In this paper we present two new mechanisms we created in VENSES, the system for semantic evaluation of the University of Venice. The first mechanism is used to match predicate-argument structures with different governors, a verb and a noun, respectively in the Hypothesis and the Text. It can be defined Augmented Finite State Automata (FSA) which are matching procedures based on tagged words in one case, and dependency relations in another. In both cases, a number of inferences – the augmentation - is fired to match different words. The second mechanism is based on the output of our module for anaphora resolution. Our system produces antecedents for pronominal expressions and equal nominal expressions. On the contrary, no decision is taken for “bridging” expressions. So the “bridging” mechanism is activated by the Semantic Evaluator and has access to the History List and the semantic features associated to each referring expression. If constraint conditions meet, the system looks for a similar association of property/entity in web ontologies like Umbel, Yago and DBPedia. The two mechanisms have been proven to contribute a 5% and 3% accuracy, respectively.

Semantic Processing for Text Entailment with VENSES

DELMONTE, Rodolfo;TONELLI, Sara;TRIPODI, ROCCO
2009-01-01

Abstract

In this paper we present two new mechanisms we created in VENSES, the system for semantic evaluation of the University of Venice. The first mechanism is used to match predicate-argument structures with different governors, a verb and a noun, respectively in the Hypothesis and the Text. It can be defined Augmented Finite State Automata (FSA) which are matching procedures based on tagged words in one case, and dependency relations in another. In both cases, a number of inferences – the augmentation - is fired to match different words. The second mechanism is based on the output of our module for anaphora resolution. Our system produces antecedents for pronominal expressions and equal nominal expressions. On the contrary, no decision is taken for “bridging” expressions. So the “bridging” mechanism is activated by the Semantic Evaluator and has access to the History List and the semantic features associated to each referring expression. If constraint conditions meet, the system looks for a similar association of property/entity in web ontologies like Umbel, Yago and DBPedia. The two mechanisms have been proven to contribute a 5% and 3% accuracy, respectively.
Proceedings of Text Analysis Conference (TAC) 2009 Workshop - Notebook Papers and Results
File in questo prodotto:
File Dimensione Formato  
10.VensesTeam.proceedings.pdf

non disponibili

Tipologia: Abstract
Licenza: Licenza non definita
Dimensione 550.03 kB
Formato Adobe PDF
550.03 kB 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/10278/39702
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact