Capacity and Trainability in Recurrent Neural Networks

Posted by Agrodotus on Wednesday June 21, 2017 @06:31PM
from the arxiv.org dept.

( 2 comments | 41 points | on HN )

Re:Capacity and Trainability in Recurrent Neural Networks by cyphar on Thursday June 22, 2017 @08:33AM (#14611758)

This is a really silly question, but why would you append [Technical] to a title? Isn't that implied by it being a link to arXiv?

[ Parent | on HN ]

Re:Capacity and Trainability in Recurrent Neural Networks by mrkgnao on Thursday June 22, 2017 @09:53AM (#14612224)

Context: the submission was originally titled "Google Brain: Experimental Limits of RNNs [Technical]".

[ Parent | on HN ]