Brain Storm Optimization Based Swarm Learning for Diabetic Retinopathy Image Classification

TL;DR

This work tackles privacy-aware diabetic retinopathy image classification in distributed settings by introducing BSO-SL, a decentralized swarm learning framework that removes the need for a central server and mitigates blockchain mining costs. By modeling each client's parameter distribution as Gaussian and clustering clients with $k$-means, the method enables intra-cluster collaboration, while a Brain Storm aggregation mechanism promotes cross-cluster exploration to avoid local optima. Empirical results on a real DR dataset show that BSO-SL achieves competitive accuracy compared to FedAvg, with better scalability and privacy preservation, and it demonstrates model-agnostic compatibility across several CNN architectures. The approach offers a practical path toward privacy-preserving, scalable medical image classification across multiple clinics, with potential applicability to other non-IID, distributed learning scenarios.

Abstract

The application of deep learning techniques to medical problems has garnered widespread research interest in recent years, such as applying convolutional neural networks to medical image classification tasks. However, data in the medical field is often highly private, preventing different hospitals from sharing data to train an accurate model. Federated learning, as a privacy-preserving machine learning architecture, has shown promising performance in balancing data privacy and model utility by keeping private data on the client's side and using a central server to coordinate a set of clients for model training through aggregating their uploaded model parameters. Yet, this architecture heavily relies on a trusted third-party server, which is challenging to achieve in real life. Swarm learning, as a specialized decentralized federated learning architecture that does not require a central server, utilizes blockchain technology to enable direct parameter exchanges between clients. However, the mining of blocks requires significant computational resources, limiting its scalability. To address this issue, this paper integrates the brain storm optimization algorithm into the swarm learning framework, named BSO-SL. This approach clusters similar clients into different groups based on their model distributions. Additionally, leveraging the architecture of BSO, clients are given the probability to engage in collaborative learning both within their cluster and with clients outside their cluster, preventing the model from converging to local optima. The proposed method has been validated on a real-world diabetic retinopathy image classification dataset, and the experimental results demonstrate the effectiveness of the proposed approach.

Figures (3)

• Figure 1: (a)-(e): different severities of diabetic retinopathy, ranging from No DR to Proliferative DR.
• Figure 2: (a) the architecture of federated learning; (b) the architecture of swarm learning.
• Figure 3: The architecture of the proposed BSO-SL.