A machine learning solver for high-dimensional integrals: Solving Kolmogorov PDEs by stochastic weighted minimization and stochastic gradient descent through a high-order weak approximation scheme of SDEs with Malliavin weights.
Riu NaitoToshihiro YamadaPublished in: CoRR (2020)
Keyphrases
- high order
- stochastic gradient descent
- machine learning
- high dimensional
- partial differential equations
- higher order
- weight vector
- pairwise
- least squares
- loss function
- matrix factorization
- low dimensional
- linear combination
- support vector machine
- step size
- random forests
- dimensionality reduction
- machine learning methods
- importance sampling
- machine learning algorithms
- model selection
- low rank
- level set
- image denoising
- markov random field
- learning tasks
- semi supervised
- knn
- state space
- data points
- computational complexity
- regularization parameter
- objective function
- numerical solution
- multiscale
- online algorithms
- feature selection