Handling the Positive-Definite Constraint in the Bayesian Learning Rule.
Wu LinMark SchmidtMohammad Emtiyaz KhanPublished in: ICML (2020)
Keyphrases
- positive definite
- learning rules
- neural network
- biologically inspired
- covariance matrix
- kernel function
- artificial neural networks
- geometric structure
- back propagation
- learning scheme
- kernel matrix
- diffusion tensor
- bayesian inference
- gaussian processes
- semi parametric
- genetic algorithm
- spatio temporal
- feature space
- learning algorithm
- reproducing kernel hilbert space
- kernel machines
- graph kernels