Dr. Aldo Glielmo, Banca d’Italia, Applied Research Team of the IT Directorate
Chairpersons: Prof. Giorgia Oggioni (University of Brescia), Prof. Maria Elena De Giuli (University of Pavia)
When: Monday, May 20th, 2024, 10:00 AM
Where: Sala della Biblioteca, San Faustino Building and Google Meet
Artificial intelligence, and computation more generally, are rapidly becoming ubiquitous within every scientific discipline, and in everyday life. The Applied Research Team (ART) of the Bank of Italy (https://www.bankit.art/) is a computer-science oriented research unit with the mission of identifying and harnessing the potential benefits of this rapid technological expansion for the institution’s objectives and operations. In my presentation, I will highlight some recent contributions of ART in computational economics. I will begin by overviewing research dedicated to constructing large-scale “agent-based” simulation models, and to calibrating such models with real data. In doing so, I will also illustrate Black-it, a calibration software recently released in open-source by Bank of Italy [Benedetti et al., JOSS, 2022] (https://github.com/ bancaditalia/black-it). Then, I will showcase two uses of machine learning techniques to enhance economic analysis: the application of reinforcement learning to analyse and extend agent-based simulations [Glielmo et al., ICAIF, 2023], and the application of a newly designed tool called the “Information Imbalance” [Glielmo et al., PNAS Nexus, 2023] for detecting hidden relationships between time series variables.