Digital twins are virtual representations of real-world products, systems, or processes, enabling simulation, integration, testing, monitoring, and maintenance. They play a pivotal role in optimizing complex systems across a wide range of domains, from industrial manufacturing and energy to environmental monitoring and healthcare.
The Engineering Digital Twin EDT program, funded by the France 2030 investment plan, is a national initiative aimed at advancing the foundations of digital twin engineering in France and Europe [2]. By bringing together leading academic and industrial partners, EDT seeks to strengthen the bases for the design, use, and deployment of digital twins, addressing key open challenges in model hybridization, composability, development methodologies, digital coupling, and human–twin interaction.
A key promise of digital twins is to enable stakeholders to explore what-if scenarios: evaluating alternative configurations, behaviours, or interventions while the system is running, so as to improve performance, reliability, and adaptability. However, enabling such exploratory interactions remains challenging in practice. Digital twins leverage diverse and heterogeneous knowledge about territories and related data. It is therefore not possible to rely solely on a single unifying model, but rather it is necessary to manage the interactions between heterogeneous representations of knowledge and various viewpoints.
The semantic web provides a set of technologies for representing and reasoning about knowledge on a web scale [3]. These technologies include RDF for representing knowledge graphs and OWL for formalising ontologies. In order to manage the heterogeneity of knowledge, alignments between ontologies make it possible to express the relationships between concepts (classes and properties) from different ontologies. At the data level, linking keys define sufficient conditions for identifying resources from different knowledge graphs.
Digital twins rely on the integration of multiple heterogeneous models and data sources, such as sensor observations, simulation models, geographic information systems, and domain knowledge bases. Ontology alignment will therefore play a central role in reconciling these heterogeneous representations and enabling consistent interpretation and integration of the data they produce.
With rapid advances in neural AI, work in the semantic web, historically based on symbolic AI (knowledge representation and reasoning), is moving towards neuro-symbolic AI [2,4]. Neuro-symbolic AI aims to combine the strengths of machine learning (noise robustness, statistical generalisation) with those of symbolic AI (explainability and logical reasoning).
The objective of this thesis is to study the contribution of neuro-symbolic to ontology alignment [5] and data linking [6] in the context of France’s digital twin.
[1]
Breit, A., Waltersdorfer, L., Ekaputra, F. J., Sabou, M., Ekelhart, A., Iana, A., Paulheim, H., Portisch, J., Revenko, A., Teije, A. T., & Harmelen, F. V. (2023). Combining Machine Learning and Semantic Web: A Systematic Mapping Study. https://doi.org/10.1145/3586163
[2]
Benoît Combemale, Pascale Vicat-Blanc, Arnaud Blouin, Hind Bril El Haouzi, Jean-Michel Bruel, Julien Deantoni, Thierry Duval, Sébastien Gérard, & Jean-Marc Jézéquel (2025). Engineering Digital Twins: A Research Roadmap. EDTconf 2025 - 2nd International Conference on Engineering Digital Twins. https://inria.hal.science/hal-05223776
[3]
Hitzler, P., Krötzsch, M., & Rudolph, S. (2009). Foundations of Semantic Web Technologies.
[4]
Janowicz, K., Hitzler, P., Bianchi, F., Ebrahimi, M., & Sarker, M. K. (2020). Neural-symbolic integration and the Semantic Web. https://doi.org/10.3233/SW-190368
[5]
Jradeh, C. K., Raoufi, E., David, J., Larmande, P., Scharffe, F., Todorov, K., & Trojahn, C. (2025). Graph Embeddings Meet Link Keys Discovery for Entity Matching. https://doi.org/10.1145/3696410.3714581
[6]
Sousa, G., Lima, R., & Trojahn, C. (2025). Results of CMatch in OAEI 2025. https://ceur-ws.org/Vol-4144/om2025-oaei-paper3.pdf
[7]
Sousa, G., Lima, R., & Trojahn, C. (2026). Survey on embedding methods applied to ontology matching.
