Nicola Farenga

MOX, DMAT, Politecnico di Milano
Office 217, Building 14 (Nave) - 2nd floor,
Piazza Leonardo da Vinci, 32,
20133, Milan, Italy
nicola.farenga at polimi.it

I am a Ph.D. candidate in Data Analytics and Decision Sciences at the Department of Mathematics, Politecnico di Milano, supervised by Prof. Andrea Manzoni, and a member of the Modeling and Scientific Computing (MOX) Laboratory. Previously, I earned a B.Sc and an M.Sc. in Mathematical Engineering at Politecnico di Milano, specializing in Computational Science and Computational Learning.

My research interests span continuous-time deep learning models, such as neural differential equations and flow-based generative models, and their use as surrogates of physical systems. My work has focused on nonlinear reduced order modeling for parametric time-dependent PDEs, addressing the formalization of latent dynamics learning and of the notion of continuous-time approximation. More recent work introduces Velocity Matching, a simulation-free framework for learning neural ordinary differential equations.


Research

For an updated list of publications, please visit my Google Scholar profile.

  1. Learning to evolve by matching velocities [arXiv, github]
    N. Farenga, A. Manzoni
    Preprint, 2026
  2. On latent dynamics learning in nonlinear reduced order modeling [doi, arXiv]
    N. Farenga, S. Fresca, S. Brivio, A. Manzoni
    Neural Networks, Volume 185, 107-146, 2025
  3. Agglomeration of polygonal grids using graph neural networks with applications to multigrid solvers [doi, arXiv, github]
    P.F. Antonietti, N. Farenga, E. Manuzzi, G. Martinelli, L. Saverio
    Computers & Mathematics with Applications, Volume 154, 45-57, 2024
  4. Neural latent dynamics models [neurips, openreview, poster]
    N. Farenga, S. Fresca, A. Manzoni
    The Symbiosis of Deep Learning and Differential Equations II, NeurIPS, 2022

Teaching


Talks & Presentations

  1. Model Reduction and Surrogate Modeling Conference (MORe) 2026, Politecnico di Milano, Milan, Italy (Oral)
  2. Prob_AI Hub Workshop 2026, University of Manchester, Manchester, UK (Poster)
  3. Minisymposium "Data-driven Dynamical Systems: Learning, Simulation, and Control", WCCM-ECCOMAS 2026, Munich, Germany (Oral)
  4. Minisymposium "Numerical Aspects of Machine Learning for PDEs", Equadiff 2026, Prague, Czech Republic (Oral)
  5. Minisymposium "Recent Advances in Data-Driven Surrogate Modeling", GIMC SIMAI Young 2026, Pisa, Italy (Oral)
  6. EUROMECH Colloquium 662 "Physics-enhanced machine learning and data-driven nonlinear dynamics", Como, Italy (Oral)
  7. SIAM Chapters meeting for Young Researchers, Politecnico di Milano, Milan, Italy (Poster)
  8. 7th International Workshop on Model Order Reduction Techniques (MORTech) 2025, Zaragoza, Spain (Oral)
  9. Minisymposium "Latent Dynamics: Learning, Simulation and Control", ENUMATH 2025, Heidelberg, Germany (Oral)
  10. 1st Annual Meeting of EMS activity group on Scientific Machine Learning, Politecnico di Milano, Milan, Italy (Poster)
  11. Minisymposium "Successes and Failures in Scientific Machine Learning", DTE & AICOMAS 2025, Paris, France (Oral)
  12. 3rd Workshop of UMI group on Mathematics for Artificial Intelligence and Machine Learning, Università degli Studi di Bari Aldo Moro, Bari, Italy (Oral)
  13. 2nd Workshop "The Symbiosis of Deep Learning and Differential Equations", NeurIPS 2022, Online (Poster)

Education