Prof. Dr. Ralf Möller
Software, Technology and Systems Group (STS),
Hamburg University of Technology (TUHH)
Publications
2011
Sofia Espinosa, Atila Kaya, and Ralf Möller.
Knowledge-Driven Multimedia Information Extraction and Ontology
Evolution, volume 6050 of LNCS, chapter Logical Formalization of
Multimedia Interpretation, pages 110–133.
Springer, 2011.
Bibtex entry Paper (PDF)
Abstract
The primary goal of this chapter is to present logical formalizations
of interpretation. The chapter presents pioneering work on logic-based
scene interpretation that has a strong in uence on multimedia
interpretation. Early approaches are discussed in more detail to
analyze the main reasoning techniques. More recent approaches, which
are more formal and therefore harder to understand, are referred to by
providing references to the literature such that the reader can get an
overview over the research field of logic-based media interpretation.
Oliver Gries, Ralf Möller, Anahita Nafissi, Maurice Rosenfeld, Kamil
Sokolski, and Sebastian Wandelt.
Dealing Efficiently with Ontology-Enhanced Linked Data for Multimedia.
Technical report, Hamburg University of Technology, 2011.
Bibtex entry Paper (PDF)
Abstract
In order to provide automatic ontology-based multimedia annotation for producing linked data, scalable high-level media interpretation processes are required. In this paper we shortly describe an abductive media interpretation agent, and based on a Multimedia Content Ontology we introduce partitioning techniques for huge sets of time-related annotation assertions such that interpretation as well as retrieval processes refer to manageable sets of metadata.
Oliver Gries, Ralf Möller, Anahita Nafissi, Maurice Rosenfeld, Kamil
Sokolski, and Michael Wessel.
Media Interpretation and Companion Feedback for Multimedia Annotation.
Technical report, Hamburg University of Technology, 2011.
Bibtex entry Paper (PDF)
Abstract
Companion technology supports, for instance, context-specific dialogues between a user and a computational system. We explore this technology in the context of cooperative computer-aided semantic annotation of multimedia (CASAM). Reasoning-based media interpretation exploits user input and multimedia analysis results to propose high-level media annotations (called interpretations). In this paper it is shown how queries are generated that are addressed to users of the system, with the goal to exploit answers such that the number of internal interpretations is substantially reduced. Queries are associated with so-called importance values specifying the expected degree of disambiguation for interpretation alternatives.
Ralf Möller.
Zur Rolle der Logik bei der Entwicklung Intelligenter Systeme.
KI-Zeitschrift, 25(4):309–311, 2011.
Bibtex entry Paper (PDF)
Özgür L. Özçep and Ralf Möller.
Combining Lightweight Description Logics with the Region Connection
Calculus.
Technical report, Institute for Softwaresystems (STS), Hamburg University of
Technology, 2011.
Available online at
http://www.sts.tu-harburg.de/tech-reports/papers.html.
Bibtex entry Paper (PDF)
Didier Verna and Ralf Möller, editors.
Proc. European Lisp Symposium. Hamburg University of Technology, 2011.
Bibtex entry Paper (PDF)
Sebastian Wandelt and Ralf Möller.
Islands and Query Answering for ALCHI-Ontologies.
In Proc. Third International Joint Conference on Knowledge Discovery,
Knowledge Engineering and Knowledge Management, Heidelberg, Germany,
volume 128 of CCIS, pages 224–236. Springer, 2011.
Bibtex entry Paper (PDF)
Sebastian Wandelt and Ralf Möller.
Sound and Complete SHI Instance Retreival for 1 Billion ABox
Assertions.
In A. Fokuoe, Th. Liebig, and Y. Guo, editors, Workshop on Scalable
Semantic Web Systems, pages 75–89, 2011.
Bibtex entry Paper (PDF)
Abstract
In the last years, reasoning over very large ontologies became
more and more important. In our contribution, we propose a reasoning infrastructure,
which can perform instance checking and instance retrieval
over ontologies with semi-expressive terminological knowledge and large
assertional parts.
The key idea is to 1) use modularizations of the assertional part, 2) use
some kind of intermediate structure to find similarities between individual
modules, and 3) store information efficiently on external memory.
For evaluation purposes, experiments on benchmark and real world ontologies
were carried out. We show that our reasoning infrastructure can
handle up to 1 billion ABox assertions. To the best of our knowledge
this is the first system to provide sound and complete reasoning over
SHI ontologies with such a huge number of assertions.