Worst configurations (Instantons) for compressed sensing over reals: A channel coding approach

Shashi Kiran Chilappagari, Michael Chertkov, Bane V Vasic

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

We consider the Linear Programming (LP) solution of the Compressed Sensing (CS) problem over reals, also known as the Basis Pursuit (BasP) algorithm. The BasP allows interpretation as a channel-coding problem, and it guarantees error-free reconstruction with a properly chosen measurement matrix and sufficiently sparse error vectors. In this manuscript, we examine how the BasP performs on a given measurement matrix and develop an algorithm to discover the sparsest vectors for which the BasP fails. The resulting algorithm is a generalization of our previous results on finding the most probable error-patterns degrading performance of a finite size Low-Density Parity-Check (LDPC) code in the error-floor regime. The BasP fails when its output is different from the actual error-pattern. We design a CS-Instanton Search Algorithm (ISA) generating a sparse vector, called a CS-instanton, such that the BasP fails on the CS-instanton, while the BasP recovery is successful for any modification of the CS-instanton replacing a nonzero element by zero. We also prove that, given a sufficiently dense random input for the error-vector, the CS-ISA converges to an instanton in a small finite number of steps. The performance of the CS-ISA is illustrated on a randomly generated 120 × 512 matrix. For this example, the CS-ISA outputs the shortest instanton (error vector) pattern of length 11.

Original languageEnglish (US)
Title of host publicationIEEE International Symposium on Information Theory - Proceedings
Pages1978-1982
Number of pages5
DOIs
StatePublished - 2010
Event2010 IEEE International Symposium on Information Theory, ISIT 2010 - Austin, TX, United States
Duration: Jun 13 2010Jun 18 2010

Other

Other2010 IEEE International Symposium on Information Theory, ISIT 2010
CountryUnited States
CityAustin, TX
Period6/13/106/18/10

Fingerprint

Compressed sensing
Channel Coding
Compressed Sensing
Channel coding
Basis Pursuit
Instantons
Configuration
Search Algorithm
Low-density Parity-check (LDPC) Codes
Output
Linear programming
Probable
Recovery
Converge

ASJC Scopus subject areas

  • Applied Mathematics
  • Modeling and Simulation
  • Theoretical Computer Science
  • Information Systems

Cite this

Chilappagari, S. K., Chertkov, M., & Vasic, B. V. (2010). Worst configurations (Instantons) for compressed sensing over reals: A channel coding approach. In IEEE International Symposium on Information Theory - Proceedings (pp. 1978-1982). [5513360] https://doi.org/10.1109/ISIT.2010.5513360

Worst configurations (Instantons) for compressed sensing over reals : A channel coding approach. / Chilappagari, Shashi Kiran; Chertkov, Michael; Vasic, Bane V.

IEEE International Symposium on Information Theory - Proceedings. 2010. p. 1978-1982 5513360.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Chilappagari, SK, Chertkov, M & Vasic, BV 2010, Worst configurations (Instantons) for compressed sensing over reals: A channel coding approach. in IEEE International Symposium on Information Theory - Proceedings., 5513360, pp. 1978-1982, 2010 IEEE International Symposium on Information Theory, ISIT 2010, Austin, TX, United States, 6/13/10. https://doi.org/10.1109/ISIT.2010.5513360
Chilappagari SK, Chertkov M, Vasic BV. Worst configurations (Instantons) for compressed sensing over reals: A channel coding approach. In IEEE International Symposium on Information Theory - Proceedings. 2010. p. 1978-1982. 5513360 https://doi.org/10.1109/ISIT.2010.5513360
Chilappagari, Shashi Kiran ; Chertkov, Michael ; Vasic, Bane V. / Worst configurations (Instantons) for compressed sensing over reals : A channel coding approach. IEEE International Symposium on Information Theory - Proceedings. 2010. pp. 1978-1982
@inproceedings{f07e7b9d6e274695aed54fc8427cb146,
title = "Worst configurations (Instantons) for compressed sensing over reals: A channel coding approach",
abstract = "We consider the Linear Programming (LP) solution of the Compressed Sensing (CS) problem over reals, also known as the Basis Pursuit (BasP) algorithm. The BasP allows interpretation as a channel-coding problem, and it guarantees error-free reconstruction with a properly chosen measurement matrix and sufficiently sparse error vectors. In this manuscript, we examine how the BasP performs on a given measurement matrix and develop an algorithm to discover the sparsest vectors for which the BasP fails. The resulting algorithm is a generalization of our previous results on finding the most probable error-patterns degrading performance of a finite size Low-Density Parity-Check (LDPC) code in the error-floor regime. The BasP fails when its output is different from the actual error-pattern. We design a CS-Instanton Search Algorithm (ISA) generating a sparse vector, called a CS-instanton, such that the BasP fails on the CS-instanton, while the BasP recovery is successful for any modification of the CS-instanton replacing a nonzero element by zero. We also prove that, given a sufficiently dense random input for the error-vector, the CS-ISA converges to an instanton in a small finite number of steps. The performance of the CS-ISA is illustrated on a randomly generated 120 × 512 matrix. For this example, the CS-ISA outputs the shortest instanton (error vector) pattern of length 11.",
author = "Chilappagari, {Shashi Kiran} and Michael Chertkov and Vasic, {Bane V}",
year = "2010",
doi = "10.1109/ISIT.2010.5513360",
language = "English (US)",
isbn = "9781424469604",
pages = "1978--1982",
booktitle = "IEEE International Symposium on Information Theory - Proceedings",

