SlideShare ist ein Scribd-Unternehmen logo
1 von 22
IEEE Conference on Computational Intelligence and Games
2018 IEEE CIG
StarCraft AI Competition
Seonghun Yoon, Kyung-Joong Kim
Cognition and Intelligence Lab (http://cilab.sejong.ac.kr)
Sejong University, Seoul, Republic of Korea
StarCraft
(Blizzard, 1998)
• StarCraft is the real-time strategy game to win the enemy with limited resource in
the specific maps.
• This real-time strategy (RTS) game has been known as one of the most difficult
video games to solve.
• People think that this game might be the next target for AI researchers after
oriental Go success (Alpha Go).
2
StarCraft Brood War StarCraft Remastered
[1] M.-J. Kim, K.-J. Kim, S.-J. Kim, and A. K. Dey, “Evaluation of StarCraft Artificial Intelligence Competition Bots by
Experienced Human Players,” ACM CHI (Late-Breaking Work), 2016
CIG 2018
~2018
Now three annual AI Competition
2011
𝟏 𝒔𝒕 Student StarCraft AI(SSCAIT)
Competition
2017
StarCraft II API released
2010
𝟏 𝒔𝒕
IEEE CIG StarCraft AI Competition
𝟏 𝒔𝒕
AIIDE StarCraft AI Competition
History of StarCraft AI
http://www.cs.mun.ca/~dchurchill/starcraftaicomp/history.shtml
3
2009
BWAPI was developed
CIG 2018
Competition
Setup
• StarCraft Brood War
• BWAPI (Brood War Programming Interface)
• C++ or JAVA or Proxy Bot (any programming
language)
• Full-round robin style competition
• Winner is the one with highest win ratio
• 5 maps used using seed number of
participants(randomly selected from official
maps)
• 125 rounds (351 games/round)
• Total 43,875 games (25 days using 14 VMs)
• Open-source policy
4
CIG 2018
5
Submissions
0
5
10
15
20
25
30
2011 2012 2013 2014 2015 2016 2017 2018
10 10
8
13 14 16
20
27
The Number of Submissions
CIG 2018
27 Entrants from 14 Countries
6
New Entries Upgrade from 2017 Version
Re-entrance of
2017 Version
# of
Entries
9 7 11
Bots
Korean (Korea)
TitanIron (China)
Locutus (Denmark)
CUNYbot
ISAMind
Stormbreaker
Ecgberht (Spain)
Microwave
Steamhammer (USA)
ZZZKBot (Australia & UK,1st in 2017)
tscmoo (Norway,2nd in 2017)
PurpleWave (USA,3rd in 2017)
LetaBot (Netherland, 4th in 2017)
McRave (Canada)
MegaBot (Brazil)
Tyr (Netherland)
Ziabot (Korea)
Iron (France)
Sling (Korea)
Aiur (France)
Bonjwa (Vietnam)
OpprimoBot (Sweden)
Salsa (Spain)
TerranUAB (Poland)
UAlbertaBot (Canada)
Overkill (China)
SRbotOne (Canada)
CIG 2018
7
Race Distribution
CIG 2018
0
1
2
3
4
5
6
7
8
9
10
2013 2014 2015 2016 2017 2018
4 4
6
10
4
6
8 8
7
8
4
7
4
3
5
7
1
2 2
Zerg Terran Protoss Random
8
Year Winner
2018 ?
2017 ZZZKBOT
2016 TSCMOO
2015 ZZZKBOT
2014 ICEBOT
2013 SkyNet
2012 SkyNet
2011 SkyNet
Race Distribution
CIG 2018
Zerg
Terran
Protoss
Random
Race of Winner
2
2
3
Zerg
Terran
Protoss
Random
# of Players Maps
Two Players
Map
(2)Destination1.1
Three Players
Map
(3)TauCross1.1 (3)GreatBarrierReef1.0
Four Players
Map
(4)Andromeda1.0 (4)Python1.3 9
Competition Maps
CIG 2018
10
0
10000
20000
30000
40000
50000
2011 2012 2013 2014 2015 2016 2017 2018
40
4050 2500 4680 2730
14800
47500
43875
Total Games Played
CIG 2018
The Trend of Bots
CIG 2018
4
198
2017 2018
# of Bots using the File I/O
Non-Use
Use
2
7
10 12
2017 2018
# of Bots using the Machine Learning
• FILE I/O is allowed AI Players can save experience over rounds to adapt strategies
33.34%
100%
36.84%
16.67%
AI Bots
12
CIG 2018
CIG 2018
Bot AI techniques
PurpleWave Task network : High-level decision-making
McRave BFS pathfinding, minor learning : choosing build orders
Korean Custom pathfinding
LetaBot MCTS, A*
CUNYbot Genetic Algorithm : Standard economic models
ISAMind Neural network : An enemy strategy prediction
Stormbreaker Deep convolution Net as the actor function and critic function
MegaBot MinimaxQ-learning
Ecgberht UCB-1, Gaussian MeanShift clustering, behavioural trees
ZZZKBot Greedy search algorithm
Tyr A few state machines
13
The Overall AI Techniques of Bots
Results Announcement
