A Comparative Study of Adaptation Strategies for Time Series Foundation Models in Anomaly Detection

TL;DR

This paper evaluates the effectiveness of time series foundation models (TSFMs) for anomaly detection, showing they outperform traditional methods and that parameter-efficient fine-tuning offers a scalable adaptation approach.

Contribution

It systematically compares adaptation strategies for TSFMs in anomaly detection, highlighting the effectiveness of PEFT methods like LoRA, OFT, and HRA.

Findings

TSFMs outperform task-specific baselines in anomaly detection.

PEFT methods match or surpass full fine-tuning performance.

TSFMs are effective even when pretrained for forecasting.

Abstract

Time series anomaly detection is essential for the reliable operation of complex systems, but most existing methods require extensive task-specific training. We explore whether time series foundation models (TSFMs), pretrained on large heterogeneous data, can serve as universal backbones for anomaly detection. Through systematic experiments across multiple benchmarks, we compare zero-shot inference, full model adaptation, and parameter-efficient fine-tuning (PEFT) strategies. Our results demonstrate that TSFMs outperform task-specific baselines, achieving notable gains in AUC-PR and VUS-PR, particularly under severe class imbalance. Moreover, PEFT methods such as LoRA, OFT, and HRA not only reduce computational cost but also match or surpass full fine-tuning in most cases, indicating that TSFMs can be efficiently adapted for anomaly detection, even when pretrained for forecasting. These findings position TSFMs as promising general-purpose models for scalable and efficient time series anomaly detection.