MaSh: Machine Learning for Sledgehammer

  • Daniel Kühlwein
  • Jasmin Christian Blanchette
  • Cezary Kaliszyk
  • Josef Urban
Conference paper

DOI: 10.1007/978-3-642-39634-2_6

Volume 7998 of the book series Lecture Notes in Computer Science (LNCS)
Cite this paper as:
Kühlwein D., Blanchette J.C., Kaliszyk C., Urban J. (2013) MaSh: Machine Learning for Sledgehammer. In: Blazy S., Paulin-Mohring C., Pichardie D. (eds) Interactive Theorem Proving. ITP 2013. Lecture Notes in Computer Science, vol 7998. Springer, Berlin, Heidelberg

Abstract

Sledgehammer integrates automatic theorem provers in the proof assistant Isabelle/HOL. A key component, the relevance filter, heuristically ranks the thousands of facts available and selects a subset, based on syntactic similarity to the current goal. We introduce MaSh, an alternative that learns from successful proofs. New challenges arose from our “zero-click” vision: MaSh should integrate seamlessly with the users’ workflow, so that they benefit from machine learning without having to install software, set up servers, or guide the learning. The underlying machinery draws on recent research in the context of Mizar and HOL Light, with a number of enhancements. MaSh outperforms the old relevance filter on large formalizations, and a particularly strong filter is obtained by combining the two filters.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  • Daniel Kühlwein
    • 1
  • Jasmin Christian Blanchette
    • 2
  • Cezary Kaliszyk
    • 3
  • Josef Urban
    • 1
  1. 1.ICISRadboud Universiteit NijmegenThe Netherlands
  2. 2.Fakultät für InformatikTechnische Universität MünchenGermany
  3. 3.Institut für InformatikUniversität InnsbruckAustria