 |
 |
Universite Paris Cite Seminar Series on Data Analytics Invited Seminar Talk |
 |
|
Prof. Andrea Cali, University of London (England) when: 23 May 2023, 5pm where: online (email the organizer for connection details), and in-person: room Turing Conseil, 7th floor, Universite Paris Cite, 45 Rue Des Saints Peres, Paris 75006 Abstract A Knowledge graph is a collection of interlinked descriptions of entities, objects and relations among them; in other terms, it is a conceptual model of a domain of interest. In this talk we introduce the topic of reasoning on knowledge graphs, where schema-level constructs are intended as inference rules rather than as constraints, and as such used to logically deduce new knowledge from a set of raw data that are potentially incomplete. Besides the general picture and its logic foundation, we show how this setting serves to carry out reasoning tasks on well-known formalisms, including the Entity-Relationship model. We show how Datalog+/-, an extension of Datalog that we proposed (or better re-proposed, as we used well-known constraints from database theory of the 1980s) in the first half of the 2000s, proves to be a very effective and natural language for representing knowledge graphs and reasoning on them. Short Bio Andrea Calí is Senior Lecturer at the Department of Computer Science of Birkbeck, University of London, as well as Director of Toorja Ltd, a consulting company in the area of Artificial Intelligence. He held positions at the University of Rome "La Sapienza", Free University of Bolzano, University of Oxford, Oxford-Man Institute of Quantitative Finance, RelationalAI Inc., University of Rijeka and a few others. His research interests include logic and databases, automated reasoning, knowledge graphs, machine learning and, in general, computational aspects of reasoning on knowledge bases. He is the recipient of the 2019 "ACM PODS Mendelzon Test-of-Time Award". Hosted by: Themis Palpanas List of past seminars |