi Estimating Hands by Opponent's Act.pdf
application/pdf
journalArticle
Cape Cod
Eger
Markus
Martens
Chris
A Browser-based Interface for the Exploration and Evaluation of Hanabi AIs
2017
en
Zotero
http://fdg2017.org/papers/FDG2017_demo_Hanabi.pdf
4
attachment
Eger and Martens - 2017 - A Browser-based Interface for the Exploration and .pdf
application/pdf
journalArticle
Gottwald
Eva Tallula
Eger
Markus
Martens
Chris
I see what you see: Integrating eye tracking into Hanabi playing agents
Humans’ eye movements convey a lot of information about their intentions, often unconsciously. Intelligent agents that cooperate with humans in various domains can benefit from interpreting this information. This paper contains a preliminary look at how eye tracking could be useful for agents that play the cooperative card game Hanabi with human players. We outline several situations in which an AI agent can utilize gaze information, and present an outlook on how we plan to integrate this with reimplementations of contemporary Hanabi agents.
en
Zotero
4
attachment
Gottwald et al. - I see what you see Integrating eye tracking into .pdf
application/pdf
computerProgram
State of the art Hanabi bots + simulation framework in rust
https://github.com/WuTheFWasThat/hanabi.rs
computerProgram
A strategy simulator for the well-known cooperative card game Hanabi
https://github.com/rjtobin/HanSim
computerProgram
A framework for writing bots that play Hanabi
https://github.com/Quuxplusone/Hanabi
journalArticle
Ludic Language Pedagogy
Ludic Language Pedagogy
deHaan
Jonathan
Jidoukan Jenga: Teaching English through remixing games and game rules
Let students play simple games in their L1. It’s ok!
Then:
You, the teacher, can help them critique the game in their L2.
You, the teacher, can help them change the game in their L2.
You, the teacher, can help them develop themselves.
#dropthestick #dropthecarrot #bringmeaning
2020-04-15
Teaching English through remixing games and game rules
https://www.llpjournal.org/2020/04/13/jidokan-jenga.html
📍 What is this? This is a recollection of a short lesson with some children. I used Jenga and a dictionary.
📍 Why did you make it? I want to show language teachers that simple games, and playing simple games in students’ first language can be a great foundation for helping students learn new vocabulary, think critically, and exercise creativity.
📍 Why is it radical? I taught using a simple board game (at a time when video games are over-focused on in research). I show what the learning looks like (I include a photo). The teaching and learning didn’t occur in a laboratory setting, but in the wild (in a community center). I focused on the learning around games.
📍 Who is it for? Language teachers can easily implement this lesson using Jenga or any other game. Language researchers can expand on the translating and remixing potential around games.
attachment
deHaan - 2020 - Jidoukan Jenga Teaching English through remixing .pdf
application/pdf
journalArticle
Heron
Michael James
Belford
Pauline Helen
Reid
Hayley
Crabb
Michael
Meeple Centred Design: A Heuristic Toolkit for Evaluating the Accessibility of Tabletop Games
6/2018
en
Meeple Centred Design
DOI.org (Crossref)
http://link.springer.com/10.1007/s40869-018-0057-8
2020-07-28 09:08:52
97-114
7
The Computer Games Journal
DOI 10.1007/s40869-018-0057-8
2
Comput Game J
ISSN 2052-773X
attachment
Full Text
https://link.springer.com/content/pdf/10.1007/s40869-018-0057-8.pdf
2020-07-28 09:08:55
1
application/pdf
journalArticle
7
The Computer Games Journal
DOI 10.1007/s40869-018-0056-9
2
Comput Game J
ISSN 2052-773X
Heron
Michael James
Belford
Pauline Helen
Reid
Hayley
Crabb
Michael
Eighteen Months of Meeple Like Us: An Exploration into the State of Board Game Accessibility
6/2018
en
Eighteen Months of Meeple Like Us
DOI.org (Crossref)
http://link.springer.com/10.1007/s40869-018-0056-9
2020-07-28 09:09:05
75-95
attachment
Full Text
https://link.springer.com/content/pdf/10.1007/s40869-018-0056-9.pdf
2020-07-28 09:09:08
1
application/pdf
thesis
Utrecht University
Andel
Daniël
On the complexity of Hive
It is shown that for an arbitrary position of a Hive game where both players have the same set of N pieces it is PSPACE-hard to determine whether one of the players has a winning strategy. The proof is done by reducing the known PSPACE-complete set of true quantified boolean formulas to a game concerning these formulas, then to the game generalised geography, then to a version of that game with the restriction of having only nodes with maximum degree 3, and finally to generalised Hive. This thesis includes a short introduction to the subject of computational complexity.
May 2020
en-US
On the complexity of Hive
https://dspace.library.uu.nl/handle/1874/396955
33
Bachelor thesis
attachment
Andel - 2020 - On the complexity of Hive.pdf
application/pdf
journalArticle
arXiv:2010.00048 [cs]
Kunda
Maithilee
Rabkina
Irina
Computer Science - Artificial Intelligence
Creative Captioning: An AI Grand Challenge Based on the Dixit Board Game
We propose a new class of "grand challenge" AI problems that we call creative captioning---generating clever, interesting, or abstract captions for images, as well as understanding such captions. Creative captioning draws on core AI research areas of vision, natural language processing, narrative reasoning, and social reasoning, and across all these areas, it requires sophisticated uses of common sense and cultural knowledge. In this paper, we analyze several specific research problems that fall under creative captioning, using the popular board game Dixit as both inspiration and proposed testing ground. We expect that Dixit could serve as an engaging and motivating benchmark for creative captioning across numerous AI research communities for the coming 1-2 decades.
2020-09-30
Creative Captioning
arXiv.org
http://arxiv.org/abs/2010.00048
2020-10-12 04:03:28
arXiv: 2010.00048
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2010.00048.pdf
2020-10-12 04:03:46
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2010.00048
2020-10-12 04:03:53
1
text/html
computerProgram
Shobu AI Playground
https://github.com/JayWalker512/Shobu
webpage
Shobu randomly played games dataset
https://www.kaggle.com/bsfoltz/shobu-randomly-played-games-104k
conferencePaper
ISBN 978-1-4503-5319-9
Proceedings of the International Conference on the Foundations of Digital Games - FDG '17
DOI 10.1145/3102071.3102105
Hyannis, Massachusetts
ACM Press
de Mesentier Silva
Fernando
Lee
Scott
Togelius
Julian
Nealen
Andy
AI-based playtesting of contemporary board games
2017
en
DOI.org (Crossref)
http://dl.acm.org/citation.cfm?doid=3102071.3102105
2020-10-12 04:09:30
1-10
the International Conference
attachment
Full Text
https://twin.sci-hub.se/6553/d80b9cdf7f993e1137d0b129dec94e6d/demesentiersilva2017.pdf#view=FitH
2020-10-12 04:09:38
1
application/pdf
attachment
PDF
http://game.engineering.nyu.edu/wp-content/uploads/2017/06/ticket-ride-fdg2017-camera-ready.pdf
2020-10-12 04:13:00
3
computerProgram
Copley
Rowan
Materials for Ticket to Ride Seattle and a framework for making more game boards
https://github.com/dovinmu/ttr_generator
report
Nguyen
Cuong
Dinjian
Daniel
The Difficulty of Learning Ticket to Ride
Ticket to Ride is a very popular, award-winning board-game where you try toscore the most points while building a railway spanning cities in America. For acomputer to learn to play this game is very difficult due to the vast state-actionspace. This project will explain why featurizing your state, and implementingcurriculum learning can help agents learn as state-action spaces grow too largefor traditional learning methods to be effective.
https://www.eecs.tufts.edu/~jsinapov/teaching/comp150_RL/reports/Nguyen_Dinjian_report.pdf
attachment
Full Text
https://www.eecs.tufts.edu/~jsinapov/teaching/comp150_RL/reports/Nguyen_Dinjian_report.pdf
2021-07-24 08:19:13
1
application/pdf
conferencePaper
ISBN 978-1-4503-6571-0
Proceedings of the 13th International Conference on the Foundations of Digital Games
DOI 10.1145/3235765.3235813
Malmö Sweden
ACM
de Mesentier Silva
Fernando
Lee
Scott
Togelius
Julian
Nealen
Andy
Evolving maps and decks for ticket to ride
2018-08-07
en
DOI.org (Crossref)
https://dl.acm.org/doi/10.1145/3235765.3235813
2020-10-12 04:12:33
1-7
FDG '18: Foundations of Digital Games 2018
attachment
Full Text
https://twin.sci-hub.se/7128/24e28b0429626f565aafd93768332e73/demesentiersilva2018.pdf#view=FitH
2020-10-12 04:12:36
1
application/pdf
journalArticle
arXiv:2008.07079 [cs, stat]
Gendre
Quentin
Kaneko
Tomoyuki
Computer Science - Artificial Intelligence
Computer Science - Machine Learning
Statistics - Machine Learning
Playing Catan with Cross-dimensional Neural Network
Catan is a strategic board game having interesting properties, including multi-player, imperfect information, stochastic, complex state space structure (hexagonal board where each vertex, edge and face has its own features, cards for each player, etc), and a large action space (including negotiation). Therefore, it is challenging to build AI agents by Reinforcement Learning (RL for short), without domain knowledge nor heuristics. In this paper, we introduce cross-dimensional neural networks to handle a mixture of information sources and a wide variety of outputs, and empirically demonstrate that the network dramatically improves RL in Catan. We also show that, for the first time, a RL agent can outperform jsettler, the best heuristic agent available.
2020-08-17
arXiv.org
http://arxiv.org/abs/2008.07079
2020-10-12 04:19:57
arXiv: 2008.07079
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2008.07079.pdf
2020-10-12 04:20:04
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2008.07079
2020-10-12 04:20:10
1
text/html
conferencePaper
ISBN 978-1-4503-8878-8
11th Hellenic Conference on Artificial Intelligence
DOI 10.1145/3411408.3411413
Athens Greece
ACM
Theodoridis
Alexios
Chalkiadakis
Georgios
Monte Carlo Tree Search for the Game of Diplomacy
2020-09-02
en
DOI.org (Crossref)
https://dl.acm.org/doi/10.1145/3411408.3411413
2020-10-12 04:20:38
16-25
SETN 2020: 11th Hellenic Conference on Artificial Intelligence
journalArticle
Eger
Markus
Martens
Chris
Sauma Chacon
Pablo
Alfaro Cordoba
Marcela
Hidalgo Cespedes
Jeisson
Operationalizing Intentionality to Play Hanabi with Human Players
2020
DOI.org (Crossref)
https://ieeexplore.ieee.org/document/9140404/
2020-11-26 08:48:44
1-1
IEEE Transactions on Games
DOI 10.1109/TG.2020.3009359
IEEE Trans. Games
ISSN 2475-1502, 2475-1510
attachment
Full Text
https://sci-hub.se/downloads/2020-08-17/f1/eger2020.pdf?rand=5fbf6bef76c6b#view=FitH
2020-11-26 08:48:52
1
application/pdf
journalArticle
16
Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment
1
AIIDE
Canaan
Rodrigo
Gao
Xianbo
Chung
Youjin
Togelius
Julian
Nealen
Andy
Menzel
Stefan
Behavioral Evaluation of Hanabi Rainbow DQN Agents and Rule-Based Agents
<p class="abstract">Hanabi is a multiplayer cooperative card game, where only your partners know your cards. All players succeed or fail together. This makes the game an excellent testbed for studying collaboration. Recently, it has been shown that deep neural networks can be trained through self-play to play the game very well. However, such agents generally do not play well with others. In this paper, we investigate the consequences of training Rainbow DQN agents with human-inspired rule-based agents. We analyze with which agents Rainbow agents learn to play well, and how well playing skill transfers to agents they were not trained with. We also analyze patterns of communication between agents to elucidate how collaboration happens. A key finding is that while most agents only learn to play well with partners seen during training, one particular agent leads the Rainbow algorithm towards a much more general policy. The metrics and hypotheses advanced in this paper can be used for further study of collaborative agents.</p>
October 1, 2020
https://ojs.aaai.org/index.php/AIIDE/article/view/7404
2020-11-26
Section: Full Oral Papers
31-37
attachment
View PDF
https://ojs.aaai.org/index.php/AIIDE/article/view/7404/7333
2020-11-26 08:52:38
3
conferencePaper
2020第82回全国大会講演論文集
ひい
とう
市来
正裕
中里
研一
Playing mini-Hanabi card game with Q-learning
February 2020
http://id.nii.ac.jp/1001/00205046/
Issue: 1
41–42
attachment
View PDF
https://ipsj.ixsq.nii.ac.jp/ej/?action=repository_uri&item_id=205142&file_id=1&file_no=1
2020-11-26 08:54:47
3
journalArticle
arXiv:2005.07156 [cs]
Reinhardt
Jack
Computer Science - Artificial Intelligence
Computer Science - Multiagent Systems
Competing in a Complex Hidden Role Game with Information Set Monte Carlo Tree Search
Advances in intelligent game playing agents have led to successes in perfect information games like Go and imperfect information games like Poker. The Information Set Monte Carlo Tree Search (ISMCTS) family of algorithms outperforms previous algorithms using Monte Carlo methods in imperfect information games. In this paper, Single Observer Information Set Monte Carlo Tree Search (SO-ISMCTS) is applied to Secret Hitler, a popular social deduction board game that combines traditional hidden role mechanics with the randomness of a card deck. This combination leads to a more complex information model than the hidden role and card deck mechanics alone. It is shown in 10108 simulated games that SO-ISMCTS plays as well as simpler rule based agents, and demonstrates the potential of ISMCTS algorithms in complicated information set domains.
2020-05-14
arXiv.org
http://arxiv.org/abs/2005.07156
2020-11-26 09:00:33
arXiv: 2005.07156
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2005.07156.pdf
2020-11-26 09:01:03
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2005.07156
2020-11-26 09:01:10
1
text/html
journalArticle
arXiv:2009.12974 [cs]
Ameneyro
Fred Valdez
Galvan
Edgar
Morales
Anger Fernando Kuri
Computer Science - Artificial Intelligence
Playing Carcassonne with Monte Carlo Tree Search
Monte Carlo Tree Search (MCTS) is a relatively new sampling method with multiple variants in the literature. They can be applied to a wide variety of challenging domains including board games, video games, and energy-based problems to mention a few. In this work, we explore the use of the vanilla MCTS and the MCTS with Rapid Action Value Estimation (MCTS-RAVE) in the game of Carcassonne, a stochastic game with a deceptive scoring system where limited research has been conducted. We compare the strengths of the MCTS-based methods with the Star2.5 algorithm, previously reported to yield competitive results in the game of Carcassonne when a domain-specific heuristic is used to evaluate the game states. We analyse the particularities of the strategies adopted by the algorithms when they share a common reward system. The MCTS-based methods consistently outperformed the Star2.5 algorithm given their ability to find and follow long-term strategies, with the vanilla MCTS exhibiting a more robust game-play than the MCTS-RAVE.
