Textual Entailment by Generality

被引：2

作者：

Pais, Sebastiao ^{[1
]}

Dias, Gael ^{[1
,2
]}

Wegrzyn-Wolska, Katarzyna ^{[3
]}

Mahl, Robert ^{[4
]}

Jouvelot, Pierre ^{[4
]}

机构：

[1] HULTIG Univ Beira Interior, Covilha, Portugal

[2] Univ Caen Basse Normandie, DLU, GREYC, F-14032 Caen, France

[3] Ecole Super dIngenieurs Informat, Genie des Telecommun, SITR, Villejuif, France

[4] Ecole Natl Superieure des Mines de Paris, CRI, Paris, France

来源：

COMPUTATIONAL LINGUISTICS AND RELATED FIELDS | 2011年 / 27卷

关键词：

Asymmetric Association Measures; Informative Asymmetric Measure; Textual Entailment;

D O I：

10.1016/j.sbspro.2011.10.606

中图分类号：

H0 [语言学];

学科分类号：

030303 ; 0501 ; 050102 ;

摘要：

Textual Entailment consists in determining if an entailment relation exists between two texts. In this paper, we present an Informative Asymmetric Measure called the Asymmetric InfoSimba (AIS), which we combine with different asym-metric association measures to recognize the specific case of Textual Entailment by Generality. In particular, the AIS proposes an unsupervised, language-independent, threshold free solution. This new measure is tested against the first Recognizing Textual Entailment dataset for an exhaustive number of asymmetric association measures and shows that the combination of the AIS with the Braun-Blanket steadily improves results against competitive measures such as the one proposed by [1]. (C) 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of PACLING Organizing Committee.

引用

页码：258 / 266

页数：9

共 21 条

[1]

[Anonymous], P 22 C ART INT AAAI

[2]

[Anonymous], P 12 INT C NAT LANG

[3]

[Anonymous], P PASCAL CHALL WORKS

[4]

[Anonymous], P 1 CHALL WORKSH REC

[5]

[Anonymous], P 20 ACM C INF KNOWL

[6]

[Anonymous], P JOINT C INT COMM C

[7]

[Anonymous], 2007, P INT C REC ADV NAT

[8]

[Anonymous], P 14 C COMP LING COL

[9]

Bayer S., 2005, P THE1ST PASCAL RECO, P41

[10]

Bos Johan., 2005, P 1 PASCAL CHALL WOR, P65

← 1 2 3 →