Bogdan Kostov

Aviation is a high risk industry where safety management is mandatory for organizations. On a global scale, safety management is hierarchical, i.e. national authorities manage safety of the sector, controlling organizations which in turn manage safety of their operations, as well as the safety of their organizational units. It is apparent that safety information management is a key factor to the success of safety management. Improving safety information management in aviation can be achieved by the systematic documentation of safety information and its seamless sharing vertically as well as horizontally through the management hierarchy. Despite attempts to tackle this problem, the industry still suffers from information management issues, e.g. usage of terminologically incompatible safety management frameworks. As a result diverse information repositories, e.g. investigation reports, safety recommendations, or audit reports managed by these frameworks, cannot be efficiently explored and compared. We apply Linked Data principles, e.g. searchability and explorability based on dataset descriptors, to help solving the problem. In this work we propose an ontology-based safety information management for the aviation industry. To achieve this we develop the Aviation Safety Ontology that is modular and covers the contexts of the organization types aerodrome, airline, air traffic management, maintenance and the state civil aviation authority. We test the appropriateness of the proposed ontology by representing safety data in terms of the ontology.

Creating applications on top of linked data and ontologies brings many difficulties. The applications are either generic (and thus not appealing to end-users), or bound to ontology structure, change of which breaks the application. We present JOPA, a tool that formalizes the contract between the application and the ontology, combining advantages of both worlds. JOPA is a persistence framework for Java applications, providing formalized object-ontological mapping, transactions, access to multiple repository contexts, and producing linked data. The system is demonstrated on a real use-case of a reporting tool that we develop for the aviation industry.

Datasets are created in the linked data ecosystem with varying data and metadata quality. Shallow (or poor-quality) metadata prevent dataset provision services (CKAN catalogs, etc.) from supporting their users in efficient search and exploration of datasets they need. In this paper we approach the problem by formalizing the dataset exploration scenario, introducing the notion of ontology-based dataset descriptors. We present a particular use-case of creating temporal dataset descriptors and their application in the field of Czech Open Data cloud. We also discuss practical problems that we came across during descriptor extraction. Experimental datasets and computed descriptors are published in a public SPARQL endpoint.

As the size of semantic data available as Linked Open Data (LOD) increases, the demand for methods for automated exploration of data sets grows as well. A data consumer needs to search for data sets meeting his interest and look into them using suitable visualization techniques to check whether the data sets are useful or not. In the recent years, particular advances have been made in the field, e.g., automated ontology matching techniques or LOD visualization platforms. However, an integrated approach to LOD exploration is still missing. On the scale of the whole web, the current approaches allow a user to discover data sets using keywords or manually through large data catalogs. Existing visualization techniques presume that a data set is of an expected type and structure. The aim of this position paper is to show the need for time and space efficient techniques for discovery of previously unknown LOD data sets on the base of a consumer’s interest and their automated visualization which we address in our ongoing work.

The MONDIS project is aimed at knowledge management in the field of monument conservation, monument damages and related interventions. The IT platform developed within the project is to be used by conservators, monument owners and other stakeholders to share their terminology, contextual knowledge, as well as experience about existing damages and their interventions.

In this paper we study the distinct count aggregation function used in queries into expressive ontologies. The main differences in this settings opposed to aggregation in relational database systems are the Open World Assumption and incomplete knowledge. We propose different interpretations useful in different practical use-cases of the distinct count function, i.e. basic count, semantic count, epistemic count and semantic tuple count some of which use knowledge derived from the ontology in order to obtain results in accordance with the Open World Assumption. We use interval semantics to model the uncertainty of the distinct count function's result induced by incomplete knowledge in the ontology. We show that interval semantics are particularly useful in aggregate queries with filtering clause and when we need the retrieval of the boundaries of the uncertainty of the distinct count function's result. We study and present relationships among the different interpretations and decidability of the semantic tuple count. We also propose a theoretical approximation of the semantic tuple count. Our results are applicable to a wide range of description logic formalisms allowing to express equality/inequality between individuals, concepts and relations, e.g.Web Ontology Language.

Web Ontology Language ontologies become more and more popular in complex domain modeling for their high expressiveness, flexibility and well defined semantics. Although query languages adequate in expressiveness to OWL reasoning capabilities were introduced before, their implementations are rather limited. In this paper, we study SPARQL-DL^NOT, an extension of one of these query languages, SPARQL-DL, and present novel evaluation and optimization techniques for efficient SPARQL-DL^NOT execution. As queries become complex easily, we also present a novel graph-based visualization that simplifies query construction and maintenance. Presented techniques and algorithms were implemented in the Pellet reasoner and in our novel Protégé plug-in OWL2Query.

High expressiveness of semantic web query languages, like SPARQL or SPARQL-DL, brings practical challenges in query author- ing and visualization for both experts and non-experts. In this paper we present the successful OWL2Query plug-in for Protégé, a simple and feature rich graphical visualization and editing tool for complex OWL 2 queries expressed in SPARQL-DL query language. Our tool is designed and implemented with the goal of inspiring and supporting users in query authoring, visualization and evaluation. The interactive query editing and the intuitive visualization are not only helping users in quick author- ing of valid queries, but also are crucial points in the efficient adoption of semantic technologies by non-experts.