Tractably modelling distributions over manifolds has long been an import...
To better conform to data geometry, recent deep generative modelling
tec...
Recent advances in deep representation learning on Riemannian manifolds
...
Research has shown that widely used deep neural networks are vulnerable ...