BIO
My name is Alex Vigani, I am from Predore (BG), and I was born in 1999. I completed both my Bachelor’s and Master’s degrees in Computer Engineering at the University of Brescia, obtaining my Master’s degree in 2024.
For my thesis, and subsequently during a research fellowship and a research grant, I focused on knowledge representation through ontologies and knowledge graphs for the integrated access to heterogeneous data sources. My work involved the design and development of knowledge graph construction pipelines and the application of reasoning techniques to support semantic integration and advanced querying.
This research was carried out in collaboration with the biopharmaceutical company Dompé and the startup Prometheux.
RESEARCH FIELD
My research aims to investigate how semantic data integration, Knowledge Graph construction, and constraint-based validation at different levels can be combined to support trustworthy and explainable decision support systems. The core idea is that explanations—whether derived from query answering mechanisms or Large Language Models—should not be generated merely as post-hoc textual justifications, but should instead be grounded in the structure of the semantic model itself, leveraging ontological axioms, reasoning mechanisms, and explicit constraints (e.g., SHACL shapes).
In this perspective, system outputs can be systematically validated against domain rules, inconsistencies can be detected at early stages, and explanations can be produced as verifiable traces that link recommendations directly to both the ontology and the underlying data sources. Moreover, validation plays a crucial role in the maintenance and evolution of Knowledge Graphs, enabling efficient updates and preserving semantic coherence even when data sources change rapidly or new information becomes available.
This research is motivated by real-world applications in critical domains such as food and health recommendation and biomedical decision support (e.g., drug discovery), where incorrect, inconsistent, or poorly justified suggestions may have serious consequences. By focusing on constraint-aware query answering and recommendation over Knowledge Graphs, my work aims to develop methods and tools that enhance Knowledge Graph quality, ensure the semantic consistency of generated answers, and produce explanations that are transparent, auditable, and firmly grounded in domain knowledge.
Tutor: Prof.ssa Anisa Rula – Prof. Devis Bianchini
Email: alex.vigani@unibs.it