Aug 2026· 2026 6th International Conference on Advanced Algorithms and Neural Networks (AANN)· pp. 1-7· 0 citations· 19 references
Abstract
Reconstruction-based multivariate time-series anomaly detection remains challenging because anomaly evidence may appear in temporal evolution, cross-variable relationships, and weak local variation patterns. To address this issue, we propose DualFormer, a dual-view Transformer framework that constructs temporal-patch and variable-token representations from the same input window. The two views are aligned into a shared variable-indexed space and coupled through feature-wise adaptive fusion before reconstruction, allowing temporal and variable evidence to jointly shape the reconstructed sequence. In addition, a first-order temporal-difference-aware objective and scoring strategy is introduced to capture local variation discrepancies beyond pointwise value errors. Experiments on a real-world aviation QAR dataset and two public benchmarks, Mars Science Laboratory (MSL) and SWaT, show competitive detection performance. Ablation and qualitative analyses further support the roles of dual-view fusion and temporal-difference-aware scoring.
Deployment Oriented Dual Path Multi View Anomaly Detection (DP-MVAD), a deployment-efficient framework designed for complex non-stationary conditions, achieves trend-perturbation disentanglement through a conditionally identifiable dual-path approximation and ensures robust cross-scenario representations.
Li-Rong Jin, Yi Liu, Qiang Jiang et al.· Cluster Computing· 0 citations
Reconstruction-based detectors for multivariate time series anomaly detection typically learn a single entangled representation of normal behavior and flag anomalies based on reconstruction error. However, real-world anomalies are not only isolated spikes in a single channel but also disruptions in the relationships be...
M. Raihan, Zahra Zamanzadeh Darban, Yue-Yang Liu et al.· 0 citations
D is proposed, a dual-encoder framework that separates reconstruction-oriented representation learning from one-class boundary modeling by imposing the SVDD constraint in a dedicated boundary space rather than directly on the reconstruction representation, which reduces the gradient conflict and scale imbalance between...
Xin Xie, Huichaoyi Zhang, Ke-Xuan Liu et al.· Applied intelligence (Boston...· 0 citations
Time-series anomaly detection (TSAD) is critical to guarantee the security of industrial control systems (ICS). However, harsh operating conditions and complex control commands may give rise to multiple intertwined feature patterns in ICS sensor signals, which are further compressed into compact 1D time-series data. Mo...
Lei Chen, Ye-Peng Xu, Ming Lu et al.· IEEE Transactions on Automat...· 0 citations
Reconstruction errors in multivariate time-series anomaly detection may not reliably distinguish abnormal behavior from benign deviations. Language-derived semantics offer complementary context, but existing multimodal approaches may rely on time-associated paired textual information that is difficult to obtain consist...
Jahyeob Koo, Kio Yun, Byoung-Mo Koo et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
