Jump to content

Future of Go Summit

From Wikipedia, the free encyclopedia

The Future of Go Summit (Chinese: 中国乌镇围棋峰会) was held in May 2017 by the Chinese Go Association, Sport Bureau of Zhejiang Province and Google in Wuzhen, Zhejiang, the permanent host of the World Internet Conference. It featured five Go games involving AlphaGo and top Chinese Go players,[1] as well as a forum on the future of AI.[2][3] It was Google’s biggest public event in partnership with the Chinese government since Google China's search engine was moved out of mainland China to Hong Kong due to the government censorship in 2010. It was seen as a charm offensive launched by Google toward Chinese officials, being part of effort to reopen China's market.[4][5]

The version of AlphaGo used in this Summit was AlphaGo Master,[6][7] using four TPUs on a single machine with an Elo rating of 4,858.[8] DeepMind claimed that this version was 3 stones stronger in games of self-play than the version used in AlphaGo v. Lee Sedol.[9]

After winning its three-game match against Chinese grandmaster Ke Jie, the world's top Go player, AlphaGo was awarded professional 9-dan by Chinese Weiqi Association.[10] DeepMind announced that AlphaGo would retire, and DeepMind would disband the team that worked on Go and spend their time exploring new AI in other areas instead of Go. After the Summit, DeepMind released 50 games AlphaGo played against itself.[11]

Schedule

[edit]

Games

[edit]

Ke Jie vs AlphaGo

[edit]
Match Date Black White Time control Result Moves
1 23 May 2017 Ke Jie AlphaGo 3 hours + five 60-second byo-yomi W+0.5 289
2 25 May 2017 AlphaGo Ke Jie 3 hours + five 60-second byo-yomi B+Res 155
3 27 May 2017 AlphaGo Ke Jie 3 hours + five 60-second byo-yomi B+Res 209
Result: AlphaGo 3 – 0 Ke Jie

Google DeepMind offered 1.5 million dollars winner prizes for this match while the losing side took 300,000 dollars for participating in the three games.[15][16] AlphaGo won all three games.[17][18]

Pair Go

[edit]
Match Date Black White Time control Result Moves
Pair Go 26 May 2017 AlphaGo +
Gu Li
AlphaGo+
Lian Xiao
1 hour +
one 60-second byo-yomi
W+Res 220

The pair Go game was played on 26 May morning. Gu Li, an eight-time world champion, and Lian Xiao, the current Chinese Mingren and Tianyuan titles holder, each had their own AlphaGo teammate, alternating moves in tag team style.[19] The move sequence was Gu Li (black), Lian Xiao (white), AlphaGo (black), and AlphaGo (white).[20]

First 99 moves
Moves 100–199
Moves 200–220

After move 198, the black AlphaGo proposed resignation, but got rejected by its teammate Gu Li.[21] The white AlphaGo and Lian Xiao finally won by resignation after move 220.

This pair Go game provided a glimpse of how human players might be able to use AI tools in the future, benefiting from the computer's insights while also relying on their own intuition.[19] After the match, David Silver, the lead researcher for AlphaGo said: "Pair Go was a beautiful game. All four players made some brilliant and creative moves. It felt like four painters working together on a shared canvas, all with different styles, all combining together to make something truly beautiful."[22]

Team Go

[edit]
Match Date Black White Time control Result Moves
Team Go 26 May 2017 Chen Yaoye/Zhou Ruiyang/Mi Yuting/
Shi Yue/Tang Weixing
AlphaGo 2.5 hours + three 60-second byo-yomi W+Res 254

