Gionni Marchetti

Gionni Marchetti

AI4Science · Quantum Matter · Explainable AI · Statistical Physics

Machine learning researcher and theoretical physicist with a PhD in condensed matter theory and experience applying AI/ML to physical and chemical problems.

About Me

I am a physicist working at the intersection of condensed matter physics, complex systems, and machine learning. My research explores how data-driven methods can be used to uncover the underlying structure and dynamics of high-dimensional physical and chemical systems. I work as an independent researcher, focusing on explainable AI (XAI), and I am building a company to make black-box models transparent and trustworthy.

My current interests include manifold learning, dimensionality reduction, topological data analysis, optimal transport, and the application of machine learning to scientific problems, such as protein spectral data. I am also developing novel approaches to charge transport in materials.

I also generate accurate scientific datasets and deposit them on Zenodo in accordance with the FAIR principles (Findable, Accessible, Interoperable, Reusable).

Research

My research combines concepts from theoretical physics with modern data-driven approaches.

Condensed Matter Physics and Dynamical Systems

Many-body interactions in materials, Boltzmann transport equation, Density Functional Theory and Thomas-Fermi potential. Dynamical systems, nonlinear phenomena, statistical physics and the emergence of complex behaviour.

Manifold Learning

Understanding the intrinsic geometry and dimensionality of high-dimensional physical trajectories using PCA, t-SNE, autoencoders and related techniques.

Topological Data Analysis

Persistent homology and topological methods for identifying geometric and dynamical structures hidden in complex datasets.

Explainable AI

Developing interpretable and scientifically meaningful machine-learning approaches for physical and scientific applications.

Selected Recent Publications

Gionni Marchetti: Learning the intrinsic dimensionality of Fermi–Pasta–Ulam–Tsingou trajectories: A nonlinear approach using a deep autoencoder model (Editor's Pick)

Chaos: An Interdisciplinary Journal of Nonlinear Science, 36, 093126 (2026)

DOI

Gionni Marchetti: Beyond Explained Variance: A Cautionary Tale of PCA

arXiv preprint (2026). To be submitted to Journal of Applied Statistics.

DOI

A. Martinez-Serra, G. Marchetti, F. D'Amico, I. Fenoglio, B. Rossi, M. P. Monopoli, G. Franzese: A Machine Learning Tool to Analyse Spectroscopic Changes in High-Dimensional Data

International Journal of Biological Macromolecules, 330, Part 3, 148095 (2025)

A. Martinez-Serra and G. Marchetti contributed equally.

DOI arXiv

Gionni Marchetti: Intrinsic Dimensionality of Fermi-Pasta-Ulam-Tsingou High-Dimensional Trajectories Through Manifold Learning: A Linear Approach (Featured)

Chaos: An Interdisciplinary Journal of Nonlinear Science, 35, 103118 (2025)

DOI

Gionni Marchetti: Revisiting the Thomas-Fermi Potential for Three-Dimensional Condensed Matter Systems

The European Physical Journal B, 97, 81 (2024)

DOI

Gionni Marchetti: Generalized Naming Game and Bayesian Naming Game as Dynamical Systems

Physical Review E, 109, 064202 (2024)

DOI
→ View all publications