Postdoc Lunch Time Talks - Computing Discriminants of Likelihood Equations

Date: October 19, 2017

Time: 12:35PM - 12:55PM

Location: BLOC 220

Speaker: Xiaoxian Tang, Texas A&M University

Description: Computing Discriminants of Likelihood Equations

Abstract: Maximum likelihood estimation (MLE) is a fundamental problem in statistics. All the complex critical points of a given statistical model can be found by solving a very well-structured system of likelihood equations. Computing the discriminant of the system is the key step to classify the real critical points. Based on the known elimination and probabilistic interpolation algorithms, we explore an more efficient method to compute discriminant s by the structures of Newton polytopes.