A new method for consequence finding and compilation in restricted languages

  • Authors:

  • Alvaro del Val

  • Affiliations:

  • -

  • Venue:
  • AAAI '99/IAAI '99 Proceedings of the sixteenth national conference on Artificial intelligence and the eleventh Innovative applications of artificial intelligence conference innovative applications of artificial intelligence
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

SFK (skip-filtered, kernel) resolution is a new method for finding "interesting" consequencoefs a first order clausal theory 驴, namely those in some restricted target language LT. In its more restrictive form, SFK resolution corresponds to a relatively efficient SAT method, directional resolution; in its more general form, to a full prime implicate algorithm, namely Tison's. It generalizes both of them by offering much more flexible search, first order completeness, and a much wider range of inferential capabilities.SFK resolution has many applications: computing "characteristic" clauses for task-specific languages in abduction, explanation and non-monotonic reasoning (Inoue 1992); obtaining LUB approximations of the input theory (Selman and Kautz 1996) which are of polynomial size; incremental and lazy exact knowledge compilation (del Val 1994); and compilation into a tractable form for restricted target languages, independently of the tractability of inference in the given target language.