Google’s AlphaGo is stronger than ever. The AI machine on Thursday beat Go world champion Ke Jie in the second out of three games in China. However, experts and netizens in China seem to take the results pretty well.
The 19-year-old prodigy from China lost his first game on Tuesday by a slim half point, and later on Thursday lost again and was forced to give up halfway. While AlphaGo has secured victory in the three-part match with the final game being scheduled on Saturday, Chinese netizens and experts seem to realize early on that Ke Jie stands little chance of winning.
#AlphaGo wins game 2. What an amazing and complex game! Ke Jie pushed AlphaGo right to the limit.
— Demis Hassabis (@demishassabis) May 25, 2017
“AI [AlphaGo] plays it totally different this year than last year. It was quite humanlike the last time it played [with Lee Sedol],” said Ke after the second game. “But this time it became more like the God of Go.”
Ke added that AI technology is constantly being improved and the gap between the machine and human brains is getting bigger. “Our [humans’] rate of winning is close to zero, and this is painful.” he said. “I’d rather play with human beings as I hold a better chance to win a human.”
Wang Xiaochuan, CEO of China’s third largest online search engine Sogou, commented on Zhihu, China’s Quora, that AlphaGo is more advanced than before. The 3,000-year-old board game is considered a more complex game than chess as it needs some level of “human intuition.” And according to Wang, the machine is much more competent than we imagine.
“AlphaGo 2.0 has more advanced techniques and can think more humanlike as compared to the previous version,” Wang wrote on Zhihu. “The style of AlphaGo’s playing has surpassed human experience.”
“In fact, it’s not scary if AlphaGo wins Ke Jie,” a Chinese netizen commented online after the game. “It is scary if it intentionally loses to Ke.”
AI expert Kai-Fu Lee, a former Microsoft and Google executive, told local media before the game that Ke’s chances of winning even one game is close to zero.
Indeed, the AlphaGo we are discussing here adopts a much more different and yet much stronger system than the one that Lee Sedol played with last year. It is powered by the Artificial Neural Network (ANN) with two kinds of networks: policy and value networks. The former helps to predict the next steps that the opponent may possibly take, and the latter evaluates the winning rates of the machine. Aside from AlphaGo, ANN has been widely used in AI related fields such as face recognition and machine learning.
Just like what Kai-Fu Lee said when he recently delivered a speech at Columbia University’s commencement, that the victory of artificial intelligence is something we should recognize and place hope on.
“The next decade will be the age of artificial intelligence,” Lee said. “Given what lies ahead, you must warmly embrace AI.”
(Top photo screenshot from YouTube)