CSE - A Automated Theorem Prover Based on Standard Contradiction Separation Dynamic Deduction
Preprint
- 21 March 2024
- preprint
- Published by Springer Nature in Research Square
Abstract
Contradiction Separation Extension (CSE in short), an automated theorem prover for first-order logic without equality, which is based on a novel standard contradiction separation (S-CS) inference rule. Different from binary resolution or its refinements, the biggest difference and characteristic of S-CS rule is that each resolution step can handle multiple (two or more) clauses for synergized deduction. This paper mainly describes problem format conversion, clause set preprocessing, deduction framework, distinctive heuristic strategies, S-CS dynamic inference mechanism in CSE. CSE is evaluated through benchmarks and the deduction characteristics of this multi-clause dynamic deduction are demonstrated, e.g., FOF division problems of CASC-26, CASC-J9. In addition, the CASC-J9 is also tested by combined systems, which combine CSE and other prover (Prover9, Eprover, Vampire), and the combined systems is applied to test on the hard problems with rating of 1. Experimental results show the CSE shows a competitive performance and can play a role in first-order logic automated theorem proving.Keywords
All Related Versions
This publication has 22 references indexed in Scilit:
- A multi-clause dynamic deduction algorithm based on standard contradiction separation ruleInformation Sciences, 2021
- Contradiction separation based dynamic multi-clause synergized automated deductionInformation Sciences, 2018
- The TPTP Problem Library and Associated InfrastructureJournal of Automated Reasoning, 2017
- The TPTP Problem Library and Associated InfrastructureJournal of Automated Reasoning, 2009
- Lightweight relevance filtering for machine-generated resolution problemsJournal of Applied Logic, 2009
- iProver – An Instantiation-Based Theorem Prover for First-Order Logic (System Description)2008
- Evaluating general purpose automated theorem proving systemsArtificial Intelligence, 2001
- Otter - The CADE-13 Competition IncarnationsJournal of Automated Reasoning, 1997
- DISCOUNT - A Distributed and Learning Equational ProverJournal of Automated Reasoning, 1997
- Complexity and related enhancements for automated theorem-proving programsComputers & Mathematics with Applications, 1976