Skip to content
Technical University of Munich

Dr. Nadiia Derevianko

Position

  • Researcher in Physics-enhanced Machine Learning at Technical University of Munich, Germany

Address

Technical University of Munich
TUM School of CIT
Department of Computer Science
Boltzmannstraße 3
85748 Garching
Germany

Office: MI 02.05.040
Email: nadiia.derevianko@tum.de
Office hours: by arrangement

Teaching

SoSe 2025

  • Lecture: Algorithms for Scientific Computing (Monday 14:00-16:00, Wednesday 08:00-10:00, 6 SWS (4V + 2Ü) / 8 Credits, English)

Supervision: Student Theses

Ongoing Theses: 

  • Yichen Tang (Master's Thesis): Neural networks with adaptive activation functions and their application to the solution of PDEs
  • Abdul Ikhlaq (Master's Thesis): Fast Algorithm for Protein-Ligand Docking

 

Publications

Recent preprints and publications:

  • Nadiia Derevianko, Gerlind Plonka, Differential approximation of the Gaussian by short cosine sums with exponential error decay, Arxiv preprint: https://arxiv.org/abs/2307.13587 (revised)
  • Nadiia Derevianko, Lennart Hübner, Parameter estimation for multivariate exponential sums via iterative rational approximation, Arxiv preprint: arxiv.org/abs/2504.19157
  • Nadiia Derevianko, Recovery of rational functions via Hankel pencil method and sensitivities of the poles, Arxiv preprint: arxiv.org/abs/2406.13192 (revised)

See www.researchgate.net/profile/Nadiia-Derevianko

Open master projects

Please write me an email if you are interested in writing your master's thesis on one of these topics.

- Graph convolutional networks for approximation jump of discontinuities (based on the paper Zhiqian Chen et all. Rational Neural Networks for Approximating Graph Convolution Operator on Jump Discontinuities, 2018 IEEE International Conference on Data Mining, DOI 10.1109/ICDM.2018.00021) Requirements: basic knowledge of graph convolutional networks, eigenvalue problem, programing skills in Python or Matlab.

 

To top