Approximating Gaussian Process Emulators with Linear Inequality Constraints and Noisy Observations via MC and MCMC

Abstract : Adding inequality constraints (e.g. boundedness, monotonicity, convexity) into Gaussian processes (GPs) can lead to more realistic stochastic emulators. Due to the truncated Gaussianity of the posterior, its distribution has to be approximated. In this work, we consider Monte Carlo (MC) and Markov chain Monte Carlo (MCMC). However, strictly interpolating the observations may entail expensive computations due to highly restrictive sample spaces. Having (constrained) GP emulators when data are actually noisy is also of interest. We introduce a noise term for the relaxation of the interpolation conditions, and we develop the corresponding approximation of GP emulators under linear inequality constraints. We show with various toy examples that the performance of MC and MCMC samplers improves when considering noisy observations. Finally, on a 5D monotonic example, we show that our framework still provides high effective sample rates with reasonable running times.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-02104840
Contributor : Andrés Lopez-Lopera <>
Submitted on : Friday, April 19, 2019 - 4:53:35 PM
Last modification on : Thursday, May 9, 2019 - 9:55:17 AM

Links full text

Identifiers

  • HAL Id : hal-02104840, version 1
  • ARXIV : 1901.04827

Citation

Andrés F. López-Lopera, François Bachoc, Nicolas Durrande, Olivier Roustant. Approximating Gaussian Process Emulators with Linear Inequality Constraints and Noisy Observations via MC and MCMC. 2019. ⟨hal-02104840⟩

Share

Metrics

Record views

64