Time–Frequency Correlation-Aware Few-Shot Learning for Time-Series Anomaly Detection
IEEE Sensors Journal · 01.06.2026 · 26 (11) · 17884-17893
DOI: 10.1109/jsen.2026.3684334 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 01.06.2026. Verfasst von Liang Xi und 4 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Gefördert von: Program for Young Talents of Basic Research in Universities of Heilongjiang Province.
Abstract (Originalsprache)
Time-series anomaly detection (TSAD) often leverages unsupervised learning to handle the extensive sensor-generated unlabeled data. However, existing methods ignore to utilize the limited labeled anomalies, and overlook the correlations between data in both the time and frequency domains, resulting in the inability to adequately capture the discrepancies between normal and anomalous patterns. To fully leverage on the exemplary role of labeled anomalies and maximize the representation learning ability, we propose a weakly time–frequency correlation aware few-shot learning (FSL) method for TSAD (TFC-FSL). Specifically, we design an FSL strategy and construct correlation graphs in the time domain and frequency domain, respectively, to fully mine the latent information in the time series. Then, we employ an autoencoder (AE) framework to extract the original and correlation representations in both time and frequency domains, and fuse them with the reconstruction representations. Finally, we design an estimation network with the guidance of few-shot anomalies to amplify the normal/abnormal distribution discrepancies for TSAD tasks. Experimental results show that TFC-FSL outperforms the other models.
Förderung
- Program for Young Talents of Basic Research in Universities of Heilongjiang Province
Zitieren
APA
Xi, L., Li, R., Liu, H., Chen, C., & Cheng, Y. (2026). Time–Frequency Correlation-Aware Few-Shot Learning for Time-Series Anomaly Detection. IEEE Sensors Journal, 26(11), 17884-17893. https://doi.org/10.1109/jsen.2026.3684334
Vancouver
Xi L, Li R, Liu H, Chen C, Cheng Y. Time–Frequency Correlation-Aware Few-Shot Learning for Time-Series Anomaly Detection. IEEE Sensors Journal. 2026;26(11):17884-17893. doi:10.1109/jsen.2026.3684334
BibTeX
@article{Xi2026,
author = {Xi, Liang and Li, Runze and Liu, Han and Chen, Chen and Cheng, Yuan},
title = {Time–Frequency Correlation-Aware Few-Shot Learning for Time-Series Anomaly Detection},
journal = {IEEE Sensors Journal},
year = {2026},
volume = {26},
number = {11},
pages = {17884--17893},
doi = {10.1109/jsen.2026.3684334}
}
EUVIMED