George Miloshevich
KU Leuven. Centre for mathematical Plasma-Astrophysics, Belgium
Short Summary
I am a physicist working on the interface between plasma physics and scientific machine learning
. My PhD from the University of Texas at Austin was on the effects of microscopic physics in turbulent plasmas, relying on analytical and numerical simulations. In the last three years I have enriched my experience by working on environmental data science with the focus on extreme events. Throughout my career I have been involved in various subjects including statistics and nonlinear dynamics.
Originally I come from a small country, nestled in Caucasus and growing up in a diverse family, I have always enjoyed and appreciated multicultural settings.
This fall I have started a postdoc at KU Leuven in a group of Giovanni Lapenta. My main project is on improving forecasting of space weaher events with machine learning.
Events that are rare present challenges, since they are hard to measure or estimate, yet they may have societal impact. In many of these fields forecasting of the events happen both with physics-based models
that are computationally expensive. A competitive approach is to rely on machine learning only. To improve the reliability and generalisability it could be advantages to merge physics based models with machine learning.
Data-driven closure
A promising approach is to learn adaptable closures which are physics-based models
that combine data-driven
component such parametrization in a way that generalizes better if conservational laws are respected
news
Nov 1, 2023 | Excited to move to Leuven, Belgium to work with Prof. Giovanni Lapenta. |
---|---|
Apr 10, 2023 | CNRS promotes our PRF paper |
May 15, 2022 | Moved to LSCE near Paris |
Dec 7, 2021 | Our work presented at APS DFD! |
Jan 15, 2019 | Moving to France! 🇫🇷 🥐 🗼 🥖 |