Skip to main content
Preprint

MizAR 60 for Mizar 50

Jakubův, JanCzech Technical University in Prague, Czech RepublicChvalovský, KarelCzech Technical University in Prague, Czech RepublicGoertzel, ZarathustraCzech Technical University in Prague, Czech RepublicKaliszyk, CezaryUniversität Innsbruck, AustriaOlšák, MirekInstitut des Hautes Études Scientifiques, Paris, FrancePiotrowski, BartoszCzech Technical University in Prague, Czech RepublicSchulz, StephanDHBW Stuttgart, GermanySuda, MartinCzech Technical University in Prague, Czech RepublicUrban, JosefCzech Technical University in Prague, Czech Republic
2023en
ABI

Abstract

As a present to Mizar on its 50th anniversary, we develop an AI/TP system that automatically proves about 60% of the Mizar theorems in the hammer setting. We also automatically prove 75% of the Mizar theorems when the automated provers are helped by using only the premises used in the human-written Mizar proofs. We describe the methods and large-scale experiments leading to these results. This includes in particular the E and Vampire provers, their ENIGMA and Deepire learning modifications, a number of learning-based premise selection methods, and the incremental loop that interleaves growing a corpus of millions of ATP proofs with training increasingly strong AI/TP systems on them. We also present a selection of Mizar problems that were proved automatically.

Identifiers

Citations and references

Cited by 270 references