Normality Constraint Learning: Adapting Foundation Models for Time Series Anomaly Detection

2026-10-05

AI Summary

This study addresses the detection failure of time-series foundation models caused by overfitting to anomalous patterns. To this end, we propose NCL, a plug-and-play adaptation framework that operates without modifying pretrained parameters. Specifically, NCL constructs a normality subspace and incorporates contrastive learning constraints to guide feature representations toward the normal manifold. It further calibrates reconstruction error scores through subspace projection and feature fusion, thereby effectively enhancing the model's sensitivity to anomalies. Extensive evaluations across multiple benchmarks demonstrate that NCL significantly improves anomaly detection performance, validating the effectiveness of this general-purpose adaptation framework.

Abstract

Time Series Foundation Models (TSFMs) achieve strong generalization by learning to reconstruct or forecast broad temporal patterns from large-scale time series during pre-training. Yet this strength can become a weakness for anomaly detection: TSFMs may model rare anomalous patterns as effectively as normal ones, allowing anomalies to be accurately reconstructed or forecasted and thus diminishing their reconstruction/forecasting error-based anomaly scores. This paper proposes $\underline{\textbf{N}}$$\textbf{ormality}$ $\underline{\textbf{C}}$$\textbf{onstraint}$ $\underline{\textbf{L}}$$\textbf{earning}$ ($\textbf{NCL}$), a lightweight plug-and-play framework that adapts pre-trained TSFMs for accurate anomaly detection without modifying their pre-trained parameters. Our key insight is to constrain the broad pattern space of TSFMs to the normal structure of a target time series, preventing their broad modeling capability from obscuring abnormal deviations. Specifically, NCL constructs a compact normality subspace from a few normal patch features and adaptively steers each patch feature toward normality within this subspace, guided by contrastive constraints that form compact and discriminative normality manifolds. The calibrated features are aggregated to reinforce normal components and fused with the original TSFM output, amplifying the discrepancy between normal and abnormal observations for the reconstruction/forecasting error-based anomaly scoring. Extensive experiments across diverse TSFM families and benchmarks show that NCL consistently improves anomaly detection performance, providing a generalizable framework for adapting TSFMs to anomaly detection.

Problem

• Time Series Foundation Models
• Anomaly Detection
• Reconstruction Error
• Generalization

Innovation

• Time Series Foundation Models
• Anomaly Detection
• Normality Constraint Learning
• Contrastive Constraints
• Plug-and-play Framework