Time series anomaly detection (TSAD) is crucial for maintaining the reliability and security of IoT-enabled service systems. Existing methods require training a specific model for each dataset, which exhibits limited generalization capability across different target datasets, hindering anomaly detection performance in scenarios with scarce training data.
To address this limitation, foundation models have emerged as a promising direction. Current approaches either repurpose large language models (LLMs) or construct large-scale time series datasets to develop general anomaly detection foundation models, but still face challenges due to severe cross-modal gaps or in-domain heterogeneity.
In this paper, we investigate the applicability of large-scale vision models to TSAD. Specifically, we adapt a visual Masked Autoencoder (MAE) pretrained on ImageNet for the TSAD task. However, directly transferring MAE to TSAD introduces two key challenges: overgeneralization and limited local perception. To tackle these challenges, we propose VAN-AD, a novel MAE-based framework for TSAD.
To alleviate the overgeneralization issue, we design an Adaptive Distribution Mapping Module (ADMM), which maps the reconstruction results before and after MAE into a unified statistical space to amplify discrepancies caused by abnormal patterns. To overcome the limitation of local perception, we further develop a Normalizing Flow Module (NFM), which combines MAE with normalizing flow to estimate the probability density of the current window under the global distribution.
Extensive experiments on nine real-world datasets demonstrate that VAN-AD consistently outperforms existing state-of-the-art methods across multiple evaluation metrics. Our code and datasets are available at GitHub.
Blogger's Review: This study successfully overcomes traditional limitations by introducing visual masked autoencoders and normalizing flow modules, showcasing the potential of foundation models in time series anomaly detection. It's worth exploring further in practical applications.