Estimating Expected Calibration Errors
This work addresses the problem of accurately assessing calibration in classifiers, which is crucial for reliable decision-making in applications like human-in-the-loop systems, but it is incremental as it focuses on improving existing metrics rather than introducing a new calibration method.
The paper evaluates different estimators of the Expected Calibration Error (ECE) for probabilistic classifiers, proposing novel estimators and using an empirical procedure to recommend which estimator to use in various settings.
Uncertainty in probabilistic classifiers predictions is a key concern when models are used to support human decision making, in broader probabilistic pipelines or when sensitive automatic decisions have to be taken. Studies have shown that most models are not intrinsically well calibrated, meaning that their decision scores are not consistent with posterior probabilities. Hence being able to calibrate these models, or enforce calibration while learning them, has regained interest in recent literature. In this context, properly assessing calibration is paramount to quantify new contributions tackling calibration. However, there is room for improvement for commonly used metrics and evaluation of calibration could benefit from deeper analyses. Thus this paper focuses on the empirical evaluation of calibration metrics in the context of classification. More specifically it evaluates different estimators of the Expected Calibration Error ($ECE$), amongst which legacy estimators and some novel ones, proposed in this paper. We build an empirical procedure to quantify the quality of these $ECE$ estimators, and use it to decide which estimator should be used in practice for different settings.