Speaker: Prof. Luca Biferale

Time: 14:00-16:00, 11 March 2025 (Tuesday)  (Beijing time)

Venue: T2-202, UIC

Abstract

We present a stochastic method for generating and reconstructing complex signals along the trajectories of small objects passively advected by turbulent flows [1]. Our approach makes use of generative Diffusion Models, a recently proposed data-driven machine learning technique. We show applications to 3D tracers and inertial particles in highly turbulent flows,  2D trajectories from NOAA’s Global Drifter Program and dynamics of charged particles in astrophysics. Supremacy against linear decomposition and Gaussian Regression Processes is analyzed in terms of statistical and point-wise metrics concerning intermittency and multi-scale properties.  Preliminary results concerning generalizability and model collapse will also be discussed.

[1] Li, T., Biferale, L., Bonaccorso, F. et al. Synthetic Lagrangian turbulence by generative diffusion models. Nat Mach Intell 6, 393–403 (2024).

About the Speaker

Luca Biferale, Full Professor of Theoretical Physics at the Venue Physics Department of the University of Rome"Tor Vergata". He is author of more than 250 scientific articles in the field of  turbulence, micro-, nano-fluidies, and dynamical systems using mainly data-driven and equation-informed numerical methods. He has been elected a fellow of Euromech and the American Physical Society (APS) and has served on the board of editors of many scientific journals including Physical Review Letter. He was twice awarded the prestigious Advanced Grant from the European Research Council.