Electrical Engineering 364m: The Mathematics of Convexity

John Duchi, Stanford University, Winter 2024

Announcements

  • Exercise set 5 solutions posted in the Exercises

  • Exercise set 6 posted in the Exercises

  • Exercise set 5 posted in the Exercises

  • Exercise set 4 solutions posted in the Exercises

  • Exercise set 3 solutions posted in the Exercises

  • Exercise set 4 posted in the Exercises

  • Exercise set 3 updated in the Exercises

  • Exercise set 2 solutions is posted in the Exercises

  • Exercise set 2 is posted in the Exercises

  • Exercise set 1 solutions posted in the Exercises

  • Exercise set 1 is posted in the Exercises

  • We are adding an additional office hour on Mondays, 10:30 - 11:30 in Packard 318.

  • If you try to enroll in the class, we currently have an enrollment cap of 50 for reasons beyond the course staff's control. We are doing our best to get that expanded as much as possible. We will likely be able to clear the waitlist, so please just remain on it if you wish to attend it.

  • Welcome to ee364m!

Lectures

Wednesdays, 10:30–11:30, 380-380W (Math Building)

Contact and communication with staff

Instructor

John Duchi

  • Office hours: Wednesdays after class (11:30 – 12:15), 126 Sequoia

  • Additional office hours: Nikhil Devanathan and Logan Bell will host an office hour Mondays, 10:30 – 11:30 in Packard 318.

Pre- or corequisites

Mathematical maturity, which in this case means mathematical analysis at the level of Math 171 (real analysis) and linear algebra (matrix analysis) at the level of Math 104. We expect students to have taken or be concurrently taking EE364a (Convex Optimization).

Description

EE364m is an extension of EE364a to help students develop the mathematics underpinning convex optimization and analysis. Convex optimization is one of the few disciplines where deep mathematical insights are frequently central to progress in the engineering and practice of optimization, whether that is through algorithmic development, formulation of new problems variants, or allowing new types of optimization problems to be solved. Convexity, of course, arises in many other disciplines, being central to statistics and the understanding of estimation uncertainty and stability of estimators; in information theory and probability through concentration inequalities and transport of measure; in variational analysis.

Grading

Your grade will be determined by a weekly problem (there will be a 1 question homework roughly each week).