Role of Explainable Edge AI to Resolve Real Time Problem

Abstract

AbstractThe growth of information technology (IT) has resulted in physical de-vices being connected to the internet and having the ability to recognize other devices. Artificial Intelligence algorithms are processed on edge or the devices of users. Edge Computing based on the same premise, stores, processes, and manages data directly at Internet of Things (IoT) endpoints. Edge artificial intelligence uses the device's hardware to process data and performs machine learning and deep learning procedures. In the model, you can troubleshoot and improve model performance while also assisting others in understanding the behavior of your models. To make the area more real, explainable edge devices come in a wide range of costs and capabilities. A decade ago, we couldn’t imagine that explainable edge artificial intelligence would be at today’s level. Now it is a part of industries and even devices for customer service. The best example of explainable edge AI is virtual assistants such as Alexa, google assistant. They learn from the user’s world and phrases and can store them directly on the device. These are just a few examples later, and we have possible applications in future works on the explainable edge artificial intelligence. Edge Computing Platform facilitates the development and elastic operation of apps and services. Its benefits the AI assisting in overcoming the technical obstacles that AI-enabled apps experience. Combination of edge and AI is buzzwords within the industry to deliver the performance and reduce the cost compared to state of arts XAI applications. Moreover edge artificial intelligence are reducing the latency, improving user experience, and reducing the necessary bandwidth, consequently reducing the costs of internet services. It surfs this movement since the need for data processing on the device themselves also represents the increasing use of artificial intelligence. Artificial intelligence edge processing focused on model which trained them in central data center using historical datasets. Compression techniques of data that enables squeezing large artificial intelligence models into small hardware form factors could push some training to the edge over time.KeywordsExplainable artificial intelligence (XAI)Edge computingDeep learningMachine learningInternet of things (IoT)Neural networkCognitive science