DeepFilter: An Instrumental Baseline for Accurate and Efficient Process Monitoring

2025-01-02

AI Summary

To address the weak semantic modeling and high computational complexity (O(N²)) of Transformer self-attention in industrial process monitoring—hindering simultaneous achievement of accuracy and real-time performance—this paper proposes GlobalFilter, a novel global filtering layer based on learnable 1D convolution and frequency-domain awareness, replacing self-attention to enable linear-complexity (O(N)) long-range and periodic temporal modeling. The method integrates lightweight positional encoding and segment-wise normalization to significantly enhance discriminative representation learning. Evaluated on multiple real-world industrial datasets, the proposed approach achieves an average 3.2% improvement in F1-score over state-of-the-art methods while accelerating inference by 5.8×. To our knowledge, this is the first work to achieve synergistic optimization of high accuracy and real-time capability in industrial process monitoring.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Natural Language Processing: Learning & Optimization for NLP
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications

Abstract

Effective process monitoring is increasingly vital in industrial automation for ensuring operational safety, necessitating both high accuracy and efficiency. Although Transformers have demonstrated success in various fields, their canonical form based on the self-attention mechanism is inadequate for process monitoring due to two primary limitations: (1) the step-wise correlations captured by self-attention mechanism are difficult to capture discriminative patterns in monitoring logs due to the lacking semantics of each step, thus compromising accuracy; (2) the quadratic computational complexity of self-attention hampers efficiency. To address these issues, we propose DeepFilter, a Transformer-style framework for process monitoring. The core innovation is an efficient filtering layer that excel capturing long-term and periodic patterns with reduced complexity. Equipping with the global filtering layer, DeepFilter enhances both accuracy and efficiency, meeting the stringent demands of process monitoring. Experimental results on real-world process monitoring datasets validate DeepFilter's superiority in terms of accuracy and efficiency compared to existing state-of-the-art models.

Problem

• Transformer Models
• Process Monitoring
• Attention Mechanism

Innovation

• Transformer-like framework
• Filtering layer
• Process monitoring