Advances in Learning Bayesian Networks of Bounded Treewidth

Siqi Nie; Denis D. Maua; Cassio P. de Campos; Qiang Ji

Advances in Learning Bayesian Networks of Bounded Treewidth

Siqi Nie, Denis D. Maua, Cassio P. de Campos, Qiang Ji

School of Electronics, Electrical Engineering and Computer Science

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

22 Citations (Scopus)

Abstract

This work presents novel algorithms for learning Bayesian networks of bounded treewidth. Both exact and approximate methods are developed. The exact method combines mixed integer linear programming formulations for structure learning and treewidth computation. The approximate method consists in sampling k-trees (maximal graphs of treewidth k), and subsequently selecting, exactly or approximately, the best structure whose moral graph is a subgraph of that k-tree. The approaches are empirically compared to each other and to state-of-the-art methods on a collection of public data sets with up to 100 variables.

Original language	English
Title of host publication	Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014
Editors	Z. Ghahramani, M. Welling, C. Cortes, N.D. Lawrence, K.Q. Weinberger
Place of Publication	New York
Publisher	Curran Associates, Inc.
Pages	2285-2293
Number of pages	9
Volume	3
Publication status	Published - Jan 2014
Event	28th Annual Conference on Neural Information Processing Systems 2014 - Montreal, Canada Duration: 08 Dec 2014 → 13 Dec 2014

Conference

Conference	28th Annual Conference on Neural Information Processing Systems 2014
Country/Territory	Canada
City	Montreal
Period	08/12/2014 → 13/12/2014

Bibliographical note

(top 4%, spotlight presentation, double-blind peer reviewed by >3 reviewers)

Access to Document

http://papers.nips.cc/paper/5530-advances-in-learning-bayesian-networks-of-bounded-treewidth.pdf

Cite this

Nie, S., Maua, D. D., de Campos, C. P., & Ji, Q. (2014). Advances in Learning Bayesian Networks of Bounded Treewidth. In Z. Ghahramani, M. Welling, C. Cortes, N. D. Lawrence, & K. Q. Weinberger (Eds.), Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014 (Vol. 3, pp. 2285-2293). Curran Associates, Inc. http://papers.nips.cc/paper/5530-advances-in-learning-bayesian-networks-of-bounded-treewidth.pdf

Nie, Siqi ; Maua, Denis D. ; de Campos, Cassio P. et al. / Advances in Learning Bayesian Networks of Bounded Treewidth. Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014. editor / Z. Ghahramani ; M. Welling ; C. Cortes ; N.D. Lawrence ; K.Q. Weinberger. Vol. 3 New York : Curran Associates, Inc. , 2014. pp. 2285-2293

@inproceedings{5bca935e05e94b2c99f7e951e02b279b,

title = "Advances in Learning Bayesian Networks of Bounded Treewidth",

abstract = "This work presents novel algorithms for learning Bayesian networks of bounded treewidth. Both exact and approximate methods are developed. The exact method combines mixed integer linear programming formulations for structure learning and treewidth computation. The approximate method consists in sampling k-trees (maximal graphs of treewidth k), and subsequently selecting, exactly or approximately, the best structure whose moral graph is a subgraph of that k-tree. The approaches are empirically compared to each other and to state-of-the-art methods on a collection of public data sets with up to 100 variables.",

author = "Siqi Nie and Maua, {Denis D.} and {de Campos}, {Cassio P.} and Qiang Ji",

note = "(top 4%, spotlight presentation, double-blind peer reviewed by >3 reviewers); 28th Annual Conference on Neural Information Processing Systems 2014 ; Conference date: 08-12-2014 Through 13-12-2014",

year = "2014",

month = jan,

language = "English",

volume = "3",

pages = "2285--2293",

editor = "Z. Ghahramani and M. Welling and C. Cortes and N.D. Lawrence and K.Q. Weinberger",

booktitle = "Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014",

publisher = "Curran Associates, Inc. ",

}

Nie, S, Maua, DD, de Campos, CP & Ji, Q 2014, Advances in Learning Bayesian Networks of Bounded Treewidth. in Z Ghahramani, M Welling, C Cortes, ND Lawrence & KQ Weinberger (eds), Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014. vol. 3, Curran Associates, Inc. , New York, pp. 2285-2293, 28th Annual Conference on Neural Information Processing Systems 2014, Montreal, Canada, 08/12/2014. <http://papers.nips.cc/paper/5530-advances-in-learning-bayesian-networks-of-bounded-treewidth.pdf>

Advances in Learning Bayesian Networks of Bounded Treewidth. / Nie, Siqi; Maua, Denis D.; de Campos, Cassio P. et al.
Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014. ed. / Z. Ghahramani; M. Welling; C. Cortes; N.D. Lawrence; K.Q. Weinberger. Vol. 3 New York: Curran Associates, Inc. , 2014. p. 2285-2293.

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

TY - GEN

T1 - Advances in Learning Bayesian Networks of Bounded Treewidth

AU - Nie, Siqi

AU - Maua, Denis D.

AU - de Campos, Cassio P.

AU - Ji, Qiang

N1 - (top 4%, spotlight presentation, double-blind peer reviewed by >3 reviewers)

PY - 2014/1

Y1 - 2014/1

N2 - This work presents novel algorithms for learning Bayesian networks of bounded treewidth. Both exact and approximate methods are developed. The exact method combines mixed integer linear programming formulations for structure learning and treewidth computation. The approximate method consists in sampling k-trees (maximal graphs of treewidth k), and subsequently selecting, exactly or approximately, the best structure whose moral graph is a subgraph of that k-tree. The approaches are empirically compared to each other and to state-of-the-art methods on a collection of public data sets with up to 100 variables.

AB - This work presents novel algorithms for learning Bayesian networks of bounded treewidth. Both exact and approximate methods are developed. The exact method combines mixed integer linear programming formulations for structure learning and treewidth computation. The approximate method consists in sampling k-trees (maximal graphs of treewidth k), and subsequently selecting, exactly or approximately, the best structure whose moral graph is a subgraph of that k-tree. The approaches are empirically compared to each other and to state-of-the-art methods on a collection of public data sets with up to 100 variables.

M3 - Conference contribution

VL - 3

SP - 2285

EP - 2293

BT - Advances in Neural Information Processing Systems 27: 28th Annual Conference on Neural Information Processing Systems 2014

A2 - Ghahramani, Z.

A2 - Welling, M.

A2 - Cortes, C.

A2 - Lawrence, N.D.

A2 - Weinberger, K.Q.

PB - Curran Associates, Inc.

CY - New York

T2 - 28th Annual Conference on Neural Information Processing Systems 2014

Y2 - 8 December 2014 through 13 December 2014

ER -

Advances in Learning Bayesian Networks of Bounded Treewidth

Abstract

Conference

Bibliographical note

Access to Document

Fingerprint

Cite this