Federated Learning in Practice: Reflections and Projections

Federated Learning (FL) is a machine learning technique that enables multiple entities to collaboratively learn a shared model without exchanging their local data. Over the past decade, FL systems have achieved substantial progress, scaling to millions of devices across various learning domains while offering meaningful differential privacy (DP) guarantees. Production systems from organizations like Google, Apple, and Meta demonstrate the real-world applicability of FL. However, key challenges remain, including verifying server-side DP guarantees and coordinating training across heterogeneous devices, limiting broader adoption. Additionally, emerging trends such as large (multi-modal) models and blurred lines between training, inference, and personalization challenge traditional FL frameworks. In response, we propose a redefined FL framework that prioritizes privacy principles rather than rigid definitions. We also chart a path forward by leveraging trusted execution environments and open-source ecosystems to address these challenges and facilitate future advancements in FL.

I. Evolution of Federated Learning

Federated Learning (FL) was introduced around 2016 as a privacy enhancing technique that directly applies the principle of data minimization by focused collection and immediate aggregation [36], which "enables mobile phones to collaboratively learn a shared prediction model while keeping all the training data on device, decoupling the ability to do machine learning from the need to store the data in the cloud." [27], [48] FL quickly became a widely-acknowledged paradigm of distributed learning from decentralized data, and has been adopted in various applications beyond the original on-device training scenarios: for example, the FL paradigm has been applied to collaborative learning across multiple institutions (silos) with richer computation resources than mobile devices, or to learning over Internet-of-Things devices with more limited resources. In 2019, out of the discussion in the Workshop on Federated Learning and Analytics at Google, Kairouz et al. [42] proposed a broader definition of FL:

Federated learning is a machine learning setting where multiple entities (clients) collaborate in solving a machine learning problem, under the coordination of a central server or service provider. Each client’s raw data is stored locally and not exchanged or transferred; instead, focused updates intended for immediate aggregation are used to achieve the learning objective.