We are happy to announce the next release of DL-Learner, a tool for learning OWL class expressions from examples and background knowledge. It extends Inductive Logic Programming (ILP) to Description Logics and the Semantic Web. The tool has matured over the past 3 years and is meanwhile used in a number of applications. Some features of this release are:

• support for OWL API 3 and OWL 2
ORE (ontology repair and enrichment) tool based on DL-Learner algorithms (soon to be migrated to an own project)
• several new heuristics, e.g. generalised F-Measure, and efficient stochastic heuristic approximation methods
• learning algorithms for the EL description logic
• support for hasValue construct in combination with string datatype
• support for refining existing definitions (instead of learning from scratch) for CELOE ontology engineering algorithm
• support for direct Pellet 2 integration and reasoners connected via OWLlink
• more unit tests, bug fixes and features

DL-Learner can be used to:

• solve general supervised Machine Learning problems using ontologies as background knowledge (given as OWL files, SPARQL endpoints, etc.), e.g. it was used to predict whether chemicals can cause cancer
• help knowledge engineers by learning definitions and subclass axioms (see the Protege plugin and OntoWiki plugin)
• generating user recommendations when browsing knowledge bases

I’d like to thank all contributors, in particular active developers and everyone who sent us valuable feedback.

The tool can be be downloaded here.

Today, we released version 0.2 of the ontology repair and enrichment (ORE) tool. It is a tool for knowledge engineers to improve an OWL ontology through a wizard like repair process and uses state-of-the-art ontology debugging methods. The main feature in version 0.2 is a mode for incrementally detecting inconsistencies in large knowledge bases available as SPARQL endpoints. Using this mode, we have detected inconsistencies and computed justifications in DBpedia Live and OpenCyc. Previously, both knowledge bases were too large to compute justifications on standard hardware to the best of our knowledge, i.e. inconsistencies could not be fixed efficiently. A screencast illustrates this process for the case of DBpedia Live. Thanks to Lorenz Bühmann for his work on ORE.

ORE Homepage | Download | Screencast | AKSW Homepage

The set of tools released by the AKSW research group has a new member: ORE. ORE stands for ontology repair and enrichment. It is a tool for knowledge engineers to improve an OWL ontology through a wizard like repair process. It uses state-of-the-art methods for fixing inconsistencies and suggesting additions to an ontology, while still being efficient for small and medium sized ontologies. A screencast, which demonstrates its functionality, is available. As usual, the tool is available as open source, so you are free to download it. More information is available on the ORE wiki page. While the initial release already offers some quite powerful features, we plan to extend the tool in the mid term future with full support for knowledge bases available as Linked Data or SPARQL endpoints (as opposed to OWL/RDF files) and the detection of many common modelling errors. Thanks to Lorenz Bühmann for implementing ORE in his master thesis.

Today, we released a major update of the DL-Learner plugin for Protege 4, the popular ontology editor. The plugin allows you to get suggestions for axioms you may want to add to your OWL ontology by analysing the instance data using the DL-Learner framework. A screencast is available to show you how it works. The plugin is installable via the standard Protégé plugin mechanism (Protege 4 stable release required). This version allows to fine tune the language features you want to use and displays the length of expressions, which have been searched. It also fixes bugs related to inconsistency handling. More information can be found on the plugin wiki page. Enjoy!

Today, we released DL-Learner Build 2009-05-06. DL-Learner is a tool for learning OWL class expressions from examples and background knowledge. It extends Inductive Logic Programming (ILP) to Description Logics and the Semantic Web. Some notable features in this release are:

• a new learning algorithm (CELOE) designed specifically for extending OWL ontologies
• a Protege plugin using CELOE
• a manual for getting started using DL-Learner
• performance improvements through stochastic methods
• more learning examples, unit tests, code quality improvements

DL-Learner can be used to:

• solve general supervised Machine Learning problems using ontologies as background knowledge (given as OWL files, SPARQL endpoints, etc.), e.g. it was used to predict whether chemicals can cause cancer
• help knowledge engineers by learning definitions and subclass axioms (see the Protege plugin and another one for OntoWiki is in progress)
• generating user recommendations when browsing knowledge bases

I’d like to thank all contributors, in particular active developers and everyone who sent us valuable feedback.