2020-10-04
arXiv.org
http://arxiv.org/abs/2009.12974
2021-01-02 18:13:09
arXiv: 2009.12974
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2009.12974.pdf
2021-01-02 18:13:12
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2009.12974
2021-01-02 18:13:17
1
text/html
journalArticle
arXiv:2007.15895 [cs]
Tanaka
Satoshi
Bonnet
François
Tixeuil
Sébastien
Tamura
Yasumasa
Computer Science - Computer Science and Game Theory
Quixo Is Solved
Quixo is a two-player game played on a 5$\times$5 grid where the players try to align five identical symbols. Specifics of the game require the usage of novel techniques. Using a combination of value iteration and backward induction, we propose the first complete analysis of the game. We describe memory-efficient data structures and algorithmic optimizations that make the game solvable within reasonable time and space constraints. Our main conclusion is that Quixo is a Draw game. The paper also contains the analysis of smaller boards and presents some interesting states extracted from our computations.
2020-07-31
arXiv.org
http://arxiv.org/abs/2007.15895
2021-01-02 18:17:10
arXiv: 2007.15895
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2007.15895.pdf
2021-01-02 18:17:17
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2007.15895
2021-01-02 18:17:21
1
text/html
journalArticle
arXiv:2006.02353 [cs]
Bertholon
Guillaume
Géraud-Stewart
Rémi
Kugelmann
Axel
Lenoir
Théo
Naccache
David
Computer Science - Computer Science and Game Theory
At Most 43 Moves, At Least 29: Optimal Strategies and Bounds for Ultimate Tic-Tac-Toe
Ultimate Tic-Tac-Toe is a variant of the well known tic-tac-toe (noughts and crosses) board game. Two players compete to win three aligned "fields", each of them being a tic-tac-toe game. Each move determines which field the next player must play in. We show that there exist a winning strategy for the first player, and therefore that there exist an optimal winning strategy taking at most 43 moves; that the second player can hold on at least 29 rounds; and identify any optimal strategy's first two moves.
2020-06-06
At Most 43 Moves, At Least 29
arXiv.org
http://arxiv.org/abs/2006.02353
2021-01-02 18:17:55
arXiv: 2006.02353
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2006.02353.pdf
2021-01-02 18:17:57
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2006.02353
2021-01-02 18:18:02
1
text/html
journalArticle
arXiv:2004.00377 [cs]
Muller-Brockhausen
Matthias
Preuss
Mike
Plaat
Aske
Computer Science - Artificial Intelligence
A New Challenge: Approaching Tetris Link with AI
Decades of research have been invested in making computer programs for playing games such as Chess and Go. This paper focuses on a new game, Tetris Link, a board game that is still lacking any scientific analysis. Tetris Link has a large branching factor, hampering a traditional heuristic planning approach. We explore heuristic planning and two other approaches: Reinforcement Learning, Monte Carlo tree search. We document our approach and report on their relative performance in a tournament. Curiously, the heuristic approach is stronger than the planning/learning approaches. However, experienced human players easily win the majority of the matches against the heuristic planning AIs. We, therefore, surmise that Tetris Link is more difficult than expected. We offer our findings to the community as a challenge to improve upon.
2020-04-01
A New Challenge
arXiv.org
http://arxiv.org/abs/2004.00377
2021-01-02 18:18:26
arXiv: 2004.00377
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2004.00377.pdf
2021-01-02 18:18:32
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2004.00377
2021-01-02 18:18:38
1
text/html
journalArticle
arXiv:1511.08099 [cs]
Cuayáhuitl
Heriberto
Keizer
Simon
Lemon
Oliver
Computer Science - Artificial Intelligence
Computer Science - Machine Learning
Strategic Dialogue Management via Deep Reinforcement Learning
Artificially intelligent agents equipped with strategic skills that can negotiate during their interactions with other natural or artificial agents are still underdeveloped. This paper describes a successful application of Deep Reinforcement Learning (DRL) for training intelligent agents with strategic conversational skills, in a situated dialogue setting. Previous studies have modelled the behaviour of strategic agents using supervised learning and traditional reinforcement learning techniques, the latter using tabular representations or learning with linear function approximation. In this study, we apply DRL with a high-dimensional state space to the strategic board game of Settlers of Catan---where players can offer resources in exchange for others and they can also reply to offers made by other players. Our experimental results report that the DRL-based learnt policies significantly outperformed several baselines including random, rule-based, and supervised-based behaviours. The DRL-based policy has a 53% win rate versus 3 automated players (`bots'), whereas a supervised player trained on a dialogue corpus in this setting achieved only 27%, versus the same 3 bots. This result supports the claim that DRL is a promising framework for training dialogue systems, and strategic agents with negotiation abilities.
2015-11-25
arXiv.org
http://arxiv.org/abs/1511.08099
2021-01-02 18:29:38
arXiv: 1511.08099
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1511.08099.pdf
2021-01-02 18:29:43
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1511.08099
2021-01-02 18:29:50
1
text/html
conferencePaper
Applying Neural Networks and Genetic Programming to the Game Lost Cities
https://minds.wisconsin.edu/bitstream/handle/1793/79080/LydeenSpr18.pdf?sequence=1&isAllowed=y
attachment
LydeenSpr18.pdf
https://minds.wisconsin.edu/bitstream/handle/1793/79080/LydeenSpr18.pdf
2021-06-12 17:03:24
3
report
A summary of a dissertation on Azul
https://old.reddit.com/r/boardgames/comments/hxodaf/update_i_wrote_my_dissertation_on_azul/
conferencePaper
Ceramic: A research environment based on the multi-player strategic board game Azul
https://ipsj.ixsq.nii.ac.jp/ej/?action=repository_action_common_download&item_id=207669&item_no=1&attribute_id=1&file_no=1
computerProgram
Ceramic: A research environment based on the multi-player strategic board game Azul
https://github.com/Swynfel/ceramic
report
Blokus Game Solver
https://digitalcommons.calpoly.edu/cpesp/290/
conferencePaper
ISBN 978-1-4799-2198-0 978-1-4799-2199-7
2013 International Conference on Field-Programmable Technology (FPT)
DOI 10.1109/FPT.2013.6718426
Kyoto, Japan
IEEE
Yoza
Takashi
Moriwaki
Retsu
Torigai
Yuki
Kamikubo
Yuki
Kubota
Takayuki
Watanabe
Takahiro
Fujimori
Takumi
Ito
Hiroyuki
Seo
Masato
Akagi
Kouta
Yamaji
Yuichiro
Watanabe
Minoru
FPGA Blokus Duo Solver using a massively parallel architecture
12/2013
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/6718426/
2021-06-28 14:38:57
494-497
2013 International Conference on Field-Programmable Technology (FPT)
attachment
Full Text
https://zero.sci-hub.se/2654/a4d3e713290066b6db7db1d9eedd194e/yoza2013.pdf#view=FitH
2021-06-28 14:39:08
1
application/pdf
conferencePaper
ISBN 978-1-4799-0565-2 978-1-4799-0562-1 978-1-4799-0563-8
The 17th CSI International Symposium on Computer Architecture & Digital Systems (CADS 2013)
DOI 10.1109/CADS.2013.6714256
Tehran, Iran
IEEE
Jahanshahi
Ali
Taram
Mohammad Kazem
Eskandari
Nariman
Blokus Duo game on FPGA
10/2013
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/6714256/
2021-06-28 14:39:04
149-152
2013 17th CSI International Symposium on Computer Architecture and Digital Systems (CADS)
attachment
Full Text
https://zero.sci-hub.se/3228/9ae6ca1efab5a2ebb63dd4e22a13bf04/jahanshahi2013.pdf#view=FitH
2021-06-28 14:39:07
1
application/pdf
journalArticle
The World Wide Web Conference
DOI 10.1145/3308558.3314131
Hsu
Chao-Chun
Chen
Yu-Hua
Chen
Zi-Yuan
Lin
Hsin-Yu
Huang
Ting-Hao 'Kenneth'
Ku
Lun-Wei
Computer Science - Computation and Language
Dixit: Interactive Visual Storytelling via Term Manipulation
In this paper, we introduce Dixit, an interactive visual storytelling system that the user interacts with iteratively to compose a short story for a photo sequence. The user initiates the process by uploading a sequence of photos. Dixit first extracts text terms from each photo which describe the objects (e.g., boy, bike) or actions (e.g., sleep) in the photo, and then allows the user to add new terms or remove existing terms. Dixit then generates a short story based on these terms. Behind the scenes, Dixit uses an LSTM-based model trained on image caption data and FrameNet to distill terms from each image and utilizes a transformer decoder to compose a context-coherent story. Users change images or terms iteratively with Dixit to create the most ideal story. Dixit also allows users to manually edit and rate stories. The proposed procedure opens up possibilities for interpretable and controllable visual storytelling, allowing users to understand the story formation rationale and to intervene in the generation process.
