Federated Survival

Federated survival analysis focuses on developing methods for training survival models on distributed, sensitive datasets—like those in healthcare—without compromising patient privacy. Current research emphasizes adapting ensemble methods, particularly random survival forests, and deep learning models to federated learning frameworks, addressing challenges posed by non-identically distributed data and partially overlapping features across contributing datasets. This field is crucial for advancing medical research and improving healthcare outcomes by enabling collaborative analysis of large, privacy-sensitive datasets while maintaining data security and regulatory compliance.