14
CIG 2018
15
RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame
Result ( 𝟏 𝒔𝒕
~ 𝟑 𝒓𝒅
)
CIG 2018
11 LocutusLocutus 3,2503,250 29922992 258258 92.06%92.06%
Congratulation!
Winner of 2018 IEEE CIG
StarCraft AI Competition
Locutus by Bruce Nielsen
from Denmark
Locutus is simple learning for choosing openings, based on what
has worked in past games. ( Using a weighting factor. )
Macro , forge expand vs. zerg are the strongest aspects.
22 PurpleWavePurpleWave 3,2503,250 29622962 288288 91.14%91.14%
Most are taken from study of pro play.
exploiting the weaknesses of the enemy bot.
Forming ad-hoc combat squads using a Fast clustering algorithm
33 McRaveMcRave 3,2503,250 26682668 582582 82.09%82.09%
Winning small battles and maintain an advantage as soon as possible.
Some minor learning for choosing build orders using multi-armed bandit.
BFS pathfinding for every combat units.
Result ( 𝟏 𝒔𝒕
~ 𝟏𝟎 𝒕𝒉
)
16
22 PurpleWavePurpleWave 3,2503,250 29622962 288288 91.1491.14
33 McRaveMcRave 3,2503,250 26682668 582582 82.0982.09
RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame
11 LocutusLocutus 3,2503,250 29922992 258258 92.0692.06
55 ISAMindISAMind 3,2503,250 26102610 640640 80.3180.31
66 IronIron 3,2503,250 24152415 835835 74.3174.31
44 tscmootscmoo 3,2503,250 26422642 608608 81.2981.29
88 MicrowaveMicrowave 3,2503,250 21072107 11431143 64.8364.83
99 LetaBotLetaBot 3,2503,250 20672067 11831183 63.663.6
77 ZZZKBotZZZKBot 3,2503,250 22452245 10051005 69.0869.08
1010 MegaBotMegaBot 3,2503,250 19881988 12621262 61.1761.17
OccupationOccupation
PublicPublic
Young
Professional
Young
Professional
PublicPublic
Young
Professional
Young
Professional
2017 bot2017 bot
PublicPublic
PublicPublic
StudentStudent
PublicPublic
StudentStudent
CIG 2018
PublicPublic
2017 bot2017 bot
Result ( 𝟏𝟏 𝒕𝒉
~ 𝟐𝟎 𝒕𝒉
)
17
1212 TyrTyr 3,2503,250 18601860 13901390 57.2357.23
1313 EcgberhtEcgberht 3,2503,250 17161716 15341534 52.852.8
RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame
1111 UAIbertaBotUAIbertaBot 3,2503,250 19691969 12811281 60.5860.58
1515 TitanIronTitanIron 3,2503,250 16721672 15781578 51.4551.45
1616 ZiabotZiabot 3,2503,250 16601660 15901590 51.0851.08
1414 AiurAiur 3,2503,250 16751675 15751575 51.5451.54
1818 OverkillOverkill 3,2503,250 11271127 21232123 34.6834.68
1919 TerranUABTerranUAB 3,2503,250 11181118 21322132 34.434.4
1717 SteamhammerSteamhammer 3,2503,250 11321132 21182118 34.8334.83
2020 CUNYbotCUNYbot 3,2503,250 959959 22912291 29.5129.51
OccupationOccupation
CIG 2018
PublicPublic
Young
Professional
Young
Professional
2017 bot2017 bot
Young
Professional
Young
Professional
2017 bot2017 bot
2017 bot2017 bot
2017 bot2017 bot
PublicPublic
Young
Professional
Young
Professional
StudentStudent
2017 bot2017 bot
2017 bot2017 bot
Result ( 𝟐𝟏 𝒔𝒕
~ 𝟐𝟕 𝒕𝒉
)
18
2222 SlingSling 3,2503,250 862862 23882388 26.5226.52
2323 SRbotOneSRbotOne 3,2503,250 792792 24582458 24.3724.37
RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame
2121 OpprimoBotOpprimoBot 3,2503,250 881881 23692369 27.1127.11
2525 StormbreakerStormbreaker 3,2503,250 690690 25602560 21.2321.23
2626 KoreanKorean 3,2503,250 250250 30003000 7.697.69
2424 BonjwaBonjwa 3,2503,250 766766 24842484 23.5723.57
2727 SalsaSalsa 3,2503,250 5050 32003200 1.541.54
OccupationOccupation
CIG 2018
StudentStudent
2017 bot2017 bot
Young
Professional
Young
Professional
2017 bot2017 bot
2017 bot2017 bot
• 추가 사항 : 라운드 별 승률 (FILE I/O) AIIDE 2017 참고
CIG 2018
Win Percentage Over Time
Locutus ( Ranked #1 )
tscmoo ( Ranked #4 )
Overkill ( Ranked #18 )
ZZZKBot ( Ranked #7 )
PurpleWave ( Ranked #2 )
Highlight Video
CIG 2018
Dave Churchill
Organizer of AIIDE StarCraft AI Competition
Acknowledgements
Michal Certicky and his team
Organizer of Student StarCraft AI Tournament
21
CIG 2018
Thank You
Email: starcraft.aic@gmail.com
www.cilab.sejong.ac.kr/sc_competition
22
CIG 2018

