2016年3月,AlphaGo在五番棋中赢了四盘,打败了世界上最好的棋手之一李世乭。AlphaGo的成功被认为是人工智能研究的一个里程碑。在围棋中,通过蛮力搜索所有可能的策略从而找出最好的走法是行不通的;落子位置的组合数比宇宙中的原子还要多,而AlphaGo所使用的处理器的计算能力与今天的超级计算机相比还算是轻量级的。其成功的秘密在于和一个特殊 In March 2016, AlphaGo won four out of five and defeated one of the best players in the world, Li Shih-t’ao. AlphaGo’s success is considered a milestone in AI research. In Go, it is not feasible to search for all possible strategies by brute force in order to find the best way to go; there are more combinations of landmarks than atoms in the universe, and the computational power of the processors that AlphaGo uses Today’s supercomputer is still relatively lightweight. The secret of success lies with a special one