Vertical Federated Learning (VFL) is a federated machine learning approach that enables multiple clients (organizations, institutions, devices) that hold complementary, non-overlapping features about the same user or entity to jointly train machine learning models without sharing their raw data. In VFL, the data of each client cover the same user base but differ in the feature set.
VFL is extremely useful in industries like healthcare and finance, where organizations possess distinct yet very informative data subsets that, when combined can greatly improve the predictive performance of machine learning models. Overall, VFL is a great solution for organizations with a shared user base but varying types of information about each user.
Vertical Federated Learning in Flower
Flower, the friendly Federated AI framework, offers numerous out-of-the-box examples to test various federated settings. For a more in-depth understanding of the VFL settings, please refer to the following example.