Using structure to solve underdetermined systems of linear equations and overdetermined systems of quadratic equations

Event Details
Contact

Holly Rush (404) 385-1043

Summaries

Summary Sentence: Big Data Chalk & Talk / Brown Bag

Full Summary: Abstract:We will start by giving a high-level overview of the fundamental results in the field that has come to be known as compressive sensing.  The central theme of this body of work is that underdetermined systems of linear equations can be meaningfully "inverted" if they have structured solutions.  Two examples of structure would be if the unknown entity is a vector which is sparse (has only a few "active" entries) or if it is a matrix which is low rank.  We discuss some of the applications of this theory in signal processing and machine learning.  In the second part of the talk, we will show how some of these structured recovery results give us new insights into solving systems of quadratic and bilinear equations.  In particular, we will show how recasting classical problems like channel separation and blind deconvolution as a structured matrix factorization gives us new insight into how to solve them.

Speaker:

Dr. Justin Romberg

Associate Professor, School of Electrical and Computer Engineering,Georgia Institute of Technology

Title:

Using structure to solve underdetermined systems of linea

Abstract:

We will start by giving a high-level overview of the fundamental results in the field that has come to be known as compressive sensing.  The central theme of this body of work is that underdetermined systems of linear equations can be meaningfully "inverted" if they have structured solutions.  Two examples of structure would be if the unknown entity is a vector which is sparse (has only a few "active" entries) or if it is a matrix which is low rank.  We discuss some of the applications of this theory in signal processing and machine learning. 

In the second part of the talk, we will show how some of these structured recovery results give us new insights into solving systems of quadratic and bilinear equations.  In particular, we will show how recasting classical problems like channel separation and blind deconvolution as a structured matrix factorization gives us new insight into how to solve them.

Bio:

Dr. Justin Romberg is an Associate Professor in the School of Electrical and Computer Engineering at the Georgia Institute of Technology.  Dr. Romberg received the B.S.E.E. (1997), M.S. (1999) and Ph.D. (2004) degrees from Rice University in Houston, Texas.  From Fall 2003 until Fall 2006, he was a Postdoctoral Scholar in Applied and Computational Mathematics at the California Institute of Technology.  In the Fall of 2006, he joined the Georgia Tech ECE faculty.  In 2009 he received a PECASE award and a Packard Fellowship, and in 2010 he was named a Rice University Outstanding Young Engineering Alumnus.  He is currently on the editorial board for the SIAM Journal on Imaging Science.

Additional Information

In Campus Calendar
No
Groups

High Performance Computing (HPC)

Invited Audience
Undergraduate students, Faculty/Staff, Graduate students
Categories
Seminar/Lecture/Colloquium
Keywords
big data
Status
  • Created By: Holly Rush
  • Workflow Status: Published
  • Created On: Nov 26, 2013 - 9:39am
  • Last Updated: Apr 13, 2017 - 5:23pm