2019-05-13
Dixit
arXiv.org
http://arxiv.org/abs/1903.02230
2021-06-28 14:40:29
arXiv: 1903.02230
3531-3535
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1903.02230.pdf
2021-06-28 14:40:38
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1903.02230
2021-06-28 14:40:43
1
text/html
computerProgram
Dominion Simulator
https://dominionsimulator.wordpress.com/f-a-q/
computerProgram
Dominion Simulator Source Code
https://github.com/mikemccllstr/dominionstats/
blogPost
Best and worst openings in Dominion
http://councilroom.com/openings
blogPost
Optimal Card Ratios in Dominion
http://councilroom.com/optimal_card_ratios
blogPost
Card Winning Stats on Dominion Server
http://councilroom.com/supply_win
forumPost
Dominion Strategy Forum
http://forum.dominionstrategy.com/index.php
journalArticle
arXiv:1811.11273 [cs]
Bendekgey
Henry
Computer Science - Artificial Intelligence
Clustering Player Strategies from Variable-Length Game Logs in Dominion
We present a method for encoding game logs as numeric features in the card game Dominion. We then run the manifold learning algorithm t-SNE on these encodings to visualize the landscape of player strategies. By quantifying game states as the relative prevalence of cards in a player's deck, we create visualizations that capture qualitative differences in player strategies. Different ways of deviating from the starting game state appear as different rays in the visualization, giving it an intuitive explanation. This is a promising new direction for understanding player strategies across games that vary in length.
2018-12-12
arXiv.org
http://arxiv.org/abs/1811.11273
2021-06-28 14:43:21
arXiv: 1811.11273
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1811.11273.pdf
2021-06-28 14:43:27
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1811.11273
2021-06-28 14:43:31
1
text/html
computerProgram
Hanabi Open Agent Dataset
https://github.com/aronsar/hoad
conferencePaper
Hanabi Open Agent Dataset
https://dl.acm.org/doi/10.5555/3463952.3464188
journalArticle
arXiv:2010.02923 [cs]
Gray
Jonathan
Lerer
Adam
Bakhtin
Anton
Brown
Noam
Computer Science - Artificial Intelligence
Computer Science - Machine Learning
Computer Science - Computer Science and Game Theory
Human-Level Performance in No-Press Diplomacy via Equilibrium Search
Prior AI breakthroughs in complex games have focused on either the purely adversarial or purely cooperative settings. In contrast, Diplomacy is a game of shifting alliances that involves both cooperation and competition. For this reason, Diplomacy has proven to be a formidable research challenge. In this paper we describe an agent for the no-press variant of Diplomacy that combines supervised learning on human data with one-step lookahead search via regret minimization. Regret minimization techniques have been behind previous AI successes in adversarial games, most notably poker, but have not previously been shown to be successful in large-scale games involving cooperation. We show that our agent greatly exceeds the performance of past no-press Diplomacy bots, is unexploitable by expert humans, and ranks in the top 2% of human players when playing anonymous games on a popular Diplomacy website.
2021-05-03
arXiv.org
http://arxiv.org/abs/2010.02923
2021-06-28 15:28:02
arXiv: 2010.02923
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2010.02923.pdf
2021-06-28 15:28:18
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2010.02923
2021-06-28 15:28:22
1
text/html
journalArticle
arXiv:1708.01503 [math]
Akiyama
Rika
Abe
Nozomi
Fujita
Hajime
Inaba
Yukie
Hataoka
Mari
Ito
Shiori
Seita
Satomi
55A20 (Primary), 05A99 (Secondary)
Mathematics - Combinatorics
Mathematics - Geometric Topology
Mathematics - History and Overview
Maximum genus of the Jenga like configurations
We treat the boundary of the union of blocks in the Jenga game as a surface with a polyhedral structure and consider its genus. We generalize the game and determine the maximum genus of the generalized game.
2018-08-31
arXiv.org
http://arxiv.org/abs/1708.01503
2021-06-28 15:28:12
arXiv: 1708.01503
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1708.01503.pdf
2021-06-28 15:28:21
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1708.01503
2021-06-28 15:28:24
1
text/html
journalArticle
arXiv:1905.08617 [cs]
Bai
Chongyang
Bolonkin
Maksim
Burgoon
Judee
Chen
Chao
Dunbar
Norah
Singh
Bharat
Subrahmanian
V. S.
Wu
Zhe
Computer Science - Artificial Intelligence
Computer Science - Computer Vision and Pattern Recognition
Automatic Long-Term Deception Detection in Group Interaction Videos
Most work on automated deception detection (ADD) in video has two restrictions: (i) it focuses on a video of one person, and (ii) it focuses on a single act of deception in a one or two minute video. In this paper, we propose a new ADD framework which captures long term deception in a group setting. We study deception in the well-known Resistance game (like Mafia and Werewolf) which consists of 5-8 players of whom 2-3 are spies. Spies are deceptive throughout the game (typically 30-65 minutes) to keep their identity hidden. We develop an ensemble predictive model to identify spies in Resistance videos. We show that features from low-level and high-level video analysis are insufficient, but when combined with a new class of features that we call LiarRank, produce the best results. We achieve AUCs of over 0.70 in a fully automated setting. Our demo can be found at http://home.cs.dartmouth.edu/~mbolonkin/scan/demo/
2019-06-15
arXiv.org
http://arxiv.org/abs/1905.08617
2021-06-28 15:32:49
arXiv: 1905.08617
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1905.08617.pdf
2021-06-28 15:32:54
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1905.08617
2021-06-28 15:32:58
1
text/html
bookSection
10068
ISBN 978-3-319-50934-1 978-3-319-50935-8
Computers and Games
Cham
Springer International Publishing
Plaat
Aske
Kosters
Walter
van den Herik
Jaap
Bi
Xiaoheng
Tanaka
Tetsuro
Human-Side Strategies in the Werewolf Game Against the Stealth Werewolf Strategy
2016
DOI.org (Crossref)
http://link.springer.com/10.1007/978-3-319-50935-8_9
2021-06-28 15:32:54
Series Title: Lecture Notes in Computer Science
DOI: 10.1007/978-3-319-50935-8_9
93-102
attachment
Full Text
https://sci-hub.se/downloads/2019-01-26//f7/bi2016.pdf#view=FitH
2021-06-28 15:33:08
1
application/pdf
journalArticle
arXiv:0804.0071 [math]
Yao
Erlin
65C20
91-01
Mathematics - Probability
A Theoretical Study of Mafia Games
Mafia can be described as an experiment in human psychology and mass hysteria, or as a game between informed minority and uninformed majority. Focus on a very restricted setting, Mossel et al. [to appear in Ann. Appl. Probab. Volume 18, Number 2] showed that in the mafia game without detectives, if the civilians and mafias both adopt the optimal randomized strategy, then the two groups have comparable probabilities of winning exactly when the total player size is R and the mafia size is of order Sqrt(R). They also proposed a conjecture which stated that this phenomenon should be valid in a more extensive framework. In this paper, we first indicate that the main theorem given by Mossel et al. [to appear in Ann. Appl. Probab. Volume 18, Number 2] can not guarantee their conclusion, i.e., the two groups have comparable winning probabilities when the mafia size is of order Sqrt(R). Then we give a theorem which validates the correctness of their conclusion. In the last, by proving the conjecture proposed by Mossel et al. [to appear in Ann. Appl. Probab. Volume 18, Number 2], we generalize the phenomenon to a more extensive framework, of which the mafia game without detectives is only a special case.
