Recent advances in machine leing are currently influen-cing the way we gather data,recognize patts,and build pre-dictive models across a wide range of scientific disciplines.No-ticeable successes include solutions in image and voice recogni-tion that have already become part of our everyday lives,mainly enabled by algorithmic developments,hardware advances,and,of course,the availability of massive data-sets.Many of such pre-dictive tasks are currently being tackled using over-parameter-ized,black-box discriminative models such as deep neural net-works,in which theoretical rigor,interpretability and adherence to first physical principles are often sacrificed in favor of flexibil-ity in representation and scalability in computation.