Federated Learning with Non-IID Data
This addresses a key statistical challenge in federated learning for edge devices, offering a practical improvement for privacy-preserving decentralized training.
The paper tackles the problem of reduced accuracy in federated learning due to non-IID data, showing a 55% drop in accuracy for neural networks with highly skewed data, and proposes a solution using a small globally shared data subset that increases accuracy by 30% on CIFAR-10 with only 5% shared data.
Federated learning enables resource-constrained edge compute devices, such as mobile phones and IoT devices, to learn a shared model for prediction, while keeping the training data local. This decentralized approach to train models provides privacy, security, regulatory and economic benefits. In this work, we focus on the statistical challenge of federated learning when local data is non-IID. We first show that the accuracy of federated learning reduces significantly, by up to 55% for neural networks trained for highly skewed non-IID data, where each client device trains only on a single class of data. We further show that this accuracy reduction can be explained by the weight divergence, which can be quantified by the earth mover's distance (EMD) between the distribution over classes on each device and the population distribution. As a solution, we propose a strategy to improve training on non-IID data by creating a small subset of data which is globally shared between all the edge devices. Experiments show that accuracy can be increased by 30% for the CIFAR-10 dataset with only 5% globally shared data.