2008-04-01
arXiv.org
http://arxiv.org/abs/0804.0071
2021-06-28 15:33:04
arXiv: 0804.0071
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/0804.0071.pdf
2021-06-28 15:33:07
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/0804.0071
2021-06-28 15:33:10
1
text/html
bookSection
11302
ISBN 978-3-030-04178-6 978-3-030-04179-3
Neural Information Processing
Cham
Springer International Publishing
Cheng
Long
Leung
Andrew Chi Sing
Ozawa
Seiichi
Zilio
Felipe
Prates
Marcelo
Lamb
Luis
Neural Networks Models for Analyzing Magic: The Gathering Cards
2018
Neural Networks Models for Analyzing Magic
DOI.org (Crossref)
http://link.springer.com/10.1007/978-3-030-04179-3_20
2021-06-28 15:33:26
Series Title: Lecture Notes in Computer Science
DOI: 10.1007/978-3-030-04179-3_20
227-239
attachment
Submitted Version
https://arxiv.org/pdf/1810.03744
2021-06-28 15:33:36
1
application/pdf
conferencePaper
The Complexity of Deciding Legality of a Single Step of Magic: The Gathering
https://livrepository.liverpool.ac.uk/3029568/
conferencePaper
Magic: The Gathering in Common Lisp
https://vixra.org/abs/2001.0065
computerProgram
Magic: The Gathering in Common Lisp
https://github.com/jeffythedragonslayer/maglisp
thesis
Mathematical programming and Magic: The Gathering
https://commons.lib.niu.edu/handle/10843/19194
conferencePaper
Deck Construction Strategies for Magic: The Gathering
https://www.doi.org/10.1685/CSC06077
thesis
Deckbuilding in Magic: The Gathering Using a Genetic Algorithm
https://doi.org/11250/2462429
report
Magic: The Gathering Deck Performance Prediction
http://cs229.stanford.edu/proj2012/HauPlotkinTran-MagicTheGatheringDeckPerformancePrediction.pdf
computerProgram
A constraint programming based solver for Modern Art
https://github.com/captn3m0/modernart
journalArticle
arXiv:2103.00683 [cs]
Haliem
Marina
Bonjour
Trevor
Alsalem
Aala
Thomas
Shilpa
Li
Hongyu
Aggarwal
Vaneet
Bhargava
Bharat
Kejriwal
Mayank
Computer Science - Artificial Intelligence
Computer Science - Machine Learning
Learning Monopoly Gameplay: A Hybrid Model-Free Deep Reinforcement Learning and Imitation Learning Approach
Learning how to adapt and make real-time informed decisions in dynamic and complex environments is a challenging problem. To learn this task, Reinforcement Learning (RL) relies on an agent interacting with an environment and learning through trial and error to maximize the cumulative sum of rewards received by it. In multi-player Monopoly game, players have to make several decisions every turn which involves complex actions, such as making trades. This makes the decision-making harder and thus, introduces a highly complicated task for an RL agent to play and learn its winning strategies. In this paper, we introduce a Hybrid Model-Free Deep RL (DRL) approach that is capable of playing and learning winning strategies of the popular board game, Monopoly. To achieve this, our DRL agent (1) starts its learning process by imitating a rule-based agent (that resembles the human logic) to initialize its policy, (2) learns the successful actions, and improves its policy using DRL. Experimental results demonstrate an intelligent behavior of our proposed agent as it shows high win rates against different types of agent-players.
2021-02-28
Learning Monopoly Gameplay
arXiv.org
http://arxiv.org/abs/2103.00683
2021-06-28 15:48:08
arXiv: 2103.00683
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2103.00683.pdf
2021-06-28 15:48:19
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2103.00683
2021-06-28 15:48:23
1
text/html
conferencePaper
ISBN 978-0-7803-7203-0
Proceedings 2001 IEEE International Symposium on Computational Intelligence in Robotics and Automation (Cat. No.01EX515)
DOI 10.1109/CIRA.2001.1013210
Banff, Alta., Canada
IEEE
Yasumura
Y.
Oguchi
K.
Nitta
K.
Negotiation strategy of agents in the MONOPOLY game
2001
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/1013210/
2021-06-28 15:49:10
277-281
2001 International Symposium on Computational Intelligence in Robotics and Automation
attachment
Full Text
https://moscow.sci-hub.se/3317/19346a5b777c1582800b51ee3a7cf5ed/negotiation-strategy-of-agents-in-the-monopoly-game.pdf#view=FitH
2021-06-28 15:49:15
1
application/pdf
conferencePaper
ISBN 978-1-4673-1194-6 978-1-4673-1193-9 978-1-4673-1192-2
2012 IEEE Conference on Computational Intelligence and Games (CIG)
DOI 10.1109/CIG.2012.6374168
Granada, Spain
IEEE
Friberger
Marie Gustafsson
Togelius
Julian
Generating interesting Monopoly boards from open data
09/2012
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/6374168/
2021-06-28 15:49:18
288-295
2012 IEEE Conference on Computational Intelligence and Games (CIG)
attachment
Submitted Version
http://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=81CA58D9ACCE8CA7412077093E520EFC?doi=10.1.1.348.6099&rep=rep1&type=pdf
2021-06-28 15:49:32
1
application/pdf
conferencePaper
ISBN 978-1-4244-5770-0 978-1-4244-5771-7
Proceedings of the 2009 Winter Simulation Conference (WSC)
DOI 10.1109/WSC.2009.5429349
Austin, TX, USA
IEEE
Friedman
Eric J.
Henderson
Shane G.
Byuen
Thomas
Gallardo
German Gutierrez
Estimating the probability that the game of Monopoly never ends
12/2009
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/5429349/
2021-06-28 15:49:23
380-391
2009 Winter Simulation Conference (WSC 2009)
attachment
Full Text
https://moscow.sci-hub.se/3233/bacac19e84c764b72c627d05f55c0ad9/friedman2009.pdf#view=FitH
2021-06-28 15:49:32
1
application/pdf
report
Learning to Play Monopoly with Monte Carlo Tree Search
https://project-archive.inf.ed.ac.uk/ug4/20181042/ug4_proj.pdf
conferencePaper
ISBN 978-1-72811-895-6
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON)
DOI 10.1109/TENCON.2019.8929523
Kochi, India
IEEE
Arun
Edupuganti
Rajesh
Harikrishna
Chakrabarti
Debarka
Cherala
Harikiran
George
Koshy
Monopoly Using Reinforcement Learning
10/2019
DOI.org (Crossref)
https://ieeexplore.ieee.org/document/8929523/
2021-06-28 15:49:50
858-862
TENCON 2019 - 2019 IEEE Region 10 Conference (TENCON)
attachment
Full Text
https://sci-hub.se/downloads/2020-04-10/35/arun2019.pdf?rand=60d9ef9f20b26#view=FitH
2021-06-28 15:50:07
1
application/pdf
report
A Markovian Exploration of Monopoly
https://pi4.math.illinois.edu/wp-content/uploads/2014/10/Gartland-Burson-Ferguson-Markovopoly.pdf
conferencePaper
Learning to play Monopoly: A Reinforcement Learning approach
https://intelligence.csd.auth.gr/publication/conference-papers/learning-to-play-monopoly-a-reinforcement-learning-approach/
presentation
What’s the Best Monopoly Strategy?
https://core.ac.uk/download/pdf/48614184.pdf
journalArticle
Nakai
Kenichiro
Takenaga
Yasuhiko
NP-Completeness of Pandemic
2012
en
DOI.org (Crossref)
https://www.jstage.jst.go.jp/article/ipsjjip/20/3/20_723/_article
2021-06-28 15:59:47
723-726
20
Journal of Information Processing
DOI 10.2197/ipsjjip.20.723
3
Journal of Information Processing
ISSN 1882-6652
attachment
Full Text
https://www.jstage.jst.go.jp/article/ipsjjip/20/3/20_723/_pdf
2021-06-28 15:59:50
1
application/pdf
thesis
On Solving Pentago
http://www.ke.tu-darmstadt.de/lehre/arbeiten/bachelor/2011/Buescher_Niklas.pdf
journalArticle
arXiv:1906.02330 [cs, stat]
Serrino
Jack
Kleiman-Weiner
Max
Parkes
David C.
