Assessing the impact of syntactic and semantic structures for answer passages reranking

Kateryna Tymoshenko, Alessandro Moschitti

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

31 Citations (Scopus)

Abstract

In this paper, we extensively study the use of syntactic and semantic structures obtained with shallow and deeper syntactic parsers in the answer passage reranking task. We propose several dependency-based structures enriched with Linked Open Data (LD) knowledge for representing pairs of questions and answer passages. We use such tree structures in learning to rank (L2R) algorithms based on tree kernel. The latter can represent questions and passages in a tree fragment space, where each substructure represents a powerful syntactic/semantic feature. Additionally since we define links between structures, tree kernels also generate relational features spanning question and passage structures. We derive very important findings, which can be useful to build state-of-the-art systems: (i) full syntactic dependencies can outperform shallow models also using external knowledge and (ii) the semantic information should be derived by effective and high-coverage resources, e.g., LD, and incorporated in syntactic structures to be effective. We demonstrate our findings by carrying out an extensive comparative experimentation on two different TREC QA corpora and one community question answer dataset, namely Answerbag. Our comparative analysis on well-defined answer selection benchmarks consistently demonstrates that our structural semantic models largely outperform the state of the art in passage reranking.

Original languageEnglish
Title of host publicationInternational Conference on Information and Knowledge Management, Proceedings
PublisherAssociation for Computing Machinery
Pages1451-1460
Number of pages10
Volume19-23-Oct-2015
ISBN (Print)9781450337946
DOIs
Publication statusPublished - 17 Oct 2015
Event24th ACM International Conference on Information and Knowledge Management, CIKM 2015 - Melbourne, Australia
Duration: 19 Oct 201523 Oct 2015

Other

Other24th ACM International Conference on Information and Knowledge Management, CIKM 2015
CountryAustralia
CityMelbourne
Period19/10/1523/10/15

Fingerprint

Reranking
Kernel
Experimentation
Resources
Benchmark
Comparative analysis
Learning to rank

Keywords

  • Kernel methods
  • Learning to rank
  • Linked data
  • Question answering
  • Structural kernels

ASJC Scopus subject areas

  • Business, Management and Accounting(all)
  • Decision Sciences(all)

Cite this

Tymoshenko, K., & Moschitti, A. (2015). Assessing the impact of syntactic and semantic structures for answer passages reranking. In International Conference on Information and Knowledge Management, Proceedings (Vol. 19-23-Oct-2015, pp. 1451-1460). Association for Computing Machinery. https://doi.org/10.1145/2806416.2806490

Assessing the impact of syntactic and semantic structures for answer passages reranking. / Tymoshenko, Kateryna; Moschitti, Alessandro.

International Conference on Information and Knowledge Management, Proceedings. Vol. 19-23-Oct-2015 Association for Computing Machinery, 2015. p. 1451-1460.

