25,161
edits
Changes
Created page with "'''Home * People * David J. Wu''' FILE:DavidJWu.jpg|border|right|thumb|link=http://www.kingpinchess.net/2015/07/arimaa-game-over/| David J. Wu <ref>Image..."
'''[[Main Page|Home]] * [[People]] * David J. Wu'''
[[FILE:DavidJWu.jpg|border|right|thumb|link=http://www.kingpinchess.net/2015/07/arimaa-game-over/| David J. Wu <ref>Image from [http://www.kingpinchess.net/2015/07/arimaa-game-over/ Arimaa: Game Over?] by [[Andy Lewis]], [http://www.kingpinchess.net/ Kingpin Chess Magazine], July 11, 2015</ref> ]]
'''David Jian Wu''',<br/>
an American computer scientist and computer games researcher and programmer, who balances a mixture of software development and research within the financial industry <ref>[http://www.kingpinchess.net/2015/07/arimaa-game-over/ Arimaa: Game Over?] by [[Andy Lewis]], [http://www.kingpinchess.net/ Kingpin Chess Magazine], July 11, 2015</ref>.
He defended his B.Sc. degree in 2011 at [https://en.wikipedia.org/wiki/Harvard_College Harvard College], [[Harvard University]], delivering the thesis ''Move Ranking and Evaluation in the Game of Arimaa'' <ref>[[David J. Wu]] ('''2011'''). ''Move Ranking and Evaluation in the Game of Arimaa''. B.Sc. thesis, [https://en.wikipedia.org/wiki/Harvard_College Harvard College], [https://en.wikipedia.org/wiki/Cambridge,_Massachusetts Cambridge, Massachusetts], [http://arimaa.com/arimaa/papers/DavidWu/djwuthesis.pdf pdf]</ref>.
David J. Wu is author of the [[Arimaa]] bot '''Sharp''', and inspired by [[AlphaZero]], the [[Go]] playing program '''KataGo''' .
=Sharp=
The [[Arimaa]] playing bot '''Sharp''' won the 2015 ''Arimaa Challenge'' and the then $12,000 USD prize by defeating each of three top-ranked human players in a three game series <ref>[http://arimaa.com/arimaa/challenge/2015/ The 2015 Arimaa Challenge]</ref>.
It already played the 2008 Arimaa computer tournament, and became runner-up behind [[David Fotland|David Fotland’s]] program ''Bomb'', and further won the 2011 and 2014 tournaments but not the contest against the best human players of that time <ref>[[Omar Syed]] ('''2015'''). ''The Arimaa Challenge: From Inception to Completion''. [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]]</ref>.
Sharp's design was elaborated by its author in the 2015 [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]] <ref>[[David J. Wu]] ('''2015'''). ''Designing a Winning Arimaa Program''. [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]]</ref>.
It follows the same fundamental design as strong Chess programs, using an [[Iterative Deepening|iterative deepening]] [[Depth|depth]] limited [[Alpha-Beta|alpha-beta search]] and various enhancements within a [[Parallel Search|parallel search]] algorithm conceptually similar to the [[Dynamic Tree Splitting|dynamic tree splitting]] described by [[Robert Hyatt]] in 1994 <ref>[[Robert Hyatt]] ('''1994'''). ''[http://www.cis.uab.edu/hyatt/search.html The DTS high-performance parallel tree search algorithm]''</ref>.
Sharp further implements several Arimaa-specific search enhancements with four steps per move, such as static goal detection and [[Captures|capture]] [[Move Generation|generation]], and continues to use and benefit greatly from a [[Move Ordering|move ordering]] function developed in 2011 as described in Wu's thesis - the move ordering function is the result of training a slightly generalized [https://en.wikipedia.org/wiki/Bradley%E2%80%93Terry_model Bradley-Terry model] over thousands of expert Arimaa games to [[Learning|learn]] to predict expert player's moves, using the same optimization procedure described by [[Rémi Coulom]] for computer [[Go]] <ref>[[Rémi Coulom]] ('''2007'''). ''[http://remi.coulom.free.fr/Amsterdam2007/ Computing Elo Ratings of Move Patterns in the Game of Go]''. [[ICGA Journal#30_4|ICGA Journal, Vol. 30, No. 4]], [[CGW 2007]], [https://www.remi-coulom.fr/Amsterdam2007/icgaj.pdf pdf]</ref>.
=KataGo=
KataGo is a [[Go]] playing entity inspired by the [AlphaZero]] approach of combining [[Deep Learning|Deep learning]] with [[Monte-Carlo Tree Search]] (MCTS) using pure [[Reinforcement Learning|reinforcement learning]] aka self play to train the [[Neural Networks|deep neural network]].
