查看文章

Mobile edge intelligence has recently emerged as a promising paradigm to support mobile artificial intelligence (AI) services at the edge of wireless networks. With the aid of mobile edge computing (MEC), mobile edge intelligence migrates the training and inference processes of AI models to the network edge, i.e., edge servers and mobile devices, which enables fast edge learning and edge inference. This thesis aims to design efficient resource management schemes and optimization algorithms to unleash the full potential of mobile edge intelligence in wireless cellular systems.First, we study the AI service placement for edge inference in a multi-user MEC system, where the edge server selectively places the most up-to-date AI service program at a subset of devices to enable local AI inference. The service placement decision is highly correlated with the system resource allocation. To minimize the computation …