PRISM: Lightweight Multivariate Time-Series Classification through Symmetric Multi-Resolution Convolutional Layers

2025-08-06

AI Summary

Multivariate time-series classification suffers from high computational overhead, weak frequency-domain modeling capability, and parameter redundancy. To address these challenges, this paper proposes PRISM—a lightweight convolutional feature extraction module that innovatively integrates classical signal processing with deep learning. PRISM employs channel-wise, multi-scale symmetric finite-impulse-response (FIR) filters for 1D convolution, enabling frequency-selective and multi-resolution temporal modeling without cross-channel computation. Coupled with a lightweight classification head, PRISM drastically reduces model size while maintaining competitive accuracy. Evaluated on human activity recognition, sleep staging, and biomedical datasets, it matches or surpasses state-of-the-art CNN- and Transformer-based methods, achieving approximately 90% reduction in both parameter count and FLOPs.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Computer Vision: Multi-modal Vision
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Multivariate time-series classification is pivotal in domains ranging from wearable sensing to biomedical monitoring. Despite recent advances, Transformer- and CNN-based models often remain computationally heavy, offer limited frequency diversity, and require extensive parameter budgets. We propose PRISM (Per-channel Resolution-Informed Symmetric Module), a convolutional-based feature extractor that applies symmetric finite-impulse-response (FIR) filters at multiple temporal scales, independently per channel. This multi-resolution, per-channel design yields highly frequency-selective embeddings without any inter-channel convolutions, greatly reducing model size and complexity. Across human-activity, sleep-stage and biomedical benchmarks, PRISM, paired with lightweight classification heads, matches or outperforms leading CNN and Transformer baselines, while using roughly an order of magnitude fewer parameters and FLOPs. By uniting classical signal processing insights with modern deep learning, PRISM offers an accurate, resource-efficient solution for multivariate time-series classification.

Problem

• Reducing computational complexity in multivariate time-series classification
• Enhancing frequency diversity without inter-channel convolutions
• Minimizing parameter budgets while maintaining accuracy

Innovation

• Symmetric multi-resolution convolutional layers
• Per-channel FIR filters design
• Lightweight classification heads integration