No dia 6 de novembro, o Departamento de Física do ISEL promove o seminário "The AI Physicist: learning equations from data".

O evento, que decorrerá pelas 14h00 no auditório Chagas Gomes (edifício F) do ISEL, terá como orador o professor Frederico Fiúza (Departamento de Física - IST/UL e Instituto de Plasmas e Fusão Nuclear - IST/UL)
 

Abstract:

Artificial intelligence and machine learning are transforming science and engineering by accelerating discovery and innovation through their ability to analyze vast datasets, uncover complex patterns, and optimize processes.

However, this power often comes at the expense of interpretability and generalizability — two core aspects of science, and physics in particular.

In this talk, I will discuss how techniques such as sparse regression can help bridge this gap by learning governing equations directly from data, thereby fostering new theoretical insights and advancing our physical understanding.

I will illustrate these ideas with examples from plasma physics, highlighting how such methods can reveal reduced models of nonlinear dynamical systems.