Weitere ähnliche Inhalte

Ähnlich wie CIG 2018 StarCraft AI Competition

Maker Faireを持続可能にするには?
Maker Faireを持続可能にするには?Maker Faireを持続可能にするには?
Maker Faireを持続可能にするには?Shigeru Kobayashi
 
Descriptive analytics in r programming language
Descriptive analytics in r programming languageDescriptive analytics in r programming language
Descriptive analytics in r programming languageAshwini Mathur
 
Geometry Friends Game AI Competition - 2013 Results
Geometry Friends Game AI Competition - 2013 ResultsGeometry Friends Game AI Competition - 2013 Results
Geometry Friends Game AI Competition - 2013 ResultsRui Prada
 
九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)
九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)
九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)Yuichi Morito
 
Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...
Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...
Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...Associate Professor in VSB Coimbatore
 
Mobile Games Introduction - Mobile Monday Amsterdam February 2009
Mobile Games Introduction - Mobile Monday Amsterdam February 2009Mobile Games Introduction - Mobile Monday Amsterdam February 2009
Mobile Games Introduction - Mobile Monday Amsterdam February 2009Ex Machina
 
2017 Fighting Game AI Competition
2017 Fighting Game AI Competition2017 Fighting Game AI Competition
2017 Fighting Game AI Competitionftgaic
 
Modeling computer networks by colored Petri nets
Modeling computer networks by colored Petri netsModeling computer networks by colored Petri nets
Modeling computer networks by colored Petri netsDmitryZaitsev5
 
