About APPFL – Advanced Privacy Preserving Federated Learning

Advanced Privacy-Preserving Federated Learning (APPFL) is an open-source software framework that allows research communities to implement, test, and validate various ideas for privacy-preserving federated learning (PPFL). With this framework, developers and/or users can:

train a user-defined neural network model on decentralized data with differential privacy,
simulate various PPFL algorithms on high-performance computing (HPC) architecture with MPI,
implement user-defined PPFL algorithms in a plug-and-play manner.

Such algorithmic components include federated learning (FL) algorithm, privacy technique, communication protocol, FL model to train, and data.

Documentation: please check out the documentation for tutorials, users guide, and developers guide.