TY - CHAP

T1 - Exact and approximate inference for annotating graphs with structural SVMs

AU - Klein, Thoralf

AU - Brefeld, Ulf

AU - Scheffer, Tobias

PY - 2008

Y1 - 2008

N2 - Training processes of structured prediction models such as structural SVMs involve frequent computations of the maximum-a-posteriori (MAP) prediction given a parameterized model. For specific output structures such as sequences or trees, MAP estimates can be computed efficiently by dynamic programming algorithms such as the Viterbi algorithm and the CKY parser. However, when the output structures can be arbitrary graphs, exact calculation of the MAP estimate is an NP-complete problem. In this paper, we compare exact inference and approximate inference for labeling graphs. We study the exact junction tree and the approximate loopy belief propagation and sampling algorithms in terms of performance and ressource requirements.

AB - Training processes of structured prediction models such as structural SVMs involve frequent computations of the maximum-a-posteriori (MAP) prediction given a parameterized model. For specific output structures such as sequences or trees, MAP estimates can be computed efficiently by dynamic programming algorithms such as the Viterbi algorithm and the CKY parser. However, when the output structures can be arbitrary graphs, exact calculation of the MAP estimate is an NP-complete problem. In this paper, we compare exact inference and approximate inference for labeling graphs. We study the exact junction tree and the approximate loopy belief propagation and sampling algorithms in terms of performance and ressource requirements.

KW - Informatics

KW - Gibbs Sampling

KW - Graph Size

KW - Junction Tree

KW - Approximate Inference

KW - Exact Inference

KW - Business informatics

UR - http://www.scopus.com/inward/record.url?scp=56049121271&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/d2368264-b85e-3034-944e-a245641cbc08/

U2 - 10.1007/978-3-540-87479-9_58

DO - 10.1007/978-3-540-87479-9_58

M3 - Article in conference proceedings

AN - SCOPUS:56049121271

SN - 978-3-540-87478-2

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 611

EP - 623

BT - Machine Learning and Knowledge Discovery in Databases

A2 - Daelemans, Walter

A2 - Goethals, Bart

A2 - Morik, Katharina

PB - Springer Verlag

CY - Berlin, Heidelberg

T2 - European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases - 2008

Y2 - 15 September 2008 through 19 September 2008

ER -