AI "DeepZenGo" Bows to Japanese Go Grandmaster
It was not so long ago when Google's AlphaGo trounced a Korean Go professional, then said to be the world's top Go player.
Now, a Japanese Go grandmaster of Korean ancestry, Cho Chikun (Jo Chi-hun), ousted a Japanese version of AI "DeepZenGo" in the first of the 3-match contest that started November 19. The matches are set without handicaps.
DeepZenGo is a joint product of a software developer and Tokyo University researchers and has reportedly adopted the latest of AI technology to acquire a high level of playing skills in the game of Go. Grandmaster Cho is regarded a formidable foe DeepZenGo to face and the 3-match contest is being watched with great enthusiasm by all Go fans at home and abroad.
The Match One started off with DeepZenGo leading much of the way in joban or the early stage, Cho grumbling most of the time trying to reverse the tide. Then, toward the end game Master Cho worked a tactful way out of the fix and finally beat DeepZenGo in a lengthy 3.5-hour match.
Earlier this year, Google's AlphaGo "made a fool" of a Korean master Lee Sedol in a five-game match to shock the world of Go - as if the computer had finally taken over man in the world's most complex game.
After the Match One, Cho had this to say:"That was such a fun! This DeepZenGo does really well in reading moves - awfully well. The artificial thing might beat us in that way. Next time, I intend to play it wilder!"
Matches 2 and 3 are slated for November 20 and 23.
The developer, Hideki Kato, recalls:"We were leading up till midway and too bad we lost. We'll fix what's missing and challenge the rest of the matches."
Allow me to remind you - I'm a great fan of Go. I'll make sure you'll hear about the outcome of the thrilling contest. Count on it. (Nathan Shiga)