Efficient convolution kernels for dependency and constituent syntactic trees

Research output: Chapter in Book/Report/Conference proceedingConference contribution

190 Citations (Scopus)

Abstract

In this paper, we provide a study on the use of tree kernels to encode syntactic parsing information in natural language learning. In particular, we propose a new convolution kernel, namely the Partial Tree (PT) kernel, to fully exploit dependency trees. We also propose an efficient algorithm for its computation which is futhermore sped-up by applying the selection of tree nodes with non-null kernel. The experiments with Support Vector Machines on the task of semantic role labeling and question classification show that (a) the kernel running time is linear on the average case and (b) the PT kernel improves on the other tree kernels when applied to the appropriate parsing paradigm.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages318-329
Number of pages12
Volume4212 LNAI
Publication statusPublished - 31 Oct 2006
Externally publishedYes
Event17th European Conference on Machine Learning, ECML 2006 - Berlin, Germany
Duration: 18 Sep 200622 Sep 2006

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4212 LNAI
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other17th European Conference on Machine Learning, ECML 2006
CountryGermany
CityBerlin
Period18/9/0622/9/06

Fingerprint

Syntactics
Convolution
Labeling
Support vector machines
Semantics
kernel
Experiments
Parsing
Partial
Language
Syntax
Dependency (Psychology)
Learning
Natural Language
Support Vector Machine
Speedup
Efficient Algorithms
Paradigm
Vertex of a graph
Experiment

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Moschitti, A. (2006). Efficient convolution kernels for dependency and constituent syntactic trees. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4212 LNAI, pp. 318-329). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4212 LNAI).

Efficient convolution kernels for dependency and constituent syntactic trees. / Moschitti, Alessandro.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4212 LNAI 2006. p. 318-329 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4212 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Moschitti, A 2006, Efficient convolution kernels for dependency and constituent syntactic trees. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 4212 LNAI, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4212 LNAI, pp. 318-329, 17th European Conference on Machine Learning, ECML 2006, Berlin, Germany, 18/9/06.
Moschitti A. Efficient convolution kernels for dependency and constituent syntactic trees. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4212 LNAI. 2006. p. 318-329. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Moschitti, Alessandro. / Efficient convolution kernels for dependency and constituent syntactic trees. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4212 LNAI 2006. pp. 318-329 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{8b519315a8004d9485c1f2ba52cbc1ca,
title = "Efficient convolution kernels for dependency and constituent syntactic trees",
abstract = "In this paper, we provide a study on the use of tree kernels to encode syntactic parsing information in natural language learning. In particular, we propose a new convolution kernel, namely the Partial Tree (PT) kernel, to fully exploit dependency trees. We also propose an efficient algorithm for its computation which is futhermore sped-up by applying the selection of tree nodes with non-null kernel. The experiments with Support Vector Machines on the task of semantic role labeling and question classification show that (a) the kernel running time is linear on the average case and (b) the PT kernel improves on the other tree kernels when applied to the appropriate parsing paradigm.",
author = "Alessandro Moschitti",
year = "2006",
month = "10",
day = "31",
language = "English",
isbn = "354045375X",
volume = "4212 LNAI",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "318--329",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Efficient convolution kernels for dependency and constituent syntactic trees

AU - Moschitti, Alessandro

PY - 2006/10/31

Y1 - 2006/10/31

N2 - In this paper, we provide a study on the use of tree kernels to encode syntactic parsing information in natural language learning. In particular, we propose a new convolution kernel, namely the Partial Tree (PT) kernel, to fully exploit dependency trees. We also propose an efficient algorithm for its computation which is futhermore sped-up by applying the selection of tree nodes with non-null kernel. The experiments with Support Vector Machines on the task of semantic role labeling and question classification show that (a) the kernel running time is linear on the average case and (b) the PT kernel improves on the other tree kernels when applied to the appropriate parsing paradigm.

AB - In this paper, we provide a study on the use of tree kernels to encode syntactic parsing information in natural language learning. In particular, we propose a new convolution kernel, namely the Partial Tree (PT) kernel, to fully exploit dependency trees. We also propose an efficient algorithm for its computation which is futhermore sped-up by applying the selection of tree nodes with non-null kernel. The experiments with Support Vector Machines on the task of semantic role labeling and question classification show that (a) the kernel running time is linear on the average case and (b) the PT kernel improves on the other tree kernels when applied to the appropriate parsing paradigm.

UR - http://www.scopus.com/inward/record.url?scp=33750333944&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33750333944&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:33750333944

SN - 354045375X

SN - 9783540453758

VL - 4212 LNAI

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

SP - 318

EP - 329

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

ER -