- meetings
- Monday 1:30pm–4:20pm, Friend Center 108, Princeton University
- instructors
- Pravesh Kothari, CS Building 219, Princeton email
- David Steurer, Simonyi Hall 010, IAS email

- discussion
- piazza
**(sign up using this google form)**

Boaz Barak is teaching a sister seminar in parallel at Harvard/MIT.

**no lecture on October 31 due to Fall break**- homework reading and exercises for next lecture on October 17
**no lecture on October 10**(due to FOCS conference in New Brunswick)- all lectures notes will be posted in advance

In this graduate seminar we cover recent research results on the use of mathematical programming for problems arising from optimization, machine learning, computational complexity and more. A particular focus is on the Parrilo–Lasserre “Sum of Squares” semidefinite programming hierarchy. We discuss both lower and upper bounds, as well as how such mathematical programs give rise to a general theory of computational difficulty, computation vs. sample size tradeoffs, and computational analogs of Bayesian probabilities.