Minimizing Regret

Logo

Google Princeton AI and Hazan Lab @ Princeton University

23 June 2019

Lecture notes: optimization for ML

by Elad Hazan

Spring semester is over, yay! To celebrate summer, I’ve compiled lecture notes from the graduate course COS 598D, a.k.a. “optimization for machine learning”.

The course is an aftermath of a few lectures and summer school tutorials given in various locations, in which  lectures goal of the course was to present the most useful methods and ideas in a rigorous-but-not-tedious way:

The most recent version can be downloaded here:
OPTtutorial

This is still work in progress, please feel free to send me typos/corrections, as well as other topics you’d like to see (on my todos already: lower bounds, quasi-convexity, and the homotopy method).

Note: zero-order/bandit optimization is an obvious topic that’s not address. The reason is purely subjective – it appears as a chapter in this textbook (that also started as lecture notes!).

tags:
Share on: