Decentralized Time Series Classification with ROCKET Features

2025-04-24

AI Summary

In privacy-sensitive, geographically distributed, and regulation-constrained time series classification (TSC) scenarios, conventional client-server federated learning (FL) suffers from single-point failure, strong reliance on a trusted central server, and high communication overhead. Method: We propose DROCKS—the first fully decentralized FL framework for TSC—eliminating the central server entirely and enabling collaborative model updates via a structured chain topology among peers. DROCKS introduces a node-adaptive ROCKET kernel selection mechanism that locally evaluates and compresses convolutional kernels based on discriminative power, jointly optimizing feature representation fidelity and communication efficiency. Contribution/Results: Evaluated on the UCR benchmark, DROCKS achieves accuracy comparable to state-of-the-art server-based FL methods while reducing total communication volume by 37%. Moreover, it demonstrates显著 robustness against node failures and adversarial attacks, validating its suitability for real-world decentralized, privacy-critical TSC applications.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Distributed Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Decentralized Web and Fediverse systems
• Security and Privacy: Blockchains and distributed ledgers

Abstract

Time series classification (TSC) is a critical task with applications in various domains, including healthcare, finance, and industrial monitoring. Due to privacy concerns and data regulations, Federated Learning has emerged as a promising approach for learning from distributed time series data without centralizing raw information. However, most FL solutions rely on a client-server architecture, which introduces robustness and confidentiality risks related to the distinguished role of the server, which is a single point of failure and can observe knowledge extracted from clients. To address these challenges, we propose DROCKS, a fully decentralized FL framework for TSC that leverages ROCKET (RandOm Convolutional KErnel Transform) features. In DROCKS, the global model is trained by sequentially traversing a structured path across federation nodes, where each node refines the model and selects the most effective local kernels before passing them to the successor. Extensive experiments on the UCR archive demonstrate that DROCKS outperforms state-of-the-art client-server FL approaches while being more resilient to node failures and malicious attacks. Our code is available at https://anonymous.4open.science/r/DROCKS-7FF3/README.md.

Problem

• Decentralized federated learning for time series classification
• Eliminating server dependency to reduce failure risks
• Improving robustness against node failures and attacks

Innovation

• Decentralized FL framework for TSC
• Uses ROCKET features for classification
• Sequential path training across nodes