PeaceMaker is an experimental Federated Learning setup to demonstrate a novel flow with incentive capabilities for cross-silo Federated Learning. To learn more about the project in depth, refer to the wiki page.
The user interface presents the data about previous training rounds, the model-data valuation, the payout value & current statistics of the global model. The Federated Learning process is started by initiating the training session via the UI by a system administrator.
The API gateway is as a service that serves information presented on the UI and acts as the middleman between the UI and the Federated Learning server.
An experimental and basic server setup to mimic Federated Learning between multiple organisations with the use of PySyft.