Béatrice MARKHOFF - PhD in computer
sciences, Université de Franche-Comté, France
HDR in computer sciences, Université
François Rabelais de Tours, France
Head of BdTln research team from October 2018 to February 2023
Research Interests
- Knowledge Engineering and Semantic Interoperability: this topic was initiated by Cheikh A. T. Niang's PhD thesis on the generation of semantic mediators from a reference ontology. It is at the heart of my collaborations with cultural heritage disciplines. It led me to co-create in late 2014 (for 2015) the international workshop Semantic Web for Cultural Heritage, to co-edit the special issue of Semantic Web Journal, and co-chair the international workshops SWODCH 2021, SWODCH 2022 and SWODCH 2023, plus another special issue to appear in JOCCH in 2023 (the next one will be in the Semantic Web Journal again). This topic requires the design and use of domain ontologies. For Cultural Heritage, the basis for semantic interoperability is the CIDOC CRMwhich does not prevent you from building your own ontology but allows you to make your production interoperable with those of other teams, see The Heritage Digital Twin: a bicycle made for two. The integration of digital methodologies into cultural heritage research.
With the LAT and the MSH Val de Loire, who leads the consortium MASA of the TGIR Huma-Num, I am responsible for the ontologies component of the ANR SESAMES (2018-2022), I participate in the European project H2020 ARIADNEplus (2019-2022) and 4CH (2021-2023). One of my current contributions is carto4CH, a semantic web solution for building dynamic open distributed mappings of actors and competences in cultural heritage.At regional level I was responsible for the IT side of ARVIVA (2014-2016), I'm head of a group on « Semantic Web and interoperability » within RTR DIAMS and I'm involved in a project on digital twins for plains observatory.
This field will remain at the heart of my european collaborations: as a member of ARIADNE AISBL I am involved in two European projects submissions (D-SCRIBE et ATRIUM). Moreover, I am contributing to set up ECHOES.
- Web of Data Querying: this topic motivated Thanh Binh Nguyen's PhD Thesis (2014-2018). He extended the set of well-known rewriting algorithms (Chase, Chase&BackChase, PerfectRef, Xrewrite) with an efficient solution for the challenge of filtering query results according to a user context. His solution generalises the conditions for triggering the rewriting in the presence of constants in the constraint and in the query, using the notion of one-way MGU. I also consider knowledge graphs querying when contributing to OpenArchaeo, implemented by MASA with Sparna. We study knowledge graphs profiling for allowing humans to query it, and Web applications like Sparnatural to be automatically set up for a given graph. The current PhD thesis of Manon Ovide is about this topic.
- Knowledge Extraction from the Web of Data: this theme was initiated in 2016 by a collaboration with Arnaud Giacometti and Arnaud Soulet, data mining specialists. The idea is to automatically discover the features of knowledge graphs in order to apply data mining algorithms, or more generally to facilitate their exploitation. We devised C3M, a method for mining significant maximum cardinalities, and Versus, which builds from Wikidata a comparison table for a given set of entities. The current PhD thesis of Hassan Abdallah is about this topic.
- Knowledge and NLP: Vivien Léonard's PhD thesis on co-reference resolution, combining NLP and LOD, revives my interest in language, which is present in knowledge graphs and closely linked to ontologies.
- Knowledge and Gender studies: the University of Tours has joined the network of the Gender Institute, within which framework a new informal research group has been formed, in which I participate. Data and knowledge in this field are shared mainly in natural language, but LOD resources can also be used to support analysis about this topic. For example, there's an important literature on the representativeness of minorities, cultures and genders in Wikidata (or in Wikipedias), and we're working on tools to support this kind of large-scale analysis.
B. Markhoff |
université de Tours - Blois |
update: 29/02/2024 |