Research

Google Scholar

Scopus

Orcid

Research

Salvatore Cuomo works in the Numerical Analysis research field with a specific focus on NA-Numerical Approximation problems (theory, practice, and applications) MAI-Mathematical aspects of Artificial Intelligence; SciML- Scientific Machine Learning. The main results of these research trajectories are: 

NA- Definition of novel fitting models that mimic the behavior of experimental data and complex systems.  Data-driven and model-based approaches have been developed to represent (multi) exponential decay, to infer new B-spline-like functions, to create novel ways to solve applicative problems.

MAI- The surge in AI advancements has indeed transformed the landscape of deep learning, machine learning, and data analysis. This approach has significantly improved the performance of classification and regression tasks across many domains. Despite their practical successes, many deep learning models are often criticized for being "black boxes," where the decisions or outputs are not easily interpretable. This opacity is problematic, especially in fields requiring high levels of trust and transparency, such as medicine and scientific applications. Therefore, establishing a robust theoretical foundation for AI is a critical area of research. Such a foundation would not only enhance our understanding of why these models are so effective but also potentially lead to improvements in their design, leading to more reliable and interpretable systems. My Recent researches are focused on developing a solid theoretical basis that involves exploring topics such as the optimization landscapes of learning algorithms, the dynamics of learning in high-dimensional spaces, and the generalization properties of neural networks.

SciML- Scientific Machine Learning (SciML) is a burgeoning field that blends traditional scientific computing methods with modern machine learning techniques to solve complex problems across various domains like physics, chemistry, biology, and engineering. One of the key innovations in SciML is the development of Physics-Informed Neural Networks (PINNs). Recent research is devoted to deeply investigating SciML from theoretical and applicative points of view both.

Professional Skills

Salvatore Cuomo aims to transfer the edge knowledge coming from the research on Numerical Analysis and AI into innovative services and applications to the real world. His main contribrution on this field are 

Co-Founder of MODAL - Mathematical Modelling and Data Analysis Laboratory

(http://www.labdma.unina.it)

Co-Founder of an Academic Spin-off  Predico s.r.l. 

(http://www.predico.eu)

Co-Inventor of a patent: PredicTS-Prediction method and related system

(https://patents.google.com/patent/US20230334283A1/en)