Graph Neural Networks gain momentum in astrophysics: student sparks debate
A RWTH Aachen CS student asks if GNNs can model galaxy formation and cosmic webs...
A future RWTH Aachen CS student, accepted for fall, seeks to merge astrophysics and ML. With no direct research group, they eye the Quantum Information Systems group and the Learning on Graphs group, wondering if Graph Neural Networks (GNNs) suit astrophysical data like galaxy formation, cosmic web structure, or particle interaction data. The post asks if GNNs are already used in astrophysics, and which other ML subfields to explore. Despite RWTH lacking a dedicated astrophysics-ML group, the student chose it for its math teaching approach.
- GNNs are already used in astrophysics for tasks like galaxy classification and cosmic web analysis
- RWTH Aachen's Learning on Graphs group offers foundational GNN research relevant to physics data
- Other subfields include geometric deep learning and equivariant neural networks for particle physics
Why It Matters
Shows how cross-disciplinary AI research can thrive even without a dedicated astrophysics lab, thanks to graph ML techniques.