Convergence Guarantees for Gaussian Process Means With Misspecified Likelihoods and Smoothness
arXiv:2001.10818
Abstract
Gaussian processes are ubiquitous in machine learning, statistics, and applied mathematics. They provide a flexible modelling framework for approximating functions, whilst simultaneously quantifying uncertainty. However, this is only true when the model is well-specified, which is often not the case in practice. In this paper, we study the properties of Gaussian process means when the smoothness of the model and the likelihood function are misspecified. In this setting, an important theoretical question of practial relevance is how accurate the Gaussian process approximations will be given the difficulty of the problem, our model and the extent of the misspecification. The answer to this problem is particularly useful since it can inform our choice of model and experimental design. In particular, we describe how the experimental design and choice of kernel and kernel hyperparameters can be adapted to alleviate model misspecification.
arXiv commentAccepted to JMLR
References in corpus (16)
- Practical Bayesian Optimization of Machine Learning Algorithms
- Reconciling modern machine learning practice and the bias-variance trade-off
- Gaussian Processes for Data-Efficient Learning in Robotics and Control
- Rates of contraction of posterior distributions based on Gaussian process priors
- Gaussian Processes and Kernel Methods: A Review on Connections and Equivalences
- Reproducing kernel Hilbert spaces of Gaussian priors
- Misspecification in infinite-dimensional Bayesian statistics
- Consistency and robustness of kernel-based regression in convex risk minimization
- Statistical emulation of a tsunami model for sensitivity analysis and uncertainty quantification
- Fast Randomized Kernel Methods With Statistical Guarantees
- Rates of Convergence for Sparse Variational Gaussian Process Regression
- A novel class of stabilized greedy kernel approximation algorithms: Convergence, stability & uniform point distribution
- Sobolev Norm Learning Rates for Regularized Least-Squares Algorithm
- A hierarchical expected improvement method for Bayesian optimization
- Approximation beats concentration? An approximation view on inference with smooth radial kernels
- Kriging prediction with isotropic Matérn correlations: Robustness and experimental design