Arthur Kosmala

Arthur Kosmala is a Doctoral Fellow at the Munich Data Science Institute (MDSI) and a PhD candidate at the Department of Informatics - I26, Technical University of Munich. His research bridges machine learning with physics, chemistry, and materials science, focusing on neural network architectures for modeling long-range interactions in molecular systems.

• Education: MSc in Theoretical and Mathematical Physics (Technical University & LMU Munich, 1.1), MSc in Mathematical and Theoretical Physics (University of Oxford, Distinction), BSc in Physics (LMU Munich, 1.0)

Research Interests include:

• Integrating physics-based long-range methods (e.g., Ewald summation) into graph neural networks
• Compressing autoregressive time series models and exploring differential privacy in ML
• Interdisciplinary collaborations at the MDSI, merging insights from applied mathematics, computational chemistry, and materials science

Scientific Awards:

• Runner-up for the MDSI Best Paper of 2023
• Recipient of the Linde/MDSI Doctoral Fellowship, Max Weber-Program scholarship, and a DAAD grant

He contributes to open-source projects like EwaldMP (Python) and Fermat (C++), with active GitHub collaborations.