Thursday, May 25, 2017

AlphaGo

Weiqi, better known in the west by its Japanese name, Go, is the oldest game of pure skill in existence and is older by far than all variants of chess. Though it's been years since chess players could stand a chance against computer chess programs — World Champion Garry Kasparov lost to IBM's Deep Blue in 1997 —  it's only recently that computer programs have begun to establish dominance in Go.



After losing a game to Google's AlphaGo in an ongoing match, Ke Jie, regarded as the best Go player humanity has to offer, said of AlphaGo that, “Last year, it was still quite humanlike when it played. But this year, it became like a god of Go.”

AlphaGo learns from itself, playing itself millions of times at micro-speeds, to improve its game. But Ke Jie learns from AlphaGo, making moves that give human onlookers pause and that seem to momentarily flummox AlphaGo.

One oddity of this match is that though Weiqi was born in China and the match is being held in Wuzhen, on the mainland, Chinese authorities are making it difficult for Chinese enthusiasts to follow the game in real time or get news of it.

Why? Because it was an American company, Google, and not a Chinese company, that programmed AlphaGo, and an insecure and irritable Chinese leadership resents that fact and exerts itself to suppress it.

Juxtapose that, if you will, with the fact that the cultural ancestors of this repressive leadership helped develop then foster and esteem the oldest game of skill known to mankind.


No comments:

Post a Comment