🤖 AI Summary
This work addresses representation distortion and noise sensitivity in multivariate time series classification, which arise from non-causal modeling and channel independence assumptions. To this end, the authors propose a Structured Manifold Preconditioning Network that jointly integrates causal self-attention, causal convolutions, and an adaptive channel recalibration mechanism—enabling the first unified modeling of temporal causal structure and channel-wise information bottlenecks. This approach effectively mitigates temporal confounding under non-stationary dynamics and suppresses noise in latent representations. The method achieves new state-of-the-art performance across four tasks spanning six heterogeneous domains, attaining 98.6% accuracy on the AWR dataset and demonstrating exceptional robustness in non-stationary scenarios.
📝 Abstract
Multivariate time series (MTS) classification is foundational to pervasive computing and financial analysis, yet existing multi-scale paradigms are often constrained by suboptimal representation fidelity. We identify two critical bottlenecks: temporal non-causality in standard encoders that induces temporal confounding in non-stationary dynamics, and the absence of explicit channel saliency mechanisms that allows noise to contaminate the latent space. To address these challenges, we propose the Causal Attention and Spatio-temporal Encoder Network (CASE-NET), an architecture designed for structural manifold pre-conditioning. CASE-NET synergizes a Causal Temporal Encoder, which enforces physical arrow-of-time constraints via masked self-attention and causal convolutions, with an Adaptive Channel Recalibration module functioning as an information bottleneck to suppress detrimental noise. Comprehensive evaluations across six heterogeneous domains demonstrate that CASE-NET establishes new state-of-the-art benchmarks on four tasks, achieving a peak accuracy of 98.6% on the AWR dataset and superior robustness in non-stationary regimes.