Due to modifications and enhancements of the AlphaZero-like training process, self-play with a only few strong [[GPU|GPUs]] of between one and several days is sufficient to reach somewhere in the range of strong-kyu up to mid-dan strength on the full 19x19 board <ref>[https://github.com/lightvector/KataGo/blob/master/README.md KataGo/README.md at master · lightvector/KataGo · GitHub]</ref>.
=Selected Publications=
<ref>[https://dblp.uni-trier.de/pers/hd/w/Wu:David_J= dblp: David J. Wu], including another [https://www.cs.virginia.edu/dwu4/ David J. Wu] ([https://en.wikipedia.org/wiki/Cryptography Cryptography], [https://en.wikipedia.org/wiki/Computer_security Computer security]) included</ref>
* [[David J. Wu]] ('''2011'''). ''Move Ranking and Evaluation in the Game of Arimaa''. B.Sc. thesis, [https://en.wikipedia.org/wiki/Harvard_College Harvard College], [https://en.wikipedia.org/wiki/Cambridge,_Massachusetts Cambridge, Massachusetts], [http://arimaa.com/arimaa/papers/DavidWu/djwuthesis.pdf pdf]
* [[David J. Wu]] ('''2015'''). ''Designing a Winning Arimaa Program''. [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]], [https://icosahedral.net/downloads/djwu2015arimaa.pdf pdf] <ref>[https://www.game-ai-forum.org/viewtopic.php?f=2&t=83 Paper describing "Sharp" the program that won the Arimaa Challenge] by ddyer, [[Computer Chess Forums|Game-AI Forum]], January 14, 2016</ref>
* [[David J. Wu]] ('''2019'''). ''Accelerating Self-Play Learning in Go''. [https://arxiv.org/abs/1902.10565 arXiv:1902.10565]
=External Links=
* [https://github.com/lightvector lightvector · GitHub]
* [http://www.kingpinchess.net/2015/07/arimaa-game-over/ Arimaa: Game Over?] by [[Andy Lewis]], [http://www.kingpinchess.net/ Kingpin Chess Magazine], July 11, 2015
* [https://play.google.com/store/apps/details?id=net.icosahedral.arimaa Arimaa - David J. Wu - Google Play]
=References=
<references />
'''[[People|Up one level]]'''
[[Category:Arimaa Programmer|Wu]]
[[Category:Go Programmer|Wu]]
[[FILE:DavidJWu.jpg|border|right|thumb|link=http://www.kingpinchess.net/2015/07/arimaa-game-over/| David J. Wu <ref>Image from [http://www.kingpinchess.net/2015/07/arimaa-game-over/ Arimaa: Game Over?] by [[Andy Lewis]], [http://www.kingpinchess.net/ Kingpin Chess Magazine], July 11, 2015</ref> ]]
'''David Jian Wu''',<br/>
an American computer scientist and computer games researcher and programmer, who balances a mixture of software development and research within the financial industry <ref>[http://www.kingpinchess.net/2015/07/arimaa-game-over/ Arimaa: Game Over?] by [[Andy Lewis]], [http://www.kingpinchess.net/ Kingpin Chess Magazine], July 11, 2015</ref>.
He defended his B.Sc. degree in 2011 at [https://en.wikipedia.org/wiki/Harvard_College Harvard College], [[Harvard University]], delivering the thesis ''Move Ranking and Evaluation in the Game of Arimaa'' <ref>[[David J. Wu]] ('''2011'''). ''Move Ranking and Evaluation in the Game of Arimaa''. B.Sc. thesis, [https://en.wikipedia.org/wiki/Harvard_College Harvard College], [https://en.wikipedia.org/wiki/Cambridge,_Massachusetts Cambridge, Massachusetts], [http://arimaa.com/arimaa/papers/DavidWu/djwuthesis.pdf pdf]</ref>.
David J. Wu is author of the [[Arimaa]] bot '''Sharp''', and inspired by [[AlphaZero]], the [[Go]] playing program '''KataGo''' .
=Sharp=
The [[Arimaa]] playing bot '''Sharp''' won the 2015 ''Arimaa Challenge'' and the then $12,000 USD prize by defeating each of three top-ranked human players in a three game series <ref>[http://arimaa.com/arimaa/challenge/2015/ The 2015 Arimaa Challenge]</ref>.
It already played the 2008 Arimaa computer tournament, and became runner-up behind [[David Fotland|David Fotland’s]] program ''Bomb'', and further won the 2011 and 2014 tournaments but not the contest against the best human players of that time <ref>[[Omar Syed]] ('''2015'''). ''The Arimaa Challenge: From Inception to Completion''. [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]]</ref>.
