The NNPDF collaboration has pioneered the usage of Machine Learning (ML) techniques in order to extract PDFs from a finite set of data, a typical example of an inverse problem. In this talk we ...