17 November 2020 in IEEE/ACM Transactions on Networking
There is an increasing interest in a fast-growing machine learning technique called Federated Learning (FL), in which the model training is distributed over mobile user equipment (UEs), exploiting UEs' local computation and training data. Despite its advantages such as preserving data privacy, FL still has challenges of heterogeneity across UEs' data and physical resources. To address these challenges, we first propose FEDL, a FL algorithm which can handle heterogeneous UE data without further assumptions except strongly convex and smooth loss functions. We provide a convergence rate characterizing the trade-off between local computation rounds of each UE to update its local model and global communication rounds to update the FL global model. We then employ FEDL in wireless networks as a resource allocation optimization problem that captures the trade-off between FEDL convergence wall clock time and energy consumption of UEs with heterogeneous computing and power resources. Even though the wireless resource allocation problem of FEDL is non-convex, we exploit this problem's structure to decompose it into three sub-problems and analyze their closed-form solutions as well as insights into problem design. Finally, we empirically evaluate the convergence of FEDL with PyTorch experiments, and provide extensive numerical results for the wireless resource allocation sub-problems. Experimental results show that FEDL outperforms the vanilla FedAvg algorithm in terms of convergence rate and test accuracy in various settings.
7 December 2022 in IEEE Transactions on Neural Networks and Learning Systems
Non-independent and identically distributed (non-IID) data distribution among clients is considered as the key factor that degrades the performance of federated learning (FL). Several approaches to handle non-IID data, such as personalized FL and federated multitask learning (FMTL), are of great interest to research communities. In this work, first, we formulate the FMTL problem using Laplacian regularization to explicitly leverage the relationships among the models of clients for multitask learning. Then, we introduce a new view of the FMTL problem, which, for the first time, shows that the formulated FMTL problem can be used for conventional FL and personalized FL. We also propose two algorithms and decentrali-zed () to solve the formulated FMTL problem in communication-centralized and decentralized schemes, respectively. Theoretically, we prove that the convergence rates of both algorithms achieve linear speedup for strongly convex and sublinear speedup of order $1/2$ for nonconvex objectives. Experimentally, we show that our algorithms outperform the conventional algorithm FedAvg, FedProx, SCAFFOLD, and AFL in FL settings, MOCHA in FMTL settings, as well as pFedMe and Per-FedAvg in personalized FL settings.
27 January 2022 in IEEE Transactions on Parallel and Distributed Systems
There is growing interest in applying distributed machine learning to edge computing, forming federated edge learning. Federated edge learning faces non-i.i.d. and heterogeneous data, and the communication between edge workers, possibly through distant locations and with unstable wireless networks, is more costly than their local computational overhead. In this work, we propose DONE, a distributed approximate Newtontype algorithm with a fast convergence rate for communication efficient federated edge learning. First, with strongly convex and smooth loss functions, DONE approximates the Newton direction in a distributed manner using the classical Richardson iteration on each edge worker. Second, we prove that DONE has linear-quadratic convergence and analyze its communication complexities. Finally, the experimental results with non-i.i.d. and heterogeneous data show that DONE attains a comparable performance to Newtons method. Notably, DONE requires fewer communication iterations compared to distributed gradient descent and outperforms DANE, FEDL, and GIANT, state-of-the-art approaches, in the case of non-quadratic loss functions.
9 August 2021 in IEEE Communications Letters
Transferring large models in federated learning (FL) networks is often hindered by clients’ limited bandwidth. We propose FedAA, an FL algorithm which achieves fast convergence by exploiting the regularized Anderson acceleration (AA) on the global level. First, we demonstrate that FL can benefit from acceleration methods in numerical analysis. Second, FedAA improves the convergence rate for quadratic losses and improves the empirical performance for smooth and strongly convex objectives, compared to FedAvg, an FL algorithm using gradient descent (GD) local updates. Experimental results demonstrate that employing AA can significantly improve the performance of FedAvg, even when the objective is non-convex.