Sum-Rate-Distortion Function for Indirect Multiterminal Source Coding in Federated Learning
Not on the map. Why?
Abstract (arXiv)
One of the main focus in federated learning (FL) is the communication efficiency since a large number of participating edge devices send their updates to the edge server at each round of the model training. Existing works reconstruct each model update from edge devices and implicitly assume that the local model updates are independent over edge devices. In FL, however, the model update is an indirect multi-terminal source coding problem, also called as the CEO problem where each edge device cannot observe directly the gradient that is to be reconstructed at the decoder, but is rather provided only with a noisy version. The existing works do not leverage the redundancy in the information transmitted by different edges. This paper studies the rate region for the indirect multiterminal source coding problem in FL. The goal is to obtain the minimum achievable rate at a particular upper bound of gradient variance. We obtain the rate region for the quadratic vector Gaussian CEO problem under unbiased estimator and derive an explicit formula of the sum-rate-distortion function in the special case where gradient are identical over edge device and dimension. Finally, we analyse communication efficiency of convex Minibatched SGD and non-convex Minibatched SGD based on the sum-rate-distortion function, respectively.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 25 of its 38 references that are in the map. First posted on arXiv 21 Jan 2021.
Builds on
25 of 38 references.
References · 25
- 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,592
- 2019Federated Learning: Challenges, Methods, and Future Directionscited by 1,531
- 2016Federated Learning: Strategies for Improving Communication Efficiencycited by 1,461
- 2016XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networkscited by 829
- 2019Towards Federated Learning at Scale: System Designcited by 826
- 2013Stochastic First- and Zeroth-order Methods for Nonconvex Stochastic Programmingcited by 716
- 2014Convex Optimization: Algorithms and Complexitycited by 569
- 2018signSGD: Compressed Optimisation for Non-Convex Problemscited by 561
- 2016DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradientscited by 557
- 2017Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Trainingcited by 471
- 2018Adaptive Federated Learning in Resource Constrained Edge Computing Systemscited by 408
- 2018Client Selection for Federated Learning with Heterogeneous Resources in Mobile Edgecited by 347
13 more
- 2017TernGrad: Ternary Gradients to Reduce Communication in Distributed Deep Learningcited by 253
- 2017Sparse Communication for Distributed Gradient Descentcited by 243
- 2015Compressing Neural Networks with the Hashing Trickcited by 150
- 2016Exploiting Cyclic Symmetry in Convolutional Neural Networkscited by 91
- 2016Group Sparse Regularization for Deep Neural Networkscited by 89
- 2019Accelerating Federated Learning via Momentum Gradient Descentcited by 75
- 2020Convergence of Update Aware Device Scheduling for Federated Learning at the Wireless Edgecited by 71
- 2019Age-Based Scheduling Policy for Federated Learning in Mobile Edge Networkscited by 39
- 2019Robust Federated Learning with Noisy Communicationcited by 33
- 2018Variance-based Gradient Compression for Efficient Distributed Deep Learningcited by 21
- 2019Data-Importance Aware User Scheduling for Communication-Efficient Edge Machine Learningcited by 20
- 2016ASAGA: Asynchronous Parallel SAGAcited by 18
- 2014Memory Bounded Deep Convolutional Networkscited by 17
Built on by
Citing papers per year: 2022: 1; 2023: 1.
2 papers.