Convex relaxations for entropy optimization problems

Oct 19, 2021 04:00 PM Singapore (Registration will open at 03:50 PM.)

Join Zoom Meeting:
https://sutd-edu.zoom.us/j/93501197484?pwd=YXVVcmZRQVRTTE5HNGdkU2JQSDJUdz09

Meeting ID: 935 0119 7484
Passcode: @Dt@qmc1

Abstract

Entropy optimization problems arise in many areas of engineering and science, notably in information theory and communication. In some cases, these problems are nonconvex and challenging to solve. I will present in this talk new approaches based on approximation theory and semidefinite relaxations to obtain convex relaxations for such problems. I will illustrate this new method on the problem of estimating the logarithmic Sobolev constant of a Markov chain, as well as a problem in quantum communication.

Based on joint works with Oisín Faust (https://arxiv.org/abs/2101.04988) and Peter Brown and Omar Fawzi (https://arxiv.org/abs/2106.13692).

About the Speaker

Hamza Fawzi received his PhD from MIT in 2016. Since then, he has been an assistant professor at the University of Cambridge, in the department of Applied Mathematics and Theoretical Physics. His interests are in convex optimization and applications. In 2020, Hamza was awarded the SIAM activity group on optimization best paper prize.

For more information about the ESD Seminar, please email esd_invite@sutd.edu.sg