Skip to main content.
Advanced search >
<< Back to previous page Print

<< Wednesday, April 03, 2013 >>


Remind me

Tell a friend

Add to my Google calendar (bCal)

Download to my calendar

Bookmark and ShareShare


CUR Decompositions with Random Sampling: Scientific Computing and Matrix Computations Seminar

Seminar: Scientific Computing | April 3 | 12:10-1 p.m. | 380 Soda Hall


Ming Gu, UC Berkeley

Electrical Engineering and Computer Sciences (EECS)


Given a (Semi-) Positive Definite matrix A, we consider the problem of computing its low-rank CUR approximations by uniformly sampling its columns randomly. We show that for matrices with low coherence (a concept to be defined at the talk), random sampling produces rank-revealing CUR approximations that are within a constant factor from optimal.


odedsc@cs.berkeley.edu, 510-516-4321