Tenenbaum
Joshua B.
Computer Science - Machine Learning
Statistics - Machine Learning
Computer Science - Multiagent Systems
Finding Friend and Foe in Multi-Agent Games
Recent breakthroughs in AI for multi-agent games like Go, Poker, and Dota, have seen great strides in recent years. Yet none of these games address the real-life challenge of cooperation in the presence of unknown and uncertain teammates. This challenge is a key game mechanism in hidden role games. Here we develop the DeepRole algorithm, a multi-agent reinforcement learning agent that we test on The Resistance: Avalon, the most popular hidden role game. DeepRole combines counterfactual regret minimization (CFR) with deep value networks trained through self-play. Our algorithm integrates deductive reasoning into vector-form CFR to reason about joint beliefs and deduce partially observable actions. We augment deep value networks with constraints that yield interpretable representations of win probabilities. These innovations enable DeepRole to scale to the full Avalon game. Empirical game-theoretic methods show that DeepRole outperforms other hand-crafted and learned agents in five-player Avalon. DeepRole played with and against human players on the web in hybrid human-agent teams. We find that DeepRole outperforms human players as both a cooperator and a competitor.
2019-06-05
arXiv.org
http://arxiv.org/abs/1906.02330
2021-06-28 16:00:28
arXiv: 1906.02330
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1906.02330.pdf
2021-06-28 16:00:35
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1906.02330
2021-06-28 16:00:38
1
text/html
thesis
A Mathematical Analysis of the Game of Santorini
https://openworks.wooster.edu/independentstudy/8917/
computerProgram
A Mathematical Analysis of the Game of Santorini
https://github.com/carsongeissler/SantoriniIS
report
The complexity of Scotland Yard
https://eprints.illc.uva.nl/id/eprint/193/1/PP-2006-18.text.pdf
conferencePaper
ISBN 978-1-4799-3547-5
2014 IEEE Conference on Computational Intelligence and Games
DOI 10.1109/CIG.2014.6932907
Dortmund, Germany
IEEE
Szubert
Marcin
Jaskowski
Wojciech
Temporal difference learning of N-tuple networks for the game 2048
8/2014
DOI.org (Crossref)
http://ieeexplore.ieee.org/document/6932907/
2021-06-28 16:09:20
1-8
2014 IEEE Conference on Computational Intelligence and Games (CIG)
attachment
Submitted Version
https://www.cs.put.poznan.pl/mszubert/pub/szubert2014cig.pdf
2021-06-28 16:09:26
1
application/pdf
journalArticle
arXiv:1501.03837 [cs]
Abdelkader
Ahmed
Acharya
Aditya
Dasler
Philip
Computer Science - Computational Complexity
F.2.2
On the Complexity of Slide-and-Merge Games
We study the complexity of a particular class of board games, which we call `slide and merge' games. Namely, we consider 2048 and Threes, which are among the most popular games of their type. In both games, the player is required to slide all rows or columns of the board in one direction to create a high value tile by merging pairs of equal tiles into one with the sum of their values. This combines features from both block pushing and tile matching puzzles, like Push and Bejeweled, respectively. We define a number of natural decision problems on a suitable generalization of these games and prove NP-hardness for 2048 by reducing from 3SAT. Finally, we discuss the adaptation of our reduction to Threes and conjecture a similar result.
2015-01-15
arXiv.org
http://arxiv.org/abs/1501.03837
2021-06-28 16:09:34
arXiv: 1501.03837
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1501.03837.pdf
2021-06-28 16:09:48
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1501.03837
2021-06-28 16:09:52
1
text/html
journalArticle
DOI 10.4230/LIPICS.FUN.2016.1
Abdelkader
Ahmed
Acharya
Aditya
Dasler
Philip
Herbstritt
Marc
000 Computer science, knowledge, general works
Computer Science
2048 Without New Tiles Is Still Hard
2016
en
DOI.org (Datacite)
http://drops.dagstuhl.de/opus/volltexte/2016/5885/
2021-06-28 16:09:58
Artwork Size: 14 pages
Medium: application/pdf
Publisher: Schloss Dagstuhl - Leibniz-Zentrum fuer Informatik GmbH, Wadern/Saarbruecken, Germany
14 pages
conferencePaper
MDA: A Formal Approach to Game Design and Game Research
https://aaai.org/Library/Workshops/2004/ws04-04-001.php
conferencePaper
6
ISBN 2342-9666
Think Design Play
DiGRA/Utrecht School of the Arts
Exploring anonymity in cooperative board games
This study was done as a part of a larger research project where the interest was on exploring if and how gameplay design could give informative principles to the design of educational activities. The researchers conducted a series of studies trying to map game mechanics that had the special quality of being inclusive, i.e., playable by a diverse group of players. This specific study focused on designing a cooperative board game with the goal of implementing anonymity as a game mechanic. Inspired by the gameplay design patterns methodology (Björk & Holopainen 2005a; 2005b; Holopainen & Björk 2008), mechanics from existing cooperative board games were extracted and analyzed in order to inform the design process. The results from prototyping and play testing indicated that it is possible to implement anonymous actions in cooperative board games and that this mechanic made rather unique forms of gameplay possible. These design patterns can be further developed in order to address inclusive educational practices.
January 2011
http://www.digra.org/digital-library/publications/exploring-anonymity-in-cooperative-board-games/
2011 DiGRA International Conference
journalArticle
arXiv:2107.07630 [cs]
Siu
Ho Chit
Pena
Jaime D.
Chang
Kimberlee C.
Chen
Edenna
Zhou
Yutai
Lopez
Victor J.
Palko
Kyle
Allen
Ross E.
Computer Science - Artificial Intelligence
Computer Science - Human-Computer Interaction
Evaluation of Human-AI Teams for Learned and Rule-Based Agents in Hanabi
Deep reinforcement learning has generated superhuman AI in competitive games such as Go and StarCraft. Can similar learning techniques create a superior AI teammate for human-machine collaborative games? Will humans prefer AI teammates that improve objective team performance or those that improve subjective metrics of trust? In this study, we perform a single-blind evaluation of teams of humans and AI agents in the cooperative card game Hanabi, with both rule-based and learning-based agents. In addition to the game score, used as an objective metric of the human-AI team performance, we also quantify subjective measures of the human's perceived performance, teamwork, interpretability, trust, and overall preference of AI teammate. We find that humans have a clear preference toward a rule-based AI teammate (SmartBot) over a state-of-the-art learning-based AI teammate (Other-Play) across nearly all subjective metrics, and generally view the learning-based agent negatively, despite no statistical difference in the game score. This result has implications for future AI design and reinforcement learning benchmarking, highlighting the need to incorporate subjective metrics of human-AI teaming rather than a singular focus on objective task performance.
2021-07-19
arXiv.org
http://arxiv.org/abs/2107.07630
2021-07-24 06:30:44
arXiv: 2107.07630
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2107.07630.pdf
2021-07-24 06:31:01
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2107.07630
2021-07-24 06:31:06
1
text/html
journalArticle
Litwiller
Bonnie H.
