My research activities are part of a scientific continuity centered on data analysis, knowledge extraction, and their formalization, with the primary objective of understanding and structuring the underlying mechanisms of complex systems. This guiding line, developed throughout my career and consolidated in my Habilitation to Supervise Research (HDR), has progressively expanded to integrate issues related to semantic interoperability, systems engineering, and artificial intelligence.

These themes lie at the intersection of artificial intelligence, knowledge engineering, and systems modeling, with particular attention paid to transforming heterogeneous data into actionable knowledge. One of the central challenges of my work consists in moving from raw data—often massive, distributed, and unstructured—to formal representations that enable interpretation, sharing, and use in complex decision-making processes. This issue is now at the core of transformations in industrial and socio-technical environments, where systems must be capable of simultaneously handling data, information, and knowledge.

Within this framework, my research is structured around three complementary axes.

The first axis concerns the analysis and extraction of knowledge from heterogeneous data, leveraging techniques from data mining, natural language processing, and machine learning. These approaches aim to identify structures, relationships, and regularities in data, particularly in multi-relational and multi-source contexts.

The second axis focuses on knowledge formalization, notably relying on methods derived from Formal Concept Analysis (FCA) [Priss, 2006] and Relational Concept Analysis (RCA) [Rouane-Hacene, 2013]. These approaches make it possible to structure knowledge in terms of concepts and relationships, ensuring semantic consistency and interpretability. They also provide a framework for studying semantic interoperability between systems by linking distinct data contexts and extracting enriched knowledge based on explicit knowledge representations such as knowledge graphs [Hogan, 2021]. In this perspective, my work contributes to extending these methods to address complex multi-relational data and the challenges posed by distributed environments.

The third axis concerns the exploitation of knowledge in complex systems, particularly in contexts where decision-making relies on the integration of multiple information sources. The objective here is to develop approaches that connect knowledge formalization, systems modeling, and artificial intelligence, in order to design systems capable of reasoning, explaining their decisions, and adapting to evolving contexts.

These three axes are part of a unified vision in which knowledge extraction, formalization, and exploitation constitute different stages of the same process. This vision finds particularly strong expression in the context of cyber-physical systems (CPS), where data from sensors and distributed systems must be transformed into actionable knowledge for monitoring, decision-making, and optimization. In such environments, the integration of perception, communication, learning, and reasoning mechanisms enables the emergence of a new generation of intelligent systems, capable of combining computational capabilities with contextual understanding.

The applications of this work are mainly situated in domains such as Industry 4.0, Agriculture 4.0, and, more recently, systems related to healthcare and infrastructures. In the industrial context, data analysis and formalization contribute to improving production processes, product quality, and decision-making. In the agricultural domain, they support more precise management of crops and resources by integrating spatial, temporal, and environmental data. Although distinct, these domains share common challenges related to the management of complex systems, data interoperability, and process optimization, and can be addressed through similar methodological frameworks.

In recent years, these themes have been developed and validated through the supervision of several doctoral works, focusing in particular on semantic annotation, process interoperability, sustainability formalization, invariants of cyber-physical systems, and multi-relational knowledge extraction. These contributions have progressively built a coherent scientific framework in which different approaches complement and enrich each other.

Today, this research is evolving toward a stronger integration with issues related to complex industrial systems and hybrid artificial intelligence. The objective is to go beyond purely data-driven approaches by proposing systems capable of explicitly integrating knowledge into their operational processes. This evolution is part of a broader perspective aimed at developing cognitive systems capable of linking data, models, and knowledge, and supporting explainable, robust, and interoperable decision-making processes.

Thus, my research themes can be summarized as the development of methods and models that transform heterogeneous data into formalized knowledge, and the use of this knowledge to improve interoperability, understanding, and decision-making in complex systems. This approach now constitutes a scientific foundation for the development of next-generation intelligent systems, particularly in industrial and socio-technical contexts characterized by complexity and heterogeneity.