Lossy computing of correlated sources with fractional sampling

Xi Liu, Osvaldo Simeone, Elza Erkip

Research output: Contribution to journalArticle

Abstract

This paper considers the problem of lossy compression for the computation of a function of two correlated sources, both of which are observed at the encoder. Due to presence of observation costs, the encoder is allowed to observe only subsets of the samples from both sources, with a fraction of such sample pairs possibly overlapping. The rate-distortion function is characterized for memoryless sources, and then specialized to Gaussian and binary sources for selected functions and with quadratic and Hamming distortion metrics, respectively. The optimal measurement overlap fraction is shown to depend on the function to be computed by the decoder, on the source statistics, including the correlation, and on the link rate. Special cases are discussed in which the optimal overlap fraction is the maximum or minimum possible value given the sampling budget, illustrating non-trivial performance trade-offs in the design of the sampling strategy. Finally, the analysis is extended to the multi-hop set-up with jointly Gaussian sources, where each encoder can observe only one of the sources.

Original languageEnglish (US)
Article number6573236
Pages (from-to)3685-3696
Number of pages12
JournalIEEE Transactions on Communications
Volume61
Issue number9
DOIs
StatePublished - 2013

Fingerprint

Sampling
Statistics
Costs

Keywords

  • fractional sampling
  • function computation
  • Multi-terminal rate-distortion theory

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Lossy computing of correlated sources with fractional sampling. / Liu, Xi; Simeone, Osvaldo; Erkip, Elza.

In: IEEE Transactions on Communications, Vol. 61, No. 9, 6573236, 2013, p. 3685-3696.

Research output: Contribution to journalArticle

Liu, Xi ; Simeone, Osvaldo ; Erkip, Elza. / Lossy computing of correlated sources with fractional sampling. In: IEEE Transactions on Communications. 2013 ; Vol. 61, No. 9. pp. 3685-3696.
@article{51d0f08da22649cba4ee8e9e37fa725c,
title = "Lossy computing of correlated sources with fractional sampling",
abstract = "This paper considers the problem of lossy compression for the computation of a function of two correlated sources, both of which are observed at the encoder. Due to presence of observation costs, the encoder is allowed to observe only subsets of the samples from both sources, with a fraction of such sample pairs possibly overlapping. The rate-distortion function is characterized for memoryless sources, and then specialized to Gaussian and binary sources for selected functions and with quadratic and Hamming distortion metrics, respectively. The optimal measurement overlap fraction is shown to depend on the function to be computed by the decoder, on the source statistics, including the correlation, and on the link rate. Special cases are discussed in which the optimal overlap fraction is the maximum or minimum possible value given the sampling budget, illustrating non-trivial performance trade-offs in the design of the sampling strategy. Finally, the analysis is extended to the multi-hop set-up with jointly Gaussian sources, where each encoder can observe only one of the sources.",
keywords = "fractional sampling, function computation, Multi-terminal rate-distortion theory",
author = "Xi Liu and Osvaldo Simeone and Elza Erkip",
year = "2013",
doi = "10.1109/TCOMM.2013.072913.120876",
language = "English (US)",
volume = "61",
pages = "3685--3696",
journal = "IEEE Transactions on Communications",
issn = "0090-6778",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "9",

}

TY - JOUR

T1 - Lossy computing of correlated sources with fractional sampling

AU - Liu, Xi

AU - Simeone, Osvaldo

AU - Erkip, Elza

PY - 2013

Y1 - 2013

N2 - This paper considers the problem of lossy compression for the computation of a function of two correlated sources, both of which are observed at the encoder. Due to presence of observation costs, the encoder is allowed to observe only subsets of the samples from both sources, with a fraction of such sample pairs possibly overlapping. The rate-distortion function is characterized for memoryless sources, and then specialized to Gaussian and binary sources for selected functions and with quadratic and Hamming distortion metrics, respectively. The optimal measurement overlap fraction is shown to depend on the function to be computed by the decoder, on the source statistics, including the correlation, and on the link rate. Special cases are discussed in which the optimal overlap fraction is the maximum or minimum possible value given the sampling budget, illustrating non-trivial performance trade-offs in the design of the sampling strategy. Finally, the analysis is extended to the multi-hop set-up with jointly Gaussian sources, where each encoder can observe only one of the sources.

AB - This paper considers the problem of lossy compression for the computation of a function of two correlated sources, both of which are observed at the encoder. Due to presence of observation costs, the encoder is allowed to observe only subsets of the samples from both sources, with a fraction of such sample pairs possibly overlapping. The rate-distortion function is characterized for memoryless sources, and then specialized to Gaussian and binary sources for selected functions and with quadratic and Hamming distortion metrics, respectively. The optimal measurement overlap fraction is shown to depend on the function to be computed by the decoder, on the source statistics, including the correlation, and on the link rate. Special cases are discussed in which the optimal overlap fraction is the maximum or minimum possible value given the sampling budget, illustrating non-trivial performance trade-offs in the design of the sampling strategy. Finally, the analysis is extended to the multi-hop set-up with jointly Gaussian sources, where each encoder can observe only one of the sources.

KW - fractional sampling

KW - function computation

KW - Multi-terminal rate-distortion theory

UR - http://www.scopus.com/inward/record.url?scp=84884900699&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84884900699&partnerID=8YFLogxK

U2 - 10.1109/TCOMM.2013.072913.120876

DO - 10.1109/TCOMM.2013.072913.120876

M3 - Article

VL - 61

SP - 3685

EP - 3696

JO - IEEE Transactions on Communications

JF - IEEE Transactions on Communications

SN - 0090-6778

IS - 9

M1 - 6573236

ER -