ByzFL: Research Framework for Robust Federated Learning
This work provides a tool for researchers and practitioners to facilitate reproducible research and rapid prototyping in robust federated learning, though it is incremental as it builds on existing methods without introducing new algorithmic breakthroughs.
The authors tackled the problem of developing and benchmarking robust federated learning algorithms by introducing ByzFL, an open-source Python library that provides a unified framework with implementations of state-of-the-art robust aggregators, configurable attacks, and simulation tools for various FL scenarios.
We present ByzFL, an open-source Python library for developing and benchmarking robust federated learning (FL) algorithms. ByzFL provides a unified and extensible framework that includes implementations of state-of-the-art robust aggregators, a suite of configurable attacks, and tools for simulating a variety of FL scenarios, including heterogeneous data distributions, multiple training algorithms, and adversarial threat models. The library enables systematic experimentation via a single JSON-based configuration file and includes built-in utilities for result visualization. Compatible with PyTorch tensors and NumPy arrays, ByzFL is designed to facilitate reproducible research and rapid prototyping of robust FL solutions. ByzFL is available at https://byzfl.epfl.ch/, with source code hosted on GitHub: https://github.com/LPD-EPFL/byzfl.