Duncan
David R.
Probabilites In Yahtzee
Teachers of units in probability are often interested in providing examples of probabilistic situations in a nonclassroom setting. Games are a rich source of such probabilities. Many people enjoy playing a commercial game called Yahtzee. A Yahtzee player receives points for achieving various specified numerical combinations of five dice during the three rolls that constitute a turn.
12/1982
DOI.org (Crossref)
https://pubs.nctm.org/view/journals/mt/75/9/article-p751.xml
2021-07-24 07:53:57
751-754
75
The Mathematics Teacher
DOI 10.5951/MT.75.9.0751
9
MT
ISSN 0025-5769, 2330-0582
presentation
Verhoeff
Tom
Optimal Solitaire Yahtzee Strategies
http://www.yahtzee.org.uk/optimal_yahtzee_TV.pdf
journalArticle
Bonarini
Andrea
Lazaric
Alessandro
Restelli
Marcello
Yahtzee: a Large Stochastic Environment for RL Benchmarks
Yahtzee is a game that is regularly played by more than 100 million people in the world. We
propose a simplified version of Yahtzee as a benchmark for RL algorithms. We have already
used it for this purpose, and an implementation is available.
http://researchers.lille.inria.fr/~lazaric/Webpage/PublicationsByTopic_files/bonarini2005yahtzee.pdf
1
thesis
KTH, School of Computer Science and Communication (CSC)
Serra
Andreas
Niigata
Kai Widell
Optimal Yahtzee performance in multi-player games
Yahtzee is a game with a moderately large search space, dependent on the factor of luck. This makes it not quite trivial to implement an optimal strategy for it. Using the optimal strategy for single-player
use, comparisons against other algorithms are made and the results are analyzed for hints on what it could take to make an algorithm that could beat the single-player optimal strategy.
April 12, 2013
en
http://www.diva-portal.org/smash/get/diva2:668705/FULLTEXT01.pdf
https://www.csc.kth.se/utbildning/kth/kurser/DD143X/dkand13/Group4Per/report/12-serra-widell-nigata.pdf
17
Independent thesis Basic level (degree of Bachelor)
manuscript
Verhoeff
Tom
How to Maximize Your Score in Solitaire Yahtzee
Yahtzee is a well-known game played with five dice. Players take turns at assembling and scoring dice patterns. The player with the highest score wins. Solitaire Yahtzee is a single-player version of Yahtzee aimed at maximizing one’s score. A strategy for playing Yahtzee determines which choice to make in each situation of the game. We show that the maximum expected score over all Solitaire Yahtzee strategies is 254.5896. . . .
en
http://www-set.win.tue.nl/~wstomv/misc/yahtzee/yahtzee-report-unfinished.pdf
18
Incomplete Draft
thesis
Yale University, Department of Computer Science
Vasseur
Philip
Using Deep Q-Learning to Compare Strategy Ladders of Yahtzee
“Bots” playing games is not a new concept,
likely going back to the first video games. However,
there has been a new wave recently using machine
learning to learn to play games at a near optimal
level - essentially using neural networks to “solve”
games. Depending on the game, this can be relatively
straight forward using supervised learning. However,
this requires having data for optimal play, which is
often not possible due to the sheer complexity of many
games. For example, solitaire Yahtzee has this data
available, but two player Yahtzee does not due to the
massive state space. A recent trend in response to this
started with Google Deep Mind in 2013, who used Deep
Reinforcement Learning to play various Atari games
[4].
This project will apply Deep Reinforcement Learning
(specifically Deep Q-Learning) and measure how an
agent learns to play Yahtzee in the form of a strategy
ladder. A strategy ladder is a way of looking at how
the performance of an AI varies with the computational
resources it uses. Different sets of rules changes how the
the AI learns which varies the strategy ladder itself. This
project will vary the upper bonus threshold and then
attempt to measure how “good” the various strategy
ladders are - in essence attempting to find the set of
rules which creates the “best” version of Yahtzee. We
assume/expect that there is some correlation between
strategy ladders for AI and strategy ladders for human,
meaning that a game with a “good” strategy ladder for
an AI indicates that game is interesting and challenging
for humans.
December 12, 2019
en
https://raw.githubusercontent.com/philvasseur/Yahtzee-DQN-Thesis/dcf2bfe15c3b8c0ff3256f02dd3c0aabdbcbc9bb/webpage/final_report.pdf
12
report
KTH Royal Institute Of Technology Computer Science And Communication
Defensive Yahtzee
In this project an algorithm has been created that plays Yahtzee using rule
based heuristics. The focus is getting a high lowest score and a high 10th
percentile. All rules of Yahtzee and the probabilities for each combination
have been studied and based on this each turn is optimized to get a
guaranteed decent high score. The algorithm got a lowest score of 79 and a
10th percentile of 152 when executed 100 000 times.
https://www.diva-portal.org/smash/get/diva2:817838/FULLTEXT01.pdf
http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-168668
22
report
Glenn
James
An Optimal Strategy for Yahtzee
http://www.cs.loyola.edu/~jglenn/research/optimal_yahtzee.pdf
presentation
Middlebury College
R. Teal Witter
Alex Lyford
Applications of Graph Theory and Probability in the Board Game Ticket to Ride
January 16, 2020
https://www.rtealwitter.com/slides/2020-JMM.pdf
attachment
Full Text
https://www.rtealwitter.com/slides/2020-JMM.pdf
2021-07-24 08:18:37
1
application/pdf
journalArticle
arXiv:1511.08099 [cs]
Cuayáhuitl
Heriberto
Keizer
Simon
Lemon
Oliver
Computer Science - Artificial Intelligence
Computer Science - Machine Learning
Strategic Dialogue Management via Deep Reinforcement Learning
Artificially intelligent agents equipped with strategic skills that can negotiate during their interactions with other natural or artificial agents are still underdeveloped. This paper describes a successful application of Deep Reinforcement Learning (DRL) for training intelligent agents with strategic conversational skills, in a situated dialogue setting. Previous studies have modelled the behaviour of strategic agents using supervised learning and traditional reinforcement learning techniques, the latter using tabular representations or learning with linear function approximation. In this study, we apply DRL with a high-dimensional state space to the strategic board game of Settlers of Catan---where players can offer resources in exchange for others and they can also reply to offers made by other players. Our experimental results report that the DRL-based learnt policies significantly outperformed several baselines including random, rule-based, and supervised-based behaviours. The DRL-based policy has a 53% win rate versus 3 automated players (`bots'), whereas a supervised player trained on a dialogue corpus in this setting achieved only 27%, versus the same 3 bots. This result supports the claim that DRL is a promising framework for training dialogue systems, and strategic agents with negotiation abilities.
