Rate-distortion bound for joint compression and classification

Yanting Dong; L. Carin

doi:10.1109/DCC.2003.1194042

Rate-distortion bound for joint compression and classification

Yanting Dong, L. Carin

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Summary form only given. Rate-distortion theory is applied to the problem of joint compression and classification. A Lagrangian distortion measure is used to consider both the Euclidean error in reconstructing the original data as well as the classification performance. The bound is calculated based on an alternating-minimization procedure, representing an extension of the Blahut-Arimoto algorithm. A hidden Markov model (HMM) source was considered as an example application and the objective is to quantize the source outputs and estimate the underlying HMM state sequence. Bounds on the minimum rate are required was presented to achieve desired average distortion on signal reconstruction and state-estimation accuracy.

Original language	English (US)
Title of host publication	Data Compression Conference, 2003. Proceedings. DCC 2003
Publisher	IEEE
ISBN (Print)	0-7695-1896-6
DOIs	https://doi.org/10.1109/DCC.2003.1194042
State	Published - Mar 27 2003
Externally published	Yes
Event	Data Compression Conference, 2003. Proceedings. DCC 2003 - Snowbird, UT, USA Duration: Mar 25 2003 → Mar 27 2003

Conference

Conference	Data Compression Conference, 2003. Proceedings. DCC 2003
Period	03/25/03 → 03/27/03

Keywords

Rate-distortion
Hidden Markov models
Distortion measurement
Lagrangian functions
Computer errors
State estimation
Signal reconstruction
Data compression

Access to Document

10.1109/DCC.2003.1194042

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=1194042

Cite this

@inproceedings{1de747e865a44f4cb499f6a6378409d3,

title = "Rate-distortion bound for joint compression and classification",

abstract = "Summary form only given. Rate-distortion theory is applied to the problem of joint compression and classification. A Lagrangian distortion measure is used to consider both the Euclidean error in reconstructing the original data as well as the classification performance. The bound is calculated based on an alternating-minimization procedure, representing an extension of the Blahut-Arimoto algorithm. A hidden Markov model (HMM) source was considered as an example application and the objective is to quantize the source outputs and estimate the underlying HMM state sequence. Bounds on the minimum rate are required was presented to achieve desired average distortion on signal reconstruction and state-estimation accuracy.",

keywords = "Rate-distortion, Hidden Markov models, Distortion measurement, Lagrangian functions, Computer errors, State estimation, Signal reconstruction, Data compression",

author = "Yanting Dong and L. Carin",

year = "2003",

month = mar,

day = "27",

doi = "10.1109/DCC.2003.1194042",

language = "English (US)",

isbn = "0-7695-1896-6",

booktitle = "Data Compression Conference, 2003. Proceedings. DCC 2003",

publisher = "IEEE",

note = "Data Compression Conference, 2003. Proceedings. DCC 2003 ; Conference date: 25-03-2003 Through 27-03-2003",

}

TY - GEN

T1 - Rate-distortion bound for joint compression and classification

AU - Dong, Yanting

AU - Carin, L.

PY - 2003/3/27

Y1 - 2003/3/27

N2 - Summary form only given. Rate-distortion theory is applied to the problem of joint compression and classification. A Lagrangian distortion measure is used to consider both the Euclidean error in reconstructing the original data as well as the classification performance. The bound is calculated based on an alternating-minimization procedure, representing an extension of the Blahut-Arimoto algorithm. A hidden Markov model (HMM) source was considered as an example application and the objective is to quantize the source outputs and estimate the underlying HMM state sequence. Bounds on the minimum rate are required was presented to achieve desired average distortion on signal reconstruction and state-estimation accuracy.

AB - Summary form only given. Rate-distortion theory is applied to the problem of joint compression and classification. A Lagrangian distortion measure is used to consider both the Euclidean error in reconstructing the original data as well as the classification performance. The bound is calculated based on an alternating-minimization procedure, representing an extension of the Blahut-Arimoto algorithm. A hidden Markov model (HMM) source was considered as an example application and the objective is to quantize the source outputs and estimate the underlying HMM state sequence. Bounds on the minimum rate are required was presented to achieve desired average distortion on signal reconstruction and state-estimation accuracy.

KW - Rate-distortion

KW - Hidden Markov models

KW - Distortion measurement

KW - Lagrangian functions

KW - Computer errors

KW - State estimation

KW - Signal reconstruction

KW - Data compression

UR - https://ieeexplore.ieee.org/document/1194042/

U2 - 10.1109/DCC.2003.1194042

DO - 10.1109/DCC.2003.1194042

M3 - Conference contribution

SN - 0-7695-1896-6

BT - Data Compression Conference, 2003. Proceedings. DCC 2003

PB - IEEE

T2 - Data Compression Conference, 2003. Proceedings. DCC 2003

Y2 - 25 March 2003 through 27 March 2003

ER -

Rate-distortion bound for joint compression and classification

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this