给定一个训练好的全局模型,如何把该全局模型进行个性化?其中一种方法便是借助元学习的概念。元学习是模型调整的一种流行范式。在标准的Meta-learning 中,假设多个任务分布背后有一个潜在的元分布。通过挖掘这个元分布的信息来帮助在新任务上进行学习。对应到联邦学习中,各个客户端的数据对应不同的任务分布(由数据收集/采样引起的),期望全局模型拟合这些任务分布背后潜在的元分布。标准的联邦学习事实上就等于是在学习这个背后的潜在分布。Model-agnostic meta-learning (MAML)是一种经典的meta-learning方法,它期望得到一个能快速适应到新任务的全局模型。Chen等人在文章Federated meta-learning with fast convergence and efficient communication中将MAML应用到FL场景中,利用MAML得到一个全局模型,之后每个客户端把该全局模型作为本地训练的初始化,经过一步或几步梯度更新后得到表现良好的本地模型。Fallah等人在Chen等人工作的基础上进一步为此种方法提供了收敛性等方面的理论保证。结合MAML的FL目标函数如下所示:
[1] Kairouz P, McMahan H B, Avent B, et al. Advances and open problems in federated learning[J]. Foundations and Trends® in Machine Learning, 2021, 14(1–2): 1-210.
[2] Li X, Huang K, Yang W, et al. On the convergence of fedavg on non-iid data[J]. arXiv preprint arXiv:1907.02189, 2019.
[3] Tongzhou Wang, Jun-Yan Zhu, Antonio Torralba, and Alexei A Efros. Dataset distillation. arXiv preprint arXiv:1811.10959, 2018.
[4] Hao Yu, Rong Jin, and Sen Yang. On the linear speedup analysis of communication efficient momentum SGD for distributed non-convex optimization. arXiv preprint arXiv:1905.03817, 2019.
[5] Li T, Sahu A K, Zaheer M, et al. Federated optimization in heterogeneous networks[J]. Proceedings of Machine Learning and Systems, 2020, 2: 429-450.
[6] Karimireddy S P, Kale S, Mohri M, et al. Scaffold: Stochastic controlled averaging for federated learning[C]//International Conference on Machine Learning. PMLR, 2020: 5132-5143.
[7] Jianyu Wang, Vinayak Tantia, Nicolas Ballas, and Michael Rabbat. SlowMo: Improving communicationefficient distributed SGD with slow momentum. arXiv preprint arXiv:1910.00643, 2019.
[8] Tzu-Ming Harry Hsu, Hang Qi, and Matthew Brown. Measuring the effects of non-identical data distribution for federated visual classification. arXiv preprint arXiv:1909.06335, 2019.
[9] Mohri M, Sivek G, Suresh A T. Agnostic federated learning[C]//International Conference on Machine Learning. PMLR, 2019: 4615-4625.
[10] Li T, Sanjabi M, Beirami A, et al. Fair resource allocation in federated learning[J]. arXiv preprint arXiv:1905.10497, 2019.
[11] Smith V, Chiang C K, Sanjabi M, et al. Federated multi-task learning[J]. Advances in neural information processing systems, 2017, 30.
[12] Chen F, Luo M, Dong Z, et al. Federated meta-learning with fast convergence and efficient communication[J]. arXiv preprint arXiv:1802.07876, 2018.
[13] Jiang Y, Konečný J, Rush K, et al. Improving federated learning personalization via model agnostic meta learning[J]. arXiv preprint arXiv:1909.12488, 2019.
[14] Fallah A, Mokhtari A, Ozdaglar A. Personalized federated learning: A meta-learning approach[J]. arXiv preprint arXiv:2002.07948, 2020.
[15] Mansour Y, Mohri M, Ro J, et al. Three approaches for personalization with applications to federated learning[J]. arXiv preprint arXiv:2002.10619, 2020.
[16] Deng Y, Kamani M M, Mahdavi M. Adaptive personalized federated learning[J]. arXiv preprint arXiv:2003.13461, 2020.
[17] Kangkang Wang, Rajiv Mathews, Chloe Kiddon, Hubert Eichner, Franc¸oise Beaufays, and Daniel Ramage. ´Federated evaluation of on-device personalization. arXiv preprint arXiv:1910.10252, 2019.
[18] Divi S, Lin Y S, Farrukh H, et al. New metrics to evaluate the performance and fairness of personalized federated learning[J]. arXiv preprint arXiv:2107.13173, 2021.