Department of Computer Science
PhD defence by Olivier Pelgrin

room 2.1.005
A. C. Meyers Vænge 15, 2450 København SV
30.04.2024 13:00 - 16:00
English
On location
room 2.1.005
A. C. Meyers Vænge 15, 2450 København SV
30.04.2024 13:00 - 16:00
English
On location
Department of Computer Science
PhD defence by Olivier Pelgrin

room 2.1.005
A. C. Meyers Vænge 15, 2450 København SV
30.04.2024 13:00 - 16:00
English
On location
room 2.1.005
A. C. Meyers Vænge 15, 2450 København SV
30.04.2024 13:00 - 16:00
English
On location
Abstract
The Semantic Web has greatly grown in popularity in recent years, partly due to the popularization of collaborative datasets as well as the multiplication of applications and use cases. With this increasing popularity, building and maintaining Semantic Web datasets in the popular RDF format has become more difficult for data producers. Similarly, many RDF datasets are continuously updated and changed as potential errors are corrected and new facts are added to them. Traditional solutions for RDF management are insufficient for dealing with such applications, as existing methods cannot efficiently handle the increase in data to manage caused by having multiple versions. Similarly, tracking versioning data also implies novel ways to access them, including new kinds of queries that can be run over the entire history of an RDF graph. In this PhD lecture, Olivier will present his work on providing solutions to the problem of managing large and evolving knowledge graphs. First, a novel analysis of some of the most widely used open knowledge graphs will be presented, together with an extensive analysis of state-of-the-art solutions for the management of evolving RDF datasets. Secondly, an innovative solution for in-memory indexing of RDF datasets with versioning metadata will be presented. Subsequently, a novel hybrid storage paradigm that uses multiple snapshots and delta chains to scale to much larger datasets will be introduced. And finally, this thesis explores complex query processing on versioned RDF datasets, which existing state-of-the-art solutions either do not have the capability to do or are unable to perform on real-world sized datasets.
Members of the assessment committee are Professor Fabien Gandon, INRIA SophiaAntipolis (France), Professor Andreas Harth, University Erlangen-Nuremberg (Germany), and Professor Kristian Torp (chairman), Aalborg University (Denmark). Supervisor for the thesis has been Professor Katja Hose, Aalborg University. Co-supervisor for the thesis has been Researcher Luis Galárraga, INRIA Rennes Bretagne-Atlantique (France). Moderator has been Assistant Professor Anders Schlichtkrull, Aalborg University.
All interested parties are welcome. After the defense the department will be hosting small reception in the CS hall next to the kitchen.