Mutual Information of Neural Network Initialisations: Mean Field Approximations.
Jared TannerGiuseppe UghiPublished in: CoRR (2021)
Keyphrases
- mutual information
- neural network
- free energy
- information theoretic
- closed form
- image registration
- mutual information maximization
- variational methods
- similarity measure
- competitive learning
- feature selection
- kullback leibler divergence
- neural network model
- back propagation
- pattern recognition
- information gain
- belief networks
- variational approximation
- fuzzy logic
- genetic algorithm
- self organizing maps
- image segmentation
- multimodal image registration
- neural network is trained
- markov random field
- artificial neural networks
- bp neural network
- markov networks
- partition function
- fault diagnosis
- medical image registration
- multilayer perceptron
- fixed point
- exact inference
- statistical mechanics
- conditional mutual information
- network architecture
- bayesian inference
- feed forward
- radial basis function
- parameter estimation
- em algorithm
- markov chain
- graphical models
- knn
- decision trees
- machine learning