Deep learning-based numerical methods for high-dimensional parabolic partial differential equations and backward stochastic differential equations.
Weinan EJiequn HanArnulf JentzenPublished in: CoRR (2017)
Keyphrases
- stochastic differential equations
- deep learning
- numerical methods
- partial differential equations
- high dimensional
- brownian motion
- differential equations
- maximum a posteriori estimation
- image denoising
- anisotropic diffusion
- numerical solution
- level set
- unsupervised learning
- image processing
- finite difference method
- multiscale
- machine learning
- image enhancement
- dimensionality reduction
- numerical algorithms
- fractional brownian motion
- energy functional
- high order
- finite difference
- denoising
- natural images
- medical images
- data points
- numerical scheme
- model selection
- pairwise