First Neural Conjecturing Datasets and Experiments

05/29/2020
by   Josef Urban, et al.
0

We describe several datasets and first experiments with creating conjectures by neural methods. The datasets are based on the Mizar Mathematical Library processed in several forms and the problems extracted from it by the MPTP system and proved by the E prover using the ENIGMA guidance. The conjecturing experiments use the Transformer architecture and in particular its GPT-2 implementation.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
06/09/2011

The MathScheme Library: Some Preliminary Experiments

We present some of the experiments we have performed to best test our de...
research
02/12/2020

Visualizing modular forms

We describe practical methods to visualize modular forms. We survey seve...
research
08/08/2022

Lotse: A Practical Framework for Guidance in Visual Analytics

Co-adaptive guidance aims to enable efficient human-machine collaboratio...
research
01/24/2017

Deep Network Guided Proof Search

Deep learning techniques lie at the heart of several significant AI adva...
research
09/07/2021

Learned Provability Likelihood for Tactical Search

We present a method to estimate the provability of a mathematical formul...
research
02/06/2021

Vampire With a Brain Is a Good ITP Hammer

Vampire has been for a long time the strongest first-order automated the...
research
04/01/2018

Probabilistic Formulations of Regression with Mixed Guidance

Regression problems assume every instance is annotated (labeled) with a ...

Please sign up or login with your details

Forgot password? Click here to reset