The TALP-QA System for Spanish at CLEF 2004: Structural and Hierarchical Relaxing of Semantic Constraints

This paper describes TALP-QA, a multilingual open-domain Question Answering (QA) system that processes both factoid and definition questions. The system is described and evaluated in the context of our participation in the CLEF 2004 Spanish Monolingual QA task. Our approach to factoid questions is t...

Full description

Saved in:
Bibliographic Details
Published inLecture notes in computer science pp. 557 - 568
Main Authors Ferrés, Daniel, Kanaan, Samir, Ageno, Alicia, González, Edgar, Rodríguez, Horacio, Surdeanu, Mihai, Turmo, Jordi
Format Book Chapter Conference Proceeding
LanguageEnglish
Published Berlin, Heidelberg Springer Berlin Heidelberg 2005
Springer
SeriesLecture Notes in Computer Science
Subjects
Applied sciences
Artificial intelligence
Computer science; control theory; systems
Exact sciences and technology
Information systems. Data bases
Memory organisation. Data processing
Name Entity
Name Entity Recognition
Passage Retrieval
Question Answer
Semantic Constraint
Software
Speech and sound recognition and synthesis. Linguistics
Query formulation
Multilingualism
Language processing
Taxonomy
Semantics
Information retrieval
Information access
Text
Sentence
Responder
Online AccessGet full text

Cover

More Information
Summary:This paper describes TALP-QA, a multilingual open-domain Question Answering (QA) system that processes both factoid and definition questions. The system is described and evaluated in the context of our participation in the CLEF 2004 Spanish Monolingual QA task. Our approach to factoid questions is to build a semantic representation of the questions and the sentences in the passages retrieved for each question. A set of Semantic Constraints (SC) are extracted for each question. An answer extraction algorithm extracts and ranks sentences that satisfy the SCs of the question. If matches are not possible the algorithm relaxes the SCs structurally (removing constraints) and/or hierarchically (abstracting the constraints using a taxonomy). Answers to definition questions are generated by selecting the text fragment with more density of those terms more frequently related to the question’s target (the Named Entity (NE) that appears in the question) throughout the corpus.
ISBN:9783540274209
3540274200
ISSN:0302-9743
1611-3349
DOI:10.1007/11519645_55