Tuesday, October 27

MS30
Recent Progress in Rank Minimization

5:00 PM - 7:00 PM
Room: Laguna Grande D

Matrix rank minimization refers to the general problem of finding the matrix of lowest rank that lies in a specified convex set. The singular value decomposition is the most famous example of rank minimization, but many other variants of rank minimization are NP-hard. Nonetheless, there is considerable interest in variants of the problem because they arise in machine learning, system identification, and data mining. The speakers in this minisymposium will present recent results in rank minimization based on techniques from linear algebra and convex optimization.

Organizer: Stephen A. Vavasis
University of Waterloo, Canada

5:00-5:25 Nuclear Norm Minimization for the Maximum Clique and Biclique Problems

Brendan Ames and Stephen A. Vavasis, University of Waterloo, Canada

5:30-5:55 Explicit Sensor Network Localization using Semidefinite Programming and Clique Reductions

Nathan Krislock and Henry Wolkowicz, University of Waterloo, Canada

6:00-6:25 Fast and Near--Optimal Matrix Completion via Randomized Basis Pursuit

Zhisu Zhu, Stanford University; Anthony Man-Cho So, The Chinese University of Hong Kong, Hong Kong; Yinyu Ye, Stanford University

6:30-6:55 A Nullspace Analysis of the Nuclear Norm Heuristic for Rank Minimization

Maryam Fazel and Krishnamurthy Dvijotham, University of Washington