Research-Stochastic Gradient Descent for Inverse Problems

Research

Computational & Applied Math Seminar

Stochastic Gradient Descent for Inverse Problems

Speaker: Bangti JIN (University College London)
Time: Mar 3, 2021, 10:00-11:00
Location: Tencent Meeting ID 713 927 632

Abstract

In this talk, I will discuss the stochastic gradient descent for solving linear and nonlinear inverse problems. Such algorithms have been very popular in a number of practical inverse problems. However, the relevant mathematical theory in the context of ill-posed inverse problems remains largely missing. In this talk, I will present some recent results in the direction, and illustrate the theory with numerical examples.