Yann LeCun: Deep Learning has outlived its usefulness


Short, very interesting:

People are now building a new kind of software by assembling networks of parameterized functional blocks and by training them from examples using some form of gradient-based optimization. (...) It's really very much like a regular program, except it's parameterized, automatically differentiated, and trainable/optimizable. 

I apologize in advance for linking to a Facebook post, but Yann does work there so I guess he's partial.


Want to receive more content like this in your inbox?