When is Multicalibration Post-Processing Necessary?

TL;DR

This paper evaluates when multicalibration post-processing is necessary for predictive models, showing it benefits uncalibrated and large models, and that traditional calibration can sometimes imply multicalibration.

Contribution

It provides the first comprehensive evaluation of multicalibration post-processing across diverse datasets and models, offering practical insights and a new Python package.

Findings

Models calibrated out of the box are often multicalibrated without post-processing.

Multicalibration post-processing improves uncalibrated and large models.

Traditional calibration measures may implicitly ensure multicalibration.

Abstract

Calibration is a well-studied property of predictors which guarantees meaningful uncertainty estimates. Multicalibration is a related notion -- originating in algorithmic fairness -- which requires predictors to be simultaneously calibrated over a potentially complex and overlapping collection of protected subpopulations (such as groups defined by ethnicity, race, or income). We conduct the first comprehensive study evaluating the usefulness of multicalibration post-processing across a broad set of tabular, image, and language datasets for models spanning from simple decision trees to 90 million parameter fine-tuned LLMs. Our findings can be summarized as follows: (1) models which are calibrated out of the box tend to be relatively multicalibrated without any additional post-processing; (2) multicalibration post-processing can help inherently uncalibrated models and large vision and language models; and (3) traditional calibration measures may sometimes provide multicalibration implicitly. More generally, we also distill many independent observations which may be useful for practical and effective applications of multicalibration post-processing in real-world contexts. We also release a python package implementing multicalibration algorithms, available via `pip install multicalibration'.