Corpus-based learning of semantic relations by the ILP system, Asium

  • Authors:
  • Claire Nedellec

  • Affiliations:
  • Univ. Paris-Sud, Orsay

  • Venue:
  • Learning language in logic
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

This chapter presents the ILP method, Asium, that learns ontologies and verb subcategorization frames from parsed corpora in specific domains. The ontology is learned by bottom-up conceptual clustering from parsed corpora. The clustering method also generalizes the grammatical relations between verbs and complement heads as they are observed in the corpora. The set of grammatical relations learned for a given verb forms the verb subcategorization frame. This paper details Asium's learning algorithms in an ILP formalism and shows how the learned linguistic resources can be applied to semantic tagging, language control and syntactic disambiguation within the ILP framework.