The seventieth UQSay seminar on UQ, DACE and related topics will take place online on Thursday afternoon, March 7, 2024.
2–3 PM — Baptiste Kerleguer (CEA, DAM/DIF) — [slides]
A Bayesian neural network approach to multi-fidelity surrogate modeling
This talk deals with the surrogate modeling of computer code results that can be evaluated at different levels of accuracy and computational cost, called multi-fidelity. We propose a method combining Gaussian process (GP) regression on low-fidelity data and a Bayesian neural network (BNN) on high-fidelity data. The novelty, compared with the state of the art, is that uncertainties are taken into account at all fidelity levels. The prediction uncertainty of the low-fidelity level is transmitted by Gauss-Hermite quadrature to the high-fidelity level. In addition, this method takes into account non-nested designs of experiment and non-linear interactions between levels. The proposed approach is then compared to several multi-fidelity GP regression methods on analytic functions and on a computer code.
Reference: A Bayesian neural network approach to multi-fidelity surrogate modeling, IJUQ 14.1, 2024.
Joint work with Josselin Garnier (CMAP) & Claire Cannamela (CEA, DAM/DIF).
Organizing committee: Pierre Barbillon (MIA-Paris), Julien Bect (L2S), Nicolas Bousquet (EDF R&D), Amélie Fau (LMPS), Filippo Gatti (LMPS), Bertrand Iooss (EDF R&D), Alexandre Janon (LMO), Sidonie Lefebvre (ONERA), Didier Lucor (LISN), Emmanuel Vazquez (L2S).
Coordinators: Julien Bect (L2S) & Sidonie Lefebvre (ONERA)
Practical details: the seminar will be held online using Microsoft Teams.
If you want to attend this seminar (or any of the forthcoming online UQSay seminars), and if you do not already have access to the UQSay group on Teams, simply send an email and you will be invited. Please specify which email address the invitation must be sent to (this has to be the address associated with your Teams account).
You will find the link to the seminar on the "General" UQSay channel on Teams, approximately 15 minutes before the beginning.
The technical side of things: you can use Teams either directly from your web browser or using the "fat client", which is available for most platforms (Windows, Linux, Mac, Android & iOS). We strongly recommend the latter option whenever possible. Please give it a try before the seminar to anticipate potential problems.
