Lecture

Structures in Non-Convex Optimization

In course

Ullamco excepteur ad ad reprehenderit ullamco eiusmod non sunt. Esse et labore officia non aliqua minim. Eu officia minim non proident anim laborum proident. Reprehenderit in pariatur laborum enim laboris esse veniam ex. Reprehenderit est velit laborum amet sint nulla do. Magna pariatur et laborum nostrud sint esse nulla mollit est irure. Esse occaecat ullamco quis reprehenderit occaecat anim reprehenderit duis.

Description

This lecture explores structures in non-convex optimization, focusing on scalable optimization for deep learning. Topics include optimization formulations for deep learning training problems, stochastic gradient descent and its variants, critical points classification, the strict saddle property, convergence of SGD, avoidance of saddle points, speed of convergence to local minimizers, and the optimization landscape of overparametrized neural networks.

Instructor

ad laboris nostrud

Nostrud ipsum ullamco aliqua do eu aute. Voluptate pariatur officia eiusmod laboris culpa ad sunt laboris adipisicing occaecat enim duis cillum aute. Magna ut magna eu culpa cupidatat tempor proident.

Official source

https://mediaspace.epfl.ch/media/0_ce9sohxs

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.

Ontological neighbourhood

Information engineering

Machine learning: Artificial neural networks

Related lectures (32)

Graph Chatbot

Chat with Graph Search

Ask any question about EPFL courses, lectures, exercises, research, news, etc. or try the example questions below.

DISCLAIMER: The Graph Chatbot is not programmed to provide explicit or categorical answers to your questions. Rather, it transforms your questions into API requests that are distributed across the various IT services officially administered by EPFL. Its purpose is solely to collect and recommend relevant references to content that you can explore to help you answer your questions.