Dear colleagues, We would like to invite you to the following SPASS seminar, jointly organized by UniPi, SNS, UniFi and UniSi (abstract below): *Generative Diffusion Models for Lagrangian Turbulence* by *Tianyi Li* (Università degli Studi di Roma Tor Vergata) The seminar will take place on *Tuesday, 24.02.2026* at *14:00 CET* in Saletta Riunioni, Dipartimento di Matematica, UNIPI and streamed online at this link <https://meet.google.com/stc-xonf-yrx>. Best regards Mario Maurelli https://sites.google.com/unipi.it/spass <https://www.google.com/url?q=https://sites.google.com/unipi.it/spass&source=gmail-imap&ust=1665669490000000&usg=AOvVaw07o0tKOUGmZjDDF2Ta7pQY> --------------------------------------- *Title: Generative Diffusion Models for Lagrangian Turbulence* *Abstract: Lagrangian turbulence plays a central role in transport and dispersion, yet modeling the intermittent and non-Gaussian nature of particle trajectories remains a major challenge. In this talk, we present a generative framework based on diffusion models to learn and sample the probability distribution of turbulent particle motion directly from data. The model generates single-particle trajectories in three-dimensional turbulence at high Reynolds numbers and reproduces key statistical properties across scales, including heavy-tailed velocity increments, anomalous scaling, and intermittency near the dissipative range. We further extend the approach to inertial particles (tracers, heavy, and light) and to the probabilistic reconstruction of missing velocity data along trajectories, with applications to real ocean drifter observations. In all cases, the generated signals retain realistic scale-dependent and non-Gaussian features. These results suggest diffusion models as a new probabilistic tool for studying and reconstructing Lagrangian turbulence. This work was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 882340).*