Tackling Integration Challenges of Machine Learning in Diverse Internet of Things
Adarsha Bhattarai, Chathumi Samaraweera, Dongming Peng, Yutong Liu, Hamid Sharif · 2025
In the modern-day interconnected world, the union of artificial intelligence (AI) and the Internet of Things (IoT) has unlocked exceptional opportunities for innovation and efficiency. AI empowers machines to learn, reason, and make decisions, while IoT enables the seamless integration of sensors, devices, and systems to collect and exchange vast amounts of data. In this chapter, we present a novel AI-IoT model that considers the hardware specification of every processing node participating in the network and partitions the AI inference models accordingly. AI inference models are trained models that get delivered to the respective processing nodes securely with the help of the AI spread-out chain (ASC). Due to the challenges posed by the higher computational complexity of state-of-the-art AI algorithms, the proposed mechanism for AI distribution within IoT will not only facilitate learning but also enable the execution of AI processes based on that learning. The tested AI spread-out model demonstrated its effectiveness through K-nearest neighbor (KNN) machine learning and ASC simulation results. The performance metrics showed promising results for the KNN spread-out model, achieving a maximum accuracy rate of 94.26% for KNN model 1. Overall, ASC proves to be very beneficial in enhancing AI-based edge computing techniques in cyber-physical systems.