KrigInv: An efficient and user-friendly implementation of batch-sequential inversion strategies based on Kriging

Clément Chevalier (1) , Victor Picheny (2) , David Ginsbourger (1)

1 IMSV - Institute of Mathematical Statistics and Actuarial Science [Bern]
2 INRA - Institut National de la Recherche Agronomique

Several strategies relying on kriging have recently been proposed for adaptively estimating contour lines and excursion sets of functions under severely limited evaluation budget. The recently released R package KrigInv is presented and offers a sound implementation of various sampling criteria for those kinds of inverse problems. KrigInv is based on the DiceKriging package, and thus benefits from a number of options concerning the underlying kriging models. Six implemented sampling criteria are detailed in a tutorial and illustrated with graphical examples. Different functionalities of KrigInv are gradually explained. Additionally, two recently proposed criteria for batch-sequential inversion are presented, enabling advanced users to distribute function evaluations in parallel on clusters or clouds of machines. Finally, auxiliary problems are discussed. These include the fine tuning of numerical integration and optimization procedures used within the computation and the optimization of the considered criteria.

Mots clés: Computer experiments; Kriging; Gaussian processes; Sequential design; Probability of failure; Active learning; Inversion; R package

Domaines: Machine Learning [stat.ML]; Applications [stat.AP]; Calcul [stat.CO]; Optimisation et contrôle [math.OC]; Apprentissage [cs.LG]; Calcul parallèle, distribué et partagé [cs.DC]

Clément Chevalier, Victor Picheny, David Ginsbourger. KrigInv: An efficient and user-friendly implementation of batch-sequential inversion strategies based on Kriging. 2013. ⟨hal-00713537v2⟩