-
Publication: (Filling et al 2025) - Direct Molecular Polarizability...
The paper describes a graph neural network architecture that uses local frames to predict local geometric properties.. -
Franzen, Wand (NeurIPS 21) - General Nonlinearities in SO(2)-Equivariant...
Handling general nonlinearities in group-convolutional networks with SO(2)-invariance. -
Publication: Franzen et al (2026) - Discretizing Group-Convolutional Neural...
A paper that describes how to speed-up deep learning on 3D point clouds with (G-)CNNs by compressing its linear representation layers by a local partial symmetry analysis. -
Publication: Filling et al (2026) - Tensor Channel Equivariant Graph Neural...
This paper improves upon the previous paper published at the NeurIPS '25 ML4PhysicalScience-Workshop by using overcomplete generating sets rather than a basis to represent local...