In the past decade we have witnessed the development of machine learning and artificial intelligence algorithms with cloud computing in providing intelligence for smart applications. By collecting raw data from a multitude of mobile and wireless devices, the cloud server executes machine learning algorithms to train the data and perform data prediction. However, the growing number of devices has generated a tremendous amount of data. Transmitting massive data to centralized machine learning in the cloud has significant communication overhead and computational complexity. With the support of edge gateways, the significant burdens of massive data flow and exchanged information from large groups of devices can be efficiently alleviated in terms of both communication and computational overhead. Meanwhile, the exchanged information aims to reach the remote online system, in which the continuous and robust online services should be always guaranteed and ensured. As a promising solution, cyberinfrastructure is designed and developed to boost network intelligence by offering powerful computing systems, flexible data storage repositories, and advanced virtualization environments at both the edge and the cloud. Edge intelligence has emerged as a new trend for next generation wireless networks.

中文翻译：

---

下一代无线网络的边缘智能

在过去的十年中，我们见证了机器学习和人工智能算法以及云计算在为智能应用程序提供智能方面的发展。通过从众多移动和无线设备收集原始数据，云服务器执行机器学习算法来训练数据并执行数据预测。但是，越来越多的设备产生了大量的数据。将大量数据传输到云中的集中式机器学习具有巨大的通信开销和计算复杂性。在边缘网关的支持下，就通信和计算开销而言，可以有效地减轻海量数据流和来自大型设备组的交换信息的沉重负担。同时，交换的信息旨在到达远程在线系统，在该系统中，应始终保证并确保连续和强大的在线服务。作为一种有前途的解决方案，网络基础设施的设计和开发旨在通过在边缘和云端提供强大的计算系统，灵活的数据存储库和高级虚拟化环境来增强网络智能。边缘智能已成为下一代无线网络的新趋势。