• Howard

Fixed Point Networks

Updated: Mar 25

TD;DR - A new class of deep learning models can approximate infinite depth while using fixed memory costs and being easy to train/implement.

This blog is based upon a recent paper (preprint available here).

(On mobile devices the embedded document below may not display appropriately, in which case you may view it as a .pdf file here.)

(Because it is difficult to make clean webpages that encode LaTex code, blogs will now be given as embedded 2 page .pdf files, which you are welcome to share.)

As always, I am happy to discuss in the comments below any comments, suggestions, criticisms, and/or questions pertaining to this post.


153 views1 comment

Recent Posts

See All