Postdoctoral researcher , at Weierstrass Institute,Germany
Dr. Francesco Romor is a postdoctoral researcher at the Weierstrass Institute for Applied Analysis and Stochastics (WIAS) in Berlin, Germany. He works within the Numerical Mathematics and Scientific Computing research group led by Prof. Volker John. His expertise lies in advanced computational methods, scientific machine learning, and mathematical modeling with applications spanning from fluid dynamics to medical image analysis. Known for combining mathematical rigor with modern machine learning techniques, Dr. Romor is establishing himself as a leading voice in the next generation of scientific computing experts.
profile
Scopus
Education
Dr. Romor earned his Ph.D. in Mathematical Analysis, Modeling, and Applications from the International School for Advanced Studies (SISSA), Trieste, Italy, in 2023. His doctoral research, supervised by Prof. Gianluigi Rozza, focused on “Nonlinear Parameter Space and Model Order Reductions enhanced by Scientific Machine Learning.” He previously obtained a Master’s degree in Mathematics (2019) and a Bachelor’s degree in Mathematics (2017), both with the highest honors (110/110 cum laude) from the University of Trieste. His academic path reflects a strong and consistent foundation in theoretical and applied mathematics.
Experience
Since November 2023, Dr. Romor has been serving as a postdoctoral researcher at WIAS, where he collaborates closely with experts such as Dr. Alfonso Caiazzo. In 2023, he completed a short research visit to the Massachusetts Institute of Technology (MIT) as part of the MISTI MIT-Italy FVG Project, working with Prof. Youssef Marzouk. His professional engagements reflect a growing international footprint and recognition in high-level research environments.
Research interest
Dr. Romor’s research centers on reduced-order modeling, nonlinear parameter space reduction, scientific machine learning, and high-dimensional numerical simulations. He is particularly interested in hybridizing classical numerical methods with deep learning tools, such as convolutional autoencoders and graph neural networks, to solve complex partial differential equations (PDEs) efficiently. His recent work extends to data assimilation in biomedical applications, including modeling aortic coarctation using shape registration and neural networks. These innovations push the boundary of how science and AI can co-evolve for real-world problem-solving.
Awards
While specific awards are not listed, Dr. Romor’s profile includes significant professional recognitions. These include being selected for a prestigious research visit to MIT and multiple invitations to speak at international conferences such as ECCOMAS, SIAM CSE, and AICOMAS. Such invitations are indicators of esteem within the scientific community and recognition of his impactful contributions.
Publications
1. Friedrichs’ systems discretized with the DGM: domain decomposable model order reduction and Graph Neural Networks approximating vanishing viscosity solutions
Authors: Francesco Romor, Davide Torlo, Gianluigi Rozza
Journal: Journal of Computational Physics
Article ID: 113915
Year: 2025
2. Explicable hyper-reduced order models on nonlinearly approximated solution manifolds of compressible and incompressible Navier-Stokes equations
Authors: Francesco Romor, Giovanni Stabile, Gianluigi Rozza
Journal: Journal of Computational Physics
Volume: 524, Article ID: 113729
Year: 2025
3. Generative Models for the Deformation of Industrial Shapes with Linear Geometric Constraints: model order and parameter space reductions
Authors: Guglielmo Padula, Francesco Romor, Giovanni Stabile, Gianluigi Rozza
Journal: Computer Methods in Applied Mechanics and Engineering
Volume: 423, Article ID: 116823
Year: 2024
4. A local approach to parameter space reduction for regression and classification tasks
Authors: Francesco Romor, Marco Tezzele, Gianluigi Rozza
Journal: Journal of Scientific Computing
Volume: 99, Issue: 3, Article ID: 83
Year: 2024
5. Non-linear Manifold Reduced-Order Models with Convolutional Autoencoders and Reduced Over-Collocation Method
Authors: Francesco Romor, Giovanni Stabile, Gianluigi Rozza
Journal: Journal of Scientific Computing
Volume: 94, Article ID: 74
Year: 2023
DOI: 10.1007/s10915-023-02128-2
Conclusion
In summary, Dr. Francesco Romor exemplifies the qualities of a forward-thinking and high-impact researcher in computational science. With a strong mathematical foundation, international experience, innovative applications of machine learning, and a robust publication record, he is well-positioned for prestigious research honors and academic recognition. His work not only advances numerical methods but also connects disciplines—from automotive engineering to biomedical modeling—making him a valuable asset to the global research community.