analysis- man-machine war machine back to clinch a comprehensive artificial intelligence will surpass human-（分析-人机战争机器回到达成全面的人工智能将超过人类）.docVIP

Analysis: man-machine war machine back to clinch a comprehensive artificial intelligence will surpass human? (Research papers Download News) exists only in some of the previous prophecy or science fiction, the movie business, is gradually becoming a reality, as the one thing known to the world as recently as. January 28, ‘Financial Times’ quoted the ‘Nature’ magazine reported that the London-based subsidiary DeepMind Google developed AlphaGo machine with 5: 0 in the record to beat three-time winner of the European Go Fan Hui (born in China, now membership France). DeepMind 2014 by Google to buy 400 million pounds of artificial intelligence team. The game occurred in October 15, 2015, because of some special reason only now released. According to reports, the game did not let the whole chess, artificial intelligence eventually defeated humanity. In March this year, AlphaGo will last decade world’s top chess players against Shishi in Seoul, which will be complicated in the human brain and a machine Go community ultimate showdown. In the end is the human brain to win, or continue to be strengthened neural network brain machine win? Suspense will pick the cup in March this year, but for a variety of speculation and intense discussion of artificial intelligence has first started. Focus One: Google AlphaGo is how to win? All focused on this game and people interested in artificial intelligence, the most want to know is, how the machine is to win. In the January 27, DeepMind team on ‘Nature’ (Nature) magazine published a cover paper ‘Mastering the game of Go with deep neural networks and tree search’ unlock secrets. Team founder Demis Hassabis said they developed Go program AlphaGo combines advanced neural network and depth of the tree to find. The neural network 30 million step training (all professional chess players are from the human race), AlphaGo predict the correct rate of human opponents next moves has reached 57%, while the previous rec