Monday, July 7

MS23
Mathematical Methods in Data Mining

4:00 PM - 6:00 PM
Room: Royal Palm 3

This minisymposium is on recent developments of some novel mathematical techniques for data mining applications. These developments highlight the multidisciplinary nature of data analysis --- the four talks in the minisymposium represent contributions from different communities, namely, Mathematics, Optimization, Statistics, Computer Science, as well as perspectives from industry. Our two main objectives are to publicize these methods and to provide a cross-fertilization platform for researchers from different communities.

Organizer: Inderjit S. Dhillon
University of Texas at Austin
Lek-Heng Lim
University of California, Berkeley

4:00-4:25 Variable Selection: Tractable Upper Bounds on the Restricted Isometry Constant

Alexandre d'Aspremont, Princeton University

4:30-4:55 Chasing \$1,000,000: How We Won the Netflix Progress Prize

Yehuda Koren, AT&T Labs - Research

5:00-5:25 Global Positioning from Local Distances

Amit Singer, Yale University

5:30-5:55 Combinatorial Hodge Theory and a Geometric Approach to Ranking

Yuan Yao, Stanford University; Lek-Heng Lim, University of California, Berkeley