Fedoag enables efficient federated learning over varied wireless channels
Over-the-Air Federated Learning in Heterogeneous Mobile Wireless Networks
Machine LearningDistributed, Parallel, and Cluster Computing
Summary
Wireless devices often learn together by sharing updates, but different devices' connections can cause problems in combining their data fairly and correctly. Existing methods either boost noise too much or accept bias in the learning process. The authors propose FedOAG, a new method that adjusts updates automatically and mixes them fairly without needing all devices to talk at once or knowing complex network details. This approach works well even when wireless conditions change and matches the best known speed for such learning.
What this means in practice
- •For mobile network engineers: Improve distributed model training efficiency in mobile networks with varying wireless conditions without needing detailed channel knowledge.
- •For edge computing developers: Build federated learning systems that automatically adjust device contributions to maintain training quality despite heterogeneous wireless links.
Authors
Ming Xiang, Nicolò Michelusi, Yonina C. Eldar, Lili Su
Abstract
Over-the-air computation has emerged as a scalable and efficient solution for deploying federated learning algorithms in wireless networks by exploiting waveform superposition for simultaneous model aggregation. Most existing work struggles with heterogeneous fading channels. These approaches either enforce unbiased updates from all devices or allow partial device contributions, requiring careful tuning of the convergence bound to mitigate bias under specific fading models. However, the former significantly amplifies receiver noise due to the weakest channel, whereas the latter is sensitive to fading model mismatch and converges only to a biased objective. To tackle these challenges, we propose FedOAG, which employs algorithmic components to automatically satisfy energy constraints via gradient normalization and evenly mix devices' updates through implicit gossiping. Importantly, FedOAG does not require transmission from all devices, nor does it rely on a specific fading model or knowledge of time-varying statistical channel distributions. We show that FedOAG converges to a stationary point of an unbiased non-convex objective at the best possible rate $O(1/\sqrt{T})$ for any stochastic first-order method. We corroborate our analysis with numerical experiments over dynamic wireless conditions on real-world datasets.