Partial differential equations for training deep neural networks.
Pratik ChaudhariAdam M. ObermanStanley J. OsherStefano SoattoGuillaume CarlierPublished in: ACSSC (2017)
Keyphrases
- partial differential equations
- neural network
- training process
- anisotropic diffusion
- image denoising
- image processing
- level set
- numerical solution
- fourth order
- multiscale
- numerical methods
- image enhancement
- energy functional
- finite difference
- numerical algorithms
- nonlinear diffusion
- pattern recognition
- finite difference method
- conservation laws
- differential equations
- diffusion equation
- mathematical morphology
- high order
- higher order
- boundary value problem
- artificial neural networks
- ambrosio tortorelli
- nonlinear partial differential equations
- difference equations
- numerical scheme
- pairwise
- machine learning