Approximate Inference in Probabilistic Models

Manfred Opper; Ole Winther

Approximate Inference in Probabilistic Models

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

Abstract

We present a framework for approximate inference in probabilistic data models which is based on free energies. The free energy is constructed from two approximating distributions which encode different aspects of the intractable model. Consistency between distributions is required on a chosen set of moments. We find good performance using sets of moments which either specify factorized nodes or a spanning tree on the nodes. The abstract should summarize the contents of the paper using at least 70 and at most 150 words. It will be set in 9-point font size and be inset 1.0 cm from the right and left margins. There will be two blank lines before and after the Abstract.

Original language	English
Title of host publication	Proceedings of the Conference on Algorithmic Learning Theory
Volume	3244
Publisher	Springer
Publication date	2004
Pages	494-504
ISBN (Print)	3-540-23356-3
Publication status	Published - 2004
Event	Annual International Conference on Algorithmic Learning Theory - Padova, Italy Duration: 2 Oct 2004 → 5 Oct 2004 Conference number: 15

Conference

Conference	Annual International Conference on Algorithmic Learning Theory
Number	15
Country/Territory	Italy
City	Padova
Period	02/10/2004 → 05/10/2004

Series	Lecture Notes in Computer Science
ISSN	0302-9743

Access to Document

http://www.springerlink.com/app/home/contribution.asp?wasp=4h83eykqtlckpl98rlrl&referrer=parent&backto=issue,37,37;journal,80,1851;linkingpublicationresults,1:105633,1

SFX availability

Full text

Cite this

Opper, M., & Winther, O. (2004). Approximate Inference in Probabilistic Models. In Proceedings of the Conference on Algorithmic Learning Theory (Vol. 3244, pp. 494-504). Springer. Lecture Notes in Computer Science http://www.springerlink.com/app/home/contribution.asp?wasp=4h83eykqtlckpl98rlrl&referrer=parent&backto=issue,37,37;journal,80,1851;linkingpublicationresults,1:105633,1

@inproceedings{3ef20241b688418198fbc27ce642b704,

title = "Approximate Inference in Probabilistic Models",

abstract = "We present a framework for approximate inference in probabilistic data models which is based on free energies. The free energy is constructed from two approximating distributions which encode different aspects of the intractable model. Consistency between distributions is required on a chosen set of moments. We find good performance using sets of moments which either specify factorized nodes or a spanning tree on the nodes. The abstract should summarize the contents of the paper using at least 70 and at most 150 words. It will be set in 9-point font size and be inset 1.0 cm from the right and left margins. There will be two blank lines before and after the Abstract.",

author = "Manfred Opper and Ole Winther",

year = "2004",

language = "English",

isbn = "3-540-23356-3",

volume = "3244",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "494--504",

booktitle = "Proceedings of the Conference on Algorithmic Learning Theory",

note = "Annual International Conference on Algorithmic Learning Theory, ALT 2004 ; Conference date: 02-10-2004 Through 05-10-2004",

}

Opper, M & Winther, O 2004, Approximate Inference in Probabilistic Models. in Proceedings of the Conference on Algorithmic Learning Theory. vol. 3244, Springer, Lecture Notes in Computer Science, pp. 494-504, Annual International Conference on Algorithmic Learning Theory, Padova, Italy, 02/10/2004. <http://www.springerlink.com/app/home/contribution.asp?wasp=4h83eykqtlckpl98rlrl&referrer=parent&backto=issue,37,37;journal,80,1851;linkingpublicationresults,1:105633,1>

TY - GEN

T1 - Approximate Inference in Probabilistic Models

AU - Opper, Manfred

AU - Winther, Ole

N1 - Conference code: 15

PY - 2004

Y1 - 2004

N2 - We present a framework for approximate inference in probabilistic data models which is based on free energies. The free energy is constructed from two approximating distributions which encode different aspects of the intractable model. Consistency between distributions is required on a chosen set of moments. We find good performance using sets of moments which either specify factorized nodes or a spanning tree on the nodes. The abstract should summarize the contents of the paper using at least 70 and at most 150 words. It will be set in 9-point font size and be inset 1.0 cm from the right and left margins. There will be two blank lines before and after the Abstract.

AB - We present a framework for approximate inference in probabilistic data models which is based on free energies. The free energy is constructed from two approximating distributions which encode different aspects of the intractable model. Consistency between distributions is required on a chosen set of moments. We find good performance using sets of moments which either specify factorized nodes or a spanning tree on the nodes. The abstract should summarize the contents of the paper using at least 70 and at most 150 words. It will be set in 9-point font size and be inset 1.0 cm from the right and left margins. There will be two blank lines before and after the Abstract.

M3 - Article in proceedings

SN - 3-540-23356-3

VL - 3244

T3 - Lecture Notes in Computer Science

SP - 494

EP - 504

BT - Proceedings of the Conference on Algorithmic Learning Theory

PB - Springer

T2 - Annual International Conference on Algorithmic Learning Theory

Y2 - 2 October 2004 through 5 October 2004

ER -

Approximate Inference in Probabilistic Models

Abstract

Conference

Access to Document

SFX availability

Fingerprint

Cite this