Mathematics Colloquium
Date and time: Wednesday 25th January, 2023, 4:00 pm
Venue: Ramanujan Hall
Host: Sanjeev Sabnis
Speaker: Prof. Sujit Ghosh
Affiliation: Department of Statistics, North Carolina State University, Raleigh
Title: A Gambler's Journey through Monte Carlo
Abstract: In the era of 'data science' we often require solving large dimensional optimization and integration problems for parameter estimation and/or for predictive analytics. Even in moderately large dimensions, the gradient-based deterministic numerical optimization and deterministic grid-based numerical integration methods suffer from the so-called 'curse of dimensionality.' In such cases stochastic or so-called Monte Carlo methods can be shown to work reasonably well even when the dimensions become much larger. This presentation will provide a brief tour of a Gambler's journey starting with basic notions of Monte Carlo methods with simple geometry based examples to more contemporary data science problems on the use of Bayesian methods for variable selection problems.