Tangent Space Separability in Feedforward Neural Networks.
Bálint DaróczyRita AlekszievAndrás BenczúrPublished in: CoRR (2019)
Keyphrases
- feedforward neural networks
- tangent space
- neural network
- euclidean space
- back propagation
- recurrent neural networks
- riemannian manifolds
- hidden layer
- principal curves
- multilayer perceptron
- tensor field
- dimensionality reduction
- grassmann manifold
- activation function
- training algorithm
- shape deformations
- multi layer perceptron
- lie group
- genetic algorithm
- principal components
- support vector
- stochastic gradient
- feature space
- optical flow
- probabilistic model
- random variables