Two talks October 29, 2021 Aaron Defazio Leave a comment An introductory lecture on optimization for deep learning A deep dive into momentum and the MADGRAD method