Logo Università degli Studi di Milano



Combining Large Language Models and Knowledge Graphs for Explainable AI

Date & time: 08/09/2023, 11:30-12:30
Room:  Room 3016 (Lab. Laurea Magistrale) at the 3rd floor, via Celoria 18
Speaker: Ismael Sanz Blasco
Organizer: Marco Mesiti


Large Language Models (LLMs) and Knowledge Graphs represent two of the most impactful advancements in artificial intelligence and data science. While LLMs excel in natural language understanding and generation, offering a more intuitive interaction with users, Knowledge Graphs provide a structured and semantically rich representation of domain-specific knowledge. Each technology has its own set of strengths and limitations, but they share the common goal of making information more accessible and useful.

The integration of LLMs with Knowledge Graphs offers an unprecedented opportunity to achieve a higher degree of intelligence and explainability in AI systems. Large Language Models serve as a versatile interface for human-computer interaction, allowing for complex queries in natural language. Meanwhile, Knowledge Graphs act as a robust backbone, facilitating real-time, context-sensitive, and factual responses. When combined, these technologies can dynamically generate not only precise answers but also provide the rationale behind them, thus advancing the cause of explainable AI.

A case study based on the project "Explainable Artificial Intelligence for Healthy Aging and Social Wellbeing" (XAI4SOC) will be presented. XAI4SOC, currently in development, employs a three-tiered approach of Monitoring, Interpretation, and Intervention aimed at two vulnerable demographics: the elderly at risk of dementia and adolescents.

 Short bio:

Ismael Sanz is an associate professor (Titular) at the Computer Science and Engineering department at Universitat Jaume I (UJI). Previously he worked in industry and at Vrije Universiteit Brussel, Belgium. His teaching is focused mainly on Databases and Information Systems. He is a member of the TKBG research group, and his research interests include the processing of semi-and unstructured data, semantic technologies and information retrieval. He has led several research projects, and has contributed to high-impact international journals and conferences. He is a co-founder of the startup SemanticBots.

04 settembre 2023
Torna ad inizio pagina