Development of Machine Learning Approaches for Solving Large Rate-Equation Systems

 

 

 

 

 

Development of Machine Learning Approaches for Solving Large Rate-Equation Systems

Supervisors: Prof. Robin Santra (DESY/UHH), Prof. Marina Tropmann-Frick (HAW)

The enormous peak brightness of X-ray free-electron lasers, such as the European XFEL, offers exciting new opportunities in the area of atomic-resolution biomolecular imaging. However, in such applications, extreme radiation-induced modifications of the electronic structure of the investigated molecules are unavoidable. This can influence the apparent molecular structure. Quantitatively accounting for this effect is a central challenge.
A validated approach to predicting X-ray-driven electronic damage is solving a coupled set of first-order ordinary differential equations describing the time evolution of populations in electronic configuration space. The main bottleneck is the first-principles calculation of all physical parameters required. Within the proposed project, strategies for allying machine learning approaches will be developed in order to overcome this critical limitation.

Apply here!