When the first AlphaGo results came out, I got quite annoyed, and polled twitter about whether I should write a debunking.
-
Show this thread
-
Consensus was: either it’s significant, in which case you will look dumb soon, or else it isn’t, in which case everyone will forget about it, and you will have wasted your time.
1 reply 0 retweets 5 likesShow this thread -
A couple years later, it’s still bogus, and the hype tsunami has climbed even higher. Probably nothing I could have done would have punctured the bubble, but I regret not trying.
4 replies 1 retweet 13 likesShow this thread -
Replying to @Meaningness
I'd be curious to see your debunking or at least a quick summary?
1 reply 0 retweets 3 likes -
Replying to @shlevy
They showed that brute force works on another board game. That’s all.
3 replies 0 retweets 10 likes -
Replying to @Meaningness @shlevy
Isn't this a bit disingenuous portrayal of the algorithms they are using (the q learning developments wrt. deep networks, etc) to say it's just pure brute force is a bit misleading no?
1 reply 0 retweets 1 like
They don’t report any control experiments of applying similar amounts of compute power with other algorithms. (This is a pervasive problem in the field.)
-
-
Without that, we don’t know what aspects of the system are important; nor what the limits of generalization are.
0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.