Generalist Representation, Specialist Detection: TS-Router for Time-Series Anomaly Detection

2026-09-30

AI Summary

This study addresses the challenges of cross-domain generalization in time series anomaly detection and the limitations of fixed scoring schemes that overlook heterogeneous dynamics. To this end, it proposes TS-Router, a novel framework that pioneers a "generalist representation, specialist detection" paradigm. Rather than directly assigning anomaly scores, the method leverages time series foundation models to extract universal representations for coordinating specialized detectors. Furthermore, it introduces simulation-task-based relative capability soft supervision alongside a Top-k adaptive routing algorithm, enabling precise detection during label-free deployment. The proposed approach achieves the best average ranking across 16 real-world benchmarks and provides theoretical guarantees with bounded regret. Ablation studies further validate the effectiveness of pretrained representations in capability estimation.

Abstract

Time-series anomaly detection (TSAD) is difficult to generalize across datasets because heterogeneous temporal dynamics imply different notions of normality and favor different detection criteria. While time-series foundation models provide transferable representations, coupling them with a fixed anomaly-scoring mechanism can overlook this variation. This motivates a different perspective on foundation-model-based TSAD: using foundation models to coordinate specialized anomaly criteria rather than directly imposing a universal one. Based on this view, we propose \textbf{TS-Router}, a generalist-representation, specialist-detection framework that estimates the relative competence of heterogeneous anomaly detectors from pretrained temporal representations and selects suitable specialists for each target series. To avoid relying on specialist-performance labels from real tasks, we derive soft competence supervision from specialists' relative performance on labeled simulated tasks. At deployment, routing requires no target anomaly labels, and only the selected specialists are fitted unsupervisedly on the target series. We bound Top-\(k\) set-competence regret under representation coverage and conditional competence stability. Across 16 real-world benchmarks and four complementary evaluation metrics, TS-Router achieves the best overall average rank. Controlled ablations with multiple frozen TSFM encoders further support the use of pretrained representations for competence estimation and adaptive specialist selection. The code is available at https://anonymous.4open.science/r/TS-Router-D8FF.

Innovation

• Time-Series Anomaly Detection
• Foundation Models
• Mixture of Experts
• Routing Mechanism
• Competence Estimation