Skip to content

core

The exauq core package contains everything required for the creation and training of both single level and multi-level deterministic Gaussian process (GP) emulators and leave-one-out (LOO) sampling methods on customisable designs. The GP emulation is built upon the mogp_emulator while implementing the ability to bound hyperparameters. Spatial domains can be created and filled with effective sampled points alongside repulsion points across the boundary.

Tutorials on how to create a basic experimental design can be found within the Experimental Design tutorials.

Modules

  • designers: Create the experimental domain using either a simple oneshot Latin hypercube or through the LOO sampling methods for both single and multi-level GPs.

  • emulators: Provides the emulators to modify and train the GP and control hyperparameters.

  • modelling: Contains the objects required for utilising and training GP emulators alongside the LOO sampling methods for both single and multi-level GPs.

  • numerics: Numerical tolerance checks.

References

mogp_emulator: https://github.com/alan-turing-institute/mogp-emulator

LOO Sampling:

Mohammadi, H. et al. (2022) "Cross-Validation-based Adaptive Sampling for Gaussian process models". DOI: https://doi.org/10.1137/21M1404260

Kimpton, L. M. et al. (2023) "Cross-Validation Based Adaptive Sampling for Multi-Level Gaussian Process Models". arXiv: https://arxiv.org/abs/2307.09095