Hybrid Technologies for Transparency – Structured Knowledge and Generative Models

The growing demand for transparency in automated decision-making processes is driving the development of hybrid solutions that combine generative models with structured knowledge representations. This lunch seminar will present a hybrid architecture that integrates Generative Models (LLMs), Knowledge Graphs (KG), and semantic rules (SHACL) to support real-world scenarios such as regulatory compliance, conflict-of-interest detection, and multiperspective analysis of private and public actors.
By analyzing open data from the City of Chicago, the seminar will show how the incremental construction of a Knowledge Graph – from domain understanding to data fusion and model definition – enables the integration of heterogeneous sources and the automatic identification of complex entities and relationships. The use of LLMs and similarity algorithms for enrichment and analysis, combined with semantic representation techniques, makes it possible to achieve contextual explainability, automated validation, and transparent reporting.
The result will be an information pipeline oriented towards transparency, capable of supporting public decision-making responsibly, by leveraging the synergies between generative AI and semantic technologies.

Speaker: Giuseppe Futia (Fellow, Nexa Center)

Biography
Giuseppe Futia, PhD, is a Data Scientist and author specializing in Knowledge Graphs, Large Language Models, and Graph Neural Networks. He is a Fellow at the Nexa Center for Internet & Society at Politecnico di Torino. With over ten years of experience across academia and industry, he works on development, communication, and research in the field of Artificial Intelligence (AI), with a particular focus on integrating structured knowledge into intelligent systems. Giuseppe actively contributes to the Knowledge Graph community, regularly speaking at leading conferences such as the Knowledge Graph Conference (KGC) and Connected Data London (CDL), where he shares insights on the development and future perspectives of these technologies. His scientific articles, published in journals and international conference proceedings, focus on knowledge representation, machine learning, and artificial intelligence. Among his goals is making AI systems more transparent, adaptable, and capable of reasoning, while continuously exploring solutions that enhance the role of human knowledge. Early in his career, he worked as a freelance journalist for La Stampa and as Communications Manager at the Nexa Center, experiences that still shape his approach to science communication.

The seminar "Hybrid Technologies for Transparency – Structured Knowledge and Generative Models" is the 124th event of the Nexa Lunch Seminar series and will be held on September 24, 2025, at 1:00 PM. You can attend through the following options:

• In person: Nexa Center for Internet & Society, Politecnico di Torino, Via Boggio 65/a (1st floor)
• Online: click here.

For further information, please contact Valeria Bergantino, and for all the details on the event, please visit the official page.