Evaluating Distributed-learning On Real-world Obstetrics Data: Comparing Distributed, Centralized And Local Models

Distributed learning is a method of training models in different settings and aggregating them together, with two main branches: federated learning and peer-to-peer learning. This approach can be implemented in various ways, such as averaging weights or aggregating individual models into an ensemble. While distributed deep learning has received attention, there is a lack of knowledge in traditional machine learning techniques for distributed learning, particularly for tabular data like electronic health records.