1: \begin{abstract}

2: In the paper, we propose an effective and efficient Compositional Federated Learning (ComFedL) algorithm

3: for solving a new compositional Federated Learning (FL) framework,

4: which frequently appears in many data mining and machine learning problems with a hierarchical structure such as

5: distributionally robust FL

6: and model-agnostic meta learning (MAML).

7: Moreover, we study the convergence analysis of our ComFedL algorithm under some mild conditions,

8: and prove that it achieves a convergence rate of $O(\frac{1}{\sqrt{T}})$,

9: where $T$ denotes the number of iteration.

10: To the best of our knowledge, our new Compositional FL framework is the

11: first work to bridge federated learning with composition stochastic optimization.

12: In particular, we first transform the distributionally robust FL (i.e., a minimax optimization problem) into

13: a simple composition optimization problem by using KL divergence regularization.

14: At the same time, we also first transform the distribution-agnostic MAML problem (i.e., a minimax optimization problem)

15: into a simple yet effective composition optimization problem.

16: Finally, we apply two popular machine learning tasks,

17: i.e., distributionally robust FL and MAML to demonstrate

18: the effectiveness of our algorithm.

19: \end{abstract}

20: