When feed-forward networks can replace recurrent ones without loss of accuracy

This is a nice write up blog post by the Berkeley AI research group on when one should replace recurrent networks with feed-forward ones.

Leave a Reply

Your email address will not be published. Required fields are marked *