The next seminar “Modelling of materials – theory, model reduction and efficient numerical methods” will take place next Wednesday (June 2, 2021) from 9:00 till 10:00. The talk will be given by Ondřej Kincl and Martin Šípka. Please see the details below.

Speaker: **Ondřej Kincl **and** Martin Šípka**

Title: **Novel multiscale tools in the modeling of superfluid helium**

Abstract: The second talk in the liquid helium miniseries will be composed of two parts which are new for the subject of superfluidity and are to be investigated in the scope of the START project. In the beginning, Ondřej Kincl will consider the application of the SPH (Smoothed particle hydrodynamics) theory for the modelling of superfluid phenomena. The implementation of the HVBK equations within the SPH framework will be shown and possible advantages and limitations shall be discussed.

The second part will be given by Martin Šípka, who will briefly introduce the use of a very recent tool in simulations: Machine Learning (ML). Machine learning is being applied to various parts of modelling, often with great success. In this talk, we shall illustrate its use in quantum chemistry, where the trained potentials speed up the subsequent calculations by orders of magnitude. We will show how we can use ML to capture the low-dimensional structure of a chemical reaction to produce so-called collective variables. We then outline the possible applications of ML on the Vortex Filament Method in superfluid helium. The method is often too slow when the number of vortices in the domain increase and would benefit from a similar speed-up as chemistry when using ML methods.