Prototype Guided Federated Learning of Visual Feature Representations
This work addresses challenges in federated learning for vision tasks, such as data heterogeneity, by proposing a novel method that enhances model training and evaluation, including first-time application to dense prediction tasks like semantic segmentation.
The paper tackles the problem of training visual models in federated learning (FL) with non-i.i.d. data by introducing FedProto, which uses prototypical representations and an attention mechanism to improve optimization, achieving state-of-the-art accuracy and convergence rates in image classification and semantic segmentation benchmarks.
Federated Learning (FL) is a framework which enables distributed model training using a large corpus of decentralized training data. Existing methods aggregate models disregarding their internal representations, which are crucial for training models in vision tasks. System and statistical heterogeneity (e.g., highly imbalanced and non-i.i.d. data) further harm model training. To this end, we introduce a method, called FedProto, which computes client deviations using margins of prototypical representations learned on distributed data, and applies them to drive federated optimization via an attention mechanism. In addition, we propose three methods to analyse statistical properties of feature representations learned in FL, in order to elucidate the relationship between accuracy, margins and feature discrepancy of FL models. In experimental analyses, FedProto demonstrates state-of-the-art accuracy and convergence rate across image classification and semantic segmentation benchmarks by enabling maximum margin training of FL models. Moreover, FedProto reduces uncertainty of predictions of FL models compared to the baseline. To our knowledge, this is the first work evaluating FL models in dense prediction tasks, such as semantic segmentation.