Uncertainty-Guided Alignment for Unsupervised Domain Adaptation in Regression

TL;DR

This paper introduces Uncertainty-Guided Alignment (UGA), a novel method for unsupervised domain adaptation in regression that leverages predictive uncertainty to improve feature alignment and model robustness across diverse tasks.

Contribution

The paper proposes UGA, integrating evidential deep learning for uncertainty estimation to enhance feature alignment in regression domain adaptation, addressing limitations of traditional methods.

Findings

UGA outperforms state-of-the-art methods on 52 transfer tasks.

Incorporates uncertainty to prevent feature collapse in OOD scenarios.

Provides well-calibrated uncertainty estimates alongside improved performance.

Abstract

Unsupervised Domain Adaptation for Regression (UDAR) aims to adapt models from a labeled source domain to an unlabeled target domain for regression tasks. Traditional feature alignment methods, successful in classification, often prove ineffective for regression due to the correlated nature of regression features. To address this challenge, we propose Uncertainty-Guided Alignment (UGA), a novel method that integrates predictive uncertainty into the feature alignment process. UGA employs Evidential Deep Learning to predict both target values and their associated uncertainties. This uncertainty information guides the alignment process and fuses information within the embedding space, effectively mitigating issues such as feature collapse in out-of-distribution scenarios. We evaluate UGA on two computer vision benchmarks and a real-world battery state-of-charge prediction across different manufacturers and operating temperatures. Across 52 transfer tasks, UGA on average outperforms existing state-of-the-art methods. Our approach not only improves adaptation performance but also provides well-calibrated uncertainty estimates.