Reasoning for Ontology Engineering and Usage - ISWC 2008
Engineering and using large OWL ontologies is a complex task for
which impressive tool support has recently been developed. Since OWL
ontologies are based on logic and involve entailments, this tool
support may involve reasoning, e.g., for query answering, inference explanation,
etc. Our tutorial provides a comprehensive overview over such tool
support from a user perspective, and explains the benets of automated
reasoning for the user.
Slides and Videos
Slides for the tutorial can be downloaded in PDF format:
Videos are available here.
Participants
This tutorial is targeted at participants with a
basic understanding of OWL and possibly a little experience in designing or using ontologies.
Tutorial Outline
This tutorial is a
full day tutorial which is structured into four parts. In all
four parts, we will use real-life examples and ontologies.
Introduction
We will provide a brief introduction to OWL, in fact OWL2,
and the underlying Description Logic, clarifying the semantics and
providing examples to help the understanding of this admittedly complex
formalism. In particular, we will discuss common misunderstandings
around OWL and OWL2, explain the open world assumption, inferences, and
the functionality of reasoners. We will use the RacerPro reasoner to demonstrate the benefit of using reasoning for query answering over ontologies. Scalability issues with respect to expressive ontologies as well as huge assertional knowledge bases are discussed.
A Bottom-up Approach to Designing Ontologies
We will explain the bottom-up approach to designing ontologies. In
particular, we will show how one can use example instances of a class
to be defined in order to automatically generate an expression that
summarizes the commonalities of these instances. This approach is
particular useful to support ontology designers who are domain experts
but not fluent in OWL and to ensure that the classes defined cover all
aspects of that class. We will use the Sonic tool and RacerPro reasoner, and lead the user through examples that illustrate the general approach and the tool.
Understanding and repairing inferences
The first two parts will have convinced the audience of the need for
tool support to help the user understand (intended and) unintended
inferences made by the reasoner and to help the ontology engineer to
repair unintended such inferences. Protege 4 will be used to demonstrate tools that have been developed to
generate explanations for these inferences and to repair unintended such
inferences will be presented. The participants will be able to gain hands-on experience with these tools.
Note that modularisation tools and techniques will also be covered in this section.
- History of explanation in OWL ontology tools
- OWL Entailments. Brief recap and examples of entailments that can arise
- Undesirable entailments (unsatisfiable classes, inconsistent ontologies) ~ meaning and terminology
- Examining unsatisfiable classes
- Tracing unsatisfiable classes through an ontology
- Pinpointing root unsatisfiable classes (leads on to justifications)
- Justifications
- Introduction to justifications (theory)
- Methods of computing justifications
- Viewing and understanding justifications for entailments
- Justifications for unsatisfiable classes
- Justifications for arbitrary entailments
- Precise justifications
- Repairing an ontology (removing undesired entailments)
- Introduction to repair tool
- Metrics for repair
- Devising and executing a repair plan
- Outlook
For this part of the tutorial we will use two ontologies.
Data Integration through Ontologies
Finally, we show the benefits of using OWL ontologies and reasoning for
data integration, and explain the underlying assumptions and principles.
We will use a Protege plugin, OBDA,
to connect the terms in an OWL ontology to external data sources, e.g.,
standard relational databases. In addition, we will use a specific
reasoner, Quonto, to answer complex queries over these data sources and
the ontology in a way that faithfully reflects the OWL axioms in the
ontology and the connections established. For parts of this tutorial,
the participants can follow along on their computers. In general, we
will encourage active participation and involve the audience so as to
deepen understanding and to relate the content of the tutorial with the
participant's experience.
- The DL-Lite family of DLs:
- The DL DL-LiteA and its relationship to OWL
- Linking data to ontologies:
- The impedance mismatch problem;
- Connecting DL-Lite ontologies to relational
data: A mapping language for solving the
impedance mismatch problem.
- Ontology-based data access and data integration
through DL-LiteA ontologies:
- Framework and semantics
- Query answering
- Practical experience:
- Use of the Protege plugin OBDA to link external
data sources to OWL (in fact DL-LiteA) ontologies
- Connecting ODBA to QuOnto, a reasoner for DL-LiteA,
to answer complex queries over these data sources
Requirements
Participants should have downloaded and installed
Protege 4.
Presenters
Diego Calvanese is
associate professor at the KRDB Research Centre of the Faculty of
Computer Science, Free University of Bozen-Bolzano since 2003. His
research interests include formalisms for knowledge representation and
reasoning, ontology languages, Description Logics, conceptual data
modeling, and data integration. He participated in several national and
international research projects, and he is the co-ordinator of the EU
STREP FET Project TONESThinking ONtologiES . He is involved in the
organization of various workshops and is PC co-chair of the Int.
Conference on Web Reasoning and Rule Systems RR 2008. He has lectured
(both in Italian and in English) several undergraduate and graduate
courses in computer science and computer engineering. He has given
tutorials at ESSLLI 2003 and at ISWC07.
Giuseppe De Giacomo
is full professor in Computer Science and Engineering since 2006. Since
1991, he has been carrying out his research activity at the Sapienza
Universit di Roma, where he is currently part of a research group on
Software Engineering, Databases and Articial Intelligence. His main
research interests are in data modeling, information integration,
service integration, knowledge representation and reasoning, reasoning
about actions, cognitive robotics, and ob ject-oriented methodologies.
He is currently involved in national and international research
projects on information integration, knowledge representation and
reasoning, and e-service modeling and composition. Giuseppe regularly
teaches courses for undergraduate, graduated and PhD students. He gave
several tutorials in international conferences and schools, including
tutorials on information integration and service composition at ESSLLI
2003 and 2005, ICSOC 2004, WWW 2005, INFWEST 2007.
Matthew Horridge is a
research associate and PhD student at the University of Manchester, UK.
He has five years of experience in developing tools to create and
maintain OWL ontologies. Some of the popular tools that he has written
are OWLViz, OWLDoc, and The Manchester OWL Syntax Editor. He also led
the design and implementation of the Protege 4 OWL Ontology Editor,
which was written from the ground up. He also works on and maintains
the OWL API, which is a high level API for creating and manipulating
OWL ontologies. The API was the first API for dealing with OWL2
ontologies. Matthew has wide experience in teaching OWL, mainly due to
his involvement in numerous versions of the Pizza tutorial
(http://www.co-ode.org/).
Ralf Möller
is Professor for Computer Science at Hamburg University of Technology
(since 2003). His research goals encompass the development practical
inference algorithms for embedding description logic systems into
software engineering and web technology. Together with Prof. Volker
Haarslev (Concordia Univ. Montreal) and Michael Wessel he is the
principal architect of the description logic reasoner Racer, which is
being used as a core engine for building ontology development tools as
well as agent systems for the semantic web by many research groups all
around the world. Ralf was the co-organizer of several international
workshops on description logics and is the author of numerous workshop
and conference papers as well as several book and journal contributions
in this research area. Ralf has broad experience in teaching to
graduate and undergraduate students, and has presented tutorials to
various partners from industry.
Anni-Yasmin Turhan
is a post-doctoral researcher at the TU Dresden. She received her Ph.D.
in computer science for her thesis on non-standard reasoning services
for augmenting ontologies with new classes. Her research interests are
non-standard reasoning services and implementation of DL systems. She
is an active member of the DIG 2.0 interface standardisation group. She
has worked in several industrial co-operation projects where she gave
tutorials on state of the art DL reasoning technology to partners from
industry. Furthermore she is lecturer on this year's DAAD summer school
on applications of logics.