Man bites dog: looking for interesting inconsistencies in structured news reports

  • Authors:
  • Emma Byrne;Anthony Hunter

  • Affiliations:
  • Department of Computer Science, University College London, Gower Street, London WC1E 6BT, UK;Department of Computer Science, University College London, Gower Street, London WC1E 6BT, UK

  • Venue:
  • Data & Knowledge Engineering
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Much useful information in news reports is often that which is surprising or unexpected. In other words, we harbour many expectations about the world, and when any of these expectations are violated (i.e. made inconsistent) by news, we have a strong indicator of some information that is interesting for us. In this paper we present a framework for identifying interesting information in news reports by finding interesting inconsistencies. An implemented system based on this framework (1) accepts structured news reports as inputs, (2) translates each report to a logical literal, (3) identifies the story of which the report is a part, (4) looks for inconsistencies between the report, the background knowledge, and a set of expectations, (5) classifies and evaluates those inconsistencies, and (6) outputs news reports of interest to the user together with associated explanations of why they are interesting.