GraphNeuralNetworks.jl
GraphNeuralNetworks.jl copied to clipboard
Added `perturb_edges` function
Relates to Issue #412
This is the first in the list of data augmentation based functionality we can add to ensure users can augment their data in many different ways and test results.
This is also the next step in testing things in combination with DropMessage paper mentioned in #412 to see how various ways of data augmentation can improve over-fitting, over-smoothing, and non-robustness(as mentioned in the paper).