Conformal Uncertainty Indicator for Continual Test-Time Adaptation

TL;DR

This paper introduces a Conformal Uncertainty Indicator (CUI) for Continual Test-Time Adaptation that uses conformal prediction to estimate uncertainty, dynamically compensates for domain shifts, and improves model adaptation performance.

Contribution

The paper proposes a novel CUI method that enhances CTTA by reliably estimating uncertainty and selectively using pseudo-labels, addressing issues caused by domain shifts.

Findings

CUI accurately estimates uncertainty in CTTA.

CUI improves adaptation performance across various methods.

Dynamic coverage compensation enhances reliability of conformal prediction.

Abstract

Continual Test-Time Adaptation (CTTA) aims to adapt models to sequentially changing domains during testing, relying on pseudo-labels for self-adaptation. However, incorrect pseudo-labels can accumulate, leading to performance degradation. To address this, we propose a Conformal Uncertainty Indicator (CUI) for CTTA, leveraging Conformal Prediction (CP) to generate prediction sets that include the true label with a specified coverage probability. Since domain shifts can lower the coverage than expected, making CP unreliable, we dynamically compensate for the coverage by measuring both domain and data differences. Reliable pseudo-labels from CP are then selectively utilized to enhance adaptation. Experiments confirm that CUI effectively estimates uncertainty and improves adaptation performance across various existing CTTA methods.