Depth induces scale-averaging in overparameterized linear Bayesian neural networks.
Jacob A. Zavatone-VethCengiz PehlevanPublished in: ACSCC (2021)
Keyphrases
- neural network
- genetic algorithm
- fuzzy systems
- artificial neural networks
- morphological scale space
- neural network model
- back propagation
- bayesian learning
- self organizing maps
- pattern recognition
- posterior distribution
- scale invariant
- highly non linear
- bayesian estimation
- maximum likelihood
- depth map
- high quality
- bayesian inference
- recurrent neural networks
- data sets
- rule extraction
- image processing
- depth information
- closed form
- input image
- fuzzy logic
- bayesian networks