MadSGM: Multivariate Anomaly Detection with Score-based Generative Models
This work addresses the challenge of unsupervised anomaly detection in multivariate time-series, which is critical for applications like monitoring and fault detection, but it is incremental as it builds on existing score-based generative models.
The paper tackled the problem of multivariate time-series anomaly detection by proposing MadSGM, a method based on score-based generative models that integrates reconstruction-based, density-based, and gradient-based anomaly measurements, achieving the most robust and accurate predictions on five real-world benchmark datasets.
The time-series anomaly detection is one of the most fundamental tasks for time-series. Unlike the time-series forecasting and classification, the time-series anomaly detection typically requires unsupervised (or self-supervised) training since collecting and labeling anomalous observations are difficult. In addition, most existing methods resort to limited forms of anomaly measurements and therefore, it is not clear whether they are optimal in all circumstances. To this end, we present a multivariate time-series anomaly detector based on score-based generative models, called MadSGM, which considers the broadest ever set of anomaly measurement factors: i) reconstruction-based, ii) density-based, and iii) gradient-based anomaly measurements. We also design a conditional score network and its denoising score matching loss for the time-series anomaly detection. Experiments on five real-world benchmark datasets illustrate that MadSGM achieves the most robust and accurate predictions.