Data and Knowledge Group

― Knowledge Representation and Reasoning

KARMA

Overview

KARMA is a prototypical implementation of a combined approach to answering conjunctive queries over ontologies specified in the DL ELHO.

Given an ELHO ontology, an ABox, and a conjunctive query, KARMA computes the answers to the query over the ontology plus the data using a three-step process:

  1. In a preliminary step, the ontology is transformed into a datalog program which captures certain, relevant, consequences of the ontology.
  2. By using HermiT as a datalog engine, KARMA evaluates the query over the datalog program plus the data stored in the ABox.
  3. The result of the second step is filtered to eliminate unsound answers.

The third step is necessary because, to ensure termination, the datalog program computed in the first step is not equivalent to the given ontology and it may imply facts that do not follow from the ontology and the data; however, this is done in a way that allows for the third step.

KARMA is open-source and released under LGPL. The system and the evaluation data is included in the downloaded zip file.



Download

Using KARMA

Currently, KARMA is only available with a command-line interface.

Researchers

Giorgio Stefanoni, Boris Motik and Ian Horrocks

Sponsors

UK Engineering and Physical Sciences Research Council

Alcatel-Lucent Bell Labs

Key Publications


Giorgio Stefanoni, Boris Motik, and Ian Horrocks. Introducing Nominals to the Combined Query Answering Approaches for EL. AAAI 2013. To appear.