(If you're like "DL is just backprop with multiple layers" then 10 years ago today you'd never heard of ReLU, Adam, Xavier init, or the other stuff that makes backprop work with multiple layers. And you didn't expect it to produce Alpha Zero. Thx 4 playing.)
-
-
Prikaži ovu nitHvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
So all in all was did AI develop as much as you thought it would this decade? Because I imagine you had quite high expectations.
-
Why would you imagine that? I was shocked by the pace of development, and certainly wrote down no advance prediction for things breaking above trend as they did. It might have leveled off in the last few years, or maybe I'm just not reading the right papers.
Kraj razgovora
Novi razgovor -
-
-
I had, but my next thought was "I don't understand why Shane Legg, who seems like a pretty smart guy, is so ludicrously optimistic that connectionism is going to suddenly start working when it never has before, and worth quitting a good academic career for." Not my best call.
-
Same here.
Kraj razgovora
Novi razgovor -
-
-
Slightly worse than that. I'd heard about it and studied it, but it was dismissed as "not as good as HOG+SVM", therefore not worth paying attention to. And the VC dimension, Oy Vey! Ring any bells today?
- Još 3 druga odgovora
Novi razgovor -
-
-
Wait for “Deep Learning but for X” to fully develop as a marketing tactic for unrelated fields and products
-
implying it hasn't already
Kraj razgovora
Novi razgovor -
-
-
1980 be like "10 years ago you'd never heard of expert systems. Good luck with the next decade."
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.