| Abstract |
Many enterprise solutions can be driven off Machine Learning models that are created from the data belonging to an enterprise. However, many enterprises can not share data freely across different locations due to regulatory restrictions, performance issues in moving large data volumes, or requirements to maintain autonomy. In these cases, the enterprise can benefit from the concept of federated learning, in which machine learning based models are created at multiple different sites, and then combined together at a federation server without the need to share data. In this paper, we describe a web-services based implementation of the federated learning system, focusing on the problems enterprises encounter in used of distributed data, and discussing how we solved those problems. |