Validation of ML-UQ calibration statistics using simulated reference values: a sensitivity analysis

• URL: http://arxiv.org/abs/2403.00423v2
• Date: Mon, 24 Jun 2024 13:43:10 GMT
• Title: Validation of ML-UQ calibration statistics using simulated reference values: a sensitivity analysis
• Authors: Pascal Pernot,
• Abstract summary: Some popular Machine Learning Uncertainty Quantification (ML-UQ) calibration statistics do not have predefined reference values. Simulated reference values, based on synthetic calibrated datasets derived from actual uncertainties, have been proposed to palliate this problem. This study explores various facets of this problem, and shows that some statistics are excessively sensitive to the choice of generative distribution to be used for validation.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Some popular Machine Learning Uncertainty Quantification (ML-UQ) calibration statistics do not have predefined reference values and are mostly used in comparative studies. In consequence, calibration is almost never validated and the diagnostic is left to the appreciation of the reader. Simulated reference values, based on synthetic calibrated datasets derived from actual uncertainties, have been proposed to palliate this problem. As the generative probability distribution for the simulation of synthetic errors is often not constrained, the sensitivity of simulated reference values to the choice of generative distribution might be problematic, shedding a doubt on the calibration diagnostic. This study explores various facets of this problem, and shows that some statistics are excessively sensitive to the choice of generative distribution to be used for validation when the generative distribution is unknown. This is the case, for instance, of the correlation coefficient between absolute errors and uncertainties (CC) and of the expected normalized calibration error (ENCE). A robust validation workflow to deal with simulated reference values is proposed.

Related papers

• Towards Certification of Uncertainty Calibration under Adversarial Attacks [96.48317453951418]
  We show that attacks can significantly harm calibration, and thus propose certified calibration as worst-case bounds on calibration under adversarial perturbations. We propose novel calibration attacks and demonstrate how they can improve model calibration through textitadversarial calibration training
  arXiv  Detail & Related papers  (2024-05-22T18:52:09Z)
• Negative impact of heavy-tailed uncertainty and error distributions on the reliability of calibration statistics for machine learning regression tasks [0.0]
  It is shown that the estimation of MV, MSE and their confidence intervals becomes unreliable for heavy-tailed uncertainty and error distributions. The same problem is expected to affect also conditional calibrations statistics, such as the popular ENCE.
  arXiv  Detail & Related papers  (2024-02-15T16:05:35Z)
• Consistent and Asymptotically Unbiased Estimation of Proper Calibration Errors [23.819464242327257]
  We propose a method that allows consistent estimation of all proper calibration errors and refinement terms. We prove the relation between refinement and f-divergences, which implies information monotonicity in neural networks. Our experiments validate the claimed properties of the proposed estimator and suggest that the selection of a post-hoc calibration method should be determined by the particular calibration error of interest.
  arXiv  Detail & Related papers  (2023-12-14T01:20:08Z)
• Selective Nonparametric Regression via Testing [54.20569354303575]
  We develop an abstention procedure via testing the hypothesis on the value of the conditional variance at a given point. Unlike existing methods, the proposed one allows to account not only for the value of the variance itself but also for the uncertainty of the corresponding variance predictor.
  arXiv  Detail & Related papers  (2023-09-28T13:04:11Z)
• Properties of the ENCE and other MAD-based calibration metrics [0.0]
  The Expected Normalized Error (ENCE) is a popular calibration statistic used in Machine Learning. A similar behavior affects the calibration error based on the variance of z-scores (ZVE), and in both cases this property is a consequence of the use of a Mean Absolute Deviation (MAD) statistic to estimate calibration errors. A solution is proposed to infer ENCE and ZVE values that do not depend on the number of bins for assumed datasets to be calibrated.
  arXiv  Detail & Related papers  (2023-05-17T08:51:42Z)
• Sharp Calibrated Gaussian Processes [58.94710279601622]
  State-of-the-art approaches for designing calibrated models rely on inflating the Gaussian process posterior variance. We present a calibration approach that generates predictive quantiles using a computation inspired by the vanilla Gaussian process posterior variance. Our approach is shown to yield a calibrated model under reasonable assumptions.
  arXiv  Detail & Related papers  (2023-02-23T12:17:36Z)
• Parametric and Multivariate Uncertainty Calibration for Regression and Object Detection [4.630093015127541]
  We show that common detection models overestimate the spatial uncertainty in comparison to the observed error. Our experiments show that the simple Isotonic Regression recalibration method is sufficient to achieve a good calibrated uncertainty. In contrast, if normal distributions are required for subsequent processes, our GP-Normal recalibration method yields the best results.
  arXiv  Detail & Related papers  (2022-07-04T08:00:20Z)
• T-Cal: An optimal test for the calibration of predictive models [49.11538724574202]
  We consider detecting mis-calibration of predictive models using a finite validation dataset as a hypothesis testing problem. detecting mis-calibration is only possible when the conditional probabilities of the classes are sufficiently smooth functions of the predictions. We propose T-Cal, a minimax test for calibration based on a de-biased plug-in estimator of the $ell$-Expected Error (ECE)
  arXiv  Detail & Related papers  (2022-03-03T16:58:54Z)
• Theoretical characterization of uncertainty in high-dimensional linear classification [24.073221004661427]
  We show that uncertainty for learning from limited number of samples of high-dimensional input data and labels can be obtained by the approximate message passing algorithm. We discuss how over-confidence can be mitigated by appropriately regularising, and show that cross-validating with respect to the loss leads to better calibration than with the 0/1 error.
  arXiv  Detail & Related papers  (2022-02-07T15:32:07Z)
• Unsupervised Calibration under Covariate Shift [92.02278658443166]
  We introduce the problem of calibration under domain shift and propose an importance sampling based approach to address it. We evaluate and discuss the efficacy of our method on both real-world datasets and synthetic datasets.
  arXiv  Detail & Related papers  (2020-06-29T21:50:07Z)
• Calibration of Neural Networks using Splines [51.42640515410253]
  Measuring calibration error amounts to comparing two empirical distributions. We introduce a binning-free calibration measure inspired by the classical Kolmogorov-Smirnov (KS) statistical test. Our method consistently outperforms existing methods on KS error as well as other commonly used calibration measures.
  arXiv  Detail & Related papers  (2020-06-23T07:18:05Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.