We love Machine Learning. The way the computer extracts rules and patterns from big amounts of data is fascinating. But centralizing data on a central cloud comes with many problems, that involve large transfers and data privacy that can be compromised. Federated Learning comes with a set of tools that solve both of these problems at once: Data does not need to be transferred to a central cloud, so no privacy to be compromised and no large transfers.
Researching architecture approaches, we managed to find some gaps in the literature, so we designed our very own flexible and powerful architecture that will be able to do both Federated Learning and classical Machine Learning, with a click of a button. The resulting platform will answer to needs of both the industry, preserving data in the company (it will never leave the company premises) and also enable companies to share knowledge with scientists for research purposes.
If this topic is your cup of tea, read it and see what new insights you will gain from it.
Link for the publication: https://link.springer.com/chapter/10.1007/978-3-031-18050-7_64