A Deep Learning architecture designed specifically to process data that exists in n-dimensional Euclidean space.
How it differs from a Graph Neural Network (GNN):
- Explicitly distinguishes between features (h: mass, charge) and coordinates (x: position).
gnn.jpg)