Non-asymptotic error estimates for the Laplace approximation in Bayesian inverse problems.
Tapio HelinRemo KretschmannPublished in: CoRR (2020)
Keyphrases
- error estimates
- inverse problems
- marginal likelihood
- cross validation
- error estimation
- image reconstruction
- model selection
- global optimization
- convex optimization
- expectation propagation
- optimization methods
- hyperparameters
- optimization problems
- error analysis
- provably correct
- partial differential equations
- early vision
- support vector
- worst case
- computationally expensive
- machine learning
- motion segmentation
- lower bound
- motion estimation
- generalization error
- image sequences
- decision trees
- computer vision
- learning algorithm