2015-11-25
arXiv.org
http://arxiv.org/abs/1511.08099
2021-07-24 08:23:51
arXiv: 1511.08099
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1511.08099.pdf
2021-07-24 08:23:57
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1511.08099
2021-07-24 08:24:01
1
text/html
conferencePaper
ISBN 978-92-837-2336-3
14th NATO Operations Research and Analysis (OR&A) Conference: Emerging and Disruptive Technology
DOI 10.14339/STO-MP-SAS-OCS-ORA-2020-WCM-01-PDF
NATO
Christoffer Limér
Erik Kalmér
Mika Cohen
Monte Carlo Tree Search for Risk
2/16/2021
en
AC/323(SAS-ACT)TP/1017
https://www.sto.nato.int/publications/STO%20Meeting%20Proceedings/STO-MP-SAS-OCS-ORA-2020/MP-SAS-OCS-ORA-2020-WCM-01.pdf
attachment
Full Text
https://www.sto.nato.int/publications/STO%20Meeting%20Proceedings/STO-MP-SAS-OCS-ORA-2020/MP-SAS-OCS-ORA-2020-WCM-01.pdf
2021-07-24 08:34:15
1
application/pdf
presentation
Christoffer Limér
Erik Kalmér
Wargaming with Monte-Carlo Tree Search
2/16/2021
en
https://www.sto.nato.int/publications/STO%20Meeting%20Proceedings/STO-MP-SAS-OCS-ORA-2020/MP-SAS-OCS-ORA-2020-WCM-01P.pdf
attachment
Full Text
https://www.sto.nato.int/publications/STO%20Meeting%20Proceedings/STO-MP-SAS-OCS-ORA-2020/MP-SAS-OCS-ORA-2020-WCM-01P.pdf
2021-07-24 08:35:04
1
application/pdf
journalArticle
arXiv:1910.04376 [cs]
Zha
Daochen
Lai
Kwei-Herng
Cao
Yuanpu
Huang
Songyi
Wei
Ruzhe
Guo
Junyu
Hu
Xia
Computer Science - Artificial Intelligence
RLCard: A Toolkit for Reinforcement Learning in Card Games
RLCard is an open-source toolkit for reinforcement learning research in card games. It supports various card environments with easy-to-use interfaces, including Blackjack, Leduc Hold'em, Texas Hold'em, UNO, Dou Dizhu and Mahjong. The goal of RLCard is to bridge reinforcement learning and imperfect information games, and push forward the research of reinforcement learning in domains with multiple agents, large state and action space, and sparse reward. In this paper, we provide an overview of the key components in RLCard, a discussion of the design principles, a brief introduction of the interfaces, and comprehensive evaluations of the environments. The codes and documents are available at https://github.com/datamllab/rlcard
2020-02-14
RLCard
arXiv.org
http://arxiv.org/abs/1910.04376
2021-07-24 08:40:55
arXiv: 1910.04376
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/1910.04376.pdf
2021-07-24 08:40:59
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/1910.04376
2021-07-24 08:41:03
1
text/html
journalArticle
arXiv:2009.12065 [cs]
Gaina
Raluca D.
Balla
Martin
Dockhorn
Alexander
Montoliu
Raul
Perez-Liebana
Diego
Computer Science - Artificial Intelligence
Design and Implementation of TAG: A Tabletop Games Framework
This document describes the design and implementation of the Tabletop Games framework (TAG), a Java-based benchmark for developing modern board games for AI research. TAG provides a common skeleton for implementing tabletop games based on a common API for AI agents, a set of components and classes to easily add new games and an import module for defining data in JSON format. At present, this platform includes the implementation of seven different tabletop games that can also be used as an example for further developments. Additionally, TAG also incorporates logging functionality that allows the user to perform a detailed analysis of the game, in terms of action space, branching factor, hidden information, and other measures of interest for Game AI research. The objective of this document is to serve as a central point where the framework can be described at length. TAG can be downloaded at: https://github.com/GAIGResearch/TabletopGames
2020-09-25
Design and Implementation of TAG
arXiv.org
http://arxiv.org/abs/2009.12065
2021-07-24 08:41:01
arXiv: 2009.12065
attachment
arXiv Fulltext PDF
https://arxiv.org/pdf/2009.12065.pdf
2021-07-24 08:41:07
1
application/pdf
attachment
arXiv.org Snapshot
https://arxiv.org/abs/2009.12065
2021-07-24 08:41:11
1
text/html
computerProgram
Adam Stelmaszczyk
Game Tree Search Algorithms - C++ library for AI bot programming.
2015
Game Tree Search Algorithms
https://github.com/AdamStelmaszczyk/gtsa
C++
computerProgram
Raluca D. Gaina
Martin Balla
Alexander Dockhorn
Raul Montoliu
Diego Perez-Liebana
TAG: Tabletop Games Framework
The Tabletop Games Framework (TAG) is a Java-based benchmark for developing modern board games for AI research. TAG provides a common skeleton for implementing tabletop games based on a common API for AI agents, a set of components and classes to easily add new games and an import module for defining data in JSON format. At present, this platform includes the implementation of seven different tabletop games that can also be used as an example for further developments. Additionally, TAG also incorporates logging functionality that allows the user to perform a detailed analysis of the game, in terms of action space, branching factor, hidden information, and other measures of interest for Game AI research.
https://github.com/GAIGResearch/TabletopGames
MIT License
Java
thesis
Örebro University, School of Science and Technology.
Nguyen, Van Hoa
A Graphical User Interface For The Hanabi Challenge Benchmark
This report will describe the development of the Graphical User Interface (GUI) forthe Hanabi Challenge Benchmark. The benchmark is based on the popular cardgame Hanabi and presents itself as a new research frontier in artificial intelligencefor cooperative multi-agent challenges. The project’s intentions and goals are tointerpret and visualize the data output from the benchmark to give us a better understandingof it.A GUI was then developed by using knowledge within theory of mind in combinationwith theories within human-computer interaction. The results of this project wereevaluated through a small-scale usability test. Users of different ages, gender andlevels of computer knowledge tested the application and through a questionnaire,the quality of the GUI was assessed.
http://oru.diva-portal.org/smash/record.jsf?pid=diva2%3A1597503
journalArticle
Osawa
Hirotaka
Kawagoe
Atsushi
Sato
Eisuke
Kato
Takuya
Emergence of Cooperative Impression With Self-Estimation, Thinking Time, and Concordance of Risk Sensitivity in Playing Hanabi
The authors evaluate the extent to which a user’s impression of an AI agent can be improved by giving the agent the ability of self-estimation, thinking time, and coordination of risk tendency. The authors modified the algorithm of an AI agent in the cooperative game Hanabi to have all of these traits, and investigated the change in the user’s impression by playing with the user. The authors used a self-estimation task to evaluate the effect that the ability to read the intention of a user had on an impression. The authors also show thinking time of an agent influences impression for an agent. The authors also investigated the relationship between the concordance of the risk-taking tendencies of players and agents, the player’s impression of agents, and the game experience. The results of the self-estimation task experiment showed that the more accurate the estimation of the agent’s self, the more likely it is that the partner will perceive humanity, affinity, intelligence, and communication skills in the agent. The authors also found that an agent that changes the length of thinking time according to the priority of action gives the impression that it is smarter than an agent with a normal thinking time when the player notices the difference in thinking time or an agent that randomly changes the thinking time. The result of the experiment regarding concordance of the risk-taking tendency shows that influence player’s impression toward agents. These results suggest that game agent designers can improve the player’s disposition toward an agent and the game experience by adjusting the agent’s self-estimation level, thinking time, and risk-taking tendency according to the player’s personality and inner state during the game.
2021-10-12
DOI.org (Crossref)
https://www.frontiersin.org/articles/10.3389/frobt.2021.658348/full
2021-11-24 07:14:38
658348
8
Frontiers in Robotics and AI
DOI 10.3389/frobt.2021.658348
Front. Robot. AI
ISSN 2296-9144
attachment
Full Text
https://www.frontiersin.org/articles/10.3389/frobt.2021.658348/pdf
2021-11-24 07:15:06
1
application/pdf
2048
Accessibility
Azul
Blokus
Carcassonne
Diplomacy
Dixit
Dominion
Frameworks
Game Design
Hanabi
Hive
Jenga
Kingdomino
Lost Cities
Mafia
Magic: The Gathering
Mobile Games
Modern Art: The card game
Monopoly
Monopoly Deal
Nmbr9
Pandemic
Patchwork
Pentago
Quixo
Race for the Galaxy
Resistance: Avalon
RISK
Santorini
Scotland Yard
Secret Hitler
Set
Settlers of Catan
Shobu
Terra Mystica
Tetris Link
Ticket to Ride
Ultimate Tic-Tac-Toe
UNO
Yahtzee