Duplicates everywhere (Kiev)
Duplicates everywhere (Kiev)Duplicates everywhere (Kiev)
Duplicates everywhere (Kiev)Alexey Grigorev
 
運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...
運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...
運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...Herman Wu
 
Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)
Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)
Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)Hashers United
 
Applying AI in Games (GDC2019)
Applying AI in Games (GDC2019)Applying AI in Games (GDC2019)
Applying AI in Games (GDC2019)Jun Okumura
 
NVIDIA press conference at CES 2014
NVIDIA press conference at CES 2014NVIDIA press conference at CES 2014
NVIDIA press conference at CES 2014NVIDIA
 
Why biased matrix factorization works well?
Why biased matrix factorization works well?Why biased matrix factorization works well?
Why biased matrix factorization works well?Joonyoung Yi
 
Game industry Three Gates Intensivdagarna 2011
Game industry  Three Gates Intensivdagarna 2011Game industry  Three Gates Intensivdagarna 2011
Game industry Three Gates Intensivdagarna 2011Stefan_Riksutstallningar
 
Android game engine
Android game engineAndroid game engine
Android game engineJulian Chu
 

Ähnlich wie CIG 2018 StarCraft AI Competition (20)

Maker Faireを持続可能にするには?
Maker Faireを持続可能にするには?Maker Faireを持続可能にするには?
Maker Faireを持続可能にするには?
 
TIC-TAC-TOE IN C
TIC-TAC-TOE IN CTIC-TAC-TOE IN C
TIC-TAC-TOE IN C
 
Descriptive analytics in r programming language
Descriptive analytics in r programming languageDescriptive analytics in r programming language
Descriptive analytics in r programming language
 
Geometry Friends Game AI Competition - 2013 Results
Geometry Friends Game AI Competition - 2013 ResultsGeometry Friends Game AI Competition - 2013 Results
Geometry Friends Game AI Competition - 2013 Results
 
九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)
九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)
九州シェアリングサミット2018 in 熊本 第4セッション(西岡誠氏)
 
Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...
Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...
Comparative Study on DES and Triple DES Algorithms and Proposal of a New Algo...
 
The City as a Platform : Week 5
The City as a Platform : Week 5The City as a Platform : Week 5
The City as a Platform : Week 5
 
MoMo #9 - Jeroen Elfferich
MoMo #9 - Jeroen ElfferichMoMo #9 - Jeroen Elfferich
MoMo #9 - Jeroen Elfferich
 
Mobile Games Introduction - Mobile Monday Amsterdam February 2009
Mobile Games Introduction - Mobile Monday Amsterdam February 2009Mobile Games Introduction - Mobile Monday Amsterdam February 2009
Mobile Games Introduction - Mobile Monday Amsterdam February 2009
 
2017 Fighting Game AI Competition
2017 Fighting Game AI Competition2017 Fighting Game AI Competition
2017 Fighting Game AI Competition
 
Modeling computer networks by colored Petri nets
Modeling computer networks by colored Petri netsModeling computer networks by colored Petri nets
Modeling computer networks by colored Petri nets
 
Duplicates everywhere (Kiev)
Duplicates everywhere (Kiev)Duplicates everywhere (Kiev)
Duplicates everywhere (Kiev)
 
WaveEngine Dotnet 2018
WaveEngine Dotnet 2018WaveEngine Dotnet 2018
WaveEngine Dotnet 2018
 
運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...
運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...
運用CNTK 實作深度學習物件辨識 Deep Learning based Object Detection with Microsoft Cogniti...
 
Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)
Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)
Building a Cost-effective Mining Rig by Michael Carter (BitsBeTrippin)
 
Applying AI in Games (GDC2019)
Applying AI in Games (GDC2019)Applying AI in Games (GDC2019)
Applying AI in Games (GDC2019)
 
NVIDIA press conference at CES 2014
NVIDIA press conference at CES 2014NVIDIA press conference at CES 2014
NVIDIA press conference at CES 2014
 
Why biased matrix factorization works well?
Why biased matrix factorization works well?Why biased matrix factorization works well?
Why biased matrix factorization works well?
 
