search

jaehong31 / CGES

Combined Group and Exclusive Sparsity for Deep Neural Networks, ICML 2017
32 stars 9 forks source link

Question on optimization method #6

Closed lizhenstat closed 5 years ago

lizhenstat commented 5 years ago

Hi, I noticed that the paper uses proximal gradient descent to solve the optimization problem. However the neural network is already a non-convex problem, why can't we just use the normal algorithmn (adam, etc) to solve this issue?

Thanks in advance

lizhenstat commented 5 years ago

@jaehong-yoon93 Looking forward to your reply

lizhenstat commented 5 years ago

I asked one of my classmates and he says that if we do not use the optimization method(like proximal gradient descent), we can't achieve the sparsity.