WSML2Reasoner
The WSML2Reasoner framework is a highly modular architecture which combines various validation, normalization, and transformation functionalities essential to the translation of ontology descriptions in WSML to the appropriate syntax of several underlying reasoning engines.
WSML (Web Service Modeling Language) is a family of formal ontology languages that are specifically aimed at describing Semantic Web Services. It's semantics is based on Description Logics, Logic Programming and First-Order Logic, with influences from F-Logic and frame-based representation systems. Conforming to the different influences, there exist five variants of WSML: WSML-Core, WSML-DL, WSML-Flight, WSML-Rule and WSML-Full. Currently, WSML2Reasoner has been implemented and tested using various reasoning engines that support all WSML variants with the exception of WSML-Full. For more infomation on WSML and the related working groups, please refer to the following links:
The framework's flexible architecture allows for easy integration of existing external reasoning components. It first maps WSML to either generic Datalog or OWL (Web Ontology Language), and then uses facades that mediate between the generic Datalog representation and the internal representation specific to the single reasoning engines, or between the resulting OWL representation and an OWL reasoning engine (e.g. Pellet). In the following sections, we breifly discuss a few of the reasonings engines currently implemented and tested within the WSML2Reasoner framework.
IRIS, Integrated Rule Inference System, is an extensible reasoning engine for expressive rule-based languages. Currently IRIS supports the following features:
The following evaluation strategies are currently supported:
To learn more about the fundamental algorithms that the reasoner is based upon have a look at the theoretical results.
The IRIS libraries are included in the LGPL release. For further information, please visit the IRIS website.
KAON2 is an infrastructure for managing OWL-DL, SWRL, and F-Logic ontologies. It has been integrated into the WSML2Reasoner framework in order to provide reasoning support for the WSML-Core, -Flight, and -DL variants. Contrary to most currently available DL reasoners, such as FaCT, FaCT++, RACER, DLP or Pellet, KAON2 does not implement the tableaux calculus. Rather, reasoning in KAON2 is implemented by novel algorithms which reduce a SHIQ(D) knowledge base to a disjunctive datalog program. For an overview of these algorithms, please refer to Reducing SHIQ- Descrption Logic to Disjunctive Datalog Programs. A detailed (and quite lengthy) technical presentation of all algorithms is given in Reasoning in Description Logics using Resolution and Deductive Databases.
KAON2 is is free of charge for universities for noncommecial academic usage (national laboratories are not considered universities). For commercial purposes, there is a commercial version of KAON2 called OntoBroker OWL. By downloading KAON2 (which is included in the WSML2Reasoner noncommercial release), you accept these license agreements.
For developers, specifically those using the WSML2Reasoner SDK release, the necessary KAON2 libs can be retrieved using the get-kaon2-jars-with-confirmation ANT target found in the build.xml included in the release. For further information, please visit the KAON2 website.
MINS is a reasoner for Datalog programs with negation and function symbols which supports well-founded semantics. MINS stands for Mins Is Not Silri. It is based on SILRI, an inference engine developed by Jürgen Angele and Stefan Decker (Institute AIFB, University of Karlsruhe). Currently MINS supports the following evaluation methods:
The MINS libraries are included in the GPL release. For further information, please visit the MINS website.
Pellet is an open source, OWL DL reasoner in Java that is developed, and commercially supported, by Clark & Parsia LLC. Pellet is based on the tableaux algorithms for expressive Description Logics (DL) and supports the full expressivity of OWL DL, including reasoning about nominals (enumerated classes).
Pellet provides standard and cutting-edge reasoning services. It also incorporates various optimization techniques described in the DL literature and contains several novel optimizations for nominals, conjunctive query answering, and incremental reasoning. There's more detailed information about the architecture of the system and its features in the Pellet Documentation.
Pellet is integrated into the WSML2Reasoner framework in order to provide DL reasoning support for WSML-DL ontologies (via translation from WSML-DL to OWL-DL). The Pellet libraries are included in the LGPL release. For further information, please visit the Pellet website.