What is the bottleneck and weakness of AlphaGo?

Is self-play a bottleneck in theory for AlphaGo to improve? My perspective is not! The real problem with AlphaGo (and any other AI and human) is the state space of Go is much larger than the state space of its neural network, therefore no matter how we train it, it still suffers from the underfitting problem. Which means there is always a problem with its value network and policy network that, when some cases are trained very well, other cases pop up.

But in terms of supervised learning vs unsupervised learning, they only differ in training set, which means they make AlphaGo‘s neural network bias to a certain style and handles certain cases very well. Unsupervised learning can provide all the information that supervised learning can provide, imagine the board is 9*9, unsupervised learning is absolutely enough to provide a good training set. So, unsupervised learning is not really a bottleneck in theory, but in practice supervised learning makes AlphaGo bias to a certain style, and have a better chance to win a certain opponent. But when its neural network gets larger to be able to accommodates more states, the value of supervised learning is also decreased.

Because of the underfitting problem, the value network may get wrong on who‘s winning the game on some states which may look simple to human. This is why AlphaGo use MCTS to rollout for many steps for validation, only when after playing down some steps, the game is still in favor of AlphaGo, the original state is considered truly good. So, AlphaGo is really a mixture of "intuition + logic", this is very similar to human.

This design makes it very hard to catch AlphaGo‘s weakness, but it does exists. Based on the analysis above, the weakness of AlphaGo is clear to me now: its value network gets wrong on not only one state, but also many steps following the state. Although the probability is very low, but it did happened in Game 4. Brilliant Lee Sedol!

时间: 2024-10-03 17:46:17

What is the bottleneck and weakness of AlphaGo?的相关文章

李开复:AlphaGo 若打败了世界冠军,意味着什么?

创新工场董事长李开复在知乎就AlphaGo与李世石的人机大战发表了自己看法,他认为四个月前的AlphaGo击败李世石基本不可能,不过这四个月AlphaGo进步很多,比赛应该很精彩.但是,无论这次结果如何,机器1-2年之内必然完胜人类.完胜人类之后呢?可以做个通用的大脑吗?意味着机器可以思考了吗?还有什么问题是机器无法超越人类的? 以下为李开复就“AlphaGo能战胜李世石吗?”在知乎的回答原文: 先直接回答这个问题,下面再分析AlphaGo和人工智能的未来.我认为AlphaGo这次的比赛打败李世

AlphaGo深度揭秘

今日,在乌镇围棋峰会人工智能高峰论坛上,AlphaGo之父.DeepMind创始人戴密斯·哈萨比斯(Demis Hassabis)和DeepMind首席科学家大卫·席尔瓦(David Silver)在论坛上透露了关于AlphaGo的重要信息,以及AlphaGo究竟意味着什么?让人们能详细了解到AlphaGo背后的秘密. AlphaGo是什么? AlphaGo 是第一个击败人类职业围棋选手并战胜围棋世界冠军的程序,是围棋史上最具实力的选手之一.2016 年 3 月,在全世界超过一亿观众的关注下,A

Cydia Substrate based DexDumper's weakness

得益于Cydia Substrate框架,HOOK Native函数变得简单,也给脱壳带来方便. 像ijiami免费版,360,classes.dex被加密到so文件并运行时释放到内存,因此针对相关函数的HOOK脱壳就比较简单了. 但也有一些容易被针对的缺点: 1.代码可以在类加载时被修改.hook dexFileParse.__mmap2或memcmp 等可能dump出是修复前的dex 2.修改了odex的标志dey\n036,dumper在内存中搜索不到,将漏掉真正的odex 3.packe

(转) 一张图解AlphaGo原理及弱点

一张图解AlphaGo原理及弱点 2016-03-23 郑宇,张钧波 CKDD 作者简介: 郑宇,博士, Editor-in-Chief of ACM Transactions on Intelligent Systems and Technology, ACM数据挖掘中国分会秘书长. 张钧波,博士,ACM数据挖掘中国分会会员,从事深度神经网络相关研究. -------------------------------------- 近期AlphaGo在人机围棋比赛中连胜李世石3局,体现了人工智能

从马文到AlphaGo AI走过了怎样的70年?

(原标题:从马文·明斯基到AlphaGo,人工智能走过了怎样的70年?) [编者按]从19世纪中叶人工智能的萌芽时期,到现今人工智能的重生,从马文·明斯基到AlphaGo,历史上发生了哪些激动人心的故事?本文以此铺展人工智能发展近70年来背后发生的故事.作者@沐阳浸月,中科院自动化所复杂系统国家重点实验室研究生,主攻机器人与人工智能. 前不久,在人工智能领域发生了两件大事,一个就是是伟大的人工智能先驱马文·明斯基教授逝世,一个是谷歌AlphaGo击败欧洲围棋冠军,职业围棋二段樊麾. 马文·明斯基

AlphaGo 开源项目研究(1)

本文的原文连接是: http://blog.csdn.net/freewebsys/article/details/50907446 未经博主同意不得转载. 博主地址是:http://blog.csdn.net/freewebsys 1,关于alphaGo 近期这个AlpahGo很的火呢. Goolge的alphaGo赢了韩国围棋大师李世石. 一下子感觉还是google大神厉害.比卖假药的百度强不少. https://deepmind.com/alpha-go.html 并且代码也开源了. 部分

科普:alphago是什么

鉴于大部分人类对alphago的认识: 1:Alphago有什么了不起的?不就是算得快吗.ibm早在20年前就通过象棋战胜人类了.又是Google的一次营销. 2:alphago 实现人工智能了,电脑立即要超过人类占据地球了. 这两种想法哪个才是正确的那?严格地说,都不对. 所以我认为非常有必要给非计算机行业的同学们科普一下.由于是一篇科普文.所以我们不讨论卷积神经网络或MCTS这些东西.对原理感兴趣的同学能够:https://deepmind.com/alpha-go.html. 另外.欢迎加

从柯洁对战AlphaGo,看商业智能

[摘要]李开复赛前说,AlphaGo和李世石的人机大战是第一次,可能还有悬念,那今天的AlphaGo已经在围棋的世界中彻底甩开了人类,不再拥有任何其他的可能.并指出,AlphaGo和柯洁的比赛并非没有意义,而是在科学价值层面已经失去看点.他呼吁既然AlphaGo和冷扑大师,已经在游戏领域的"不可战胜"不再有悬念,那不如把更多的精力投入到人工智能的商业应用层面. 人工智能与人类智慧的再一次对决,谷歌AlphaGO给围棋世界冠军柯洁递上战书.对于这场人机大战,从对战前就被各种关注各种刷屏.

深度解读 AlphaGo 算法原理

深度解读AlphaGo