Mardi 25 février à 11h, en amphi Fermi / IUSTI
Abstract: Inverse Problems has emerged as a transformative force, reshaping countless industries—from healthcare to finance—by automating complex tasks, extracting insights from massive datasets, and making predictions with remarkable accuracy. In Machine Learning (ML), information is extracted from data in order to yield models capable of generalizing the underlying physics. As for the Inverse Problem, the general models are typically replaced by first principles-based ones, so that one seeks to reconstruct information regarding input parameters, such as boundary and initial conditions, physical properties and so forth. This seminar presents a showcase of applications of such methods in Mechanical Engineering and other areas, including heat flux measurements in hazardous scenarios, thermal management of electronics, characterization of novel materials and even some biomedical applications in oncology and epidemiology.