ABSTRACT
This chapter offers a thorough explanation of the privacy and security issues connected to federated learning and emphasizes possible uses for the technology across a range of industries. In horizontal federated learning, the updates from each device are combined to create a global model, and the data samples from each device are used to train the model in a decentralized way. By allowing model training on enormous amounts of decentralized data while protecting user privacy, horizontal federated learning has the potential to revolutionize machine learning. Each hospital can maintain its data locally and add to a shared model in a vertical federated learning setup. Verifying that the model can utilize vertically partitioned data from various organisations while adhering to their privacy requirements is the primary challenge in vertical federated learning. The process of combining these various features, calculating the training loss, and keeping everything private to create a centralized model is known as vertical federated learning.
