lp Subspace Embeddings and Regression

PhD Qualifying Examination


Title: "lp Subspace Embeddings and Regression"

by

Mr. Di CHEN


Abstract:

Subspace embeddings map a d-dimensional subspace of the n-dimensional real 
space, under an lp norm, to a much lower dimensional real number space, 
while preserving norms. They are useful for dimension reduction in 
numerical linear algebra. Important applications include linear 
regression, singular value decomposition, and low-rank approximation of 
matrices. In this survey we examine the concepts and techniques that have 
led to recent advances on computing subspace embeddings efficiently, as an 
attempt to organize insights that may be useful for further exploration.


Date:			Tuesday, 19 August 2014

Time:                  	2:00pm - 4:00pm

Venue:                  Room 4472
                         Lifts 25/26

Committee Members:	Prof. Mordecai Golin (Supervisor)
  			Dr. Ke Yi (Supervisor)
 			Prof. Siu-Wing Cheng (Chairperson)
  			Dr. Sunil Arya


**** ALL are Welcome ****