Continual Learning at the Edge: An Agnostic IIoT Architecture

TL;DR

This paper proposes a continual learning approach tailored for edge computing in industrial IoT, enabling real-time quality control by mitigating catastrophic forgetting and handling dynamic data streams effectively.

Contribution

It introduces a novel incremental learning method specifically designed for edge IoT systems in industrial environments, addressing latency, bandwidth, and data variability challenges.

Findings

Improved real-time quality control performance.

Reduced catastrophic forgetting in edge IoT scenarios.

Efficient handling of dynamic data streams.

Abstract

The exponential growth of Internet-connected devices has presented challenges to traditional centralized computing systems due to latency and bandwidth limitations. Edge computing has evolved to address these difficulties by bringing computations closer to the data source. Additionally, traditional machine learning algorithms are not suitable for edge-computing systems, where data usually arrives in a dynamic and continual way. However, incremental learning offers a good solution for these settings. We introduce a new approach that applies the incremental learning philosophy within an edge-computing scenario for the industrial sector with a specific purpose: real time quality control in a manufacturing system. Applying continual learning we reduce the impact of catastrophic forgetting and provide an efficient and effective solution.