Uncertainty quantification of the 4th kind; optimal posterior accuracy-uncertainty tradeoff with the minimum enclosing ball. (English) Zbl 07605580
Summary: Uncertainty quantification (UQ) is, broadly, the task of determining appropriate uncertainties to model predictions. There are essentially three kinds of approaches to Uncertainty Quantification: (A) robust optimization (min and max), (B) Bayesian (conditional average) and (C) decision theory (minmax). Although (A) is robust, it is unfavorable with respect to accuracy and data assimilation. (B) requires a prior, it is generally non-robust (brittle) with respect to the choice of that prior and posterior estimations can be slow. Although (C) leads to the identification of an optimal prior, its approximation suffers from the curse of dimensionality and the notion of loss/risk used to identify the prior is one that is averaged with respect to the distribution of the data. We introduce a 4th kind which is a hybrid between (A), (B), (C), and hypothesis testing. It can be summarized as, after observing a sample \(x\), (1) defining a likelihood region through the relative likelihood and (2) playing a minmax game in that region to define optimal estimators and their risk. The resulting method has several desirable properties: (a) an optimal prior is identified after measuring the data and the notion of loss/risk is a posterior one, (b) the determination of the optimal estimate and its risk can be reduced to computing the minimum enclosing ball of the image of the likelihood region under the quantity of interest map (such computations are fast and do not suffer from the curse of dimensionality). The method is characterized by a parameter in \([0, 1]\) acting as an assumed lower bound on the rarity of the observed data (the relative likelihood). When that parameter is near 1, the method produces a posterior distribution concentrated around a maximum likelihood estimate (MLE) with tight but low confidence UQ estimates. When that parameter is near 0, the method produces a maximal risk posterior distribution with high confidence UQ estimates. In addition to navigating the accuracy-uncertainty tradeoff, the proposed method addresses the brittleness of Bayesian inference by navigating the robustness-accuracy tradeoff associated with data assimilation.
Keywords:
uncertainty quantification; accuracy-robustness tradeoff; relaxing maximum likelihood estimation; minimum enclosing ball; relative likelihood region; data assimilationReferences:
| [1] | Aliprantis, C. D.; Border, K. C., Infinite Dimensional Analysis: A Hitchhiker’s Guide (2006), Springer: Springer Berlin · Zbl 1156.46001 |
| [2] | Aliprantis, C. D.; Burkinshaw, O., Principles of Real Analysis (1998), Academic Press · Zbl 1006.28001 |
| [3] | Berger, J. O., Statistical Decision Theory and Bayesian Analysis (2013), Springer Science & Business Media |
| [4] | Bertsekas, D. P.; Nedić, A.; Ozdaglar, A., Convex Analysis and Optimization. Athena Scientific Optimization and Computation Series (2003), Athena Scientific · Zbl 1140.90001 |
| [5] | Bonnans, J. F.; Shapiro, A., Perturbation Analysis of Optimization Problems (2000), Springer-Verlag · Zbl 0966.49001 |
| [6] | Bădoiu, M.; Har-Peled, S.; Indyk, P., Approximate clustering via core-sets, (Proceedings of the Thirty-Fourth Annual ACM Symposium on Theory of Computing (2002), ACM), 250-257 · Zbl 1192.68871 |
| [7] | Casella, G.; Berger, R. L., Statistical Inference (2002), Thomson Learning Inc |
| [8] | Dudley, R., Statistics for Applications, 18.443 (2009), MIT OpenCourseWare |
| [9] | England Wales Court of Appeal (Civil Division), Nulty & Ors v. Milton Keynes Borough Council (2013), [2013] EWCA Civ 15, Case No. A1/2012/0459 |
| [10] | Gärtner, B., Fast and robust smallest enclosing balls, (European Symposium on Algorithms (1999), Springer), 325-338 |
| [11] | Goel, A.; Indyk, P.; Varadarajan, K. R., Reductions among high dimensional proximity problems, (Proceedings of the Twelfth Annual ACM-SIAM Symposium on Discrete Algorithms (2001), SIAM), 769-778 · Zbl 0988.65022 |
| [12] | Karr, A. F., Extreme points of certain sets of probability measures, with applications, Math. Oper. Res., 8, 1, 74-85 (1983) · Zbl 0509.60007 |
| [13] | Kempthorne, P. J., Numerical specification of discrete least favorable prior distributions, SIAM J. Sci. Stat. Comput., 8, 2, 171-184 (1987) · Zbl 0618.65147 |
| [14] | Kingma, D. P.; Adam, J. Ba., A method for stochastic optimization (2014), arXiv preprint |
| [15] | Lim, T.; McCann, R. J., Geometrical bounds for variance and recentered moments, Math. Oper. Res. (2021) |
| [16] | Lotka, A. J., Analytical note on certain rhythmic relations in organic systems, Proc. Natl. Acad. Sci. USA, 6, 7, 410-415 (1920) |
| [17] | Nocedal, J.; Wright, S., Numerical Optimization (2006), Springer Science & Business Media · Zbl 1104.65059 |
