From Information Networking to Intelligence Networking: Motivations, Scenarios, and Challenges

Abstract

By enabling information networking among people and machines, the Internet has become one of the major foundations for our socio-economic systems. After several decades of research and development of the Internet, it is relatively easy for humans/machines to obtain information. However, there are new challenges in the post-Internet era, including information overload, fake information and the design of trustworthy, cost-effective autonomous systems. In order to address these challenges, we need to think about networking in a larger timescale. Actually, in order to facilitate humans' cooperation, we have invented technologies enabling networking for matter (grid of transportation), for energy (grid of energy), and for information (the Internet). In this article, we argue that the next networking paradigm could be intelligence networking, where intelligence can be easily obtained, like matter, energy, and information. Specifically, we present the motivations, scenarios and challenges of intelligence networking. In addition, we present a novel collective reinforcement learning scheme enabled by intelligence networking. Some simulation results are presented to show the effectiveness of the proposed intelligence networking paradigm.