Can we ease the Injectivity Bottleneck on Lorentzian Manifolds for Graph Neural Networks?

Open in new window