Man-machine war final Conference: Li Shishi reported seeing the limitations of man-Li Shishi, chess, artificial intelligence, man-machine war-IT information
The afternoon of March 15 news, man-machine chess showdown finale today, Li Shishi eventually defeats, lost aggregate Google go program AlphaGo 1:4. At a press conference held after the game, Li Shishi talked about the idea of this game, views on the entire game. DeepMind founder hassabis also attended the press conference, talked about AlphaGo errors in the game, as well as the description of the next challenge.
Said Li Shishi, hopes to end in chess today, but in the end did not achieve, expressed surprise at the strength of the AlphaGo show, and that duel and machine the most important psychologically to stay highly focused, also think through this game a lot of inspiration.
Li Shishi said of his advantage in the first half, but finally lost to the machine, "first half offer some advantages, but lose in the end. I still lack some, unfortunately. Very grateful to so many people supporting me and cheering for me. I will work harder, so that everyone sees an ever-progressing Li Shishi. ”
Founder of AlphaGo development company DeepMind hassabis said AlphaGo there are errors in the game, but later adjusted, that match was really exciting.
When talking about the next challenge, hassabis did not make it clear, he said: "after a clear plan yet, after our game of AlphaGo data were analyzed, to determine areas in which to improve him. ”
After the AlphaGo core developers had expressed a wish to challenge the number one Chinese player Ke Jie, and Ke Jie has clearly responded to weibo, and expressed the hope that the challenge AlphaGo. Qisheng NIE Weiping would like to see this scene, today to visit SINA's live commentary, NIE Weiping said: "Hopefully, SINA can fix AlphaGo and Ke Jie match. "Asked DeepMind Sina science and technology reporter at the front on this issue, the other side says: wait and Li Shishi after the game, will be further considered and planned.
At the press conference, the English commentary of the match Chris Garlock also expressed his views, he said: "the incomparable game, thanks to AlphaGo team and Li Shishi, presents 5 awesome games, and have never seen so many people go, it is the largest gift received by the chess community. This contest will go show it to the world with a long history, concerned with so many people go. Thank you very much maikeleimeng section and I'll explain, this is a very special experience for me, thank you. ”
Korean explanation Jin Chenglong section also paid tribute to the man-machine chess showdown, he said: "compared to the first game, today more people, thank you very much. This Council and the fourth different strategies have been used, very good. And calculation of AlphaGo spell, feeling very good. I had always wanted to win. But it is a pity. Forth let us see the Li Shishi super powers of imagination, I thank the AlphaGo team, and AlphaGo friend-Li Shishi. ”
Following is a record of questions questions and answers:
Q: Quan Jialong said Li Shishi as a child very much and than people with higher levels of the game, to see your enthusiasm for the game. See AlphaGo5 chess, do you feel he is fit in your understanding of the game are different?
Li Shishi: didn't think AlphaGo is my hand, I continue to believe that humans and AlphaGo under some, you can fight. Go is certainly enjoyable, whether professional or amateur chess player, it is important to derive pleasure from the game of chess. The match I would have been happy. On the understanding of chess, I felt the limitations of human creativity. See AlphaGo contains detailed explanations, I think chess understanding was it before?
Q: this 5 chess AlphaGo understand change, chess player and his where is the difference? Where is the winning?
Li Shishi: a difference, of course, first of all opponents are not human. Fresh feeling a lot, I'm going to meet him, first of all is the psychological aspect, very focused. I continued and he may not be able to win, because they can not concentrate more than he, they can't beat him.
Q: Li Shishi hard, thank you for giving such an exciting game. Through the contest see AlphaGo disabilities, future plans and prospects are?
Hassabis: after definite plans have not, after our game of AlphaGo data were analyzed, to determine areas in which to improve him.
Q: how artificial intelligence?
Hassabis: first of all, I would like to stress is that artificial intelligence development is still in its infancy, there is much to be desired. Artificial intelligence is a powerful tool in many ways, we just play a supporting role, should be able to provide a lot of help, to make progress, there is great potential, will apply to Medicaid, and many other areas. How is the direction of future efforts to better develop, will continue to be published in a scientific journal, welcome the continued attention.
人机大战终局发布会：李世石称看到人的局限 - 李世石,围棋,人工智能,人机大战 - IT资讯
3月15日下午消息，人机围棋对决今日举行收官战，最终李世石败北，总比分1:4不敌谷歌围棋程序AlphaGo。在赛后举行的发布会上，李世石谈到了本场比赛中的想法，对整个比赛的看法。DeepMind创始人哈萨比斯也出席了发布会，谈到了AlphaGo在比赛中的失误，以及对于下一步挑战计划的说明。
李世石说，原本今天希望能以赢棋收尾，但是最后没能实现，对AlphaGo展现出的战力表示惊讶，并认为和机器对决最重要的是心理上要保持高度专注，还觉得通过这次比赛收获了很多启发。
李世石谈到了自己在前半盘里的优势，但最后还是败给了机器，“前半盘有些优势，但最后还是输掉了。可见我还是有些不足，个人来说非常可惜。非常感谢这么多人支持我，为我加油。我会更加努力，让大家看到一个不断在进步的李世石。”
AlphaGo开发公司DeepMind创始人哈萨比斯则表示，AlphaGo在这场比赛中有失误，但后来又调整了过来，说明比赛非常精彩。
在谈到下一步挑战计划时，哈萨比斯并没有做出明确表态，他说：“之后明确的计划还没有，我们对AlphaGo的比赛数据等进行分析之后，再来决定从哪些方面来提高改善他。”
此前，这位AlphaGo最核心的开发者曾表示希望挑战现排名第一的中国选手柯洁，而且柯洁已在微博明确做出了回应，表示希望挑战AlphaGo。棋圣聂卫平也希望看到这一情景，在今日做客新浪的直播解说中，聂卫平表示：“希望新浪可以撮合AlphaGo和柯洁的比赛。”新浪科技记者在前方就此问题询问了DeepMind方面，对方表示：等到和李世石比赛结束后，会进一步作考虑和计划。
在发布会上，此次比赛的英语解说Chris Garlock也发表了看法，他说：“无与伦比的比赛，感谢李世石和AlphaGo团队等，展现了5盘这么棒的比赛，也从来没有看到过这么多人关注围棋，这是棋界收到的最大礼物。这次比赛将历史悠久的围棋展现给全世界，这么多人一起关注围棋。非常感谢麦克雷蒙九段一起和我来解说，这对我来说是非常特别的经历，非常感谢。”
韩语解说金成龙九段也称赞了此次人机围棋对决，他认为：“比起第一局，今天来的人更多，非常感谢大家。本局和第四局采用了不同战略，非常精彩。与AlphaGo拼计算，感觉非常棒。我一直想如果能赢下来就好了。但很可惜。第四局让我们看到了李世石超强的想象力，我非常感谢AlphaGo团队，还有AlphaGo的朋友——李世石。”
以下为提问环节问答实录：
问：听权甲龙说李世石小时候，非常喜欢和比自己水平高的人比赛，看到你对围棋的热情。看到与AlphaGo5局棋，是否感觉他是上手，你对围棋的理解是否有不同？
李世石：没想过AlphaGo是我的上手，我依然认为人类和AlphaGo是有的下，可以对抗的。围棋当然是令人享受的，无论职业还是业余棋手，首要是从棋中得到快乐。这次比赛我就得到了快乐。关于对棋的理解，我感到了人的创意的局限性。看到AlphaGo的招法，我会想，以前对棋的理解是否是对的呢？
问：通过这5局棋是否对AlphaGo理解有变化，棋手和他的差别在哪里？胜机在哪里？
李世石：当然有差别，首先对手不是人类。新鲜的感觉有很多，我要去适应他，首先就是心理方面，非常的专注。我继续和他下不一定能赢，因为无法比他更专注，这些方面是赢不了他的。
问：李世石辛苦了，感谢奉献如此精彩的比赛。通过比赛是否看到AlphaGo的缺陷，今后的计划和展望是？
哈萨比斯：之后明确的计划还没有，我们对AlphaGo的比赛数据等进行分析之后，再来决定从哪些方面来提高改善他。
问：如何理解人工智能？
哈萨比斯：首先我想强调的是，人工智能开发尚在初期，还有很多不足。在很多方面人工智能是强有力的工具，我们只是起辅助作用，应能会为人类提供很多帮助，取得进步，有很大潜力，以后会运用到医疗等很多领域当中。我们如何更好开发是今后努力的方向，将会继续在科学杂志上发表，欢迎大家继续关注。