The team Go match was held on 26 May afternoon. In this five-on-one match, five world champion players, Chen Yaoye, Zhou Ruiyang, Mi Yuting, Shi Yue, and Tang Weixing (black), joined forces to play against AlphaGo (white).[23] There was a two-and-a-half-hour set time limit for each side followed by three 60-second byo-yomi overtime periods.[citation needed] Consulting each other on every move, the human team approached the challenge in a light-hearted manner, obviously enjoying the experience of playing together.[19] Chosen as the team leader, Zhou Ruiyang positioned the stones on the board during regular time.[citation needed] Tang Weixing put down the stones in overtime, almost working on his own,[citation needed] and finally resigned in the face of AlphaGo’s certain victory.[24] Eric Schmidt, the Executive Chairman of DeepMind's parent company Alphabet, Inc. tweeted: "This speaks volumes about where AI is headed - human players are teaming up with AlphaGo to have even more fun with the game!".[25]

First 99 moves (87/93/99 at 65, 90/96 at 66)
Moves 100–199 (105/111/119/125/131 at ,
108/114/122/128 at 102, 172 at 115)
Moves 200–254 (219 at 213)

White 54 is a surprising move to human players.[26][27] After move 58, Ke Jie pointed out that White could play 60, but was teased by other professional players in the research room until AlphaGo actually played this move.[28] Tan Xiao commented that it's difficult for Black to win after move 64.[26]

Chinese coverage

[edit]

The game was widely reported by Chinese media and became one of the hottest topics on social networks.[29][30][31] However, due to the instruction of the Chinese authorities, the state owned China Central Television cancelled the plan to live telecast the game, and live streaming on a number of Chinese websites was cut off.[32][33] The only official live broadcast was on YouTube, which is blocked in China. The restrictions caused widespread complaints among Chinese netizens on the social platform Sina Weibo.[34] Chinese media was instructed to play down the mention of Google or Google's products in their reports, as its services are blocked in China.[35][36]

