Kate Haitsiukevich

I am a PhD student at Aalto University School of Science supervised by Prof. Alexander Ilin.

I am passionate about deep learning applications in science and engineering. My research lies in the intersection of partial differential equations and deep neural networks. Prior to starting my PhD studies I've worked as a Data Scientist and a Quantitative Analyst in the financial industry. I've earned my Bachelor's and Master's degrees in Applied Mathematics. My framework of choice is Pytorch 🔥.

Email  /  CV  /  Google Scholar  /  Github  /  Twitter

profile photo

I study the ways to efficiently combine Partial Differential Equations (PDEs) with neural networks for systems modeling.

PDEs are the established mathematical models for many real-world problems. However, only a handful of equations have an analytical solution while for the rest the solution is approximated numerically. Traditional numerical solvers usually provide an accurate approximation but tend to be quite slow. Other limitation of traditional numerical algorithms is their inability to tune the solution according to observations from the real process.

Alternatively, neural network-based solvers are data-driven in nature and have faster inference. These properties allow extension of neural solvers for applications where the equation itself is partially unknown but can be inferred from data. Finally, when the governing equations of the process are unknown, inductive biases from differential equations or traditional numerical solvers baked into neural network architecture significantly reduce data requirements and improve the results.

prl Improved Training of Physics-Informed Neural Networks with Model Ensembles
Katsiaryna Haitsiukevich, Alexander Ilin
International Joint Conference on Neural Networks (IJCNN), 2023
arXiv / Code / BibTex

Neural PDE solvers benefit from gradual expansion of the solution domain starting from points with supervision signal (e.g. initial conditions). We propose to automate this expansion based on the agreement of PDE solver ensemble.

blind-date Learning Trajectories of Hamiltonian Systems with Neural Networks
Katsiaryna Haitsiukevich, Alexander Ilin
Artificial Neural Networks and Machine Learning (ICANN), 2022
arXiv / BibTex

A neural network architecture with encoded energy conservation inductive bias that enables direct calculation of the derivatives needed to satisfy the conservation constrains.

clean-usnob A Deep Learning Model of Tubular Reactors
Katsiaryna Haitsiukevich, Samuli Bergman, Cesar de Araujo Filho, Francesco Corona, Alexander Ilin
IEEE 19th International Conference on Industrial Informatics (INDIN), 2021
arXiv / BibTex

We developed a PDE-motivated neural network architecture for modeling tubular reactor states. The project was done in collaboration with Neste.


During my studies I participated as a Teaching Assistant in the following courses:

Webpage design courtesy of Jon Barron.