An Adversarial Filtering Framework With Time–Frequency Cross-Domain Consistency for Multivariate Time-Series Anomaly Detection

Authors:

Abstract
Most data collected in real-world scenarios is contaminated, adversely affecting the performance of existing anomaly detection methods that typically rely on clean training data to learn normal patterns. While some approaches have made progress in addressing data contamination, they still face the following challenges: 1) most methods use pixel-level constraints between the contaminated and reconstructed data as the training loss, and the contaminated data is preferentially learned due to the larger error; and 2) current anomaly detection methods based on the frequency domain do not fully utilize the complex spectral characteristics nor the cross-domain consistency of time-series data, which limits the further improvement of the anomaly detection performance. This article proposes a filter-enhanced reconstruction network with domain-adversarial training for multivariate time-series anomaly detection (FERDA), which enables the model to extract normal patterns and high-level semantic information more accurately by training with the error of filtered and reconstructed data. FERDA designs a dual-domain contamination filtering module consisting of the multihead memory module and the complex linear network. The module is trained to learn normal patterns and cross-domain information through adversarial learning of intradomain semantic consistency and interdomain distinguishability. Moreover, a two-stage contamination suppression strategy is proposed. The contamination degree of data is quantified by the difference between pre- and postfiltering, and the impact of contamination is further mitigated by two-stage training with different losses. The proposed method outperforms 23 state-of-the-art baselines in experiments on five benchmark datasets across different domains.
Funding Information
  • Science and Technology Foundation of State Grid Corporation of China (5216A025000M-057-ZN)
  • Beijing University of Posts and Telecommunications

This publication has 52 references indexed in Scilit: