摘要 Abstract 

This talk analyzes the performance metrics of different Artificial Intelligent Internet of Things (AIoT) lightweight botnet attack detection models. After which, we deploy meta-learning ensemble botnet detection models and evaluate the capability of a single-board system in addressing cyber-attack threats. Then, we optimize the detection model to fit the constraints of low-end IoT devices and compare the detection and inference performance metrics. The Aposemat IoT-23 [1], UC Irvine KDD99 [2], and UNSW TON [3] datasets provide IoT and network traffic flow captures, which are used to evaluate the proposed meta-learning methodologies.