Multilevel Stein variational gradient descent with applications to Bayesian inverse problems.
Terrence AlsupLuca VenturiBenjamin PeherstorferPublished in: MSML (2021)
Keyphrases
- inverse problems
- image reconstruction
- global optimization
- convex optimization
- optimization methods
- cost function
- optimization problems
- emission tomography
- regularization methods
- regularization method
- image segmentation
- optical flow
- smoothness constraint
- bayesian networks
- early vision
- partial differential equations
- computationally expensive
- loss function
- simulated annealing
- super resolution
- parameter space
- differential evolution
- evolutionary algorithm
- objective function