Confidence Calibration for Domain Generalization under Covariate Shift

TL;DR

This paper introduces a novel calibration method for domain generalization that reduces reliance on target domain data and improves calibration transfer across domains, supported by theoretical and empirical results.

Contribution

It proposes calibration algorithms leveraging multiple source domains to enhance calibration transfer without needing target domain data, addressing key limitations of existing methods.

Findings

Achieved an 8.86 percentage point reduction in expected calibration error.

Demonstrated improved calibration transfer across multiple domains.

Validated effectiveness through theoretical analysis and experiments.

Abstract

Existing calibration algorithms address the problem of covariate shift via unsupervised domain adaptation. However, these methods suffer from the following limitations: 1) they require unlabeled data from the target domain, which may not be available at the stage of calibration in real-world applications and 2) their performance depends heavily on the disparity between the distributions of the source and target domains. To address these two limitations, we present novel calibration solutions via domain generalization. Our core idea is to leverage multiple calibration domains to reduce the effective distribution disparity between the target and calibration domains for improved calibration transfer without needing any data from the target domain. We provide theoretical justification and empirical experimental results to demonstrate the effectiveness of our proposed algorithms. Compared against state-of-the-art calibration methods designed for domain adaptation, we observe a decrease of 8.86 percentage points in expected calibration error or, equivalently, an increase of 35 percentage points in improvement ratio for multi-class classification on the Office-Home dataset.