Link-Aware Energy-Frugal Continual Learning for Fault Detection in IoT Networks

2025-12-15

AI Summary

To address the accuracy degradation of lightweight machine learning models on IoT devices caused by non-stationary wireless environments and limited initial data, this paper proposes an event-driven collaborative continual learning framework. The framework introduces a novel dual-aware event-triggering mechanism that jointly perceives link-state dynamics and energy budgets, enabling co-optimization of edge-model updates, lightweight model inference, channel-adaptive communication, and energy-constrained scheduling. This design overcomes fundamental limitations of conventional periodic sampling and static continual learning in terms of energy efficiency and environmental adaptability. Evaluated on real-world IoT datasets, the proposed method achieves up to a 42.8% improvement in fault detection recall over periodic sampling and non-adaptive continual learning baselines, demonstrating superior trade-offs between inference accuracy and energy efficiency.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Intelligent Robots: Learning & Optimization for ROB
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The use of lightweight machine learning (ML) models in internet of things (IoT) networks enables resource constrained IoT devices to perform on-device inference for several critical applications. However, the inference accuracy deteriorates due to the non-stationarity in the IoT environment and limited initial training data. To counteract this, the deployed models can be updated occasionally with new observed data samples. However, this approach consumes additional energy, which is undesirable for energy constrained IoT devices. This letter introduces an event-driven communication framework that strategically integrates continual learning (CL) in IoT networks for energy-efficient fault detection. Our framework enables the IoT device and the edge server (ES) to collaboratively update the lightweight ML model by adapting to the wireless link conditions for communication and the available energy budget. Evaluation on real-world datasets show that the proposed approach can outperform both periodic sampling and non-adaptive CL in terms of inference recall; our proposed approach achieves up to a 42.8% improvement, even under tight energy and bandwidth constraint.

Problem

• Develops an event-driven communication framework for IoT networks
• Integrates continual learning to adapt to environmental non-stationarity
• Optimizes energy use and link conditions for fault detection

Innovation

• Event-driven communication framework for continual learning
• Adaptive model updates based on link conditions and energy
• Collaborative lightweight ML updates between IoT device and edge