Solving Bayesian Inverse Problems With Expensive Likelihoods Using Constrained Gaussian Processes and Active Learning.
Maximilian DinkelCarolin M. GeitnerGil Robalo ReiJonas NitzlerWolfgang A. WallPublished in: CoRR (2023)
Keyphrases
- gaussian processes
- inverse problems
- active learning
- image reconstruction
- gaussian process
- convex optimization
- global optimization
- gaussian process regression
- optimization problems
- optimization methods
- computationally expensive
- learning algorithm
- maximum likelihood
- semi supervised
- early vision
- random variables
- labeled data
- supervised learning
- partial differential equations
- learning process
- hyperparameters
- smoothness constraint
- machine learning
- semi supervised learning
- training set
- multiscale
- image processing