Title:
Recent advances in scaling problems
Abstract:
Scaling problems are linear algebraic problems commonly appearing in various fields.
In this talk, I will survey matrix andoperator scaling problems and their applications in machine learning, combinatorial optimization, and theoretical computer science.
The main focus of the talk is on a simple iterative algorithm called the Sinkhorn algorithm and its connection with convex or geodesically-convex optimization.
In the second part of the talk, I will discuss our recent result on how to find shrunk subspaces using the Sinkhorn algorithm.
This talk is based on joint work with Cole Franks and Michel Goemans.
