Description of PhD student project
Topic: Completing and Debugging Large Ontologies and their Networks
Ontologies are a key semantic technology that is used in different computer science contexts (e.g., semantic web, data mining, machine learning) and application areas (e.g., drug discovery, materials design, health surveillance).
They are used for defining metadata or as formalized knowledge of a domain of interest. They lead to a better understanding of a domain and to more effective and efficient handling, search and integration of information.
Many initiatives have developed ontologies and mappings between ontologies.
Semantically-enabled applications need high-quality ontologies and mappings without defects and as complete as possible. In practice, this is not the case, especially in large ontologies. Incomplete and defective ontologies and
mappings may lead to deriving wrong conclusions and missing valid conclusions. Modeling defects and semantic defects are the most severe and represent well-known challenges that need to be solved for applications to fully
use the power of semantic technology. Unfortunately, current approaches for dealing with these defects do not deal well with large ontologies and cannot be extended for large ontologies.
In our group we create an understanding of and solve research problems by developing theory (what are the fundamentals and how can they used to obtain theoretical solutions?), algorithms (based on the theory, and if necessary, explain and develop a clear understanding of why theory and practice differ), tools (based on the algorithms) and experiments (e.g., performance of the algorithms and user studies for the tools).
Therefore, in this project we will formalize, study and develop algorithms and tools for dealing with completion and debugging problems for large ontologies and mappings.
The results of this project may be used in our collaboration projects with researchers in different domains, such as
- AHSO (Animal Health Surveillance Ontology); the ontology in BioPortal
- DCMD (Data-driven computational materials design)
- SPIRIT (EU project on Scalable privacy preserving intelligence analysis for resolving identities)
To work on this project it is recommended to have background in at least one of database technologies, Web technologies, artificial intelligence and HCI (i.e., having taken courses or other relevant experience). A good background in logic is a special merit. Other important properties for working on this project are good programming and problem analysis skills, and a desire to implement solutions into systems and perform comprehensive experimentations.
Related work by our group
Some of the proposed work will be extensions of former work such as
- RepOSE, a unique system for completing and debugging ontologies (since 2009)
- SAMBO, ontology alignment system (since 2006)
- Alignment Cubes, a novel ontology alignment evaluation system (2017)
- KitAMO, one of the first ontology alignment evaluation systems (2007)
Community efforts by our group
Our group (co-)organizes the Anatomy track (since 2013) and the Interactive track (since 2015) of the Ontology Alignment Evaluation Initiative (OAEI). OAEI is a yearly event for evaluating ontology alignment systems on different benchmarks. Patrick Lambrix is also steering committee member of this initiative.
Our group co-organized the International Workshop on Visualizations and User Interfaces for Knowledge Engineering and Linked Data Analytics in 2014. As a follow up we founded the VOILA series (Workshop on Visualizations and User Interfaces for Ontologies and Linked Data / International Workshop on Visualization and Interaction for Ontologies and Linked Data) in 2015 which is yearly running as an ISWC workshop. We were guest editors of a special issue on these topics for the Journal of Web Semantics.
We co-founded the International Workshop on Debugging Ontologies and Ontology Mappings series (2012-2014). The scope was extended in 2016 and we co-organized the International Workshop on Completing and Debugging the Semantic Web.
|
