Error Bounds for a Matrix-Vector Product Approximation with Deep ReLU Neural Networks.
Tilahun Melkamu GetuPublished in: CoRR (2021)
Keyphrases
- error bounds
- neural network
- sparse matrix
- theoretical analysis
- pattern recognition
- matrix representation
- worst case
- artificial neural networks
- hopfield neural network
- eigenvalues and eigenvectors
- rows and columns
- linearly independent
- product information
- singular value decomposition
- back propagation
- symmetric matrix
- multi layer
- genetic algorithm
- wavelet synopses
- transformation matrix
- euclidean norm
- recurrent neural networks
- life cycle
- covariance matrix
- singular values
- activation function
- finite sample
- polynomial time approximation
- fuzzy logic
- symmetric matrices
- upper bound
- feature vectors