Offloading Evaluation of machine learning serverless functions over 5G

There are several challenges to using Artificial Intelligence (AI) at the edge network, whether in distributing AI models or constructing specific models for IoT devices. Conscious of that challenge, this work seeks to advance the knowledge of Edge Intelligence (AI on Edge) by evaluating the offloading of machine learning models through serverless functions executed between Fog and Edge. Thus, this work aims to provide an overview of how the 5G network can improve the application performance of serverless functions when deployed to Edge Intelligence. Applying with an image face recognition AI application, we evaluate parameters like battery consumption, and latency over 4G and 5G networks. The results indicate that the use of 5G can improve solutions that uses serverless functions on Edge.