Experimenting a "general purpose" textual entailment learner in AVE

Fabio Massimo Zanzotto, Alessandro Moschitti

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

Abstract

In this paper we present the use of a "general purpose" textual entaiment recognizer in the Answer Validation Exercise (AVE) task. Our system has been developed to learn entailment rules from annotated examples. The main idea of the system is the cross-pair similirity measure we defined. This similarity allows us to define an implicit feature space using kernel functions in SVM learners. We experimented with our system using different training and testing sets: RTE data sets and AVE data sets. The comparative results show that entailment rules can be learned from data sets, e.g. RTE, that are different from AVE. Moreover, it seems that better results are obtained using more controlled training data (the RTE set) that less controlled ones (the AVE development set). Although, the high variability of the outcome prevents us to derive definitive conclusions, the results of our system show that our approach is quite promising and improvable in the future.

Original languageEnglish
Title of host publicationCEUR Workshop Proceedings
PublisherCEUR-WS
Volume1172
Publication statusPublished - 2006
Externally publishedYes
Event2006 Working Notes for CLEF Workshop, CLEF 2006 - Co-located with the 10th European Conference on Digital Libraries, ECDL 2006 - Alicante, Spain
Duration: 20 Sep 200622 Sep 2006

Other

Other2006 Working Notes for CLEF Workshop, CLEF 2006 - Co-located with the 10th European Conference on Digital Libraries, ECDL 2006
CountrySpain
CityAlicante
Period20/9/0622/9/06

Fingerprint

Testing

Keywords

  • Question answering
  • Textual entailment recognition

ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Zanzotto, F. M., & Moschitti, A. (2006). Experimenting a "general purpose" textual entailment learner in AVE. In CEUR Workshop Proceedings (Vol. 1172). CEUR-WS.

Experimenting a "general purpose" textual entailment learner in AVE. / Zanzotto, Fabio Massimo; Moschitti, Alessandro.

CEUR Workshop Proceedings. Vol. 1172 CEUR-WS, 2006.

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

Zanzotto, FM & Moschitti, A 2006, Experimenting a "general purpose" textual entailment learner in AVE. in CEUR Workshop Proceedings. vol. 1172, CEUR-WS, 2006 Working Notes for CLEF Workshop, CLEF 2006 - Co-located with the 10th European Conference on Digital Libraries, ECDL 2006, Alicante, Spain, 20/9/06.
Zanzotto FM, Moschitti A. Experimenting a "general purpose" textual entailment learner in AVE. In CEUR Workshop Proceedings. Vol. 1172. CEUR-WS. 2006
Zanzotto, Fabio Massimo ; Moschitti, Alessandro. / Experimenting a "general purpose" textual entailment learner in AVE. CEUR Workshop Proceedings. Vol. 1172 CEUR-WS, 2006.
@inproceedings{0376ba771ef04fc0b5427035ccebb9cf,
title = "Experimenting a {"}general purpose{"} textual entailment learner in AVE",
abstract = "In this paper we present the use of a {"}general purpose{"} textual entaiment recognizer in the Answer Validation Exercise (AVE) task. Our system has been developed to learn entailment rules from annotated examples. The main idea of the system is the cross-pair similirity measure we defined. This similarity allows us to define an implicit feature space using kernel functions in SVM learners. We experimented with our system using different training and testing sets: RTE data sets and AVE data sets. The comparative results show that entailment rules can be learned from data sets, e.g. RTE, that are different from AVE. Moreover, it seems that better results are obtained using more controlled training data (the RTE set) that less controlled ones (the AVE development set). Although, the high variability of the outcome prevents us to derive definitive conclusions, the results of our system show that our approach is quite promising and improvable in the future.",
keywords = "Question answering, Textual entailment recognition",
author = "Zanzotto, {Fabio Massimo} and Alessandro Moschitti",
year = "2006",
language = "English",
volume = "1172",
booktitle = "CEUR Workshop Proceedings",
publisher = "CEUR-WS",

}

TY - GEN

T1 - Experimenting a "general purpose" textual entailment learner in AVE

AU - Zanzotto, Fabio Massimo

AU - Moschitti, Alessandro

PY - 2006

Y1 - 2006

N2 - In this paper we present the use of a "general purpose" textual entaiment recognizer in the Answer Validation Exercise (AVE) task. Our system has been developed to learn entailment rules from annotated examples. The main idea of the system is the cross-pair similirity measure we defined. This similarity allows us to define an implicit feature space using kernel functions in SVM learners. We experimented with our system using different training and testing sets: RTE data sets and AVE data sets. The comparative results show that entailment rules can be learned from data sets, e.g. RTE, that are different from AVE. Moreover, it seems that better results are obtained using more controlled training data (the RTE set) that less controlled ones (the AVE development set). Although, the high variability of the outcome prevents us to derive definitive conclusions, the results of our system show that our approach is quite promising and improvable in the future.

AB - In this paper we present the use of a "general purpose" textual entaiment recognizer in the Answer Validation Exercise (AVE) task. Our system has been developed to learn entailment rules from annotated examples. The main idea of the system is the cross-pair similirity measure we defined. This similarity allows us to define an implicit feature space using kernel functions in SVM learners. We experimented with our system using different training and testing sets: RTE data sets and AVE data sets. The comparative results show that entailment rules can be learned from data sets, e.g. RTE, that are different from AVE. Moreover, it seems that better results are obtained using more controlled training data (the RTE set) that less controlled ones (the AVE development set). Although, the high variability of the outcome prevents us to derive definitive conclusions, the results of our system show that our approach is quite promising and improvable in the future.

KW - Question answering

KW - Textual entailment recognition

UR - http://www.scopus.com/inward/record.url?scp=84922032243&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84922032243&partnerID=8YFLogxK

M3 - Conference contribution

VL - 1172

BT - CEUR Workshop Proceedings

PB - CEUR-WS

ER -