8 0Zkzbz0z 7 Opoqz0z0 6 Po0zps0z 5 Z0s0apop 4 0Zpzpzpz 3 Zko0mpz0 2 Naroqz0z 1 Znz0z0z0 a B C D E F G H
Total Page:16
File Type:pdf, Size:1020Kb

Load more
Recommended publications
-
Elements of DSAI: Game Tree Search, Learning Architectures
Introduction Games Game Search Evaluation Fns AlphaGo/Zero Summary Quiz References Elements of DSAI AlphaGo Part 2: Game Tree Search, Learning Architectures Search & Learn: A Recipe for AI Action Decisions J¨orgHoffmann Winter Term 2019/20 Hoffmann Elements of DSAI Game Tree Search, Learning Architectures 1/49 Introduction Games Game Search Evaluation Fns AlphaGo/Zero Summary Quiz References Competitive Agents? Quote AI Introduction: \Single agent vs. multi-agent: One agent or several? Competitive or collaborative?" ! Single agent! Several koalas, several gorillas trying to beat these up. BUT there is only a single acting entity { one player decides which moves to take (who gets into the boat). Hoffmann Elements of DSAI Game Tree Search, Learning Architectures 3/49 Introduction Games Game Search Evaluation Fns AlphaGo/Zero Summary Quiz References Competitive Agents! Quote AI Introduction: \Single agent vs. multi-agent: One agent or several? Competitive or collaborative?" ! Multi-agent competitive! TWO players deciding which moves to take. Conflicting interests. Hoffmann Elements of DSAI Game Tree Search, Learning Architectures 4/49 Introduction Games Game Search Evaluation Fns AlphaGo/Zero Summary Quiz References Agenda: Game Search, AlphaGo Architecture Games: What is that? ! Game categories, game solutions. Game Search: How to solve a game? ! Searching the game tree. Evaluation Functions: How to evaluate a game position? ! Heuristic functions for games. AlphaGo: How does it work? ! Overview of AlphaGo architecture, and changes in Alpha(Go) Zero. Hoffmann Elements of DSAI Game Tree Search, Learning Architectures 5/49 Introduction Games Game Search Evaluation Fns AlphaGo/Zero Summary Quiz References Positioning in the DSAI Phase Model Hoffmann Elements of DSAI Game Tree Search, Learning Architectures 6/49 Introduction Games Game Search Evaluation Fns AlphaGo/Zero Summary Quiz References Which Games? ! No chance element. -
Drago Document
C:\home\kuroki\Dropbox\Go\Master60.sgf Printed by Drago ⿊ NN ⽩ Magist ⽇時 2016-12-29 19:01:34 結果 ⽩の 中押し勝ち A B C D E F G H J K L M N O P Q R S T A B C D E F G H J K L M N O P Q R S T 19 19 19 19 18 18 1849 48 41 51 37 33 39 18 1723 3 20 15 14 10 9 11 17 17 1750 35 34 38 40 17 1619 18 22 28 29 26 12 8 16 1 16 16 16 1521 24 27 30 15 1531 15 1425 13 14 1436 32 43 56 55 14 13 13 1344 58 54 13 127 12 1245 57 12 11 11 1152 46 53 11 10 10 1047 10 9999 8888 7777 665 6642 5555 444 2 4459 60 336 33 2222 1111 A B C D E F G H J K L M N O P Q R S T A B C D E F G H J K L M N O P Q R S T 棋譜 1 棋譜 1 A B C D E F G H J K L M N O P Q R S T A B C D E F G H J K L M N O P Q R S T 19 19 19 19 18 18 18 18 17 17 17 17 16 16 16 16 15 15 15 15 1479 77 14 14 14 1376 13 13 13 1273 81 82 12 12 12 1171 85 80 75 84 11 11114 115 11 1083 86 10 10116 10 9969 70 74 78 87 99117 118 8867 68 72 88 88119 7765 66 77102 120 6664 6696 101 98 5563 62 5595 108 106 444493 110 97 104 3361 90 3392 91 107 103 2289 2294 105 109 99 100 1111112 111 113 A B C D E F G H J K L M N O P Q R S T A B C D E F G H J K L M N O P Q R S T 棋譜 1 棋譜 1 18/01/2017 C:\home\kuroki\Dropbox\Go\Master60.sgf Printed by Drago A B C D E F G H J K L M N O P Q R S T 19 19 18 18 17 17 16 16 15 15 14138 136 140 14 13139 137 135 130 134 133 144 125 121 13 12124 123 131 132 127 122 12 11129 126 11 10143 146 10 99141 145 88142 77 66 55 44 33128 22 11 A B C D E F G H J K L M N O P Q R S T 棋譜 1 18/01/2017 C:\home\kuroki\Dropbox\Go\Master60.sgf Printed by Drago ⿊ NN ⽩ Magist ⽇時 2016-12-29 19:20:57 結果 ⽩の 中押し勝ち A B C D E F G -
ELF: an Extensive, Lightweight and Flexible Research Platform for Real-Time Strategy Games
