The "deep learning" revolution consists in just one thing --the ability to learn multi-layered representations jointly rather than greedily.
-
The best 1st layer in a 2-layer model isn't the same as in 20-layer model. Greedy representation learning is fundamentally broken.
And deep learning changed everything because it offered a computationally practical way to learn all layers at the same time.
-
-
At last --if you believe that SGD via backprop of a loss function is the only way to achieve this, or the best way, you are sorely mistaken.
-
@fchollet we have a pretty good example that it's not the case since our brains are pretty good at learning and they don't backprop
End of conversation
New conversation -
-
-
@fchollet what changed since 15 years ago? Only compute power, or also algorithms? -
@yoavgo more than hardware, it took a while to achieve "practical maturity" (refined set of best practices, and confidence in their result). - Show replies
New conversation -
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.