This talk has been automatically captioned. You can remove these by pressing CC on the video toolbar.

Name: Vladas Pipiras

Talk Title: Multifidelity Monte Carlo estimation for extremes

Abstract: When modeling a random phenomenon, data are often available from multiple sources, or models, of varying fidelity, those with higher fidelity carrying higher costs. Multifidelity Monte Carlo (MFMC) methods offer tools that allow combining the data from multiple sources for better estimation of quantities for high-fidelity models. With a few exceptions though, much of the focus of the MFMC literature has been on characterizing uncertainty related to averages, in the context of non-rare problems where data are available to estimate these averages directly. In this work, we extend some MFMC methods to estimation of extremal quantities, for example, probabilities of rare events, possibly those that have not been observed in high-fidelity data. The suggested approaches are based on statistical extreme value theory, applied to simultaneously extreme observations from low-fidelity and high-fidelity models. The ideas are illustrated with synthetic data examples and the application to extreme ship motions. 

This talk is a contributed talk at EVA 2021