Improving Federated Learning Personalization via Model Agnostic Meta Learning

TLDR

Federated Learning trains a global model from decentralized data, such as mobile phone activity, but data heterogeneity makes device‑specific personalization challenging. The authors argue that Model‑Agnostic Meta‑Learning, which optimizes for rapid few‑shot adaptation across tasks, naturally aligns with FL personalization goals and present FL as a practical setting for MAML. They show that Federated Averaging can be viewed as a meta‑learning algorithm, that fine‑tuning a Federated Averaging model improves accuracy and personalization, whereas a datacenter‑trained model is harder to personalize, highlighting trade‑offs that motivate further research.

Abstract

Federated Learning (FL) refers to learning a high quality global model based on decentralized data storage, without ever copying the raw data. A natural scenario arises with data created on mobile phones by the activity of their users. Given the typical data heterogeneity in such situations, it is natural to ask how can the global model be personalized for every such device, individually. In this work, we point out that the setting of Model Agnostic Meta Learning (MAML), where one optimizes for a fast, gradient-based, few-shot adaptation to a heterogeneous distribution of tasks, has a number of similarities with the objective of personalization for FL. We present FL as a natural source of practical applications for MAML algorithms, and make the following observations. 1) The popular FL algorithm, Federated Averaging, can be interpreted as a meta learning algorithm. 2) Careful fine-tuning can yield a global model with higher accuracy, which is at the same time easier to personalize. However, solely optimizing for the global model accuracy yields a weaker personalization result. 3) A model trained using a standard datacenter optimization method is much harder to personalize, compared to one trained using Federated Averaging, supporting the first claim. These results raise new questions for FL, MAML, and broader ML research.