mlsanityData to 7 Oct 2026

CGSTA: Cross-Scale Graph Contrast with Stability-Aware Alignment for Multivariate Time-Series Anomaly Detection

arXiv 2602.20468 ↗First posted 24 Feb 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Multivariate time-series anomaly detection is essential for reliable industrial control, telemetry, and service monitoring. However, the evolving inter-variable dependencies and inevitable noise render it challenging. Existing methods often use single-scale graphs or instance-level contrast. Moreover, learned dynamic graphs can overfit noise without a stable anchor, causing false alarms or misses. To address these challenges, we propose the CGSTA framework with two key innovations. First, Dynamic Layered Graph Construction (DLGC) forms local, regional, and global views of variable relations for each sliding window; rather than contrasting whole windows, Contrastive Discrimination across Scales (CDS) contrasts graph representations within each view and aligns the same window across views to make learning structure-aware. Second, Stability-Aware Alignment (SAA) maintains a per-scale stable reference learned from normal data and guides the current window's fast-changing graphs toward it to suppress noise. We fuse the multi-scale and temporal features and use a conditional density estimator to produce per-time-step anomaly scores. Across four benchmarks, CGSTA delivers optimal performance on PSM and WADI, and is comparable to the baseline methods on SWaT and SMAP.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 20 of its 28 references that are in the map. First posted on arXiv 24 Feb 2026.

Builds on

20 of 28 references.

References · 20

  1. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
  2. 2018Detecting Spacecraft Anomalies Using LSTMs and Nonparametric Dynamic Thresholdingcited by 407
  3. 2021Anomaly Transformer: Time Series Anomaly Detection with Association Discrepancycited by 387
  4. 2021Graph Neural Network-Based Anomaly Detection in Multivariate Time Seriescited by 300
  5. 2023DCdetector: Dual Attention Contrastive Representation Learning for Time Series Anomaly Detectioncited by 103
  6. 2025TAB: Unified Benchmarking of Time Series Anomaly Detection Methodscited by 58
  7. 2022Graph Anomaly Detection via Multi-Scale Contrastive Learning Networks with Augmented Viewcited by 54
  8. 2023Time Series Contrastive Learning with Information-Aware Augmentationscited by 52
  9. 2023TimesURL: Self-supervised Contrastive Learning for Universal Time Series Representation Learningcited by 51
  10. 2023Soft Contrastive Learning for Time Seriescited by 49
  11. 2022Graph-Augmented Normalizing Flows for Anomaly Detection of Multiple Time Seriescited by 42
  12. 2022Calibrated One-class Classification for Unsupervised Time Series Anomaly Detectioncited by 39
8 more
  1. 2023Fully-Connected Spatial-Temporal Graph for Multivariate Time-Series Datacited by 23
  2. 2022Detecting Multivariate Time Series Anomalies with Zero Known Labelcited by 21
  3. 2023Correlation-aware Spatial-Temporal Graph Learning for Multivariate Time-series Anomaly Detectioncited by 17
  4. 2023CARLA: Self-supervised Contrastive Representation Learning for Time Series Anomaly Detectioncited by 15
  5. 2023Label-Free Multivariate Time Series Anomaly Detectioncited by 14
  6. 2023Rethinking Sensors Modeling: Hierarchical Information Enhanced Traffic Forecastingcited by 11
  7. 2024Multivariate Time-Series Anomaly Detection based on Enhancing Graph Attention Networks with Topological Analysiscited by 11
  8. 2025GCAD: Anomaly Detection in Multivariate Time Series from the Perspective of Granger Causalitycited by 10

Built on by

Not cited yet.