Title: Second Order Methods for Large-Scale Machine Learning

Gillian Chin
PhD Candidate, Department of Industrial Engineering and Management Sciences

Date and Time: 

Friday, November 16, 2012 - 2:30pm

Event Location: 

Mohler Lab Room 451

We will explore the development of efficient batch optimization algorithms for solving large-scale statistical learning applications; particularly those that can be formulated as a nonlinear programming problem. We rst investigate smooth, unconstrained problems, with applications in speech recognition. To reduce the computational cost of the optimization process, we will introduce two effective strategies, being: stochastic Hessian information and dynamic gradient sampling. We will then focus on developing second order methods for non-smooth optimization problems, specically in devising a semi-smooth Newton framework that can be used to generate several popular methods for machine learning.

Bio Sketch: 

Gillian Chin is a Ph.D. Candidate at Northwestern University in the Department of Industrial Engineering and Management Sciences in Evanston, Illinois. She received her BASc. in Industrial Engineering from the University of Toronto, Canada in 2008, with Honours. Her Ph.D. adviser is Dr. Jorge Nocedal, and the focus of her dissertation is developing efficient optimization algorithms for large{scale machine learning and nonlinear programming, with additional emphasis on sparse problems. She has held internships at Google Research and Argonne National Laboratory, and has been supported by fellowships from the Natural Sciences and Engineering Research Council of Canada (NSERC), and the Northwestern Cabell Fellowship.