Edge Intelligence: A Robust Reinforcement of Edge Computing and Artificial Intelligence

Due to the development of faster and improved modes of communication, technologies, as well as customers, get benefited. The world is moving rapidly toward digitization, and connectivity has been placed under tremendous pressure. The development and implementation of IoT devices benefit all industry sectors, stimulating more areas in terms of convenience, productivity, and communication. But such a huge amount of data generated by IoT devices could result in a breakdown of IT infrastructure. To reach to the desired destination, this massive data travels via some intermediator. When the so-called intermediator that is cloud database is based in a remote location, the data can experience some kind of delay before it reaches the cloud for processing. So in recent years, the IT industry is attracted tremendous attention to improving communication between these technologies. And this is what the aim of edge computing (EC) is. Meanwhile also artificial intelligence (AI) algorithms and models have made breakthrough progress to accelerate the successful deployment of intelligence in the cloud services. By and large, AI services are executed in cloud for dealing with demands, because of the way that most AI models are intricate and difficult to process their induction results in favor of resource-limited devices. Nonetheless, such sort of ‘end–cloud’ architecture cannot address the issues of real-time AI services such as real-time analytics and smart manufacturing. Accordingly, deploying AI applications on the edge can widen the application situations of AI especially as for the low-latency characteristic. Combining the above two-mentioned paradigms, i.e., EC and AI can give rise to a new outlook: Edge Intelligence (EI). This paper provides insights into this new outlook by discussing core definitions, concepts, components, and frameworks. It also describes some necessary background in future research areas and challenges.