Game industry Three Gates Intensivdagarna 2011
Game industry  Three Gates Intensivdagarna 2011Game industry  Three Gates Intensivdagarna 2011
Game industry Three Gates Intensivdagarna 2011
 
Android game engine
Android game engineAndroid game engine
Android game engine
 

Kürzlich hochgeladen

Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...Boston Institute of Analytics
 
Cyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded dataCyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded dataTecnoIncentive
 
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...Amil Baba Dawood bangali
 
Decoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis ProjectDecoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis ProjectBoston Institute of Analytics
 
Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Cathrine Wilhelmsen
 
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...Boston Institute of Analytics
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPTBoston Institute of Analytics
 
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...Milind Agarwal
 
INTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processingINTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processingsocarem879
 
SMOTE and K-Fold Cross Validation-Presentation.pptx
SMOTE and K-Fold Cross Validation-Presentation.pptxSMOTE and K-Fold Cross Validation-Presentation.pptx
SMOTE and K-Fold Cross Validation-Presentation.pptxHaritikaChhatwal1
 
Bank Loan Approval Analysis: A Comprehensive Data Analysis Project
Bank Loan Approval Analysis: A Comprehensive Data Analysis ProjectBank Loan Approval Analysis: A Comprehensive Data Analysis Project
Bank Loan Approval Analysis: A Comprehensive Data Analysis ProjectBoston Institute of Analytics
 
wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...
wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...
wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...KarteekMane1
 
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...Thomas Poetter
 
Principles and Practices of Data Visualization
Principles and Practices of Data VisualizationPrinciples and Practices of Data Visualization
Principles and Practices of Data VisualizationKianJazayeri1
 
modul pembelajaran robotic Workshop _ by Slidesgo.pptx
modul pembelajaran robotic Workshop _ by Slidesgo.pptxmodul pembelajaran robotic Workshop _ by Slidesgo.pptx
modul pembelajaran robotic Workshop _ by Slidesgo.pptxaleedritatuxx
 
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...Boston Institute of Analytics
 
Learn How Data Science Changes Our World
Learn How Data Science Changes Our WorldLearn How Data Science Changes Our World
Learn How Data Science Changes Our WorldEduminds Learning
 
The Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptx
The Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptxThe Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptx
The Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptxTasha Penwell
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryJeremy Anderson
 
convolutional neural network and its applications.pdf
convolutional neural network and its applications.pdfconvolutional neural network and its applications.pdf
convolutional neural network and its applications.pdfSubhamKumar3239
 

Kürzlich hochgeladen (20)

Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
Decoding the Heart: Student Presentation on Heart Attack Prediction with Data...
 
Cyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded dataCyber awareness ppt on the recorded data
Cyber awareness ppt on the recorded data
 
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
NO1 Certified Black Magic Specialist Expert Amil baba in Lahore Islamabad Raw...
 
Decoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis ProjectDecoding Patterns: Customer Churn Prediction Data Analysis Project
Decoding Patterns: Customer Churn Prediction Data Analysis Project
 
Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)Data Factory in Microsoft Fabric (MsBIP #82)
Data Factory in Microsoft Fabric (MsBIP #82)
 
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
Data Analysis Project Presentation: Unveiling Your Ideal Customer, Bank Custo...
 
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default  Presentation : Data Analysis Project PPTPredictive Analysis for Loan Default  Presentation : Data Analysis Project PPT
Predictive Analysis for Loan Default Presentation : Data Analysis Project PPT
 
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
Unveiling the Role of Social Media Suspect Investigators in Preventing Online...
 
INTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processingINTRODUCTION TO Natural language processing
INTRODUCTION TO Natural language processing
 
SMOTE and K-Fold Cross Validation-Presentation.pptx
SMOTE and K-Fold Cross Validation-Presentation.pptxSMOTE and K-Fold Cross Validation-Presentation.pptx
SMOTE and K-Fold Cross Validation-Presentation.pptx
 
Bank Loan Approval Analysis: A Comprehensive Data Analysis Project
Bank Loan Approval Analysis: A Comprehensive Data Analysis ProjectBank Loan Approval Analysis: A Comprehensive Data Analysis Project
Bank Loan Approval Analysis: A Comprehensive Data Analysis Project
 
wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...
wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...
wepik-insightful-infographics-a-data-visualization-overview-20240401133220kwr...
 
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
Minimizing AI Hallucinations/Confabulations and the Path towards AGI with Exa...
 
Principles and Practices of Data Visualization
Principles and Practices of Data VisualizationPrinciples and Practices of Data Visualization
Principles and Practices of Data Visualization
 
modul pembelajaran robotic Workshop _ by Slidesgo.pptx
modul pembelajaran robotic Workshop _ by Slidesgo.pptxmodul pembelajaran robotic Workshop _ by Slidesgo.pptx
modul pembelajaran robotic Workshop _ by Slidesgo.pptx
 
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
Data Analysis Project : Targeting the Right Customers, Presentation on Bank M...
 
Learn How Data Science Changes Our World
Learn How Data Science Changes Our WorldLearn How Data Science Changes Our World
Learn How Data Science Changes Our World
 
The Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptx
The Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptxThe Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptx
The Power of Data-Driven Storytelling_ Unveiling the Layers of Insight.pptx
 
Defining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data StoryDefining Constituents, Data Vizzes and Telling a Data Story
Defining Constituents, Data Vizzes and Telling a Data Story
 
convolutional neural network and its applications.pdf
convolutional neural network and its applications.pdfconvolutional neural network and its applications.pdf
convolutional neural network and its applications.pdf
 