| [18] | Owhadi, H.; Scovel, C., Brittleness of Bayesian inference and new Selberg formulas, Commun. Math. Sci., 14, 1, 83-145 (2016) · Zbl 1357.62125 |
| [19] | Owhadi, H.; Scovel, C., Extreme points of a ball about a measure with finite support, Commun. Math. Sci., 15, 1, 77-96 (2017) · Zbl 1376.60010 |
| [20] | Owhadi, H.; Scovel, C., Qualitative robustness in Bayesian inference, ESAIM Probab. Stat., 21, 251-274 (2017) · Zbl 1395.62059 |
| [21] | Owhadi, H.; Scovel, C., Toward machine Wald, (Owhadi, H.; Ghanem, R.; Higdon, D., Handbook of Uncertainty Quantification (2017), Springer), 157-191 · Zbl 1372.60001 |
| [22] | Owhadi, H.; Scovel, C.; Sullivan, T., Brittleness of Bayesian inference under finite information in a continuous world, Electron. J. Stat., 9, 1, 1-79 (2015) · Zbl 1305.62123 |
| [23] | Owhadi, H.; Scovel, C.; Sullivan, T., On the brittleness of Bayesian inference, SIAM Rev., 57, 4, 566-582 (2015) · Zbl 1341.62094 |
| [24] | Owhadi, H.; Scovel, C.; Sullivan, T. J.; McKerns, M.; Ortiz, M., Optimal uncertainty quantification, SIAM Rev., 55, 2, 271-345 (2013) · Zbl 1278.60040 |
| [25] | Pass, B., Generalized barycenters and variance maximization on metric spaces (2020) |
| [26] | Paszke, A.; Gross, S.; Massa, F.; Lerer, A.; Bradbury, J.; Chanan, G.; Killeen, T.; Lin, Z.; Gimelshein, N.; Antiga, L.; Desmaison, A.; Kopf, A.; Yang, E.; DeVito, Z.; Raison, M.; Tejani, A.; Chilamkurthy, S.; Steiner, B.; Fang, L.; Bai, J.; Chintala, S., Pytorch: an imperative style, high-performance deep learning library, (Wallach, H.; Larochelle, H.; Beygelzimer, A.; d’Alché-Buc, F.; Fox, E.; Garnett, R., Advances in Neural Information Processing Systems, vol. 32 (2019), Curran Associates, Inc.), 8024-8035 |
| [27] | Popoviciu, T., Sur les équations algébriques ayant toutes leurs racines réelles, Mathematica, 129-145 (1935) · Zbl 0014.10003 |
| [28] | Rockafellar, R. T., Convex Analysis (1970), Princeton University Press · Zbl 0193.18401 |
| [29] | Rockafellar, R. T.; Wets, R. J.-B., Variational Analysis, Grundlehren der Mathematischen Wissenschaften., vol. 317 (1998), Springer-Verlag: Springer-Verlag Berlin · Zbl 0888.49001 |
| [30] | Rojo, J., Optimality: The Second Erich L. Lehmann Symposium (2006), IMS |
| [31] | Rojo, J., Optimality: The Third Erich L. Lehmann Symposium (2009), IMS |
| [32] | Rojo, J.; Pérez-Abreu, V., The First Erich L. Lehmann Symposium: Optimality (2004), IMS |
| [33] | Rossi, R. J., Mathematical Statistics: an Introduction to Likelihood Based Inference (2018), John Wiley & Sons · Zbl 1407.62006 |
| [34] | Schott, J. R., Matrix Analysis for Statistics (2016), John Wiley & Sons |
| [35] | Shapiro, A.; Kleywegt, A., Minimax analysis of stochastic problems, Optim. Methods Softw., 17, 3, 523-542 (2002) · Zbl 1040.90030 |
| [36] | Sion, M., On general minimax theorems, Pac. J. Math., 8, 1, 171-176 (1958) · Zbl 0081.11502 |
| [37] | Sprott, D. A., Statistical Inference in Science (2008), Springer Verlag |
| [38] | Stark, P., Your prior can bite you on the posterior: contrasting Bayesian and frequentist measures of uncertainty, (JPL Science Visitor and Colloquium Program - Earth Science Seminar, Sept. 1, 2020 (2020)) |
| [39] | Storn, R.; Price, K., Differential evolution-a simple and efficient heuristic for global optimization over continuous spaces, J. Glob. Optim., 11, 4, 341-359 (1997) · Zbl 0888.90135 |
| [40] | Sylvester, J. J., A question in the geometry of situation, Q. J. Pure Appl. Math., 1, 1, 79-80 (1857) |
| [41] | Virtanen, P.; Gommers, R.; Oliphant, T. E.; Haberland, M.; Reddy, T.; Cournapeau, D.; Burovski, E.; Peterson, P.; Weckesser, W.; Bright, J.; van der Walt, S. J.; Brett, M.; Wilson, J.; Millman, K. J.; Mayorov, N.; Nelson, A. R.J.; Jones, E.; Kern, R.; Larson, E.; Carey, C. J.; Polat, I.; Feng, Y.; Moore, E. W.; VanderPlas, J.; Laxalde, D.; Perktold, J.; Cimrman, R.; Henriksen, I.; Quintero, E. A.; Harris, C. R.; Archibald, A. M.; Ribeiro, A. H.; Pedregosa, F.; van Mulbregt, P., SciPy 1.0: fundamental algorithms for scientific computing in python, Nat. Methods, 17, 261-272 (2020) |
| [42] | von Neumann, J., Zur Theorie der Gesellschaftsspiele, Math. Ann., 100, 1, 295-320 (1928) · JFM 54.0543.02 |
| [43] | von Weizsäcker, H.; Winkler, G., Integral representation in the set of solutions of a generalized moment problem, Math. Ann., 246, 1, 23-32 (1979) · Zbl 0403.46015 |
| [44] | Wald, A., Statistical Decision Functions (1950), John Wiley & Sons Inc.: John Wiley & Sons Inc. New York, NY · Zbl 0040.36402 |
| [45] | Welzl, E., Smallest enclosing disks (balls and ellipsoids), (New Results and New Trends in Computer Science (1991), Springer), 359-370 |
| [46] | Yildirim, E. A., Two algorithms for the minimum enclosing ball problem, SIAM J. Optim., 19, 3, 1368-1391 (2008) · Zbl 1180.90240 |
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. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