References

[edit]
  1. ^ "Exploring the mysteries of Go with AlphaGo and China's top players". DeepMind. 10 April 2017. Archived from the original on 11 April 2017. Retrieved 10 April 2017.
  2. ^ "DeepMind's AlphaGo is back..and this time it's taking on five humans at once". Wired.com. 10 April 2017. Retrieved 10 April 2017.
  3. ^ "AlphaGo China | DeepMind". DeepMind. Archived from the original on 2019-06-06. Retrieved 2017-05-25.
  4. ^ "Google deploys AI for Go tournament in China charm offensive AlphaGo". Financial Times. 23 May 2017. Retrieved 2017-05-23.
  5. ^ "Google tries again in China using AlphaGo event that targets the young". Market Watch. Retrieved 2017-05-27.
  6. ^ "AlphaGo官方解读让三子 对人类高手没这种优势" (in Chinese). Sina.com. 25 May 2017. Retrieved 1 June 2017.
  7. ^ "各版alphago实力对比 master能让李世石版3子" (in Chinese). Sina.com. 24 May 2017. Retrieved 2 June 2017.
  8. ^ Silver, David; Schrittwieser, Julian; Simonyan, Karen; Antonoglou, Ioannis; Huang, Aja; Guez, Arthur; Hubert, Thomas; Baker, Lucas; Lai, Matthew; Bolton, Adrian; Chen, Yutian; Lillicrap, Timothy; Fan, Hui; Sifre, Laurent; Driessche, George van den; Graepel, Thore; Hassabis, Demis (19 October 2017). "Mastering the game of Go without human knowledge". Nature. 550 (7676): 354–359. Bibcode:2017Natur.550..354S. doi:10.1038/nature24270. ISSN 0028-0836. PMID 29052630. S2CID 205261034. Retrieved 10 December 2017.Closed access icon
  9. ^ "AlphaGo棋力比去年涨三子 柯洁想赢它可能只剩一招" (in Chinese). Sina.com. 25 May 2017. Retrieved 1 June 2017.
  10. ^ "中国围棋协会授予AlphaGo职业九段 并颁发证书" (in Chinese). 27 May 2017. Retrieved 28 May 2017.
  11. ^ Metz, Cade (2017-05-27). "After Win in China, AlphaGo's Designers Explore New AI". Wired.
  12. ^ a b c "The Future of Go Summit". Archived from the original on 2017-06-15. Retrieved 2017-05-25.
  13. ^ "人机大战第二季5月乌镇见! 慢棋+团队配对赛" (in Chinese). Sina.com. 10 April 2017. Retrieved 20 May 2017.
  14. ^ "柯洁AlphaGo大战发布会举行 胜者获150万美金" (in Chinese). Sina.com. 10 April 2017. Retrieved 20 May 2017.
  15. ^ "World No.1 Go player Ke Jie takes on upgraded AlphaGo in May". 2017-04-10.
  16. ^ "Ke Jie vs. AlphaGo: 8 things you must know". 2017-05-27. Archived from the original on 2017-12-14. Retrieved 2017-05-29.
  17. ^ Metz, Cade (2017-05-23). "Revamped AlphaGo Wins First Game Against Chinese Go Grandmaster". Wired.
  18. ^ Metz, Cade (2017-05-25). "Google's AlphaGo Continues Dominance With Second Win in China". Wired.
  19. ^ a b c "AlphaGo Pair and Team Go wrap up". American Go Association. 26 May 2017. Archived from the original on 5 November 2019. Retrieved 29 May 2017.
  20. ^ "配对赛解析:人类出彩 连笑成阿法狗"大腿"?" (in Chinese). QQ.com. 26 May 2017. Retrieved 29 May 2017.
  21. ^ "连笑队赢人机配对赛 AlphaGo欲认输遭古力拒绝" (in Chinese). QQ.com. 26 May 2017. Retrieved 9 June 2017.
  22. ^ "AlphaGo Triumphs In China". I Programmer. 27 May 2017. Retrieved 30 May 2017.
  23. ^ "Google AlphaGo AI took on 5 champion Go players at once and still won". TechRepublic. 26 May 2017. Retrieved 29 May 2017.
  24. ^ "Google's AlphaGo AI defeats team of five leading Go players". TechCrunch. 26 May 2017. Retrieved 30 May 2017.
  25. ^ "Google's AI-powered AlphaGo defeats five leading Go players". The Economic Times. 27 May 2017. Archived from the original on 2 June 2017. Retrieved 4 June 2017.
  26. ^ a b "人类再败!中国五顶尖棋手联手 仍不敌AlphaGo" (in Chinese). QQ.com. 26 May 2017. Retrieved 9 June 2017.
  27. ^ "围棋峰会团队赛无悬念 AlphaGo执白中盘胜中国队" (in Chinese). 弈城围棋网. 26 May 2017. Retrieved 9 June 2017.
  28. ^ "详解5名顶尖棋手输在哪?开局不利意见存分歧" (in Chinese). QQ.com. 26 May 2017. Retrieved 9 June 2017.
  29. ^ "Pessimism looms over Chinese Go master's upcoming match against AlphaGo". www.ecns.cn.
  30. ^ "AlphaGo 1, human 0 in first of 3 games". www.ecns.cn.
  31. ^ F_200804. "Chinese Go master Ke Jie loses to Google's AlphaGo in first of three-game match - People's Daily Online". en.people.cn.{{cite web}}: CS1 maint: numeric names: authors list (link)
  32. ^ 中国限制直播柯洁与 AlphaGo 的人机大战 Archived 2017-05-22 at the Wayback Machine.Solidot.
  33. ^ 柯洁和AlphaGo对弈 Archived 2017-08-20 at the Wayback Machine.
  34. ^ "The world's top Go player was beaten by Google's AI — and Chinese authorities banned the livestream". Mashable.
  35. ^ "AlphaGo wins first match of three over Chinese Go master with ability to surprise". South China Morning Post.
  36. ^ "An epic man-against-machine tournament is taking place in China but Chinese viewers can’t watch it live". Quartz.
[edit]