Short Bio

Since February 2025, Jasper De Landsheere has been a PhD student at the Heid Lab, TU Wien. He works as a research assistant for the ERC-funded DeepRxn project, focusing on deep learning for chemical reactions. He holds BSc and MSc degrees (Magna Cum Laude) in Bioscience Engineering Technology from Ghent University, Belgium (2022). He also holds an MSc (Dipl.-Ing.) degree in Data Science from the TU Wien (2025).

PhD Project - Deep Learning for Chemical Reactions

Supervised by Esther Heid

Knowing certain chemical reaction properties are very important to industrial applications. Currently, obtaining these properties are very time-consuming as they are either measured experimentally or calculated using expensive quantum mechanical calculations. Jasper’s research focuses on applying deep learning to predict these reaction properties. Specifically, he is developing new neural network architectures and reaction representations for reaction barrier height prediction.

Publications and Conferences

Journal Papers

• Kovács, P., Heid, E., De Landsheere, J., & Madsen, G. K. (2024). LoGAN: local generative adversarial network for novel structure prediction. Machine Learning: Science and Technology, 5(3), 035079.
• Karwounopoulos J., De Landsheere J., Galustian L., Jechtl T., Heid E. Graph-based prediction of reaction barrier heights with on-the-fly prediction of transition states. ChemRxiv. 2025; doi:10.26434/chemrxiv-2025-w2kgt-v2.

Presentations

• FemChem Scientific Workshop 2025: Talk, Deep Learning for Reaction Barrier Height Prediction