Boosting variant recognition with light semantics

  • Authors:
  • Cécile Fabre;Christian Jacquemin

  • Affiliations:
  • Univ. Toulouse-Le Mirail, Toulouse, France;CNRS-LIMSI, Orsay, France

  • Venue:
  • COLING '00 Proceedings of the 18th conference on Computational linguistics - Volume 1
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

A reasonably simple, domain-independent, large-scale approach of lexical semantics to paraphrase recognition is presented in this paper. It relies on the enrichment of morphosyntactic rules and the addition of four boolean syntactico-semantic features to a set of 1,023 words. It results in a significant enhancement of precision of 30% with a slight decrease in recall of 10%.