}

TY - GEN

T1 - Worst configurations (Instantons) for compressed sensing over reals

T2 - A channel coding approach

AU - Chilappagari, Shashi Kiran

AU - Chertkov, Michael

AU - Vasic, Bane V

PY - 2010

Y1 - 2010

N2 - We consider the Linear Programming (LP) solution of the Compressed Sensing (CS) problem over reals, also known as the Basis Pursuit (BasP) algorithm. The BasP allows interpretation as a channel-coding problem, and it guarantees error-free reconstruction with a properly chosen measurement matrix and sufficiently sparse error vectors. In this manuscript, we examine how the BasP performs on a given measurement matrix and develop an algorithm to discover the sparsest vectors for which the BasP fails. The resulting algorithm is a generalization of our previous results on finding the most probable error-patterns degrading performance of a finite size Low-Density Parity-Check (LDPC) code in the error-floor regime. The BasP fails when its output is different from the actual error-pattern. We design a CS-Instanton Search Algorithm (ISA) generating a sparse vector, called a CS-instanton, such that the BasP fails on the CS-instanton, while the BasP recovery is successful for any modification of the CS-instanton replacing a nonzero element by zero. We also prove that, given a sufficiently dense random input for the error-vector, the CS-ISA converges to an instanton in a small finite number of steps. The performance of the CS-ISA is illustrated on a randomly generated 120 × 512 matrix. For this example, the CS-ISA outputs the shortest instanton (error vector) pattern of length 11.

AB - We consider the Linear Programming (LP) solution of the Compressed Sensing (CS) problem over reals, also known as the Basis Pursuit (BasP) algorithm. The BasP allows interpretation as a channel-coding problem, and it guarantees error-free reconstruction with a properly chosen measurement matrix and sufficiently sparse error vectors. In this manuscript, we examine how the BasP performs on a given measurement matrix and develop an algorithm to discover the sparsest vectors for which the BasP fails. The resulting algorithm is a generalization of our previous results on finding the most probable error-patterns degrading performance of a finite size Low-Density Parity-Check (LDPC) code in the error-floor regime. The BasP fails when its output is different from the actual error-pattern. We design a CS-Instanton Search Algorithm (ISA) generating a sparse vector, called a CS-instanton, such that the BasP fails on the CS-instanton, while the BasP recovery is successful for any modification of the CS-instanton replacing a nonzero element by zero. We also prove that, given a sufficiently dense random input for the error-vector, the CS-ISA converges to an instanton in a small finite number of steps. The performance of the CS-ISA is illustrated on a randomly generated 120 × 512 matrix. For this example, the CS-ISA outputs the shortest instanton (error vector) pattern of length 11.

UR - http://www.scopus.com/inward/record.url?scp=77955699057&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77955699057&partnerID=8YFLogxK

U2 - 10.1109/ISIT.2010.5513360

DO - 10.1109/ISIT.2010.5513360

M3 - Conference contribution

AN - SCOPUS:77955699057

SN - 9781424469604

SP - 1978

EP - 1982

BT - IEEE International Symposium on Information Theory - Proceedings

ER -