CANDI: Curated Test-Time Adaptation for Multivariate Time-Series Anomaly Detection Under Distribution Shift

2026-04-02

AI Summary

This work addresses the significant performance degradation of multivariate time series anomaly detection under distribution shifts. To tackle this challenge, the authors propose a test-time adaptation framework that leverages a false positive mining (FPM) strategy to identify potential false alarm samples and introduces a plug-and-play spatio-temporal awareness normality adaptation (SANA) module. This module enables efficient adaptation while preserving pre-trained knowledge. Requiring only unlabeled test data, the method substantially improves detection performance across various distribution shift scenarios, achieving up to a 14% gain in AUROC. Moreover, it attains this enhancement with fewer adaptation samples, demonstrating superior robustness and practicality.

Abstract

Multivariate time-series anomaly detection (MTSAD) aims to identify deviations from normality in multivariate time-series and is critical in real-world applications. However, in real-world deployments, distribution shifts are ubiquitous and cause severe performance degradation in pre-trained anomaly detector. Test-time adaptation (TTA) updates a pre-trained model on-the-fly using only unlabeled test data, making it promising for addressing this challenge. In this study, we propose CANDI (Curated test-time adaptation for multivariate time-series ANomaly detection under DIstribution shift), a novel TTA framework that selectively identifies and adapts to potential false positives while preserving pre-trained knowledge. CANDI introduces a False Positive Mining (FPM) strategy to curate adaptation samples based on anomaly scores and latent similarity, and incorporates a plug-and-play Spatiotemporally-Aware Normality Adaptation (SANA) module for structurally informed model updates. Extensive experiments demonstrate that CANDI significantly improves the performance of MTSAD under distribution shift, improving AUROC up to 14% while using fewer adaptation samples.

Problem

• multivariate time-series anomaly detection
• distribution shift
• test-time adaptation
• false positives

Innovation

• Test-Time Adaptation
• Multivariate Time-Series Anomaly Detection
• Distribution Shift
• False Positive Mining
• Spatiotemporal Adaptation