FIS Starting Grant awarded to Andrea Manzoni

Andrea Manzoni has been awarded a FIS Starting Grant for his proposal on Deep Learning for the real-time approximation of partial differential equations (DREAM). Launched with a first call in 2021, the Italian Science Fund (FIS) aims at promoting the development of fundamental research on the model of ERC – the most prestigious program dedicated to basic research at the European level – supporting pioneering projects across all scientific fields.

Started on April 1st, 2024, the DREAM project will deal with the mathematical and computational setting of accurate and efficient reduced order models for the numerical approximation of systems governed by parametrized partial differential equations in real-time. Foreseen methodologies will leverage the accuracy of physics-based techniques and the versatility of data-driven methods based on neural networks. These strategies are motivated by several applications in Engineering, ranging from the simulation-assisted design & calibration of micro-electro-mechanical systems, to the optimal transport of nano-particle swarms in personalized medicine. These are just two examples of problems where, besides numerical approximation, efficient strategies are required to handle PDE-constrained optimization and uncertainty quantification problems.

Funded with 1 M€, the DREAM project is one of the 9 starting grants awarded at the national level, out of 462 submitted proposals, in the Physical Sciences and Engineering (PE) macrosector. Hosted by Politecnico di Milano, DREAM will last for 5 years and will involve a (dream) team of PhD students, Postdoctoral Fellows, and Young Researchers at the forefront in Computational Sciences and Engineering.