We present a system that couples techniques belonging to Information Extraction and deep linguistic processing for Question Answering. The system presented in the paper has undergone extensive testing and the parser has been trained on available testsuites. The system uses text entailment processing to select best sentences to match with each question. Both sentences and questions need to parsed syntactically and semantically and a logical form has to be produced with predicate argument structures and propositional level analysis. In order to pick the right answer from a set of five, after extracting the best sentence/s from the text, we organized different strategies according to question type and semantic propositional type. The system has access to a wide range of computational lexica, ontologies and datasets to carry out the task: for common sense knowledge we used ConceptNet.

VENSES GetAsk: a System for Hybrid Question Answering And Answer Recovery using Text Entailment

DELMONTE, Rodolfo
2013-01-01

Abstract

We present a system that couples techniques belonging to Information Extraction and deep linguistic processing for Question Answering. The system presented in the paper has undergone extensive testing and the parser has been trained on available testsuites. The system uses text entailment processing to select best sentences to match with each question. Both sentences and questions need to parsed syntactically and semantically and a logical form has to be produced with predicate argument structures and propositional level analysis. In order to pick the right answer from a set of five, after extracting the best sentence/s from the text, we organized different strategies according to question type and semantic propositional type. The system has access to a wide range of computational lexica, ontologies and datasets to carry out the task: for common sense knowledge we used ConceptNet.
File in questo prodotto:
File Dimensione Formato  
newvalencia.pdf

non disponibili

Tipologia: Abstract
Licenza: Licenza non definita
Dimensione 1.05 MB
Formato Adobe PDF
1.05 MB Adobe PDF   Visualizza/Apri

I documenti in ARCA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10278/39522
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact