A structure preserving numerical scheme for Fokker-Planck equations of structured neural networks with learning rules.
Qing HeJingwei HuZhennan ZhouPublished in: CoRR (2021)
Keyphrases
- learning rules
- numerical scheme
- structure preserving
- neural network
- partial differential equations
- anisotropic diffusion
- finite difference
- color image processing
- artificial neural networks
- numerical solution
- level set
- back propagation
- variational methods
- biologically inspired
- energy functional
- fixed point
- dimensionality reduction
- lie group
- color images
- finite element
- pattern recognition
- learning scheme
- differential equations
- learning algorithm
- control system
- multiscale
- machine learning