Data-driven methods in engineering

The course is composed of 12 hours of lectures and aims to provide the main elements regarding "machine learning" techniques for modeling complex dynamic systems.
The course is divided into the following main phases
1) Introduction. The singular-value decomposition (SVD)
2) Compressed sensing, optimal sensor placement
3) Robust principal component analysis and dynamic-mode decomposition (DMD)
4) Implementation of modal decompositions
5) Deep-learning applications and developments
Theoretical/practical exercises are proposed during the course. Applications are also developed for the application of "machine learning" techniques to simplified models. At the end of the course, a project is assigned which serves as a final assessment./ Il corso è articolato in 12 ore di lezione frontale e ha lo scopo di fornire i principali elementi riguardanti tecniche “machine learning” per la modellazione di sistemi dinamici complessi.
Il corso è suddiviso nelle seguenti fasi principali
1) Introduction. The singular-value decomposition (SVD)
2) Compressed sensing, optimal sensor placement
3) Robust principal component analysis and dynamic-mode decomposition (DMD)
4) Implementation of modal decompositions
5) Deep-learning applications and developments
Durante il corso vengono proposti esercizi teorico/pratici. Vengono inoltre sviluppati degli applicativi per l'applicazione di tecniche “machine learning” a dei modelli semplificati. Al termine del corso viene assegnato un  progetto che funge da verifica finale.