Metaphors have always represented an obstacle to Natural Language Processing systems, because of their unpredictable behavior. Since the sixties, Artificial Intelligence researchers focused on language analysis have started working on metaphors and analogy by developing methods and approaches mainly based on the compositional view. In the last decade, the failure of these approaches caused a growing interest in corpus based and qualitative approaches to metaphor identification. These new approaches are mainly based on statistics and often oriented to quantitative analysis. In this work we present a system that detects metaphorical meanings through an analogy-based engine, and it is able to correctly disambiguate among alternative word meanings, even if they are non literal. An Italian case-study is illustrated, but, due to its nature, the system can be applied both to Italian and English text corpora.

Words on the edge: conceptual rules and usage variability

GOLA, ELISABETTA;FEDERICI, STEFANO
2009-01-01

Abstract

Metaphors have always represented an obstacle to Natural Language Processing systems, because of their unpredictable behavior. Since the sixties, Artificial Intelligence researchers focused on language analysis have started working on metaphors and analogy by developing methods and approaches mainly based on the compositional view. In the last decade, the failure of these approaches caused a growing interest in corpus based and qualitative approaches to metaphor identification. These new approaches are mainly based on statistics and often oriented to quantitative analysis. In this work we present a system that detects metaphorical meanings through an analogy-based engine, and it is able to correctly disambiguate among alternative word meanings, even if they are non literal. An Italian case-study is illustrated, but, due to its nature, the system can be applied both to Italian and English text corpora.
2009
1368-9223, 2009
metaphor; corpus linguistics; machine learning
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS 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/11584/31260
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact