Machine 1, Man 0: AlphaGo slams world's best Go player in the first round

AlphaGo yesterday one-upped man as it won the first out of three games against Ke Jie, the world’s number one player in Go. AlphaGo is currently taking another pop at humans as it plays its second competitive match during The Future of Go Summit – a five-day festival celebrating the ancient Chinese game. Last year, it won …

[Nick Kew]

Hats off :)

I'm a very bad Go player. That is to say, I know the rules, and have the rudimentary insight that comes with a mathematical mind, but have never put in serious effort to the game.

Twenty years ago, in an era when Deep Blue had already thrashed Kasparov at Chess, I could still easily beat a leading AI at Go. It was considered an immeasurably harder problem than Chess.

And now they've cracked it!

[Mage]

Cracked and good PR

However it's not AI and little of the actual application is useful for anything else. Maybe the search of big data.

Can the system play any other game not programmed in?

Did the computer "learn" the previous matches? No, they were loaded into a database.

[LionelB]

Re: Cracked and good PR

However it's not AI

According to ... what/whose definition of AI? (not a rhetorical question).

Can the system play any other game not programmed in?

I play a fair game of chess, but am absolutely rubbish at Go. Never had the time or motivation to program it in.

Did the computer "learn" the previous matches? No, they were loaded into a database.

Correction: it learned from previous matches. Perhaps those matches were "loaded from a database" during the training phase. I used to load matches from databases for training during my chess days - we called them "chess books" back then.

Hint: why not find out how AlphaGo really works.

[inmypjs]

And how much hardware it takes....

to beat a human seems to be a secret. I see old figures like 1920 CPUs plus 280 GPUs and probably now google's custom tensor processors.

Are we talking about silicon running on 85kW (30W/CPU + 100W/GPU) to beat meat running on 20W?

How much of the success is down program improvement and how much down to throwing large amounts of hardware at it?

[LionelB]

Re: And how much hardware it takes....

To be fair, human brains throw massively more hardware than any computer system in existence at doing ... just about anything. Plus they have had the benefit of aeons of evolutionary time to hone their algorithms and heuristics.

Looked at that way, it hardly seems like a fair contest.

[allthecoolshortnamesweretaken]

Newsflash

Machine specially designed for one highly specific task is really good at it!

Yes, it's without any doubt a remarkable achievement. Just don't call it "AI" until it can design a game like Go by itself. From scratch.

[Nick Kew]

Re: Newsflash

Whoa! That's a high bar to set. You'd have to lower it quite a long way to find BI (Biological Intelligence) in the world today, even among our greatest and most famous game creators.

[LionelB]

Re: Newsflash

Just don't call it "AI" until it can design a game like Go by itself. From scratch.

You can do that? Hats off, sir/madam.

[Jay 2]

Yes I recall when Deep Blue was winning against humans one of the higher-ups involved with it dismissed thoughts of us being crushed by computers (or whatever) due to the fact that aside from doing one thing extremely well they were "as thick as two short planks".