ELF: An Extensive, Lightweight and Flexible Research Platform for Real-time Strategy Games Yuandong Tian1 Qucheng Gong1 Wenling Shang2 Yuxin Wu1 C. Lawrence Zitnick1 1Facebook AI Research 2Oculus 1fyuandong, qucheng, yuxinwu, [email protected] [email protected] Abstract In this paper, we propose ELF, an Extensive, Lightweight and Flexible platform for fundamental reinforcement learning research. Using ELF, we implement a highly customizable real-time strategy (RTS) engine with three game environ- ments (Mini-RTS, Capture the Flag and Tower Defense). Mini-RTS, as a minia- ture version of StarCraft, captures key game dynamics and runs at 40K frame- per-second (FPS) per core on a laptop. When coupled with modern reinforcement learning methods, the system can train a full-game bot against built-in AIs end- to-end in one day with 6 CPUs and 1 GPU. In addition, our platform is flexible in terms of environment-agent communication topologies, choices of RL methods, changes in game parameters, and can host existing C/C++-based game environ- ments like ALE [4]. Using ELF, we thoroughly explore training parameters and show that a network with Leaky ReLU [17] and Batch Normalization [11] cou- pled with long-horizon training and progressive curriculum beats the rule-based built-in AI more than 70% of the time in the full game of Mini-RTS. Strong per- formance is also achieved on the other two games. In game replays, we show our agents learn interesting strategies. ELF, along with its RL platform, is open sourced at https://github.com/facebookresearch/ELF. 1 Introduction Game environments are commonly used for research in Reinforcement Learning (RL), i.e. -
Weiqi in Australia
Weiqi in Australia Neville Smythe, Australian Go Association australiango.asn.au Weiqi in Australia • Earliest game played in Australia? • 1960 – Go played at Sydney Chess Club • by 1970, Sydney Go Club formed • 1972 Canberra club • 1977 Brisbane club • 1980 Melbourne club An early Sydney player… John Power (right) with Richard Bozulich and friend (WAGC 2008) Weiqi in Australia • 1978 – Sydney Brisbane and Canberra Clubs meet to form Australian Go Association • 1978 –1st National Championship • 1979 – played in first WAGC • 1982 – foundation member IGF • National, State and other tournaments since then Tournaments National Championship 2007 (Sydney Go Club) Tournaments NEC Cup 2009 Australian Go Congress 2nd Australian Go Congress 2016 [4 colour Go – madness!] World Collegiate Championship Sydney University, July 2019 International participation 1st WAGC 1979: Dae Hahn (Australia) v Shin-Auk Kang (USA) International participation Team Australia WMSG Beijing 2008 (we claim JiaJia as Australian!) 2019 status • Over 500 members, 15 clubs • Strong participation and support by the expatriate Chinese and Korean communities • Not enough female players The future Internet – blessing or curse? • Beginners now progress very quickly and can play anyone anywhere any time • But • We have no idea how many people in Australia only play on-line • Viability of physical clubs is under threat • But it can be an opportunity… Internet – blessing or curse? http://club.artofgo.org/ National Treasures Wu Soong Shen at 2004 Australian National Championship Canberra National Treasures An Younggil, National Coach Visiting Professionals Kobayashi Chizu meeting emus (Canberra, 1981) Visiting Professionals • Chizu was one of the first pro’s to visit • (Before Chizu 3 players from the Nihon Kiin in 1975) • Since then many professionals from Japan, Korea and China have visited • .. -
Recent Articles from the China Journal of System Engineering Prepared
Recent Articles from the China Journal of System Engineering Prepared by the University of Washington Quantum System Engineering (QSE) Group.1 Bibliography [1] Mu A-Hua, Zhou Shao-Lei, and Yu Xiao-Li. Research on fast self-adaptive genetic algorithm and its simulation. Journal of System Simulation, 16(1):122 – 5, 2004. [2] Guan Ai-Jie, Yu Da-Tai, Wang Yun-Ji, An Yue-Sheng, and Lan Rong-Qin. Simulation of recon-sat reconing process and evaluation of reconing effect. Journal of System Simulation, 16(10):2261 – 3, 2004. [3] Hao Ai-Min, Pang Guo-Feng, and Ji Yu-Chun. Study and implementation for fidelity of air roaming system above the virtual mount qomolangma. Journal of System Simulation, 12(4):356 – 9, 2000. [4] Sui Ai-Na, Wu Wei, and Zhao Qin-Ping. The analysis of the theory and technology on virtual assembly and virtual prototype. Journal of System Simulation, 12(4):386 – 8, 2000. [5] Xu An, Fan Xiu-Min, Hong Xin, Cheng Jian, and Huang Wei-Dong. Research and development on interactive simulation system for astronauts walking in the outer space. Journal of System Simulation, 16(9):1953 – 6, Sept. 2004. [6] Zhang An and Zhang Yao-Zhong. Study on effectiveness top analysis of group air-to-ground aviation weapon system. Journal of System Simulation, 14(9):1225 – 8, Sept. 2002. [7] Zhang An, He Sheng-Qiang, and Lv Ming-Qiang. Modeling simulation of group air-to-ground attack-defense confrontation system. Journal of System Simulation, 16(6):1245 – 8, 2004. [8] Wu An-Bo, Wang Jian-Hua, Geng Ying-San, and Wang Xiao-Feng. -
