Time: 12:45pm-1:45pm, Friday, October 23
Place: Room 4422, CUNY Graduate Center, 365 Fifth Ave (34str&35str).
Speaker: Andrew Rosenberg (CUNY)

Expectation Maximization Tutorial

Expectation Maximization (EM) is prevalent and powerful parameter
estimation technique. In this tutorial, I will describe the basic
structure of this iterative estimation technique. We will derive
the mathematical foundation for fitting Gaussian Mixture Models. A
brief survey of applications of EM in natural language processing
will be presented. Time permitting, we will discuss the general
framework that EM provides for estimating other parameters over other
models involving latent variables.

Andrew Rosenberg is an Assistant Professor at Queens College and CUNY
Graduate Center. His research focuses on developing techniques for
machines to process intonational information in speech. More
broadly, his research is in the fields of Natural Language
Processing, Spoken Language Processing and Machine Learning. He
received his Ph.D. from Columbia University in 2009.