What is the relationship between mathematical proofs and efficient algorithms?
We will study this question through the lens of the sum of squares method, a powerful framework for algorithm design linking mathematical proofs, polynomials, and optimization, which generalizes linear programming and spectral methods. Along the way, we will describe and analyze state-of-the-art algorithms for foundational problems from combinatorial optimization, high-dimensional statistics, quantum information, cryptoanalysis, and beyond, and touch on questions from worst-case and average-case computational complexity.
Prerequisites: mathematical maturity, especially in linear algebra and probability.