A probabilistic setting and lexical cooccurrence model for textual entailment

  • Authors:
  • Oren Glickman;Ido Dagan

  • Affiliations:
  • Bar Ilan University;Bar Ilan University

  • Venue:
  • EMSEE '05 Proceedings of the ACL Workshop on Empirical Modeling of Semantic Equivalence and Entailment
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper proposes a general probabilistic setting that formalizes a probabilistic notion of textual entailment. We further describe a particular preliminary model for lexical-level entailment, based on document cooccurrence probabilities, which follows the general setting. The model was evaluated on two application independent datasets, suggesting the relevance of such probabilistic approaches for entailment modeling.