Rural AI: Serverless-Powered Federated Learning for Remote Applications
Panos Patros, Melanie Po‐Leen Ooi, Victoria Huang, Michael Mayo, Chris Anderson, S. M. Burroughs, Matt Baughman, Osama Almurshed, Omer Rana, Ryan Chard, Kyle Chard, Ian Foster
Abstract
With increasing connectivity to support digital services in urban areas, there is a realization that demand for offering similar capability in rural communities is still limited. To unlock the potential of artificial intelligence (AI) within rural economies, we propose rural AI—the mobilization of serverless computing to enable AI in austere environments. Inspired by problems observed in New Zealand, we analyze major challenges in agrarian communities and define their requirements. We demonstrate a proof-of-concept rural AI system for cross-field pasture weed detection that illustrates the capabilities serverless computing offers to traditional federated learning.