Alpha Go vs. Lee Sedol

Jussslic

Ancient Alien
Joined
Apr 5, 2012
Messages
351
Are you guys following this?

I am finding the whole thing completely fascinating and kind of astonishing. Not just because AlphaGo won the first two matches, but the way it won them. All the pros watching and commenting seem completely taken aback, because the strategy is like nothing they've ever seen before.

Definitely feels like a different sort of achievement than Deep Blue beating Kasparov because Deep Blue was still playing in a way that was recognizably "human."

Anyway, there's a lot of press, but the Wired article is pretty good: http://www.wired.com/2016/03/sadness-beauty-watching-googles-ai-play-go/?mbid=social_fb

AlphaGo is obviously a niche application, but watching a computer totally shake up a 3,000 year old game in the span of a single match really makes you wonder where machine learning is going to take us in the next couple decades.
 
Last edited:
I preferthe term "Simulated Intelligence" to "Artificial Intelligence". Somputers manipulate symbols, they do not understand the symbols.

http://www.milesbrundage.com/blog-posts/alphago-and-ai-progress

AlphaGo has won with 3 of 5 and 2 to go, so I am curious about the real world applications. Will Lee Sedol win any or does he need more time to adjust to AG's style. Can any human make that adjustment?

Is the architecture of AlphaGo anything like that of Deep Blue? Would AG's architecture need to be changed fo chess?

psik
 
Last edited:

Sponsors


We try to keep the forum as free of ads as possible, please consider supporting SFFWorld on Patreon


Your ad here.
Back
Top