“人类棋王”对弈谷歌人工智能

韩国九段顶尖中国围棋高手李世石（Lee Se-dol）与谷歌人工智能项目AlphaGo展开对决。

测试中可能遇到的词汇和知识：

yield 屈服[jiːld]

intuition 直觉[ɪntjʊ'ɪʃ(ə)n]

mimic 模仿['mɪmɪk]

ultimately 最后['ʌltɪmətlɪ]

Nature 科学杂志《自然》

robotics 机器人学[rəʊ'bɒtɪks]

阅读即将开始，建议您计算一下阅读整篇文章所用时间，并对照我们在文章最后给出的参考值来估算您的阅读速度。

Google DeepMind takes on Go grandmaster in AI match（383 words）

By Song Jung-a in Seoul

* * *

The world champion of ancient Chinese board game Go is confident he can beat Google’s AlphaGo machine, ahead of a high-profile match that could yield a milestone for artificial intelligence.

Lee Se-dol, the South Korean who has been the top Go player in the world for the past decade, predicted he would win easily in the five-game match, which will be played over the course of five days starting on Wednesday.

“It would be a computer’s victory if it wins even one game,” Mr Lee told the Financial Times. “I believe human intuition and human senses are too advanced for artificial intelligence to catch up. I doubt how far AlphaGo can mimic such things.”

AlphaGo, developed by Demis Hassabis and colleagues at Google DeepMind in London, already beat Fan Hui, the European Go champion, by five games to zero in October — a feat that experts had predicted would take a decade.

Mr Hassabis stressed that AlphaGo has made many improvements since the October match, which was played behind closed doors with the results published in the scientific journal Nature.

“I saw the record of the match with Fan Hui and was surprised to see the level that AlphaGo has achieved, but it is not yet up to my level,” Mr Lee said. “I don’t think that five months are enough time for it to improve that much to beat me.”

DeepMind uses games as a testing ground for AI algorithms that could have real-world applications.

“We think it is very efficient to develop AI algorithms and test them in game settings,” Mr Hassabis said.

He predicted that it would still take many decades to create human-level AI but said the match will provide a chance for researchers to assess the learning algorithm’s flexibility and ability to solve complex problems that could be applied to a range of industries.

“Ultimately, we want to use this technology for more than just playing games,” he said. “What we hope to do is to apply [it] to a lot of real-world problems in the future, including healthcare . . . , household robots, robots that care for the elderly and the smart assistant on your phone.”

If Mr Lee wins, Google will pay him $1m and if he loses the company will give $1m to charity.

请根据你所读到的文章内容，完成以下自测题目：

1. Where is Lee Se-dol from?

a. South Korea

b. North Korea

c. China

d. Japan

2. How many day is there for this AI match?

a. 3

b. 7

c. 1

d. 5

3. How about the last AI match with the European Go champion Fan Hui in January?

a. Fan was only slightly ahead of AlphaGo machine.

b. Ended in a draw.

c. The machine won a landslide victory.

d. Fan won a landslide victory.

4. Which one is not include as the ultimate goal from AlphaGo’s developer?

a. healthcare

b. as a smart assistant on the phone

c. caring for the elderly

d. playing games

[1] 答案a. South Korea

解释：Lee Se-dol(李世石)，韩国围棋九段棋手。截至如今，他已经获得了14个世界慢棋个人赛冠军和5个亚军，以及3个世界快棋个人赛冠军和1个亚军。是最近10年中获得世界第一头衔最多的棋手。

[2] 答案d. 5

解释：比赛有5轮，将进行5天：3月9日、3月10日、3月12日、3月13日、3月15日。

[3] 答案c. The machine won a landslide victory.

解释：1月谷歌人工智能AlphaGo 5:0击败欧洲冠军樊辉职业二段。

[4] 答案d. playing games

解释：AlphaGode 开发者Demis Hassabis表示，人工智能技术的最终目标并不是在游戏中取胜。