UCLA Mathnet Login

Distinguished Lecture Series

Every year, the Distinguished Lecture Series (DLS) brings two to four eminent mathematicians to UCLA for a week or more to give a lecture series on their field, and to meet with faculty and graduate students.

The first lecture of each series is aimed at a general mathematical audience, and offers a rare opportunity to see the state of an area of mathematics from the perspective of one of its leaders.  The remaining lectures in the series are usually more advanced, concerning recent developments in the area.

Previous speakers of the DLS include: Peter Sarnak, Peter Schneider, Zhengan Weng, Etienne Ghys, Goro Shimura, Jean Bellissard, Andrei Suslin, Pierre Deligne, Michael Harris, Alexander Lubotzky, Shing-Tau Yau, Hillel Furstenberg, Robert R. Langlands, Clifford Taubes, Louis Nirenberg, Oded Schramm, Louis Nirenberg, I.M. Singer, Jesper Lutzen, L.H. Eliasson, Raoul Bott, Dennis Gaitsgory, Gilles Pisier, Gregg Zuckerman, Freydoon Shahidi, Alain Connes, Jöran Friberg, David Mumford, Sir Michael Atiyah, Jean-Michel Bismut, Jean-Pierre Serre, G. Tian, N. Sibony, C. Deninger, Peter Lax, and Nikolai Reshetikhin.

The DLS is currently supported by the Larry M. Wiener fund. 

Past Lectures

University of Strasbourg
Visit: 04/03/2013 to 04/20/2013
Rheinische Friedrich-Wilhelms-Universität Bonn
Visit: 05/07/2013 to 05/09/2013
Eötvös Loránd University
Visit: 05/28/2013 to 05/30/2013
Texas A&M
Visit: 10/22/2013 to 10/26/2013
IAS, Princeton
Visit: 10/30/2013 to 11/06/2013
Duke University / UC Berkeley
Visit: 05/19/2014 to 05/23/2014
Cambridge University
Visit: 10/04/2014 to 10/10/2014
Microsoft Research
Visit: 11/03/2014 to 11/06/2014
Columbia University
Visit: 02/17/2015 to 02/19/2015
Princeton University
Visit: 05/19/2015 to 05/21/2015
Harvard University
Visit: 04/25/2016 to 04/28/2016

Upcoming Lectures

Donald Goldfarb

Columbia University

Visit: 05/17/2016 to 05/19/2016


Optimization for Learning and Big Data

Abstract:  Many problems in both supervised and unsupervised machine learning (e.g., logistic regression, support vector machines, deep neural networks, robust principal component analysis, dictionary learning, latent variable models) and signal processing (e.g., face recognition and compressed sensing) are solved by optimization and related algorithms.  In today's age of big data, the size of these problems is often formidable.  E.g., in logistic regression the objective function may be expressed as the sum of  ~10^9  functions (one for each data point) involving  ~10^6  variables (features).  In this series of talks, we will review current optimization approaches for addressing this challenge from the following classes of methods:  first-order (and accelerated variants), stochastic gradient and second-order extensions, alternating direction methods for structured problems (including proximal and conditional gradient and multiplier methods), tensor decomposition, randomized methods for linear systems, and parallel and distributed variants.

Lecture 1:  Tuesday, May 17;  3:00 pm, MS 6627

Lecture 2:  Wednesday, May 18; 3:00 pm, MS 6627

Lecture 3:  Thursday, May 19; 3:00 pm, MS 6627