Sharp's design was elaborated by its author in the 2015 [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]] <ref>[[David J. Wu]] ('''2015'''). ''Designing a Winning Arimaa Program''. [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]]</ref>.
It follows the same fundamental design as strong Chess programs, using an [[Iterative Deepening|iterative deepening]] [[Depth|depth]] limited [[Alpha-Beta|alpha-beta search]] and various enhancements within a [[Parallel Search|parallel search]] algorithm conceptually similar to the [[Dynamic Tree Splitting|dynamic tree splitting]] described by [[Robert Hyatt]] in 1994 <ref>[[Robert Hyatt]] ('''1994'''). ''[http://www.cis.uab.edu/hyatt/search.html The DTS high-performance parallel tree search algorithm]''</ref>.
Sharp further implements several Arimaa-specific search enhancements with four steps per move, such as static goal detection and [[Captures|capture]] [[Move Generation|generation]], and continues to use and benefit greatly from a [[Move Ordering|move ordering]] function developed in 2011 as described in Wu's thesis - the move ordering function is the result of training a slightly generalized [https://en.wikipedia.org/wiki/Bradley%E2%80%93Terry_model Bradley-Terry model] over thousands of expert Arimaa games to [[Learning|learn]] to predict expert player's moves, using the same optimization procedure described by [[Rémi Coulom]] for computer [[Go]] <ref>[[Rémi Coulom]] ('''2007'''). ''[http://remi.coulom.free.fr/Amsterdam2007/ Computing Elo Ratings of Move Patterns in the Game of Go]''. [[ICGA Journal#30_4|ICGA Journal, Vol. 30, No. 4]], [[CGW 2007]], [https://www.remi-coulom.fr/Amsterdam2007/icgaj.pdf pdf]</ref>.
=KataGo=
KataGo is a [[Go]] playing entity inspired by the [AlphaZero]] approach of combining [[Deep Learning|Deep learning]] with [[Monte-Carlo Tree Search]] (MCTS) using pure [[Reinforcement Learning|reinforcement learning]] aka self play to train the [[Neural Networks|deep neural network]].
Due to modifications and enhancements of the AlphaZero-like training process, self-play with a only few strong [[GPU|GPUs]] of between one and several days is sufficient to reach somewhere in the range of strong-kyu up to mid-dan strength on the full 19x19 board <ref>[https://github.com/lightvector/KataGo/blob/master/README.md KataGo/README.md at master · lightvector/KataGo · GitHub]</ref>.
=Selected Publications=
<ref>[https://dblp.uni-trier.de/pers/hd/w/Wu:David_J= dblp: David J. Wu], including another [https://www.cs.virginia.edu/dwu4/ David J. Wu] ([https://en.wikipedia.org/wiki/Cryptography Cryptography], [https://en.wikipedia.org/wiki/Computer_security Computer security]) included</ref>
* [[David J. Wu]] ('''2011'''). ''Move Ranking and Evaluation in the Game of Arimaa''. B.Sc. thesis, [https://en.wikipedia.org/wiki/Harvard_College Harvard College], [https://en.wikipedia.org/wiki/Cambridge,_Massachusetts Cambridge, Massachusetts], [http://arimaa.com/arimaa/papers/DavidWu/djwuthesis.pdf pdf]
* [[David J. Wu]] ('''2015'''). ''Designing a Winning Arimaa Program''. [[ICGA Journal#38_1|ICGA Journal, Vol. 38, No. 1]], [https://icosahedral.net/downloads/djwu2015arimaa.pdf pdf] <ref>[https://www.game-ai-forum.org/viewtopic.php?f=2&t=83 Paper describing "Sharp" the program that won the Arimaa Challenge] by ddyer, [[Computer Chess Forums|Game-AI Forum]], January 14, 2016</ref>
* [[David J. Wu]] ('''2019'''). ''Accelerating Self-Play Learning in Go''. [https://arxiv.org/abs/1902.10565 arXiv:1902.10565]
=External Links=
* [https://github.com/lightvector lightvector · GitHub]
* [http://www.kingpinchess.net/2015/07/arimaa-game-over/ Arimaa: Game Over?] by [[Andy Lewis]], [http://www.kingpinchess.net/ Kingpin Chess Magazine], July 11, 2015
* [https://play.google.com/store/apps/details?id=net.icosahedral.arimaa Arimaa - David J. Wu - Google Play]
=References=
<references />
'''[[People|Up one level]]'''
[[Category:Arimaa Programmer|Wu]]
[[Category:Go Programmer|Wu]]
