Premise Selection and External Provers for HOL4

@article{Gauthier2015PremiseSA,
  title={Premise Selection and External Provers for HOL4},
  author={Thibault Gauthier and C. Kaliszyk},
  journal={Proceedings of the 2015 Conference on Certified Programs and Proofs},
  year={2015}
}
Learning-assisted automated reasoning has recently gained popularity among the users of Isabelle/HOL, HOL Light, and Mizar. In this paper, we present an add-on to the HOL4 proof assistant and an adaptation of the HOL(y)Hammer system that provides machine learning-based premise selection and automated reasoning also for HOL4. We efficiently record the HOL4 dependencies and extract features from the theorem statements, which form a basis for premise selection. HOL(y)Hammer transforms the HOL4… 

Tables from this paper

Sharing HOL4 and HOL Light Proof Knowledge
TLDR
A number of methods are proposed and evaluated, which strengthen proof automation by learning from proof libraries of different provers, which can be proved directly from the dependencies induced by similar proofs in the other library.
TacticToe: Learning to Reason with HOL4 Tactics
TLDR
A unified proof assistant automation approach which attempts to automate the selection of appropriate tactics and tactic-sequences with an optimized small-scale hammering approach, implemented as a tactic-level automation for HOL4: TacticToe.
Thor: Wielding Hammers to Integrate Language Models and Automated Theorem Provers
TLDR
Thor is introduced, a framework integrating language models and automated theorem provers to overcome this challenge of selecting useful premises from a large library to unlock the proof of a given conjecture.
Improving automation in interactive theorem provers by efficient encoding of lambda-abstractions
TLDR
Developing algorithms which outperform both lambda-lifting and the simple Scho ̈nfinkel’s algorithm used in Sledgehammer for Isabelle/HOL, thus enhancing au- tomation in proof assistants.
Certified Connection Tableaux Proofs for HOL Light and TPTP
TLDR
This paper shows how theLeanCoP's core algorithm can be implemented inside HOL Light, providing an implementation of leanCoP that certifies its proofs, and discusses the differences between the direct implementation using an explicit Prolog stack, to the continuation passing implementation of MESON present in HOL Light.
A Learning-Based Fact Selector for Isabelle/HOL
TLDR
This work introduces MaSh, an alternative that learns from successful proofs, and outperforms the old fact selector on large formalizations.
Proof searching and prediction in HOL4 with evolutionary/heuristic and deep learning techniques
TLDR
Experimental results suggest that combining evolutionary/heuristic and deep learning techniques with proof assistants can greatly facilitate proof finding/optimization and proof prediction.
Hammering towards QED
This paper surveys the emerging methods to automate reasoning over large libraries developed with formal proof assistants. We call these methods hammers. They give the authors of formal proofs a
Property Invariant Embedding for Automated Reasoning
TLDR
This work proposes a novel encoding of formulas that extends existing graph neural network models and provides additional links between such nodes that allow the network to recover the meaning and therefore correctly embed such nodes irrespective of the given labels.
GRUNGE: A Grand Unified ATP Challenge
TLDR
A large set of related theorem proving problems obtained by translating theorems from the HOL4 standard library into multiple logical formalisms results in a new "grand unified" large theory benchmark that emulates the ITP/ATP hammer setting.
...
...

References

SHOWING 1-10 OF 43 REFERENCES
Learning-Assisted Automated Reasoning with Flyspeck
TLDR
It is shown that 39 % of the 14185 theorems could be proved in a push-button mode (without any high-level advice and user interaction) in 30 seconds of real time on a fourteen-CPU workstation.
MaLARea SG1- Machine Learner for Automated Reasoning with Semantic Guidance
This paper describes a system combining model-based and learning-based methods for automated reasoning in large theories, i.e. on a large number of problems that use many axioms, lemmas, theorems,
HOL(y)Hammer: Online ATP Service for HOL Light
TLDR
HOL(y)Hammer is an online AI/ATP service for formal mathematics encoded in the HOL Light system that allows its users to upload and automatically process an arbitrary formal development based on HOL Light, and to attack arbitrary conjectures that use the concepts defined in some of the uploaded projects.
MaLARea: a Metasystem for Automated Reasoning in Large Theories
TLDR
A simple metasystem iteratively combining deductive Automated Reasoning tools (now the E and the SPASS ATP systems) with a machine learning component, intended use is in large theories, i.e. on a large number of problems which in a consistent fashion use many axioms, lemmas, theorems, definitions and symbols.
Automated Reasoning in Higher-Order Logic using the TPTP THF Infrastructure
TLDR
Key developments have been the specification of the THF language, the addition of higher-order problems to theTPTP, the development of the TPTP THF infrastructure, several ATP systems for higher- order logic, and the use of higher -order ATP in a range of domains.
MaSh: Machine Learning for Sledgehammer
TLDR
MaSh, an alternative that learns from successful proofs, outperforms the old relevance filter on large formalizations, and a particularly strong filter is obtained by combining the two filters.
SMT solvers: new oracles for the HOL theorem prover
  • Tjark Weber
  • Computer Science
    International Journal on Software Tools for Technology Transfer
  • 2011
TLDR
This paper describes an integration of Satisfiability Modulo Theories (SMT) solvers with the HOL4 theorem prover, which makes state-of-the-art SMT solving techniques available to users of theHol4 system, thereby increasing the degree of automation for a substantial fragment of its logic.
Automated Deduction – CADE-24
  • M. Bonacina
  • Computer Science
    Lecture Notes in Computer Science
  • 2013
TLDR
This paper defends the idea that a collaborative use of several provers is a key to easier and faster verification, and introduces a logic that is designed to target a wide set of theorem provers.
Sine Qua Non for Large Theory Reasoning
TLDR
This paper contains the presentation of the method for axiom selection in the Sumo INference Engine system and describes experiments with it in the theorem prover Vampire.
Beagle as a HOL4 external ATP method
TLDR
A translation of the higher-order goals to the TFA format of TPTP and add trace output to Beagle to reconstruct the intermediate steps derived by the ATP in HOL4.
...
...