K-Means Based TinyML Anomaly Detection and Distributed Model Reuse via the Distributed Internet of Learning (DIoL)

2026-03-28

AI Summary

This work addresses anomaly detection on resource-constrained microcontrollers by proposing a “Train Once, Share Everywhere” (TOSE) paradigm that eliminates the need for per-device retraining. The approach leverages a lightweight K-Means clustering algorithm to perform local feature extraction and adaptive threshold estimation, integrated within a Distributed Internet-of-Learning (DIoL) framework. By encoding models into a textual representation, the method enables direct cross-device model reuse without additional training. Experimental validation on a dual-device prototype demonstrates that the shared model maintains consistent detection performance, achieves inference speeds comparable to independently deployed models, and incurs negligible parsing overhead. This significantly enhances deployment efficiency and system scalability in edge environments with limited computational resources.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Search and Optimization: Distributed Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

This paper presents a lightweight K-Means anomaly detection model and a distributed model-sharing workflow designed for resource-constrained microcontrollers (MCUs). Using real power measurements from a mini-fridge appliance, the system performs on-device feature extraction, clustering, and threshold estimation to identify abnormal appliance behavior. To avoid retraining models on every device, we introduce the Distributed Internet of Learning (DIoL), which enables a model trained on one MCU to be exported as a portable, text-based representation and reused directly on other devices. A two-device prototype demonstrates the feasibility of the "Train Once, Share Everywhere" (TOSE) approach using a real-world appliance case study, where Device A trains the model and Device B performs inference without retraining. Experimental results show consistent anomaly detection behavior, negligible parsing overhead, and identical inference runtimes between standalone and DIoL-based operation. The proposed framework enables scalable, low-cost TinyML deployment across fleets of embedded devices.

Problem

• TinyML
• anomaly detection
• model reuse
• resource-constrained devices
• distributed learning

Innovation

• TinyML
• K-Means anomaly detection
• Distributed Internet of Learning
• model reuse
• microcontroller