GraphNeuralNetworks.jl icon indicating copy to clipboard operation
GraphNeuralNetworks.jl copied to clipboard

Allow for additonal features in GNNGraph

Open Dsantra92 opened this issue 2 years ago • 4 comments

There are cases where the shape of the feature array might not have the last dim = number of nodes and edges. For example: In a mesh represented by a graph where the vertices are nodes. Currently, there is no way to include triangular faces, except for graph data (which they are not).

GNNGraph should be allowed to contain such features which are not either node or edge. cc: @CarloLucibello

Dsantra92 avatar Oct 01 '22 20:10 Dsantra92

How does PyG works in this setting?

CarloLucibello avatar Oct 03 '22 16:10 CarloLucibello

PyG doesn't force checks on features.

Dsantra92 avatar Oct 03 '22 20:10 Dsantra92

Why we can't store them as graph data?

CarloLucibello avatar Oct 04 '22 14:10 CarloLucibello

We can do it given our current implementation. But at the end of the day, they are not exactly graph level features.

Dsantra92 avatar Oct 04 '22 15:10 Dsantra92

closing as it not clear how to act on this

CarloLucibello avatar Nov 07 '22 20:11 CarloLucibello