Scholarlink

<p>Multivariate time series (MTS) anomaly detection faces critical challenges, including complex sensor interdependencies, environmental noise, and the inefficiency of modeling long-range dependencies. Existing methods often fail to balance computational efficiency with fine-grained modeling of normal patterns. To address these issues, we propose P-ALIGN, a framework that integrates patch-based feature extraction with prototypical alignment and contrastive learning. Specifically, a patching mechanism is employed to capture long-term context with linear complexity, while an EmbedPatch encoder learns a set of normal prototypes as global reference points. By aligning latent features with these prototypes, P-ALIGN effectively suppresses noise and prevents the model from over-reconstructing anomalies. Furthermore, a Contrastive Fusion module is introduced to enlarge the discriminative gap between normal and abnormal distributions. Extensive experiments on six real-world benchmarks demonstrate that P-ALIGN consistently outperforms state-of-the-art methods, achieving an 11% improvement in F1-score and a 12.23% increase in the Normalized Affinity (NAff) metric.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prototypical contrastive learning with patch-based spatio-temporal alignment for multivariate time series anomaly detection

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月11日

查看全文
  • Chaoyi Yang,
  • Xuewu Li,
  • Kunhuan Xu,
  • Xijun Lin,
  • Gehua Fu

摘要

Multivariate time series (MTS) anomaly detection faces critical challenges, including complex sensor interdependencies, environmental noise, and the inefficiency of modeling long-range dependencies. Existing methods often fail to balance computational efficiency with fine-grained modeling of normal patterns. To address these issues, we propose P-ALIGN, a framework that integrates patch-based feature extraction with prototypical alignment and contrastive learning. Specifically, a patching mechanism is employed to capture long-term context with linear complexity, while an EmbedPatch encoder learns a set of normal prototypes as global reference points. By aligning latent features with these prototypes, P-ALIGN effectively suppresses noise and prevents the model from over-reconstructing anomalies. Furthermore, a Contrastive Fusion module is introduced to enlarge the discriminative gap between normal and abnormal distributions. Extensive experiments on six real-world benchmarks demonstrate that P-ALIGN consistently outperforms state-of-the-art methods, achieving an 11% improvement in F1-score and a 12.23% increase in the Normalized Affinity (NAff) metric.