Generating functionals for computational intelligence: the Fisher information as an objective function for self-limiting Hebbian learning rules.
Rodrigo EchevesteClaudius GrosPublished in: CoRR (2014)
Keyphrases
- genetic algorithm
- learning rules
- fisher information
- artificial neural networks
- objective function
- neural network
- fuzzy logic
- biologically inspired
- back propagation
- probability density function
- information geometry
- euclidean space
- constrained optimization
- noise level
- learning algorithm
- optimal solution
- positive definite
- special case
- lower bound
- estimation error
- learning scheme
- multidimensional scaling
- multiscale
- fixed point
- kullback leibler divergence
- kl divergence