Agnostic Multi-Fidelity Gaussian Process Regression for Modeling Complex Systems in Aerospace
Friday 10.9.2021, h.17 - Sala Consiglio DAER - Edificio B12 - Second Floor, Via La Masa, 34 - Milano
In a multi-fidelity setting, it is fundamental to establish the correct hierarchy in terms of data credibility w.r.t. the reality of interest. Unfortunately, the complexity characterizing aerospace applications generally makes the direct estimation of data fidelity difficult, if not intractable, leaving ground to modeling biases. In this context, we aim at developing an agnostic framework which is capable of recognizing the coherence of models predictions w.r.t. the reality of interest and, at the same time, capable of neglecting inaccurate, or wrong, information. It will be shown that this capability is particularly relevant whenever the Modeler does not select the appropriate ordering for the sequential construction of the multi-level surrogate. Moreover, we will show how the proposed formulation can also be exploited to obtain invaluable insights about the physics underlying the reality of interest. Eventually, we also provide perspectives concerning the future deployment of the proposed methodology, with particular reference to Bayesian optimization methods and to the efficient construction of databases.
Speaker: Dr. Giulio Gori
Dr. Giulio Gori graduated in Aeronautical Engineering in 2013 at Politecnico di Milano. He obtained his Ph.D. in 2019 in the same institution. He has been UTOPIAE MSCA-ITN Early-Stage Researcher at INRIA/ Centre de Mathématiquées Appliqueé, École Polytechnique, IPP, France from 2017 to 2020 and he is currently a Post-doc researcher in the CS2 H2020 MONNALISA Project at Department of Aerospace Science and Technology, Politecnico di Milano, Italy. He has been visiting fellow in several highly reputed research centers, such as CTR at Stanford University and VKI in Brussels. He actively contributes to a wide variety of topics related both to fundamental science and to industrial applications, including non-ideal gas dynamics, ice accretion and multi-fidelity modeling methods under epistemic uncertainty. He his an active developer of the SU2 open-source code.
To access the event you must show the green pass at the entrance.
LinkedIn | Twitter | Instagram | YouTube - @AerospacePoliMi