Improved Policy Networks for Computer Go
Improved Policy Networks for Computer Go Tristan Cazenave Universite´ Paris-Dauphine, PSL Research University, CNRS, LAMSADE, PARIS, FRANCE Abstract. Golois uses residual policy networks to play Go. Two improvements to these residual policy networks are proposed and tested. The first one is to use three output planes. The second one is to add Spatial Batch Normalization. 1 Introduction Deep Learning for the game of Go with convolutional neural networks has been ad- dressed by [2]. It has been further improved using larger networks [7, 10]. AlphaGo [9] combines Monte Carlo Tree Search with a policy and a value network. Residual Networks improve the training of very deep networks [4]. These networks can gain accuracy from considerably increased depth. On the ImageNet dataset a 152 layers networks achieves 3.57% error. It won the 1st place on the ILSVRC 2015 classifi- cation task. The principle of residual nets is to add the input of the layer to the output of each layer. With this simple modification training is faster and enables deeper networks. Residual networks were recently successfully adapted to computer Go [1]. As a follow up to this paper, we propose improvements to residual networks for computer Go. The second section details different proposed improvements to policy networks for computer Go, the third section gives experimental results, and the last section con- cludes. 2 Proposed Improvements We propose two improvements for policy networks. The first improvement is to use multiple output planes as in DarkForest. The second improvement is to use Spatial Batch Normalization. 2.1 Multiple Output Planes In DarkForest [10] training with multiple output planes containing the next three moves to play has been shown to improve the level of play of a usual policy network with 13 layers. -
Heft 1/2014 89. Jahrgang
Heft 1/2014 89. Jahrgang 1 DGoZ 1/2014 Vorwort Fangen und Retten 15 von Yilun Yang In dieser Ausgabe seht die Berichterstattung zur Kisei-Partie im spanischen Alcalá de Henares klar im Zentrum – mit einem spannenden, bilderreichen Reisebericht von Yvonne Limbach und Axel Köhler sowie einem ausführlichen Partiekommentar (fast 30 Erläuterungsdiagramme!) von Yoon Young Sun 8p. Lesenswert auch für Erwachsene sind die beiden Interviews auf den Kinderseiten, die einen Einblick in das Leben künftiger Go-Profis in China erlauben. Schwarz am Zug. Es sieht so aus, als könne man die Absolut spannend ist Victor Lin 6d wieder seine drei schwarzen Steine nicht mehr retten. Simmt das? Serie „Der etwas andere Zug“ geraten – Highlevel- Lösung auf S. 41. Go dieses Mal mit Blick auf die Sicherheit. Über den Beginn des Jubango zwischen Gu Li und Lee Sedol wird bisher nur in den Fernost- Liebe Go-Spieler in Deutschland, Nachrichten berichtet. Dafür folgt voraussichtlich ein Abdruck der ersten Partie inklusive eines aus- ich hoffe sehr, für euch hat das neue Jahr gut führlichen Kommentars im nächsten Heft. angefangen. Für eure persönlichen Projekte und Tobias Berben euren Go-lichen Fortschritt wünsche ich viel Erfolg. Den wünsche ich auch uns Go-Spielern für unsere Projekte. Denn da ist im letzten Jahr Inhalt einiges in Europa und Deutschland auf die Beine gestellt worden. Wie immer gilt jedoch auch hier: Go-Foto .......................................................1 Es kommt auf euch an, damit der DGoB als Verein Vorwort, Inhalt, Retten und Fangen ..............2 und damit auch das deutsche Go voran kommen. Nachrichten ..........................................2–11 Und 2013 war ein Jahr mit vielen Aktivitäten, aber Ausschreibung: DM-Vorrunde ..................10 auch ein Jahr des personellen Umbruchs. -
Heft 1/2019 94. Jahrgang
