no code implementations • 21 Nov 2023 • Cédric Travelletti, Jörg Franke, David Ginsbourger, Stefan Brönnimann
This work introduces a new, distributed implementation of the Ensemble Kalman Filter (EnKF) that allows for non-sequential assimilation of large datasets in high-dimensional problems.
no code implementations • 11 Oct 2023 • Philip Stange, David Ginsbourger
We tackle the extension to the vector-valued case of consistency results for Stepwise Uncertainty Reduction sequential experimental design strategies established in [Bect et al., A supermartingale approach to Gaussian process based sequential design of experiments, Bernoulli 25, 2019].
no code implementations • 15 Jun 2022 • Johanna Ziegel, David Ginsbourger, Lutz Dümbgen
We present new classes of positive definite kernels on non-standard spaces that are integrally strictly positive definite or characteristic.
no code implementations • 8 Sep 2021 • Cédric Travelletti, David Ginsbourger, Niklas Linde
Furthermore, in that context, covariance matrices can become too large to be stored.
no code implementations • 16 Apr 2021 • Eliane Maalouf, David Ginsbourger, Niklas Linde
We propose a novel approach for solving inverse-problems with high-dimensional inputs and an expensive forward mapping.
no code implementations • 15 Feb 2021 • Athénaïs Gautier, David Ginsbourger, Guillaume Pirot
In the study of natural and artificial complex systems, responses that are not completely determined by the considered decision variables are commonly modelled probabilistically, resulting in response distributions varying across decision space.
no code implementations • 8 Jan 2021 • David Ginsbourger, Cedric Schärer
We further establish in the case of noiseless observations that correcting for covariances between residuals in cross-validation-based estimation of the scale parameter leads back to MLE.
1 code implementation • 7 Jul 2020 • Trygve Olav Fossum, Cédric Travelletti, Jo Eidsvik, David Ginsbourger, Kanna Rajan
Improving and optimizing oceanographic sampling is a crucial task for marine science and maritime resource management.
no code implementations • 11 Oct 2019 • Noémie Jaquier, David Ginsbourger, Sylvain Calinon
In learning from demonstrations, it is often desirable to adapt the behavior of the robot as a function of the variability retrieved from human demonstrations and the (un)certainty encoded in different parts of the task.
no code implementations • 9 Oct 2019 • Poompol Buathong, David Ginsbourger, Tipaluck Krityakierne
We investigate two classes of set kernels that both rely on Reproducing Kernel Hilbert Space embeddings, namely the ``Double Sum'' (DS) kernels recently considered in Bayesian set optimization, and a class introduced here called ``Deep Embedding'' (DE) kernels that essentially consists in applying a radial kernel on Hilbert space on top of the canonical distance induced by another kernel such as a DS kernel.
1 code implementation • 18 Apr 2017 • Mickaël Binois, David Ginsbourger, Olivier Roustant
Then, the search of solutions can be reduced to the random embedding of a low dimensional space into the original one, resulting in a more manageable optimization problem.
no code implementations • 22 Nov 2016 • Dario Azzimonti, David Ginsbourger, Clément Chevalier, Julien Bect, Yann Richet
The system is modeled by an expensive-to-evaluate function, such as a computer experiment, and we are interested in its excursion set, i. e. the set of points where the function takes values above or below some prescribed threshold.
no code implementations • 9 Sep 2016 • Sébastien Marmin, Clément Chevalier, David Ginsbourger
We deal with the efficient parallelization of Bayesian global optimization algorithms, and more specifically of those based on the expected improvement criterion and its variants.
no code implementations • 3 Aug 2016 • Julien Bect, François Bachoc, David Ginsbourger
Thisobservation enables us to establish generic consistency results for abroad class of SUR strategies.
no code implementations • 18 Mar 2015 • Sébastien Marmin, Clément Chevalier, David Ginsbourger
The computational burden of this selection rule being still an issue in application, we derive a closed-form expression for the gradient of the multipoint Expected Improvement, which aims at facilitating its maximization using gradient-based ascent algorithms.
no code implementations • 15 Jan 2015 • Dario Azzimonti, Julien Bect, Clément Chevalier, David Ginsbourger
In this setting, the posterior distribution on the objective function gives rise to a posterior distribution on excursion sets.
no code implementations • 13 Nov 2014 • Mickaël Binois, David Ginsbourger, Olivier Roustant
This works extends the Random Embedding Bayesian Optimization approach by integrating a warping of the high dimensional subspace within the covariance kernel.
no code implementations • 6 Aug 2013 • David Ginsbourger, Olivier Roustant, Nicolas Durrande
We study pathwise invariances of centred random fields that can be controlled through the covariance.