Discretization-free Multicalibration through Loss Minimization over Tree Ensembles

TL;DR

This paper introduces a novel discretization-free multicalibration method that directly optimizes risk over tree ensembles, avoiding discretization errors and hyperparameters, and demonstrates superior empirical performance.

Contribution

The authors propose a new ERM-based multicalibration approach using decision trees, eliminating the need for output discretization and hyperparameter tuning, with theoretical guarantees and practical effectiveness.

Findings

Method achieves multicalibration under loss saturation condition.

Empirical results show consistent outperforming of existing methods.

Approach integrates with standard tree ensemble algorithms like LightGBM.

Abstract

In recent years, multicalibration has emerged as a desirable learning objective for ensuring that a predictor is calibrated across a rich collection of overlapping subpopulations. Existing approaches typically achieve multicalibration by discretizing the predictor's output space and iteratively adjusting its output values. However, this discretization approach departs from the standard empirical risk minimization (ERM) pipeline, introduces rounding error and additional sensitive hyperparameter, and may distort the predictor's outputs in ways that hinder downstream decision-making. In this work, we propose a discretization-free multicalibration method that directly optimizes an empirical risk objective over an ensemble of depth-two decision trees. Our ERM approach can be implemented using off-the-shelf tree ensemble learning methods such as LightGBM. Our algorithm provably achieves multicalibration, provided that the data distribution satisfies a technical condition we term as loss saturation. Across multiple datasets, our empirical evaluation shows that this condition is always met in practice. Our discretization-free algorithm consistently matches or outperforms existing multicalibration approaches--even when evaluated using a discretization-based multicalibration metric that shares its discretization granularity with the baselines.