×

zbMATH — the first resource for mathematics

On degeneracy and invariances of random fields paths with applications in Gaussian process modelling. (English) Zbl 1329.60135
Summary: We study pathwise invariances and degeneracies of random fields with motivating applications in Gaussian process modelling. The key idea is that a number of structural properties one may wish to impose a priori on functions boil down to degeneracy properties under well-chosen linear operators. We first show in a second order set-up that almost sure degeneracy of random field paths under some class of linear operators defined in terms of signed measures can be controlled through the two first moments. A special focus is then put on the Gaussian case, where these results are revisited and extended to further linear operators thanks to state-of-the-art representations. Several degeneracy properties are tackled, including random fields with symmetric paths, centred paths, harmonic paths, or sparse paths. The proposed approach delivers a number of promising results and perspectives in Gaussian process modelling. In a first numerical experiment, it is shown that dedicated kernels can be used to infer an axis of symmetry. Our second numerical experiment deals with conditional simulations of a solution to the heat equation, and it is found that adapted kernels notably enable improved predictions of non-linear functionals of the field such as its maximum.

MSC:
60G60 Random fields
60G17 Sample path properties
60G15 Gaussian processes
62J02 General nonlinear regression
65C20 Probabilistic models, generic numerical methods in probability and statistics
Software:
DiceDesign; hgam; kergp
PDF BibTeX XML Cite
Full Text: DOI
References:
[1] Adler, R. J., (An Introduction to Continuity, Extrema, and Related Topics for General Gaussian Processes, Lecture Notes-Monograph Series, vol. 12, (1990), Published by: Institute of Mathematical Statistics) · Zbl 0747.60039
[2] Adler, R.; Taylor, J., Random fields and geometry, (2007), Springer Boston · Zbl 1149.60003
[3] Berlinet, A.; Thomas-Agnan, C., Reproducing kernel Hilbert spaces in probability and statistics, (2004), Kluwer Academic Publishers · Zbl 1145.62002
[4] Cramér, H.; Leadbetter, M. R., Stationary and related stochastic processes: sample function properties and their applications, (1967), Wiley New York · Zbl 0162.21102
[5] Cressie, N., (Statistics for Spatial Data, Wiley Series in Probability and Mathematical Statistics, (1993))
[6] Deheuvels, P., Karhunen-Loève expansions of mean-centered Wiener processes, (High Dimensional Probability, Lecture Notes-Monograph Series, vol. 51, (2006)), 62-76 · Zbl 1130.60045
[7] Deville, Y., Ginsbourger, D., Roustant, O., 2015. kergp: Gaussian Process Laboratory, R package version 0.1.0.
[8] Durrande, N.; Ginsbourger, D.; Roustant, O., Additive covariance kernels for high-dimensional Gaussian process modeling, Ann. Fac. Sci. Toulouse, 21, 481-499, (2012) · Zbl 1266.60068
[9] Durrande, N.; Ginsbourger, D.; Roustant, O.; Carraro, L., ANOVA kernels and RKHS of zero mean functions for model-based sensitivity analysis, J. Multivariate Anal., 115, 57-67, (2013) · Zbl 1259.49066
[10] Duvenaud, D., Automatic model construction with Gaussian processes, (2014), University of Cambridge, (Ph.D. thesis)
[11] Duvenaud, D. K.; Nickisch, H.; Rasmussen, C. E., Additive Gaussian processes, (Shawe-Taylor, J.; Zemel, R.; Bartlett, P.; Pereira, F.; Weinberger, K., Advances in Neural Information Processing Systems 24, (2011), Curran Associates, Inc.), 226-234
[12] Franco, J., Dupuy, D., Roustant, O., Damblin, G., Iooss, B., 2013. DiceDesign: Designs of Computer Experiments, R package version 1.3.
[13] Fricker, T. E.; Oakley, J. E.; Urban, N. M., Multivariate Gaussian process emulators with nonseparable covariance structures, Technometrics, 55, 1, 47-56, (2013)
