A Geometric Insight into Equivariant Message Passing Neural Networks on Riemannian Manifolds.
Ilyes BatatiaPublished in: TAG-ML (2023)
Keyphrases
- message passing
- riemannian manifolds
- neural network
- geometric structure
- euclidean space
- belief propagation
- distributed systems
- mean shift
- geodesic distance
- vector space
- multi class classification
- markov random field
- manifold learning
- graphical models
- feature space
- dimensionality reduction
- covariance matrices
- reproducing kernel hilbert space
- parameter space
- visual data
- probability density function
- stereo matching
- object recognition