谷歌机器击败华裔围棋冠军樊麾

A computer has comprehensively beaten a human champion playing Go, the world’s most complex board game — achieving “a historic milestone¹ in artificial intelligence”, according to Nature, the journal publishing the science behind the feat².

《自然》(Nature)杂志表示，一台电脑击败了一位人类围棋(Go)冠军，达到了“人工智能的历史性里程碑”。围棋是世界上最复杂的棋盘游戏。《自然》发布了这一成就背后的科学进步。

The AlphaGo machine, developed by Demis Hassabis and colleagues at Google’s DeepMind subsidiary in London, triumphed over Fan Hui, the European Go champion, by five games to zero. Jon Diamond, president of the British Go Association, commented: “Before this match I was expecting it to be at least five to 10 years before a program would be able to beat the top human players.”

由谷歌(Google)伦敦子公司DeepMind的杰米斯哈萨比斯(Demis Hassabis)及其同事开发的AlphaGo机器，以五比零的战绩击败了欧洲围棋冠军樊麾(Fan Hui)。英国围棋协会(British Go Association)主席乔恩戴蒙德(Jon Diamond)表示：“这场比赛之前，我曾预计，计算机程序能够击败人类顶级围棋选手，起码要等到5至10年后。”

Of course Google, a company at the forefront of the drive to transform business through AI, is not just playing for fun. “While games are the perfect platform for developing and testing AI algorithms quickly and efficiently³, ultimately we want to apply these techniques to important real-world problems,” Mr Hassabis said.

当然，置身于借助人工智能转变商业模式前沿的谷歌，并不仅仅是为了好玩。“尽管游戏是快速和高效率开发和检验人工智能算法的完美平台，但最终我们希望把这些方法应用于现实世界的重要问题，”哈萨比斯说。

Applications will combine pattern recognition and forward planning. One early example will be an improved “intelligent personal assistant” for smartphones. Another application will be in medicine, combining images from scans and other patient data to diagnose disease and decide treatment.

应用将结合模式识别和提前规划。一个较早的例子将是改进智能手机的“智能个人助理”。其他应用领域将包括医药，综合利用扫描图像和其他患者数据来诊断疾病和决定治疗方案。

The ancient game of Go, which originated in China more than 2,500 years ago, has an estimated 40m regular players worldwide and 1,000 professionals, mainly in east Asia.

源于2500多年前的中国的古老游戏围棋，在全世界拥有大约4000万玩家和1000名专业棋手——他们主要位于东亚地区。

Although IBM’s Deep Blue computer famously beat Garry Kasparov at chess in 1997, Go has a complexity⁴ far exceeding that of chess, said Mr Hassabis who is a keen player of both games. “There are more positions in Go than the number of atoms in the universe,” he added. “Go has more subtlety⁵ and intellectual depth than chess.”

哈萨比斯说，尽管IBM的“深蓝”(Deep Blue)在1997年击败了国际象棋大师加里愠斯帕罗夫(Garry Kasparov)，但围棋的复杂程度要远远超过国际象棋。哈萨比斯是一名非常热衷于围棋和国际象棋的选手。“围棋的棋局比宇宙中的原子数量还要多，”他接着说，“与国际象棋相比，围棋更为微妙，体现更深的智力水平。”

Google has beaten Facebook, whose chief executive Mark Zuckerberg posted yesterday that his team was close to “building an AI” to win at Go.

谷歌赢了Facebook。昨天Facebook的首席执行官马克丠克伯格(Mark Zuckerberg)公布称，其团队接近打造出赢得围棋比赛的人工智能。

Google’s system defeated Fan Hui, a Chinese-born Frenchman who has played Go professionally since his teens, by 5 games to 0. “In March, AlphaGo will face its ultimate challenge, a five-game match in Seoul against the legendary⁶ Lee Sedol — the top Go player in the world over the past decade,” Mr Hassabis said.

谷歌研发的系统以五比零的战绩打败了在中国出生、现籍法国的樊麾。樊麾从青少年时期开始专业下围棋。“3月，AlphaGo将面临终极挑战，在首尔与传奇人物李世石(Lee Sedol)——过去十年全世界最顶级的围棋选手——进行一场五局制比赛，”哈萨比斯说。