Skip to main content
notice

Doctoral Thesis Defense: Zixi Quan

April 23, 2019
|


Speaker: Zixi Quan

Supervisor: Dr. V. Haarslev

Examining Committee: Drs. W. Du, D. Goswami, C. Wang, Y. Yan,
S. Samuel Li (Chair)

Title: A Framework for Parallelizing OWL Classification in
Description Logic Reasoners

Date: Tuesday, April 23, 2019

Time: 10:00 a.m.

Place: EV 3.309

ABSTRACT

The Web Ontology Language (OWL) is a widely used knowledge representation language for describing knowledge in application domains by using classes, properties, and individuals. Ontology classification is an important and widely used service that computes a taxonomy of all classes occurring in an ontology. It can require significant amounts of runtime, but most OWL reasoners do not support any kind of parallel processing.

This thesis reports on a black-box approach to parallelize existing description logic (DL) reasoners for the Web Ontology Language. We focus on OWL ontology classification, which is an important inference service and supported by every major OWL/DL reasoner. To the best of our knowledge, we are the first to propose a flexible parallel framework which can be applied to existing OWL reasoners in order to speed up their classification process. There are two versions of our methods discussed: (i) the first version implements a novel thread-level parallel architecture with two parallel strategies to achieve a good speedup factor with an increasing number of threads, but does not rely on locking techniques and thus avoids possible race conditions. (ii) The improved version implements an improved data structure and various parallel computing techniques for precomputing and classification to reduce the overhead of processing ontologies and compete with other DL reasoners based on the wall clock time for classification.

In order to test the performance of both versions of our approaches, we use a real-world repository for choosing the tested ontologies. For the first version of our approach, we evaluated our prototype implementation with a set of selected real-world ontologies. Our experiments demonstrate a very good scalability resulting in a speedup that is linear to the number of available cores. For the second version, its performance is evaluated by parallelizing major OWL reasoners for concept classification. Currently, we mainly focus on comparison with
two popular DL reasoners: Hermit and JFact. In comparison to the selected black-box reasoners, our results demonstrate that the wall clock time of ontology classification can be improved by one order of magnitude for most real-world ontologies in the repository.




Back to top

© Concordia University