Bayesian neural networks increasingly sparsify their units with depth.
Mariia VladimirovaJulyan ArbelPablo MesejoPublished in: CoRR (2018)
Keyphrases
- neural network
- bayesian networks
- pattern recognition
- neural network model
- depth map
- depth information
- fuzzy logic
- bayesian estimation
- artificial neural networks
- back propagation
- genetic algorithm
- hidden units
- bayesian learning
- activation function
- maximum likelihood
- multilayer perceptron
- feed forward
- posterior probability
- self organizing maps
- bayesian inference
- multi layer
- fault diagnosis
- parameter estimation
- cellular neural networks
- multi view