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

Tymoshenko, K & Moschitti, A 2015, Assessing the impact of syntactic and semantic structures for answer passages reranking. in International Conference on Information and Knowledge Management, Proceedings. vol. 19-23-Oct-2015, Association for Computing Machinery, pp. 1451-1460, 24th ACM International Conference on Information and Knowledge Management, CIKM 2015, Melbourne, Australia, 19/10/15. https://doi.org/10.1145/2806416.2806490
Tymoshenko K, Moschitti A. Assessing the impact of syntactic and semantic structures for answer passages reranking. In International Conference on Information and Knowledge Management, Proceedings. Vol. 19-23-Oct-2015. Association for Computing Machinery. 2015. p. 1451-1460 https://doi.org/10.1145/2806416.2806490
Tymoshenko, Kateryna ; Moschitti, Alessandro. / Assessing the impact of syntactic and semantic structures for answer passages reranking. International Conference on Information and Knowledge Management, Proceedings. Vol. 19-23-Oct-2015 Association for Computing Machinery, 2015. pp. 1451-1460
@inproceedings{ace4e91a42724076820af4a5f6455960,
title = "Assessing the impact of syntactic and semantic structures for answer passages reranking",
abstract = "In this paper, we extensively study the use of syntactic and semantic structures obtained with shallow and deeper syntactic parsers in the answer passage reranking task. We propose several dependency-based structures enriched with Linked Open Data (LD) knowledge for representing pairs of questions and answer passages. We use such tree structures in learning to rank (L2R) algorithms based on tree kernel. The latter can represent questions and passages in a tree fragment space, where each substructure represents a powerful syntactic/semantic feature. Additionally since we define links between structures, tree kernels also generate relational features spanning question and passage structures. We derive very important findings, which can be useful to build state-of-the-art systems: (i) full syntactic dependencies can outperform shallow models also using external knowledge and (ii) the semantic information should be derived by effective and high-coverage resources, e.g., LD, and incorporated in syntactic structures to be effective. We demonstrate our findings by carrying out an extensive comparative experimentation on two different TREC QA corpora and one community question answer dataset, namely Answerbag. Our comparative analysis on well-defined answer selection benchmarks consistently demonstrates that our structural semantic models largely outperform the state of the art in passage reranking.",
keywords = "Kernel methods, Learning to rank, Linked data, Question answering, Structural kernels",
author = "Kateryna Tymoshenko and Alessandro Moschitti",
year = "2015",
month = "10",
day = "17",
doi = "10.1145/2806416.2806490",
language = "English",
isbn = "9781450337946",
volume = "19-23-Oct-2015",
pages = "1451--1460",
booktitle = "International Conference on Information and Knowledge Management, Proceedings",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Assessing the impact of syntactic and semantic structures for answer passages reranking

AU - Tymoshenko, Kateryna

AU - Moschitti, Alessandro

PY - 2015/10/17

Y1 - 2015/10/17

N2 - In this paper, we extensively study the use of syntactic and semantic structures obtained with shallow and deeper syntactic parsers in the answer passage reranking task. We propose several dependency-based structures enriched with Linked Open Data (LD) knowledge for representing pairs of questions and answer passages. We use such tree structures in learning to rank (L2R) algorithms based on tree kernel. The latter can represent questions and passages in a tree fragment space, where each substructure represents a powerful syntactic/semantic feature. Additionally since we define links between structures, tree kernels also generate relational features spanning question and passage structures. We derive very important findings, which can be useful to build state-of-the-art systems: (i) full syntactic dependencies can outperform shallow models also using external knowledge and (ii) the semantic information should be derived by effective and high-coverage resources, e.g., LD, and incorporated in syntactic structures to be effective. We demonstrate our findings by carrying out an extensive comparative experimentation on two different TREC QA corpora and one community question answer dataset, namely Answerbag. Our comparative analysis on well-defined answer selection benchmarks consistently demonstrates that our structural semantic models largely outperform the state of the art in passage reranking.

AB - In this paper, we extensively study the use of syntactic and semantic structures obtained with shallow and deeper syntactic parsers in the answer passage reranking task. We propose several dependency-based structures enriched with Linked Open Data (LD) knowledge for representing pairs of questions and answer passages. We use such tree structures in learning to rank (L2R) algorithms based on tree kernel. The latter can represent questions and passages in a tree fragment space, where each substructure represents a powerful syntactic/semantic feature. Additionally since we define links between structures, tree kernels also generate relational features spanning question and passage structures. We derive very important findings, which can be useful to build state-of-the-art systems: (i) full syntactic dependencies can outperform shallow models also using external knowledge and (ii) the semantic information should be derived by effective and high-coverage resources, e.g., LD, and incorporated in syntactic structures to be effective. We demonstrate our findings by carrying out an extensive comparative experimentation on two different TREC QA corpora and one community question answer dataset, namely Answerbag. Our comparative analysis on well-defined answer selection benchmarks consistently demonstrates that our structural semantic models largely outperform the state of the art in passage reranking.

KW - Kernel methods

KW - Learning to rank

KW - Linked data

KW - Question answering

KW - Structural kernels

UR - http://www.scopus.com/inward/record.url?scp=84958254447&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958254447&partnerID=8YFLogxK

U2 - 10.1145/2806416.2806490

DO - 10.1145/2806416.2806490

M3 - Conference contribution

SN - 9781450337946

VL - 19-23-Oct-2015

SP - 1451

EP - 1460

BT - International Conference on Information and Knowledge Management, Proceedings

PB - Association for Computing Machinery

ER -