Federated Koopman-Reservoir Learning for Large-Scale Multivariate Time-Series Anomaly Detection

2025-03-14

AI Summary

To address the challenge of distributed anomaly detection for large-scale heterogeneous multivariate time series (MVTS) in edge computing environments, this paper proposes FedKO—the first federated unsupervised framework integrating Koopman operator theory with reservoir computing. FedKO employs a bi-level optimization formulation to enable privacy-preserving sharing of cross-domain dynamic linearization models, while performing lightweight local anomaly detection. Compared to state-of-the-art methods, FedKO achieves significant improvements in detection accuracy across multiple benchmark datasets, reduces communication overhead by 87.5% (to 1/8), cuts memory consumption by 50%, and supports real-time deployment on edge devices. Its core innovation lies in embedding nonlinear dynamical system modeling (via the Koopman operator) and low-complexity recurrent neural computation (via reservoir computing) into the federated learning paradigm—thereby jointly ensuring expressive model capacity, computational efficiency, and rigorous privacy preservation.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Search and Optimization: Distributed Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

The proliferation of edge devices has dramatically increased the generation of multivariate time-series (MVTS) data, essential for applications from healthcare to smart cities. Such data streams, however, are vulnerable to anomalies that signal crucial problems like system failures or security incidents. Traditional MVTS anomaly detection methods, encompassing statistical and centralized machine learning approaches, struggle with the heterogeneity, variability, and privacy concerns of large-scale, distributed environments. In response, we introduce FedKO, a novel unsupervised Federated Learning framework that leverages the linear predictive capabilities of Koopman operator theory along with the dynamic adaptability of Reservoir Computing. This enables effective spatiotemporal processing and privacy preservation for MVTS data. FedKO is formulated as a bi-level optimization problem, utilizing a specific federated algorithm to explore a shared Reservoir-Koopman model across diverse datasets. Such a model is then deployable on edge devices for efficient detection of anomalies in local MVTS streams. Experimental results across various datasets showcase FedKO's superior performance against state-of-the-art methods in MVTS anomaly detection. Moreover, FedKO reduces up to 8x communication size and 2x memory usage, making it highly suitable for large-scale systems.

Problem

• Detects anomalies in large-scale multivariate time-series data.
• Addresses privacy and heterogeneity in distributed environments.
• Reduces communication size and memory usage for edge devices.

Innovation

• Federated Learning for privacy-preserving MVTS analysis
• Koopman operator theory for linear predictive capabilities
• Reservoir Computing for dynamic adaptability in anomaly detection