Proof-Pattern Recognition and Lemma Discovery in ACL2

  title={Proof-Pattern Recognition and Lemma Discovery in ACL2},
  author={J{\'o}nathan Heras and Ekaterina Komendantskaya and Moa Johansson and Ewen Maclean},
  booktitle={Logic Programming and Automated Reasoning},
We present a novel technique for combining statistical machine learning for proof-pattern recognition with symbolic methods for lemma discovery. The resulting tool, ACL2(ml), gathers proof statistics and uses statistical pattern-recognition to pre-processes data from libraries, and then suggests auxiliary lemmas in new proofs by analogy with already seen examples. This paper presents the implementation of ACL2(ml) alongside theoretical descriptions of the proof-pattern recognition and lemma… 

Proof Pattern Search in Coq/SSReflect

Three new improvements to ML4PG are presented, including a new method of "recurrent clustering" that is introduced to collect statistical features from Coq terms and an automaton-shape representation for proof patterns.

Learning and Exploration in Automated Theorem Proving

This note describes a new project with the purpose of combining the advantages of statistical machine learning with those of symbolic methods for lemma discovery, such as theory exploration. We can

Proof Mining with Dependent Types

This paper presents a method that combines statistical data mining and theory exploration in order to analyse and automate proofs in dependently typed language of Coq.

Lemma Discovery for Induction - A Survey

Various techniques for automating the discovery of auxiliary lemmas are surveyed, including both top-down techniques attempting to generate a lemma from an ongoing proof attempt, as well as bottom-up theory exploration techniques trying to construct interesting lemma about available functions and datatypes, thus constructing a richer background theory.

Proof output and machine learning for inductive theorem provers

HipSpec is an inductive theorem prover that has not yet explored this area, and a foundation for supporting machine learning implementations within HipSpec is laid, which allows HipSpec to remember already-proven theorems in between executions.

Machine Learning for Inductive Theorem Proving

This work explores a combination of machine learning, a simple Boyer-Moore model and ATPs as a means of improving the automation of inductive proofs in the proof assistant HOL Light, and evaluates the framework using a number of induction proof corpora.

Mining State-Based Models from Proof Corpora

Preliminary experiments show that the inferred models are generally accurate (contain few false-positive sequences) and that representing existing proofs in such a way can be very useful when guiding new ones.

ACL2(ml): Machine-Learning for ACL2

The two most recent extensions for ACL2(ml) can suggest now families of similar function definitions, in addition to the families ofsimilar theorems, and the lemma generation tool implemented in ACL2 (ml) has been improved with a method to generate preconditions using the guard mechanism of ACL2.

Towards neuro-symbolic conjecturing

This work describes ongoing work in combining data-driven and symbolic methods for automated conjecturing, where the data- driven part should identify which kinds of conjectures are likely to be useful and restrict the symbolic search to those ones.

Automating Event-B invariant proofs by rippling and proof patching

This work combines a scheme-based theory-exploration system with rippling to develop a proof patch via lemma discovery, and develops two new proof patches to unfold operator definitions and to suggest case-splits, respectively.



Machine Learning in Proof General: Interfacing Interfaces

ML4PG allows users to gather proof statistics related to shapes of goals, sequences of applied tactics, and proof tree structures from the libraries of interactive higher-order proofs written in Coq and SSReflect.

Scheme-based theorem discovery and concept invention

The use of data-mining for the automatic formation of tactics

The aim of this project is to evaluate the applicability of data-mining techniques to the automatic formation of tactics from large corpuses of proofs to find commonly occurring patterns.

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,

Towards Algorithmic Cut-Introduction

We describe a method for abbreviating an analytic proof in classical first-order logic by the introduction of a lemma. Our algorithm is based on first computing a compressed representation of the

Automating Inductive Proofs Using Theory Exploration

HipSpec automatically generates a set of equational theorems about the available recursive functions of a program that make up an algebraic specification for the program and can in addition be used as a background theory for proving additional user-stated properties.

Automatic Acquisition of Search Control Knowledge from Multiple Proof Attempts

Two inference control heuristics for equational deduction that are based on the evaluation of previous successful proof attempts in domains of interest are presented and their usefulness is demonstrated.

Ascertaining Mathematical Theorems

The HR Program for Theorem Generation

The applications of HR to automated reasoning include thegeneration of constraints for constraint satisfaction problems, the generation of lemmas for automated theorem proving, and the production of benchmark theorems for the TPTP library of test problems for ATP systems.

Proving Theorems about Java and the JVM with ACL2

A methodology for proving theorems mechanically about Java methods and classes by compiling them with javac and then proving the corresponding theorem about the JVM.