Speaker: Luca Biferale
Data-driven and equations-informed tools for generation of turbulent flows
Our ability to collect data is rapidly increasing thanks to computational power and the unprecedented diversity of sensors. But how good are we at extracting, reconstructing, and understanding information from them? We present a short overview of some recent advancements for data-assimilation and modelling of turbulent multi-scale flows using both data-driven and equations-informed tools, starting from sparse and heterogeneous observations of complex fluid systems. Issues connected to validations and benchmarks in the presence of full or partial observability will be discussed. A few examples of data-augmentation based on Generative Adversarial Learning [1], Nudging, Gappy-Proper Orthogonal Decomposition [2] and Generative Diffusion Models will be quantitatively discussed [3].
[1] M. Buzzicotti, F. Bonaccorso, P. Clark Di Leoni, L. Biferale. Reconstruction of turbulent data with deep generative models for semantic inpainting from TURB-Rot database. Physical Review Fluids 6 (5), 050503, (2021).
[2] T. Li, M. Buzzicotti, L. Biferale, F. Bonaccorso, S. Chen, M. Wan. Data reconstruction of turbulent flows with Gappy POD, Extended POD and Generative Adversarial Networks. arXiv:2210.11921v1(2023) Journal of Fluid Mechanics 971, A3
[3] T Li, L Biferale, F Bonaccorso. M Scarpolini and M Buzzicotti, Synthetic Lagrangian Turbulence by Generative Diffusion Models, arxive 2307.08529 Nat Mach Intell (2024)