Federated learning (FL) is an emerging machine learning (ML) training paradigm where clients own their data and collaborate to train a global model, without revealing any data to the server and other participants. Researchers commonly perform experiments in a simulation environment to quickly iterate on ideas. However, existing open-source tools do not offer the efficiency required to simulate FL on larger and more realistic FL datasets. We introduce pfl-research, a fast, modular, and easy-to-use Python framework for simulating FL. It supports TensorFlow, PyTorch, and non-neural network models, and is tightly integrated with state-of-the-art privacy algorithms. We study the speed of open-source FL frameworks and show that pfl-research is 7-72$\times$ faster than alternative open-source frameworks on common cross-device setups. Such speedup will significantly boost the productivity of the FL research community and enable testing hypotheses on realistic FL datasets that were previously too resource intensive. We release a suite of benchmarks that evaluates an algorithm's overall performance on a diverse set of realistic scenarios. The code is available on GitHub at https://github.com/apple/pfl-research.

Learning aggregate population trends can allow for better data-driven decisions, and application of machine learning can improve user experience. Compared to learning from public curated datasets, learning from a larger population offers several benefits. As an example, a next-word prediction model trained on words typed by users (a) can better fit the actual distribution of language used on devices, (b) can adapt faster to shifts in distribution, and (c) can more faithfully represent smaller sub-populations that may not be well-represented in curated datasets. At the same time, training such models may involve sensitive user data.