by John Power, Japan Correspondent for the E-Journal
Motoki does well in Kisei knock-out: The first game in the irregular knock-out among the various league winners in the 42nd Kisei tournament was played on October 2. Motoki Katsuya 8P, winner of the C League, beat Yo Seiki (Yu Chengqi) 7P, winner of the play-off between the winners of the two B Leagues. Motoki followed up this success by beating the winner of the A League, Takao Shinji Meijin, on October 12. Taking black, Motoki won by resignation. Next, he will meet the second-place-getter in the S League, Yamashita Keigo.
Chinese pair wins world championship: The second part of the Pair Go World Championship 2017 was held at the Cerulean Tower Tokyu Hotel in Shibuya, Tokyo, on October 5. The winning pair in the Stars Tournament, held at the same venue on August 12 and 13, that is, Xie Yimin 6P and Iyama Yuta 9P, representing Japan, met the reigning Pair Go world champions, Yu Zhiying 5P and Ke Jie 9P of China, in the Masters Match. The Chinese pair earned their title in the Pair Go World Cup 2016 Tokyo. Playing white, Yu and Ke defended their title, forcing Xie and Iyama to resign after 198 moves. Like the final in the August event, the game was played in the Noh Theater in the basement of the hotel.
A unique event, the Go AI Research and Goodwill Game by the Pair Go Format, was held on the previous day. Xie, Iyama and the program DeepZenGo were matched against Yu, Ke and DeepZenGo. To fit the Pair Go format, the human players on each team played as one member of the pair with the AI program; they were free to consult each other and together played every second move for their side. This was a lighthearted event, with the players occasionally bursting into laughter. Xie, of course, could also follow what the Chinese players were saying, which she said made the game even more fun. Just for the record, this game was also won by the Chinese pair.
New Honinbo league starts: The first two games in the 73rd Honinbo League were played on October 5. Yamashita Keigo 9P (B) beat Kobayashi Satoru 9P by resig. and Motoki Katsuya 8P (W) beat Shibano Toramaru 7P, also by resig. So far, October has been a good month for Motoki, the previous Honinbo challenger.
Tomorrow: Record age gap in women’s game; Mutsuura wins Agon Kiriyama Cup; Tengen title match starts
Self-taught AlphaGo Zero bests all previous versions in record time
Wednesday October 18, 2017
AlphaGo Zero, the latest version of the go-playing AlphaGo AI, defeated all previous versions of AlphaGo in just forty days. More importantly, AlphaGo Zero was taught go’s rules, but given no additional instructions, instead learning the best moves by playing
millions of games against itself. Details of the new program were published Wednesday in the journal Nature.
“AlphaGo Zero independently found, used and occasionally transcended many established sequences of moves used by human players,” write the AGA’s Andy Okun and Andrew Jackson in an accompanying article. “In particular, the AI’s opening choices and end-game methods have converged on ours — seeing it arrive at our sequences from first principles suggests that we haven’t been on entirely the wrong track. On the other hand, some of its middle-game judgements are truly mysterious and give observing human players the feeling that they are seeing a strong human play, rather than watching a computer calculate.”
The ability to self-train without human input is a crucial step towards the dream of creating a general AI that can tackle any task, reports Nature. In the nearer-term, though, it could enable programs to take on scientific challenges such as protein folding or materials research, said DeepMind chief executive Demis Hassabis at a press briefing. “We’re quite excited because we think this is now good enough to make some real progress on some real problems.”
Prof Satinder Singh, a computer scientist at Michigan University, who reviewed the findings for the journal said: “The AI massively outperforms the already superhuman AlphaGo and, in my view, is one of the biggest advances, in terms of applications, for the field of reinforcement learning so far.”
– Chris Garlock; Image credit: Saran_Poroong Getty Images
Read more…
The AI That Has Nothing to Learn From Humans
DeepMind’s Go-playing AI doesn’t need human help to beat us anymore
AlphaGo Zero: Google DeepMind supercomputer learns 3,000 years of human knowledge in 40 days
DeepMind has a bigger plan for its newest Go-playing AI
AI versus AI: Self-Taught AlphaGo Zero Vanquishes Its Predecessor
AlphaGo Zero Goes From Rank Beginner to Grandmaster in Three Days—Without Any Help
‘It’s able to create knowledge itself’: Google unveils AI that learns on its own
DeepMind AlphaGo Zero learns on its own without meatbag intervention
This more powerful version of AlphaGo learns on its own