Federated Compositional Deep AUC Maximization
This addresses the challenge of poor prediction performance in real-world applications with class imbalance, offering a solution for privacy-preserving large-scale learning.
The paper tackles the problem of federated learning with imbalanced data by developing a novel method that directly optimizes the AUC score, achieving favorable theoretical bounds on computational and communication complexities.
Federated learning has attracted increasing attention due to the promise of balancing privacy and large-scale learning; numerous approaches have been proposed. However, most existing approaches focus on problems with balanced data, and prediction performance is far from satisfactory for many real-world applications where the number of samples in different classes is highly imbalanced. To address this challenging problem, we developed a novel federated learning method for imbalanced data by directly optimizing the area under curve (AUC) score. In particular, we formulate the AUC maximization problem as a federated compositional minimax optimization problem, develop a local stochastic compositional gradient descent ascent with momentum algorithm, and provide bounds on the computational and communication complexities of our algorithm. To the best of our knowledge, this is the first work to achieve such favorable theoretical results. Finally, extensive experimental results confirm the efficacy of our method.