Heft 1/2019 94. Jahrgang 1 DGoZ 1/2019 Inhalt Vorwort Dieses Winter-Titelbild ist wohl eine würdige Go-Winter (Fundstück) .........................Titel Fortsetzung zum Titelbild der letzten Ausgabe. Was Vorwort, Inhalt, Nachrichten ......................2–4 mich besonders freut: Neu in dieser Ausgabe ist die „Steinzeit” von Pierre-Alain Chamot (siehe rechts) Turnierberichte ...................................... 5–7 als ebenso würdige Nachfolge für die „Stones” von Die Go-vember-Herausforderung ...........8–9 Andreas Fecke. Ansonsten bleibt mir nur, allen Go in Chemnitz ..................................10–12 Lesern viel Go-Spaß im chinesischen Jahr des Auschreibungen: DPGM, DBGM ............ 13 Schweins zu wünschen! Interview mit Philipp Neubert ............14–15 Tobias Berben Kinderseite(n) ....................................16–17 Yoon Young Sun kommentiert (44) ..........18–23 Jugend-Nationalmannschaft Hwang In-Seongs Winter-Go-Camp ........ 24 Franz-Josef Dickhut (6d, Lippstadt) hat nach zwei Mein Weg zum „Dan-Spieler” .............25–29 Jahren die Leitung der deutschen Jugendmannschaft Durchbruch zum 12. Kyu (8) .............30–32 abgegeben. Neu wird sie gemeinsam von Cheng Deutschlandpokal ..................................... 33 Ying (4d, Düsseldorf) und Marc Oliver Rieger (6k, Trier) geleitet werden. Fernostnachrichten ............................. 34–37 Anfängerprobleme ..............................38–39 Oliver Nebelung gestorben Der etwas andere Zug (28) ..................40–46 Impressum ............................................... -
Sydney Go Journal Issue Date – February 2007
Author – David Mitchell on behalf of The Sydney Go Club Sydney Go Journal Issue Date – February 2007 Dr. Geoffrey Gray’s antique Go Ban (picture courtesy of Dr Gray) Up coming events Queensland Go Championship Saturday 17th and Sunday 18th February in Brisbane. Venue: Brisbane Bridge Centre Registration and other details on page 33 For the latest details visit www.uq.net.au/~zzjhardy/brisgo.html Contributions, comments and suggestions for the SGJ to: [email protected] Special thanks to Devon Bailey and Geoffrey Gray for proof reading this edition and correcting my mistakes. © Copyright 2007 – David Mitchell Page 1 February 2007 Author – David Mitchell on behalf of The Sydney Go Club Sydney Lightning Tournament report 3 Changqi Cup 4 3rd Changqi Cup – 1st Qualifier 6 3rd Changqi Cup – 2nd Qualifier 10 Problems 14 Handicap Strategy 15 Four Corners 29 Two page Joseki lesson 35 Answers 37 Korean Go Terms 39 The Sydney Go Club Meets Friday nights at :- At Philas House 17 Brisbane St Surry Hills From 5.00pm Entrance fee - $5 per head; Concession $3; Children free - includes tea and coffee. For further information from Robert [email protected] © Copyright 2007 – David Mitchell Page 2 February 2007 Lightning Tournament The lightning tournament was held on the January 12th and a good time was had by all, thanks to Robert Vadas organising skills. The final was between Max Latey and David Mitchell, the latter managing another lucky win. The following pictures tell the story David Mitchell (foreground); Max more eloquently than words. Latey (background); the two finalists Robert giving some sage advice. -
Chinese Health App Arrives Access to a Large Population Used to Sharing Data Could Give Icarbonx an Edge Over Rivals
NEWS IN FOCUS ASTROPHYSICS Legendary CHEMISTRY Deceptive spice POLITICS Scientists spy ECOLOGY New Zealand Arecibo telescope faces molecule offers cautionary chance to green UK plans to kill off all uncertain future p.143 tale p.144 after Brexit p.145 invasive predators p.148 ICARBONX Jun Wang, founder of digital biotechnology firm iCarbonX, showcases the Meum app that will use reams of health data to provide customized medical advice. BIOTECHNOLOGY Chinese health app arrives Access to a large population used to sharing data could give iCarbonX an edge over rivals. BY DAVID CYRANOSKI, SHENZHEN medical advice directly to consumers through another $400 million had been invested in the an app. alliance members, but he declined to name the ne of China’s most intriguing biotech- The announcement was a long-anticipated source. Wang also demonstrated the smart- nology companies has fleshed out an debut for iCarbonX, which Wang