Where were you when Petey Norvig's book proved that Alphago can't play 25x25 go and AI is impossible?
AI fanatics make me sick. cLEARLY they haven't read the works of NORVIG
>>7921703
And humans can? Only a few people in the world can play top level go on a 19*19 no reason to think it does not drop near to zero on a 25*25
>>7921706
You clearly don't know what you're talking about. Have you ever played GO? I bet you haven't. well, just to let you know, you don't need lot of time to become a GO top player, it's not like chess where you should spend literally all your life to become GM.
Six-seven months of study should be enough to become a GO top player, I heard that lee sedol started to play in its late twenties.
>>7921706
humans will always be better than AI at 25x25. clearly you don't understand deep learning since you havent read norvig
>>7921703
Norvig is senile as proven we he ran that AI MOOC a few years ago.
> muh GOFAI
>>7921719
> FAGGOT TALKING BIG ABOUT ARTIFICIAL INTELLIGENCE
> HAVEN'T READ YET THE MASTERPIECE OF PETER NORVIG
FUCK OFF. people who talk about scientific field without any knowledge make me sick
>>7921724
I think you misunderstood me. You made a statement regarding the impossibly of computers playing 25*25. As far as I'm concerned it may very well be true.
What I'm asking is what are the implications.
I'm not saying that such statement does not imply that ai is impossible.
I'm asking to explain the implication
>>7921703
And where does it say that? From page 194 Russel + Norvig say: "Rapid advances are likely as experimentation continues with new forms of Monte Carlo search."
>>7921720
>>7921728
you unscientific ai fanatics make me puke. you can't prove that he is senile because you don't have enough citations
even if it barely mentions go it still proves my point precisely.
the fact that you don't think norvig's book is a MASTERPEICE is SCIENTIFIC PROOF that you HAvent actually read it.
>>7921732
the fact that you are unscientifically taking quotes out of context means that you obviously have not read norvig
In Minksy 3:16 the prophet Norvig foretells a great AI winter after alphago fails at 25x25
>>7921732
> "Rapid advances are likely as experimentation continues with new forms of Monte Carlo search."
Monte Carlo methods doesn't work with 25x25 go, and peter norvig never said they would.
>>7921728
> claiming that the peter norvig's work is an introductory book
do you know that this so called introductory book has been the most influential book about artificial intelligence ever written? do you know that it's still used in almost every AI courses? do you know that it predicted every artificial intelligence winter? but I bet that you don't even know what an artificial intelligence winter is. It takes at least 3 years to fully grasp half of the peter norvig's work
>>7921732
Monte Carlo scales terribly with increased input matrix size because of the curse of dimensionality. If you had properly understood Norvig you would know that.
"Where were you when I laid the earth's foundation? Tell me, if you understand."
peter norvig 38:4
>>7921703
>modern approach
>last edition released 1970
This memebook is like Loomis in /ic/
>>7921820
it has been updated every five years or so, you brainlet faggot
>>7921751
Except that AlphaGo is a bit smarter in the way it is doing MCTS. It generates new board states in a smarter way.
It is quite similar to the TD-gammon work mentioned in Russel and Norvig
>>7921834
Almost forgot, I'll bet that AlphaGo won't generalize to 25x25 and they'll have to almost completely retrain.
So it's not skynet
>>7921833
Adding a prologue that says
>This book is outdated as shit.
Doesn't qualify as an update.
>>7921838
>I'll bet
who the hell are you, fag?