Contextual language models for ranking answers to natural language definition questions

Alejandro Figueroa, John Atkinson

Resultado de la investigación: Article

5 Citas (Scopus)

Resumen

Question-answering systems make good use of knowledge bases (KBs, e.g., Wikipedia) for responding to definition queries. Typically, systems extract relevant facts from articles regarding the question across KBs, and then they are projected into the candidate answers. However, studies have shown that the performance of this kind of method suddenly drops, whenever KBs supply narrow coverage. This work describes a new approach to deal with this problem by constructing context models for scoring candidate answers, which are, more precisely, statistical n-gram language models inferred from lexicalized dependency paths extracted from Wikipedia abstracts. Unlike state-of-the-art approaches, context models are created by capturing the semantics of candidate answers (e.g., "novel,""singer,""coach," and "city"). This work is extended by investigating the impact on context models of extra linguistic knowledge such as part-of-speech tagging and named-entity recognition. Results showed the effectiveness of context models as n-gram lexicalized dependency paths and promising context indicators for the presence of definitions in natural language texts.

Idioma originalEnglish
Páginas (desde-hasta)528-548
Número de páginas21
PublicaciónComputational Intelligence
Volumen28
N.º4
DOI
EstadoPublished - nov 2012

Huella dactilar

Language Model
Natural Language
Ranking
N-gram
Wikipedia
Named Entity Recognition
Question Answering System
Path
Tagging
Scoring
Linguistics
Knowledge Base
Model
Coverage
Semantics
Context
Query

ASJC Scopus subject areas

  • Computational Mathematics
  • Artificial Intelligence

Citar esto

@article{9ac98e0835ca4448a5dd46bd8005548d,
title = "Contextual language models for ranking answers to natural language definition questions",
abstract = "Question-answering systems make good use of knowledge bases (KBs, e.g., Wikipedia) for responding to definition queries. Typically, systems extract relevant facts from articles regarding the question across KBs, and then they are projected into the candidate answers. However, studies have shown that the performance of this kind of method suddenly drops, whenever KBs supply narrow coverage. This work describes a new approach to deal with this problem by constructing context models for scoring candidate answers, which are, more precisely, statistical n-gram language models inferred from lexicalized dependency paths extracted from Wikipedia abstracts. Unlike state-of-the-art approaches, context models are created by capturing the semantics of candidate answers (e.g., {"}novel,{"}{"}singer,{"}{"}coach,{"} and {"}city{"}). This work is extended by investigating the impact on context models of extra linguistic knowledge such as part-of-speech tagging and named-entity recognition. Results showed the effectiveness of context models as n-gram lexicalized dependency paths and promising context indicators for the presence of definitions in natural language texts.",
keywords = "context definition models, definition questions, feature analysis, lexicalized dependency paths, question answering, statistical language models",
author = "Alejandro Figueroa and John Atkinson",
year = "2012",
month = "11",
doi = "10.1111/j.1467-8640.2012.00426.x",
language = "English",
volume = "28",
pages = "528--548",
journal = "Computational Intelligence",
issn = "0824-7935",
publisher = "Wiley-Blackwell",
number = "4",

}

Contextual language models for ranking answers to natural language definition questions. / Figueroa, Alejandro; Atkinson, John.

En: Computational Intelligence, Vol. 28, N.º 4, 11.2012, p. 528-548.

Resultado de la investigación: Article

TY - JOUR

T1 - Contextual language models for ranking answers to natural language definition questions

AU - Figueroa, Alejandro

AU - Atkinson, John

PY - 2012/11

Y1 - 2012/11

N2 - Question-answering systems make good use of knowledge bases (KBs, e.g., Wikipedia) for responding to definition queries. Typically, systems extract relevant facts from articles regarding the question across KBs, and then they are projected into the candidate answers. However, studies have shown that the performance of this kind of method suddenly drops, whenever KBs supply narrow coverage. This work describes a new approach to deal with this problem by constructing context models for scoring candidate answers, which are, more precisely, statistical n-gram language models inferred from lexicalized dependency paths extracted from Wikipedia abstracts. Unlike state-of-the-art approaches, context models are created by capturing the semantics of candidate answers (e.g., "novel,""singer,""coach," and "city"). This work is extended by investigating the impact on context models of extra linguistic knowledge such as part-of-speech tagging and named-entity recognition. Results showed the effectiveness of context models as n-gram lexicalized dependency paths and promising context indicators for the presence of definitions in natural language texts.

AB - Question-answering systems make good use of knowledge bases (KBs, e.g., Wikipedia) for responding to definition queries. Typically, systems extract relevant facts from articles regarding the question across KBs, and then they are projected into the candidate answers. However, studies have shown that the performance of this kind of method suddenly drops, whenever KBs supply narrow coverage. This work describes a new approach to deal with this problem by constructing context models for scoring candidate answers, which are, more precisely, statistical n-gram language models inferred from lexicalized dependency paths extracted from Wikipedia abstracts. Unlike state-of-the-art approaches, context models are created by capturing the semantics of candidate answers (e.g., "novel,""singer,""coach," and "city"). This work is extended by investigating the impact on context models of extra linguistic knowledge such as part-of-speech tagging and named-entity recognition. Results showed the effectiveness of context models as n-gram lexicalized dependency paths and promising context indicators for the presence of definitions in natural language texts.

KW - context definition models

KW - definition questions

KW - feature analysis

KW - lexicalized dependency paths

KW - question answering

KW - statistical language models

UR - http://www.scopus.com/inward/record.url?scp=84868088027&partnerID=8YFLogxK

U2 - 10.1111/j.1467-8640.2012.00426.x

DO - 10.1111/j.1467-8640.2012.00426.x

M3 - Article

AN - SCOPUS:84868088027

VL - 28

SP - 528

EP - 548

JO - Computational Intelligence

JF - Computational Intelligence

SN - 0824-7935

IS - 4

ER -