founded phone app, called Meum after the Latin for earlier quixotic promise to use artificial in October 2015 shortly after he left his lead- ‘my’, that customers would use to enter data Ointelligence (AI) to revolutionize health care. ership position at China’s genomics pow- and receive advice. The Shenzhen firm iCarbonX has formed erhouse, BGI, also in Shenzhen. The firm As well as Google, IBM and various smaller an ambitious alliance with seven technology has now raised more than US$600 million in companies, such as Arivale of Seattle, Wash- companies from around the world that special- investment — this contrasts with the tens of ington, are working on similar technology. But ize in gathering different types of health-care millions that most of its rivals are thought Wang says that the iCarbonX alliance will be data, said the company’s founder, Jun Wang, to have invested (although several big play- able to collect data more cheaply and quickly. -
Fml-Based Dynamic Assessment Agent for Human-Machine Cooperative System on Game of Go
Accepted for publication in International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems in July, 2017 FML-BASED DYNAMIC ASSESSMENT AGENT FOR HUMAN-MACHINE COOPERATIVE SYSTEM ON GAME OF GO CHANG-SHING LEE* MEI-HUI WANG, SHENG-CHI YANG Department of Computer Science and Information Engineering, National University of Tainan, Tainan, Taiwan *[email protected], [email protected], [email protected] PI-HSIA HUNG, SU-WEI LIN Department of Education, National University of Tainan, Tainan, Taiwan [email protected], [email protected] NAN SHUO, NAOYUKI KUBOTA Dept. of System Design, Tokyo Metropolitan University, Japan [email protected], [email protected] CHUN-HSUN CHOU, PING-CHIANG CHOU Haifong Weiqi Academy, Taiwan [email protected], [email protected] CHIA-HSIU KAO Department of Computer Science and Information Engineering, National University of Tainan, Tainan, Taiwan [email protected] Received (received date) Revised (revised date) Accepted (accepted date) In this paper, we demonstrate the application of Fuzzy Markup Language (FML) to construct an FML- based Dynamic Assessment Agent (FDAA), and we present an FML-based Human–Machine Cooperative System (FHMCS) for the game of Go. The proposed FDAA comprises an intelligent decision-making and learning mechanism, an intelligent game bot, a proximal development agent, and an intelligent agent. The intelligent game bot is based on the open-source code of Facebook’s Darkforest, and it features a representational state transfer application programming interface mechanism. The proximal development agent contains a dynamic assessment mechanism, a GoSocket mechanism, and an FML engine with a fuzzy knowledge base and rule base. -
Complete Journal (PDF)
B RITISH J OURNAL Number 140 Spring 2007 Price £3.50 British Go Journal 140 Spring 2007 CONTENTS EDITORIAL 2 UK NEWS Tony Atkins 3 OUTGOING PRESIDENT’S REPORT, 2005-6 Simon Goss 9 10 YEARS AGO Tony Atkins 11 THREE PEAKS Ian Marsh 13 REFEREEING QUESTION Nick Wedd 14 BRITISH CHAMPIONSHIP 2006 17 COUNTING BY AREA AND TERRITORY Andrei Sokolov 18 WORLD NEWS Tony Atkins 21 REVIEW:BADUK MADE FUN AND EASY Ian Davis 25 AMATEUR PAIR GO CHAMPIONSHIPS Natasha Regan 26 VISIT TO KOREA -OCTOBER 2006 T Mark Hall 29 KOREA Tony Atkins 33 HOW TO GIVE A GAME REVIEW Nick Wedd 35 WEAK GROUP THEORY Ron Bell 37 UK CLUBS LIST 43 GLOSSARY OF GO TERMS 44 Copyright c 2007 British Go Association. Articles may be reproduced for the purposes of promoting Go and ’not for profit’ providing the British Go Journal is attributed as the source and the permission of the Editor and of the articles’ author(s) have been sought and obtained. Views expressed are not necessarily those of the BGA or of the Editor. USEFUL WEB AND EMAIL ADDRESSES Journal comments and contributions: [email protected] Email for general BGA enquiries: [email protected] BGA website: http://www.britgo.org/ BGA email list: [email protected] used for general discussion and announcements - control from: http://three.pairlist.net/mailman/listinfo/gotalk For discussion of how the BGA operates: [email protected] http://two.pairlist.net/mailman/listinfo/bga-policy 1 EDITORIAL [email protected] Welcome to my first attempt at producing your Journal.