Seminar at SMU Delhi
July 17, 2013 (Wednesday) ,
3:30 PM at Webinar
Speaker:
Partha Sarathi Dey,
New York Univerity
Title:
A new approach to Stein's method for Normal Approximation
Abstract of Talk
Stein's method is a semi-classical tool for establishing
distributional convergence with explicit rates, particularly effective in
problems involving complex dependencies. Currently there are several
approaches for using Stein's method for Gaussian Central Limit theorems,
including dependency graphs, exchangeable pairs, zero- and size-biasing.
In this talk I will briefly explain the main idea behind the method and
describe a new approach for applying the method in proving superdiffusive
Central Limit Theorems. The new approach utilizes the fact that each
individual summand depends only on a small part of the environment. Our
main example will be the number of small subgraphs in ER random graph and
number of local maximas in random energy landscapes.