More Like this

1. Communication Efficient Asynchronous Stochastic Gradient Descent

   Ahmed, Y; Ghosh, A; Wang, C-C; Shroff, NB (May 2025, IEEE)

   In this paper, we address the challenges of asynchronous gradient descent in distributed learning environments, particularly focusing on addressing the challenges of stale gradients and the need for extensive communication resources. We develop a novel communication efficient framework that incorporates a gradient evaluation algorithm to assess and utilize delayed gradients based on their quality, ensuring efficient and effective model updates while significantly reducing communication overhead. Our proposed algorithm requires agents to only send the norm of the gradients rather than the computed gradient. The server then decides whether to accept the gradient if the ratio between the norm of the gradient and the distance between the global model parameter and the local model parameter exceeds a certain threshold. With the proper choice of the threshold, we show that the convergence rate achieves the same order as the synchronous stochastic gradient without depending on the staleness value unlike most of the existing works. Given the computational complexity of the initial algorithm, we introduce a simplified variant that prioritizes the practical applicability without compromising on the convergence rates. Our simulations demonstrate that our proposed algorithms outperform existing state-of-the-art methods, offering improved convergence rates, stability, accuracy, and resource consumption. 

   more » « less
2. Analog Compression and Communication for Federated Learning over Wireless MAC

   https://doi.org/10.1109/SPAWC48557.2020.9154309  

   Abdi, Afshin; Saidutta, Yashas Malur; Fekri, Faramarz (August 2020, IEEE 21st International Workshop on Signal Processing Advances in Wireless Communications (SPAWC))

   null (Ed.)

   In this paper, we consider federated learning in wireless edge networks. Transmitting stochastic gradients (SG) or deep model's parameters over a limited-bandwidth wireless channel can incur large training latency and excessive power consumption. Hence, data compressing is often used to reduce the communication overhead. However, efficient communication requires the compression algorithm to satisfy the constraints imposed by the communication medium and take advantage of its characteristics, such as over-the-air computations inherent in wireless multiple-access channels (MAC), unreliable transmission and idle nodes in the edge network, limited transmission power, and preserving the privacy of data. To achieve these goals, we propose a novel framework based on Random Linear Coding (RLC) and develop efficient power management and channel usage techniques to manage the trade-offs between power consumption, communication bit-rate and convergence rate of federated learning over wireless MAC. We show that the proposed encoding/decoding results in an unbiased compression of SG, hence guaranteeing the convergence of the training algorithm without requiring error-feedback. Finally, through simulations, we show the superior performance of the proposed method over other existing techniques. 

   more » « less
3. Training Recommender Systems at Scale: Communication-Efficient Model and Data Parallelism

   https://doi.org/10.1145/3447548.3467080  

   Gupta, Vipul; Choudhary, Dhruv; Tang, Peter; Wei, Xiaohan; Wang, Xing; Huang, Yuzhen; Kejariwal, Arun; Ramchandran, Kannan; Mahoney, Michael W. (August 2021, KDD '21: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining)

   In this paper, we consider hybrid parallelism—a paradigm that em- ploys both Data Parallelism (DP) and Model Parallelism (MP)—to scale distributed training of large recommendation models. We propose a compression framework called Dynamic Communication Thresholding (DCT) for communication-efficient hybrid training. DCT filters the entities to be communicated across the network through a simple hard-thresholding function, allowing only the most relevant information to pass through. For communication efficient DP, DCT compresses the parameter gradients sent to the parameter server during model synchronization. The threshold is updated only once every few thousand iterations to reduce the computational overhead of compression. For communication efficient MP, DCT incorporates a novel technique to compress the activations and gradients sent across the network during the forward and backward propagation, respectively. This is done by identifying and updating only the most relevant neurons of the neural network for each training sample in the data. We evaluate DCT on publicly available natural language processing and recommender models and datasets, as well as recommendation systems used in production at Facebook. DCT reduces communication by at least 100× and 20× during DP and MP, respectively. The algorithm has been deployed in production, and it improves end-to-end training time for a state-of-the-art industrial recommender model by 37%, without any loss in performance. 

   more » « less
4. Shuffled model of differential privacy in federated learnin

   Girgis, Antonious M; Data, Deepesh; Diggavi, Suhas; Kairouz, Peter; and Suresh, Theerta. (January 2021, International Conference on Artificial Intelligence and Statistics, AISTATS)

   We consider a distributed empirical risk minimization (ERM) optimization problem with communication efficiency and privacy requirements, motivated by the federated learn- ing (FL) framework. We propose a distributed communication-efficient and local differentially private stochastic gradient descent (CLDP-SGD) algorithm and analyze its communication, privacy, and convergence trade-offs. Since each iteration of the CLDP- SGD aggregates the client-side local gradients, we develop (optimal) communication-efficient schemes for mean estimation for several lp spaces under local differential privacy (LDP). To overcome performance limitation of LDP, CLDP-SGD takes advantage of the inherent privacy amplification provided by client sub- sampling and data subsampling at each se- lected client (through SGD) as well as the recently developed shuffled model of privacy. For convex loss functions, we prove that the proposed CLDP-SGD algorithm matches the known lower bounds on the centralized private ERM while using a finite number of bits per iteration for each client, i.e., effectively get- ting communication efficiency for “free”. We also provide preliminary experimental results supporting the theory. 

   more » « less
5. Learn Locally, Correct Globally: A Distributed Algorithm for Training Graph Neural Networks

   Morteza Ramezani, Weilin Cong (April 2022, International Conference on Learning (ICLR))

   Despite the recent success of Graph Neural Networks (GNNs), training GNNs on large graphs remains challenging. The limited resource capacities of the existing servers, the dependency between nodes in a graph, and the privacy concern due to the centralized storage and model learning have spurred the need to design an effective distributed algorithm for GNN training. However, existing distributed GNN training methods impose either excessive communication costs or large memory overheads that hinders their scalability. To overcome these issues, we propose a communication-efficient distributed GNN training technique named (LLCG). To reduce the communication and memory overhead, each local machine in LLCG first trains a GNN on its local data by ignoring the dependency between nodes among different machines, then sends the locally trained model to the server for periodic model averaging. However, ignoring node dependency could result in significant performance degradation. To solve the performance degradation, we propose to apply on the server to refine the locally learned models. We rigorously analyze the convergence of distributed methods with periodic model averaging for training GNNs and show that naively applying periodic model averaging but ignoring the dependency between nodes will suffer from an irreducible residual error. However, this residual error can be eliminated by utilizing the proposed global corrections to entail fast convergence rate. Extensive experiments on real-world datasets show that LLCG can significantly improve the efficiency without hurting the performance. One-sentence Summary: We propose LLCG a communication efficient distributed algorithm for training GNNs. 

   more » « less