Knowledge Aware State-Space Capsule Network for Multivariate Time Series Classification

作者：Zhiwen Xiao, Qian Wan, Wei-Ping Ding, Fuhong Song, Huagang Tong · 发表于：IEEE Transactions on Knowledge and Data Engineering · 年份：2026 · DOI：10.1109/tkde.2026.3692700 · 被引用次数：2 · 研究领域：Computer Science

Multivariate time series classification (MTSC) requires simultaneous modeling of local temporal patterns, long-range dependencies, and complex inter-variable relationships. However, existing convolutional capsule models suffer from limited receptive fields, while transformer-based variants struggle to preserve local structures in long sequences. To overcome these limitations, we propose KACapMamba, a Knowledge-Aware State-Space Capsule Network, which integrates three attentive Mamba blocks with a routing layer to achieve hierarchical temporal modeling. Unlike conventional transformer-based methods that primarily depend on self-attention for global dependency modeling, each attentive Mamba block in KACapMamba fuses 1-dimensional convolutional neural networks (CNNs), self-attention, state-space module (SSM), and mutual cross-attention, enabling a more structured and adaptive feature representation. Self-attention ensures effective long-range dependency modeling, while SSM provides a recurrent-state mechanism, inherently better suited for sequential processing compared to purely attention-based architectures, thereby enhancing temporal continuity and long-term pattern retention. Notably, mutual crossattention addresses the limitations of traditional fusion strategies such as element-wise addition or multiplication, which lack the capacity to selectively enhance relevant features. By dynamically reweighting interactions between features, mutual crossattention enables more express...