Keynote Speakers: BigCom 2022
2022
The availability of massive amounts of data at mobile edge devices has led to a surge of interest in developing artificial intelligence (AI) services at the edge of next generation wireless systems. To facilitate distributed machine learning while keeping data local at the devices, federated learning has emerged as a new paradigm where distributed devices collaboratively train a shared AI model based on their local data. In practice, unreliable wireless propagation environment is the main bottleneck that limits the convergence and accuracy of wireless federated learning. In particular, the channel heterogeneity across mobile devices causes severe straggler issues that inevitably complicates the learning system design. In this talk, I am going to discuss the exploitation of advanced wireless technologies, such as reconfigurable intelligent surface, massive MIMO, and adaptive power control to enable high learning accuracy, fast convergence, and better data privacy in federated learning.