Linearly constrained Gaussian processes
arXiv:1703.00787
Abstract
We consider a modification of the covariance function in Gaussian processes to correctly account for known linear constraints. By modelling the target function as a transformation of an underlying function, the constraints are explicitly incorporated in the model such that they are guaranteed to be fulfilled by any sample drawn or prediction made. We also propose a constructive procedure for designing the transformation operator and illustrate the result on both simulated and real-data examples.
A few fixes and added citation inforomation
Cited by in corpus (19)
- Probabilistic modelling and reconstruction of strain
- Gaussian process regression for forest attribute estimation from airborne laser scanning data
- Algorithmic Linearly Constrained Gaussian Processes
- Linearly Constrained Neural Networks
- Gaussian processes with linear operator inequality constraints
- Neutron Transmission Strain Tomography for Non-Constant Stress-Free Lattice Spacing
- Linearly Constrained Gaussian Processes with Boundary Conditions
- Posterior Projection for Inference in Constrained Spaces
- NIPS - Not Even Wrong? A Systematic Review of Empirically Complete Demonstrations of Algorithmic Effectiveness in the Machine Learning and Artificial Intelligence Literature
- Sparse Spectrum Warped Input Measures for Nonstationary Kernel Learning
- High-Dimensional Gaussian Process Inference with Derivatives
- Efficient methods for Gaussian Markov random fields under sparse linear constraints
- Nonnegativity-Enforced Gaussian Process Regression
- Continuous-time Value Function Approximation in Reproducing Kernel Hilbert Spaces
- Deep kernel learning for integral measurements
- Sequential Subspace Search for Functional Bayesian Optimization Incorporating Experimenter Intuition
- Stochastic Processes Under Linear Differential Constraints : Application to Gaussian Process Regression for the 3 Dimensional Free Space Wave Equation
- Hierarchical Non-Stationary Temporal Gaussian Processes With -Regularization
- Conditionally Independent Multiresolution Gaussian Processes