23 January 2026
Machine Learning (ML) techniques have found great success in modeling dynamical and physical systems, including Partial Differential Equations (PDEs) and Physics-Informed Neural Networks (PINNs) emerged as an effective paradigm to learn the dynamics of a PDE. Unfortunately, PINNs suffer from optimization issues which can lead to poor generalization.
In this study, EMERGE partners from University of Pisa propose Derivative Learning (DERL), a supervised approach that models physical systems by learning their partial derivatives. The authors also leverage DERL to build physical models incrementally, by designing a distillation protocol that effectively transfers knowledge from a pre-trained model to a student one.
Read the paper in the link below.
More Information
