Arthur Kosmala
Technical University of Munich
Department of Informatics - I26
Boltzmannstr. 3
85748 Garching b. München
Germany
Room: 00.11.062
E-Mail: a [dot] kosmala [at] tum [dot] de
GitHub: arthurkosmala
Google Scholar: Google Scholar
Research Interests
- Machine learning for graphs, molecules and materials
- Learning of long-range interactions with graph neural networks
- Compression of autoregressive time series models
- Differential Privacy
Selected Publications
- Simon Geisler*, Arthur Kosmala*, Daniel Herbst, Stephan Günnemann
Spatio-Spectral Graph Neural Networks
[Paper | Code | Project Page]
Advances in Neural Information Processing Systems (NeurIPS), 2024 - Jan Schuchardt, Mihail Stoian*, Arthur Kosmala*, Stephan Günnemann
Unified Mechanism-Specific Amplification by Subsampling and Group Privacy Amplification
Advances in Neural Information Processing Systems (NeurIPS), 2024
[Paper | Code | Project Page] - Arthur Kosmala, Johannes Gasteiger, Nicholas Gao, Stephan Günnemann
Ewald-based Long-Range Message Passing for Molecular Graphs
International Conference on Machine Learning (ICML), 2023
[Paper | Code | Project Page]
* equal contribution
Education
- 2020-2023: MSc in Theoretical and Mathematical Physics, Technical University & Ludwig Maximilian University of Munich, 1.1
- 2019-2020: MSc in Mathematical and Theoretical Physics, University of Oxford, Distinction
- 2016-2019: BSc in Physics, Ludwig-Maximilian University of Munich, 1.0
Academic Honors and Awards
- Runner-up, MDSI Best Paper of 2023 Award (for Ewald Message Passing)
- Linde / MDSI Doctoral Fellowship
- Scholarship awarded by the Max Weber-Program of the State of Bavaria
- One-year grant awarded by the German Academic Exchange Service (DAAD)