RoseCDL: Robust and Scalable Convolutional Dictionary Learning for Rare-event Detection

2025-09-09

AI Summary

Addressing the challenge of detecting rare events and anomalous patterns in large-scale signals, this paper proposes a robust and scalable convolutional dictionary learning (CDL) framework. To overcome the high computational cost and sensitivity to noise and outliers inherent in conventional CDL, we introduce stochastic window sampling to reduce time complexity and integrate an online anomaly detection mechanism that dynamically identifies and suppresses anomalous responses during sparse coding. This work is the first to explicitly reformulate CDL as an unsupervised learning paradigm tailored for rare-event discovery, achieving both efficiency and robustness. Experiments demonstrate substantial improvements in anomaly localization accuracy and computational efficiency on long-duration time-series signals. The method has been successfully deployed in real-world applications, including pulsar detection in astronomical data and abnormal pattern identification in neural electrophysiological recordings.

Technology Category

• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Online Learning & Bandits
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Identifying recurring patterns and rare events in large-scale signals is a fundamental challenge in fields such as astronomy, physical simulations, and biomedical science. Convolutional Dictionary Learning (CDL) offers a powerful framework for modeling local structures in signals, but its use for detecting rare or anomalous events remains largely unexplored. In particular, CDL faces two key challenges in this setting: high computational cost and sensitivity to artifacts and outliers. In this paper, we introduce RoseCDL, a scalable and robust CDL algorithm designed for unsupervised rare event detection in long signals. RoseCDL combines stochastic windowing for efficient training on large datasets with inline outlier detection to enhance robustness and isolate anomalous patterns. This reframes CDL as a practical tool for event discovery and characterization in real-world signals, extending its role beyond traditional tasks like compression or denoising.

Problem

• Detects rare events in large-scale signals efficiently
• Addresses computational cost and outlier sensitivity in CDL
• Enables unsupervised pattern discovery in real-world data

Innovation

• Stochastic windowing for efficient large-scale training
• Inline outlier detection to enhance robustness
• Unsupervised rare event detection in long signals