Gradient-free Stein variational gradient descent with kernel approximation.
Liang YanXiling ZouPublished in: Appl. Math. Lett. (2021)
Keyphrases
- support vector
- cost function
- objective function
- approximation algorithms
- free energy
- image segmentation
- kernel function
- optical flow
- closed form
- polynomial kernels
- approximation methods
- piecewise constant
- update rule
- feature selection
- optical flow computation
- variational methods
- kernel matrix
- error bounds
- kernel methods
- support vector machine