Viktor Zaverkin

Contact

E-Mail: echo "ue.balcen@nikrevaz.rotkiv" | rev

Viktor Zaverkin
NEC Laboratories Europe
Kurfürsten-Anlage 36
69115 Heidelberg
Germany

About

I am a Research Scientist at NEC Laboratories Europe in Heidelberg, Germany, specializing in the intersection of machine learning, chemistry, materials science, and physics. I earned my doctoral degree from the University of Stuttgart, where I introduced local equivariant atomic representation (Gaussian moments) [1, 2] and ensemble-free gradient-based uncertainty quantification methods [3, 4, 5]. Originally from Luhansk in eastern Ukraine.

Publications

Here are links to my Google Scholar and ORCID.

Resume

Here is link to my LinkedIn.

Invited talks

2023, International Conference on Data-Integrated Simulation Science (SimTech2023), Stuttgart, Germany
2023, 59th Symposium on Theoretical Chemistry (STC2023, Sigrid Peyerimhoff Doctoral Award), Zürich, Switzerland
2022, NEC Laboratories Europe, Heidelberg, Germany

Software

During my doctoral study together with David Holzmüller we published a code for (active, transfer) learning equivariant interatomic potentials:

GM-NN, MIT license,

and active learning for regression in general:

BMDAL-REG, Apache license.