Lecture
Loss Landscape: Minima and Saddle Points
Description
This lecture explores the concept of minima in the error function of deep neural networks, discussing the difference between acceptable and poor local minima, the presence of multiple minima and saddle points, and the implications of weight space symmetry. It also covers the idea of near-equivalent good solutions and concludes with a summary quiz on the number of minima in deep networks.
Official source
About this result
This page is automatically generated and may contain information that is not correct, complete, up-to-date, or relevant to your search query. The same applies to every other page on this website. Please make sure to verify the information with EPFL's official sources.