CIG 2018 StarCraft AI Competition

  • 1. IEEE Conference on Computational Intelligence and Games 2018 IEEE CIG StarCraft AI Competition Seonghun Yoon, Kyung-Joong Kim Cognition and Intelligence Lab (http://cilab.sejong.ac.kr) Sejong University, Seoul, Republic of Korea
  • 2. StarCraft (Blizzard, 1998) • StarCraft is the real-time strategy game to win the enemy with limited resource in the specific maps. • This real-time strategy (RTS) game has been known as one of the most difficult video games to solve. • People think that this game might be the next target for AI researchers after oriental Go success (Alpha Go). 2 StarCraft Brood War StarCraft Remastered [1] M.-J. Kim, K.-J. Kim, S.-J. Kim, and A. K. Dey, “Evaluation of StarCraft Artificial Intelligence Competition Bots by Experienced Human Players,” ACM CHI (Late-Breaking Work), 2016 CIG 2018
  • 3. ~2018 Now three annual AI Competition 2011 𝟏 𝒔𝒕 Student StarCraft AI(SSCAIT) Competition 2017 StarCraft II API released 2010 𝟏 𝒔𝒕 IEEE CIG StarCraft AI Competition 𝟏 𝒔𝒕 AIIDE StarCraft AI Competition History of StarCraft AI http://www.cs.mun.ca/~dchurchill/starcraftaicomp/history.shtml 3 2009 BWAPI was developed CIG 2018
  • 4. Competition Setup • StarCraft Brood War • BWAPI (Brood War Programming Interface) • C++ or JAVA or Proxy Bot (any programming language) • Full-round robin style competition • Winner is the one with highest win ratio • 5 maps used using seed number of participants(randomly selected from official maps) • 125 rounds (351 games/round) • Total 43,875 games (25 days using 14 VMs) • Open-source policy 4 CIG 2018
  • 5. 5 Submissions 0 5 10 15 20 25 30 2011 2012 2013 2014 2015 2016 2017 2018 10 10 8 13 14 16 20 27 The Number of Submissions CIG 2018
  • 6. 27 Entrants from 14 Countries 6 New Entries Upgrade from 2017 Version Re-entrance of 2017 Version # of Entries 9 7 11 Bots Korean (Korea) TitanIron (China) Locutus (Denmark) CUNYbot ISAMind Stormbreaker Ecgberht (Spain) Microwave Steamhammer (USA) ZZZKBot (Australia & UK,1st in 2017) tscmoo (Norway,2nd in 2017) PurpleWave (USA,3rd in 2017) LetaBot (Netherland, 4th in 2017) McRave (Canada) MegaBot (Brazil) Tyr (Netherland) Ziabot (Korea) Iron (France) Sling (Korea) Aiur (France) Bonjwa (Vietnam) OpprimoBot (Sweden) Salsa (Spain) TerranUAB (Poland) UAlbertaBot (Canada) Overkill (China) SRbotOne (Canada) CIG 2018
  • 7. 7 Race Distribution CIG 2018 0 1 2 3 4 5 6 7 8 9 10 2013 2014 2015 2016 2017 2018 4 4 6 10 4 6 8 8 7 8 4 7 4 3 5 7 1 2 2 Zerg Terran Protoss Random
  • 8. 8 Year Winner 2018 ? 2017 ZZZKBOT 2016 TSCMOO 2015 ZZZKBOT 2014 ICEBOT 2013 SkyNet 2012 SkyNet 2011 SkyNet Race Distribution CIG 2018 Zerg Terran Protoss Random Race of Winner 2 2 3 Zerg Terran Protoss Random
  • 9. # of Players Maps Two Players Map (2)Destination1.1 Three Players Map (3)TauCross1.1 (3)GreatBarrierReef1.0 Four Players Map (4)Andromeda1.0 (4)Python1.3 9 Competition Maps CIG 2018
  • 10. 10 0 10000 20000 30000 40000 50000 2011 2012 2013 2014 2015 2016 2017 2018 40 4050 2500 4680 2730 14800 47500 43875 Total Games Played CIG 2018
  • 11. The Trend of Bots CIG 2018 4 198 2017 2018 # of Bots using the File I/O Non-Use Use 2 7 10 12 2017 2018 # of Bots using the Machine Learning • FILE I/O is allowed AI Players can save experience over rounds to adapt strategies 33.34% 100% 36.84% 16.67%
  • 13. CIG 2018 Bot AI techniques PurpleWave Task network : High-level decision-making McRave BFS pathfinding, minor learning : choosing build orders Korean Custom pathfinding LetaBot MCTS, A* CUNYbot Genetic Algorithm : Standard economic models ISAMind Neural network : An enemy strategy prediction Stormbreaker Deep convolution Net as the actor function and critic function MegaBot MinimaxQ-learning Ecgberht UCB-1, Gaussian MeanShift clustering, behavioural trees ZZZKBot Greedy search algorithm Tyr A few state machines 13 The Overall AI Techniques of Bots
  • 15. 15 RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame Result ( 𝟏 𝒔𝒕 ~ 𝟑 𝒓𝒅 ) CIG 2018 11 LocutusLocutus 3,2503,250 29922992 258258 92.06%92.06% Congratulation! Winner of 2018 IEEE CIG StarCraft AI Competition Locutus by Bruce Nielsen from Denmark Locutus is simple learning for choosing openings, based on what has worked in past games. ( Using a weighting factor. ) Macro , forge expand vs. zerg are the strongest aspects. 22 PurpleWavePurpleWave 3,2503,250 29622962 288288 91.14%91.14% Most are taken from study of pro play. exploiting the weaknesses of the enemy bot. Forming ad-hoc combat squads using a Fast clustering algorithm 33 McRaveMcRave 3,2503,250 26682668 582582 82.09%82.09% Winning small battles and maintain an advantage as soon as possible. Some minor learning for choosing build orders using multi-armed bandit. BFS pathfinding for every combat units.
  • 16. Result ( 𝟏 𝒔𝒕 ~ 𝟏𝟎 𝒕𝒉 ) 16 22 PurpleWavePurpleWave 3,2503,250 29622962 288288 91.1491.14 33 McRaveMcRave 3,2503,250 26682668 582582 82.0982.09 RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame 11 LocutusLocutus 3,2503,250 29922992 258258 92.0692.06 55 ISAMindISAMind 3,2503,250 26102610 640640 80.3180.31 66 IronIron 3,2503,250 24152415 835835 74.3174.31 44 tscmootscmoo 3,2503,250 26422642 608608 81.2981.29 88 MicrowaveMicrowave 3,2503,250 21072107 11431143 64.8364.83 99 LetaBotLetaBot 3,2503,250 20672067 11831183 63.663.6 77 ZZZKBotZZZKBot 3,2503,250 22452245 10051005 69.0869.08 1010 MegaBotMegaBot 3,2503,250 19881988 12621262 61.1761.17 OccupationOccupation PublicPublic Young Professional Young Professional PublicPublic Young Professional Young Professional 2017 bot2017 bot PublicPublic PublicPublic StudentStudent PublicPublic StudentStudent CIG 2018
  • 17. PublicPublic 2017 bot2017 bot Result ( 𝟏𝟏 𝒕𝒉 ~ 𝟐𝟎 𝒕𝒉 ) 17 1212 TyrTyr 3,2503,250 18601860 13901390 57.2357.23 1313 EcgberhtEcgberht 3,2503,250 17161716 15341534 52.852.8 RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame 1111 UAIbertaBotUAIbertaBot 3,2503,250 19691969 12811281 60.5860.58 1515 TitanIronTitanIron 3,2503,250 16721672 15781578 51.4551.45 1616 ZiabotZiabot 3,2503,250 16601660 15901590 51.0851.08 1414 AiurAiur 3,2503,250 16751675 15751575 51.5451.54 1818 OverkillOverkill 3,2503,250 11271127 21232123 34.6834.68 1919 TerranUABTerranUAB 3,2503,250 11181118 21322132 34.434.4 1717 SteamhammerSteamhammer 3,2503,250 11321132 21182118 34.8334.83 2020 CUNYbotCUNYbot 3,2503,250 959959 22912291 29.5129.51 OccupationOccupation CIG 2018 PublicPublic Young Professional Young Professional 2017 bot2017 bot Young Professional Young Professional 2017 bot2017 bot 2017 bot2017 bot 2017 bot2017 bot PublicPublic Young Professional Young Professional StudentStudent
  • 18. 2017 bot2017 bot 2017 bot2017 bot Result ( 𝟐𝟏 𝒔𝒕 ~ 𝟐𝟕 𝒕𝒉 ) 18 2222 SlingSling 3,2503,250 862862 23882388 26.5226.52 2323 SRbotOneSRbotOne 3,2503,250 792792 24582458 24.3724.37 RankRank Bot NameBot Name WinWinRaceRace LoseLose Win RateWin RateGameGame 2121 OpprimoBotOpprimoBot 3,2503,250 881881 23692369 27.1127.11 2525 StormbreakerStormbreaker 3,2503,250 690690 25602560 21.2321.23 2626 KoreanKorean 3,2503,250 250250 30003000 7.697.69 2424 BonjwaBonjwa 3,2503,250 766766 24842484 23.5723.57 2727 SalsaSalsa 3,2503,250 5050 32003200 1.541.54 OccupationOccupation CIG 2018 StudentStudent 2017 bot2017 bot Young Professional Young Professional 2017 bot2017 bot 2017 bot2017 bot
  • 19. • 추가 사항 : 라운드 별 승률 (FILE I/O) AIIDE 2017 참고 CIG 2018 Win Percentage Over Time Locutus ( Ranked #1 ) tscmoo ( Ranked #4 ) Overkill ( Ranked #18 ) ZZZKBot ( Ranked #7 ) PurpleWave ( Ranked #2 )
  • 21. Dave Churchill Organizer of AIIDE StarCraft AI Competition Acknowledgements Michal Certicky and his team Organizer of Student StarCraft AI Tournament 21 CIG 2018