Revisiting Boltzmann learning: parameter estimation in Markov random fields

Lars Kai Hansen; Lars Nonboe Andersen; Ulrik Kjems; Jan Larsen

doi:10.1109/ICASSP.1996.550606

Revisiting Boltzmann learning: parameter estimation in Markov random fields

Lars Kai Hansen
, Lars Nonboe Andersen
, Ulrik Kjems
, Jan Larsen

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

518 Downloads (Orbit)

Abstract

This article presents a generalization of the Boltzmann machine that allows us to use the learning rule for a much wider class of maximum likelihood and maximum a posteriori problems, including both supervised and unsupervised learning. Furthermore, the approach allows us to discuss regularization and generalization in the context of Boltzmann machines. We provide an illustrative example concerning parameter estimation in an inhomogeneous Markov field. The regularized adaptation produces a parameter set that closely resembles the “teacher” parameters, hence, will produce segmentations that closely reproduce those of the inhomogeneous teacher network

Original language	English
Title of host publication	Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing
Volume	Volume 6
Publisher	IEEE
Publication date	1996
Pages	3394-3397
ISBN (Print)	07-80-33192-3
DOIs	https://doi.org/10.1109/ICASSP.1996.550606
Publication status	Published - 1996
Event	1996 IEEE International Conference on Acoustics, Speech and Signal Processing - Atlanta, United States Duration: 7 May 1996 → 10 May 1996 Conference number: 21 http://www.eng.auburn.edu/~sjreeves/ICASSP/

Conference

Conference	1996 IEEE International Conference on Acoustics, Speech and Signal Processing
Number	21
Country/Territory	United States
City	Atlanta
Period	07/05/1996 → 10/05/1996
Internet address	http://www.eng.auburn.edu/~sjreeves/ICASSP/

Bibliographical note

Copyright 1996 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Access to Document

10.1109/ICASSP.1996.550606

KjemsFinal published version, 453 KB

OpenUrl availability

Check availability in DTU Findit

Cite this

@inproceedings{569397c751514fb6b979c5bcab8280a9,

title = "Revisiting Boltzmann learning: parameter estimation in Markov random fields",

abstract = "This article presents a generalization of the Boltzmann machine that allows us to use the learning rule for a much wider class of maximum likelihood and maximum a posteriori problems, including both supervised and unsupervised learning. Furthermore, the approach allows us to discuss regularization and generalization in the context of Boltzmann machines. We provide an illustrative example concerning parameter estimation in an inhomogeneous Markov field. The regularized adaptation produces a parameter set that closely resembles the “teacher” parameters, hence, will produce segmentations that closely reproduce those of the inhomogeneous teacher network",

author = "Hansen, \{Lars Kai\} and Andersen, \{Lars Nonboe\} and Ulrik Kjems and Jan Larsen",

note = "Copyright 1996 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.; 1996 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 1996 ; Conference date: 07-05-1996 Through 10-05-1996",

year = "1996",

doi = "10.1109/ICASSP.1996.550606",

language = "English",

isbn = "07-80-33192-3",

volume = "Volume 6",

pages = "3394--3397",

booktitle = "Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing",

publisher = "IEEE",

address = "United States",

url = "http://www.eng.auburn.edu/\textasciitilde{}sjreeves/ICASSP/",

}

Hansen, LK, Andersen, LN, Kjems, U & Larsen, J 1996, Revisiting Boltzmann learning: parameter estimation in Markov random fields. in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing. vol. Volume 6, IEEE, pp. 3394-3397, 1996 IEEE International Conference on Acoustics, Speech and Signal Processing, Atlanta, Georgia, United States, 07/05/1996. https://doi.org/10.1109/ICASSP.1996.550606

Revisiting Boltzmann learning: parameter estimation in Markov random fields. / Hansen, Lars Kai; Andersen, Lars Nonboe; Kjems, Ulrik et al.
Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing. Vol. Volume 6 IEEE, 1996. p. 3394-3397.

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

TY - GEN

T1 - Revisiting Boltzmann learning: parameter estimation in Markov random fields

AU - Hansen, Lars Kai

AU - Andersen, Lars Nonboe

AU - Kjems, Ulrik

AU - Larsen, Jan

N1 - Conference code: 21

PY - 1996

Y1 - 1996

N2 - This article presents a generalization of the Boltzmann machine that allows us to use the learning rule for a much wider class of maximum likelihood and maximum a posteriori problems, including both supervised and unsupervised learning. Furthermore, the approach allows us to discuss regularization and generalization in the context of Boltzmann machines. We provide an illustrative example concerning parameter estimation in an inhomogeneous Markov field. The regularized adaptation produces a parameter set that closely resembles the “teacher” parameters, hence, will produce segmentations that closely reproduce those of the inhomogeneous teacher network

AB - This article presents a generalization of the Boltzmann machine that allows us to use the learning rule for a much wider class of maximum likelihood and maximum a posteriori problems, including both supervised and unsupervised learning. Furthermore, the approach allows us to discuss regularization and generalization in the context of Boltzmann machines. We provide an illustrative example concerning parameter estimation in an inhomogeneous Markov field. The regularized adaptation produces a parameter set that closely resembles the “teacher” parameters, hence, will produce segmentations that closely reproduce those of the inhomogeneous teacher network

U2 - 10.1109/ICASSP.1996.550606

DO - 10.1109/ICASSP.1996.550606

M3 - Article in proceedings

SN - 07-80-33192-3

VL - Volume 6

SP - 3394

EP - 3397

BT - Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing

PB - IEEE

T2 - 1996 IEEE International Conference on Acoustics, Speech and Signal Processing

Y2 - 7 May 1996 through 10 May 1996

ER -

Revisiting Boltzmann learning: parameter estimation in Markov random fields

Abstract

Conference

Bibliographical note

Access to Document

OpenUrl availability

Fingerprint

Cite this