[14] Gihman, I.; Skorohod, A., The theory of stochastic processes I, (1974), Springer Berlin, Heidelberg, New York · Zbl 0291.60019
[15] Ginsbourger, D.; Bay, X.; Roustant, O.; Carraro, L., Argumentwise invariant kernels for the approximation of invariant functions, Ann. Fac. Sci. Toulouse, 21, 3, 501-527, (2012) · Zbl 1291.60105
[16] Ginsbourger, D.; Durrande, N.; Roustant, O., Kernels and designs for modelling invariant functions: from group invariance to additivity, (Uciński, D.; Atkinson, A.; Patan, C., mODa 10 Advances in Model-Oriented Design and Analysis, (2013), Physica-Verlag HD) · Zbl 1259.49066
[17] Ginsbourger, D., Roustant, O., Schuhmacher, D., Durrande, N., Lenz, N., 2014. On ANOVA decompositions of kernels and Gaussian random field paths. http://arxiv.org/abs/1409.6008. · Zbl 1356.65027
[18] Haasdonk, B.; Burkhardt, H., Invariant kernels for pattern analysis and machine learning, Mach. Learn., 68, 35-61, (2007)
[19] Jones, D., A taxonomy of global optimization methods based on response surfaces, J. Global Optim., 21, 21, 345-383, (2001) · Zbl 1172.90492
[20] Matheron, G., Principles of geostatistics, Econ. Geol., 58, 1246-1266, (1963)
[21] Meier, L.; Van de Geer, S.; Bühlmann, P., High-dimensional additive modeling, Ann. Statist., 37, 3779-3821, (2009) · Zbl 1360.62186
[22] Mercer, J., Functions of positive and negative type and their connection with the theory of integral equations, Phil. Trans. R. Soc. A, 209, 415-446, (1909) · JFM 40.0408.02
[23] O’Hagan, A., Bayesian analysis of computer code outputs: A tutorial, Reliab. Eng. Syst. Saf., 91, 10-11, 1290-1300, (2006)
[24] Rajput, B.; Cambanis, S., Gaussian processes and Gaussian measures, Ann. Math. Statist., 43, 6, 1944-1952, (1972) · Zbl 0255.60004
[25] Rasmussen, C. R.; Williams, C. K.I., Gaussian processes for machine learning, (2006), MIT Press · Zbl 1177.68165
[26] Schaback, R., Solving the Laplace equation by meshless collocation using harmonic kernels, Adv. Comput. Math., 31, 4, 457-470, (2009) · Zbl 1180.65161
[27] Scheuerer, M., A comparison of models and methods for spatial interpolation in statistics and numerical analysis, (2009), University of Göttingen, (Ph.D. thesis)
[28] Scheuerer, M., Regularity of the sample paths of a general second order random field, Stochastic Process. Appl., 120, 1879-1897, (2010) · Zbl 1196.60061
[29] Scheuerer, M.; Schlather, M., Covariance models for divergence-free and curl-free random vector fields, Stoch. Models, 28, 3, 433-451, (2012) · Zbl 1266.60093
[30] Stein, M., Interpolation of spatial data, some theory for Kriging, (1999), Springer · Zbl 0924.62100
[31] Steinwart, I., Convergence types and rates in generic Karhunen-Loève expansions with applications to sample path properties, tech. rep. 2014-007, (2014), Fakultät für Mathematik und Physik, Universität Stuttgart
[32] Steinwart, I.; Scovel, C., Mercer’s theorem on general domains: on the interaction between measures, kernels, and rkhss, Constr. Approx., 35, 363-417, (2012) · Zbl 1252.46018
[33] Tarieladze, V.; Vakhania, N., Disintegration of Gaussian measures and average-case optimal algorithms, J. Complexity, 23, 851-866, (2007) · Zbl 1131.41010
[34] Van der Vaart, A.; Van Zanten, J., Rates of contraction of posterior distributions based on Gaussian process priors, Ann. Statist., 36, 1435-1463, (2008) · Zbl 1141.60018
[35] Van der Vaart, A.; Van Zanten, J., Reproducing kernel Hilbert spaces of Gaussian priors, (Pushing the Limits of Contemporary Statistics: Contributions in Honor of Jayanta K. Ghosh, Institute of Mathematical Statistics Collect, 3, (2008)), 200-222
[36] Van der Vaart, A.; van Zanten, H., Information rates of nonparametric Gaussian process methods, J. Mach. Learn. Res., 12, 2095-2119, (2011) · Zbl 1280.68228
[37] Welch, W. J.; Buck, R. J.; Sacks, J.; Wynn, H. P.; Mitchell, T. J.; Morris, M. D., Screening, predicting, and computer experiments, Technometrics, 34, 15-25, (1992)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.