Patrick Wischnewski

On the saturation of YAGO

YAGO is an automatically generated ontology out of Wikipedia and WordNet. It is eventually represented in a proprietary flat text file format and a core comprises 10 million facts and formulas. We present a translation of YAGO into the Bernays-Schonfinkel Horn class with equality. A new variant of the superposition calculus is sound, complete and terminating for this class. Together with extended term indexing data structures the new calculus is implemented in Spass-YAGO. YAGO can be finitely saturated by Spass-YAGO in about 1 hour. We have found 49 inconsistencies in the original generated ontology which we have fixed. Spass-YAGO can then prove non-trivial conjectures with respect to the resulting saturated and consistent clause set of about 1.4 GB in less than one second.

External sources of axioms in automated theorem proving

In recent years there has been a growing demand for Automated Theorem Proving (ATP) in large theories, which often have more axioms than can be handled effectively as normal internal axioms. This work addresses the issues of accessing external sources of axioms from a first-order logic ATP system, and presents an implemented ATP system that retrieves external axioms asynchronously, on demand.

Automated verification of interactive rule-based configuration systems

Rule-based specifications of systems have again become common in the context of product line variability modeling and configuration systems. In this paper, we define a logical foundation for rule-based specifications that has enough expressivity and operational behavior to be practically useful and at the same time enables decidability of important overall properties such as consistency or cycle-freeness. Our logic supports rule-based interactive user transitions as well as the definition of a domain theory via rule transitions. As a running example, we model DOPLER, a rule-based configuration system currently in use at Siemens.

Subterm contextual rewriting

Sophisticated reductions are an important means to achieve progress in automated theorem proving. We consider the powerful reduction rule Contextual Rewriting in connection with the superposition calculus. If considered in its most general form the applicability of contextual rewriting is not decidable. We develop an instance of contextual rewriting where applicability becomes decidable while preserving a great deal of its simplification power. A sophisticated implementation of the rule in SPASS reveals its application potential. Our contextual rewriting instance called subterm contextual rewriting is feasible in the sense that it can be executed on the overall TPTP resulting in a gain of solved problems and new solutions to a number of problems that could not be solved by theorem provers so far.

Compliance, Functional Safety and Fault Detection by Formal Methods

With the increasing complexity of todays cars functional safety and compliance guarantees are more and more dicult to obtain. During the life time of a vehicle the detection of malfunctioning non-mechanical components requires meanwhile more attention than the maintenance of its mechanical counterparts. A full fledged formal verification of the overall car is not realistic and even hard to obtain for single non-trivial components such as assistant systems. Furthermore, it does not support fault detection at run time. We suggest an approach towards formal safety, compliance and fault detection at run time via an auditor. The auditor is automatically fed out of the engineering and production process by a suitable abstract specification and respective model of the car and can detect then detect violations and faulty components.