FedBiCross: A Bi-Level Optimization Framework to Tackle Non-IID Challenges in Data-Free One-Shot Federated Learning on Medical Data
This addresses privacy-sensitive medical applications by enhancing model performance in federated learning with non-IID data, though it is incremental as it builds on existing one-shot federated learning methods.
The paper tackled the problem of conflicting predictions in data-free one-shot federated learning under non-IID medical data by proposing FedBiCross, a personalized framework that clusters clients and uses bi-level optimization to improve knowledge transfer, resulting in consistent outperformance over state-of-the-art baselines across four medical image datasets.
Data-free knowledge distillation-based one-shot federated learning (OSFL) trains a model in a single communication round without sharing raw data, making OSFL attractive for privacy-sensitive medical applications. However, existing methods aggregate predictions from all clients to form a global teacher. Under non-IID data, conflicting predictions cancel out during averaging, yielding near-uniform soft labels that provide weak supervision for distillation. We propose FedBiCross, a personalized OSFL framework with three stages: (1) clustering clients by model output similarity to form coherent sub-ensembles, (2) bi-level cross-cluster optimization that learns adaptive weights to selectively leverage beneficial cross-cluster knowledge while suppressing negative transfer, and (3) personalized distillation for client-specific adaptation. Experiments on four medical image datasets demonstrate that FedBiCross consistently outperforms state-of-the-art baselines across different non-IID degrees.