1056 papers with code • 12 benchmarks • 10 datasets
Federated Learning is a machine learning approach that allows multiple devices or entities to collaboratively train a shared model without exchanging their data with each other. Instead of sending data to a central server for training, the model is trained locally on each device, and only the model updates are sent to the central server, where they are aggregated to improve the shared model.
This approach allows for privacy-preserving machine learning, as each device keeps its data locally and only shares the information needed to improve the model.
Modern mobile devices have access to a wealth of data suitable for learning models, which in turn can greatly improve the user experience on the device.
Theoretically, we provide convergence guarantees for our framework when learning over data from non-identical distributions (statistical heterogeneity), and while adhering to device-level systems constraints by allowing each participating device to perform a variable amount of work (systems heterogeneity).
In this work, we look at the effect such non-identical data distributions has on visual classification via Federated Learning.
We obtain tight convergence rates for FedAvg and prove that it suffers from `client-drift' when the data is heterogeneous (non-iid), resulting in unstable and slow convergence.
In such an attack, a client's contribution during training and information about their data set is revealed through analyzing the distributed model.