A simple connection from loss flatness to compressed representations in neural networks

Add code
Oct 03, 2023

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: