$20M NSF AI-EDGE INSTITUTE AIMS TO TRANSFORM 5G AND BEYOND NETWORKS Next-generation networks, including 6G and beyond, are changing how and where AI applications are trained and implemented. By providing much higher bandwidth, faster speed, lower latency, and broader coverage, next-generation networks invite a new way of configuring networks. This is especially true at the edges of where most of the growth of network intelligence and distributed AI is expected. These edge networks will encompass mobile and stationary end devices, wireless and wired access points, and computing and sensing devices. Professors Mingyan Liu and Lei Ying are core members of a newly-established $20M NSF AI Institute for Future Edge Networks and Distributed Intelligence (AI-EDGE), led by The Ohio State University (OSU). AI-EDGE is expected to make AI more efficient, interactive, and secure for applications in sectors such as intelligent transportation, remote health care, distributed robotics and smart aerospace. “The Michigan team will be developing theories and algorithms for AI-aware networks that deliver the right information at the right time and place to support distributed AI in dynamic, heterogeneous, and non-stationary wireless edge networks,” said Liu. “We will also co-direct a substantial effort in education and workforce development aimed at the joint education in AI and networks.” “Our interest is to design new network architecture and algorithms that can support future AI applications and distributed intelligence,” 18
said Prof. Lei Ying, who will lead the research team focusing on Network Operation for Distributed AI-Applications. Next-generation networks, including 6G and beyond, are changing how and where AI applications are trained and implemented. By providing much higher bandwidth, faster speed, lower latency, and broader coverage, next-generation networks invite a new way of configuring networks. This is especially true at the edges of where most of the growth of network intelligence and distributed AI is expected. These edge networks will encompass mobile and stationary end devices, wireless and wired access points, and computing and sensing devices. “In the future, we envision we can move the intelligence from a centralized cloud center to distributed network edges,” said Ying. The primary benefits of doing this are to satisfy the requirements of real-time applications, such as autonomous driving, and to enhance data security and privacy. Another major challenge is to design a flexible network able to adapt to uncertain and changing conditions. While it may sound like this network would require massive resources, in fact they must also be able to service devices which are severely energy constrained, such as battery-powered IoT devices. Ying’s group has been doing related research looking at distributed computing in networks and how to do dynamic resource allocation in communication networks, but the scale has been smaller and the tasks relatively simple.