Relay-Assisted Cooperative Federated Learning

Federated learning (FL) has recently emerged as a promising technology to enable artificial intelligence (AI) at the network edge, where distributed mobile devices collaboratively train shared AI model under coordination of an edge server. To significantly improve communication efficiency FL, over-the-air computation allows large number concurrently upload their local models by exploiting superposition property wireless multi-access channels. Due channel fading, aggregation error server is dominated weakest among all devices, causing severe straggler issues. In this paper, we propose relay-assisted cooperative FL scheme effectively address issue. particular, deploy multiple half-duplex relays cooperatively assist in uploading updates The nature poses system objectives and constraints that are distinct from those traditional relay systems. Moreover, strong coupling between design variables renders optimization such challenging. tackle issue, alternating-optimization-based algorithm optimize transceiver operation with low complexity. Then, analyze single-relay case show our achieves smaller than one without provided transmit power gains sufficiently large. analysis provides critical insights on deployment implementation FL. Extensive numerical results faster convergence compared state-of-the-art schemes.