Federated Learning Over Wireless Fading Channels
Open accessThe DOI matches its Crossref record; 41 of the 44 references it cites were found in a registry or an index.
Abstract
We study federated machine learning at the wireless network edge, where limited power wireless devices, each with its own dataset, build a joint model with the help of a remote parameter server (PS). We consider a bandwidth-limited fading multiple access channel (MAC) from the wireless devices to the PS, and propose various techniques to implement distributed stochastic gradient descent (DSGD) over this shared noisy wireless channel. We first propose a digital DSGD (D-DSGD) scheme, in which one device is selected opportunistically for transmission at each iteration based on the channel conditions; the scheduled device quantizes its gradient estimate to a finite number of bits imposed by the channel condition, and transmits these bits to the PS in a reliable manner. Next, motivated by the additive nature of the wireless MAC, we propose a novel analog communication scheme, referred to as the compressed analog DSGD (CA-DSGD), where the devices first sparsify their gradient estimates while accumulating error from previous iterations, and project the resultant sparse vector into a low-dimensional vector for bandwidth reduction. We also design a power allocation scheme to align the received gradient vectors at the PS in an efficient manner. Numerical results show that D-DSGD outperforms other digital approaches in the literature; however, in general the proposed CA-DSGD algorithm converges faster than the D-DSGD scheme, and reaches a higher level of accuracy. We have observed that the gap between the analog and digital schemes increases when the datasets of devices are not independent and identically distributed (i.i.d.). Furthermore, the performance of the CA-DSGD scheme is shown to be robust against imperfect channel state information (CSI) at the devices. Overall these results show clear advantages for the proposed analog over-the-air DSGD scheme, which suggests that learning and communication algorithms should be designed jointly to achieve the best end-to-end performance in machine learning applications at the wireless edge.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The convergence behavior demonstrates that the noisy iterates typically introduce non-negligible optimality gap in various FL algorithms, e.g., vanilla gradient method [174], quantized gradient method [175], sparsified gradient method [173], and operator splitting method [87].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [173]
In the citing paper
“The optimality gap can be further controlled by transmit power allocation [41], [173], [176], model aggregation receiver beamforming design [31], [177], [178], and device scheduling [31], [178], [179].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [173]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Communication-Efficient Learning of Deep Networks from Decentralized Data
McMahan, H. B., Moore, E., Ramage, D., Hampson, S., & Arcas, B. A. Y. (2016) · arXiv · doi:10.48550/arxiv.1602.05629 · Cited by 14,215
Fundamentals of Wireless Communication
Tse, D., & Viswanath, P. (2005) · doi:10.1017/cbo9780511807213 · Cited by 13,162 · cited 1×
“For a transmit power P m ∗ ( t ) ( t ) , the capacity of this parallel Gaussian channel is the result of the following optimization problem [32, Section 5.4.6]: The optimization problem in (9) is solved through waterfilling, and the optimal power allocation is given by (cid:12)(cid:12) where ζ is…”
BackgroundInfluentialChannel Coding Rate in the Finite Blocklength Regime
Polyanskiy, Y., Poor, H. V., & Verdu, S. (2010) · IEEE Transactions on Information Theory · doi:10.1109/tit.2010.2043769 · Cited by 3,885 · cited 1×
“In practice, this would result in a higher error probability or reduced transmission rate [33].”
BackgroundFederated Learning: Strategies for Improving Communication Efficiency
Konečný, J., McMahan, H. B., Yu, F. X., Richtárik, P., Suresh, A. T., & Bacon, D. (2016) · arXiv · doi:10.48550/arxiv.1610.05492 · Cited by 3,766
Message-passing algorithms for compressed sensing
Donoho, D. L., Maleki, A., & Montanari, A. (2009) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.0909892106 · Cited by 2,416 · cited 1×
“For this, using its knowledge of matrix A and the CSI, PS employs the approximate message passing (AMP) algorithm [29].”
MethodDoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
Zhou, S., Wu, Y., Ni, Z., Zhou, X., Wen, H., & Zou, Y. (2016) · arXiv · doi:10.48550/arxiv.1606.06160 · Cited by 1,690
Deep Learning with Limited Numerical Precision
Gupta, S., Agrawal, A., Gopalakrishnan, K., & Narayanan, P. (2015) · arXiv · doi:10.48550/arxiv.1502.02551 · Cited by 1,364
Federated Multi-Task LearningPreprint
Smith, V., Chiang, C.-K., Sanjabi, M., & Talwalkar, A. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1705.10467 · Cited by 1,320
Deep Gradient Compression: Reducing the Communication Bandwidth for Distributed Training
Lin, Y., Han, S., Mao, H., Wang, Y., & Dally, W. J. (2017) · arXiv · doi:10.48550/arxiv.1712.01887 · Cited by 1,134
Research works printed but not linked: 3
- lecun, 1998, The MNIST Database of Handwritten Digits
- alistarh, 2017, QSGD: Communication-efficient SGD via randomized quantization and encoding, Proc NIPS, 1709
- mcmahan, 2017, Federated learning Collaborative machine learning without centralized training data
Cited by
Federated Learning in Mobile Edge Networks: A Comprehensive Survey
Lim, W. Y. B., Luong, N. C., Hoang, D. T., Jiao, Y., Liang, Y.-C., Yang, Q., Niyato, D., & Miao, C. (2020) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2020.2986024 · Cited by 2,565
“As an extension, the authors in [129] introduce error accumulation and gradient sparsification in addition to over-the-air computation.”
BackgroundMethodA survey on federated learning
Zhang, C., Xie, Y., Bai, H., Yu, B., Li, W., & Gao, Y. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.106775 · Cited by 1,899
Federated Learning for Internet of Things: Recent Advances, Taxonomy, and Open Challenges
Khan, L. U., Saad, W., Han, Z., Hossain, E., & Hong, C. S. (2021) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2021.3090430 · Cited by 892
Edge artificial intelligence for 6g: vision, enabling technologies, and applications
Letaief, K. B., Shi, Y., Lu, J., & Lu, J. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126076 · Cited by 829
“The convergence behavior demonstrates that the noisy iterates typically introduce non-negligible optimality gap in various FL algorithms, e.g., vanilla gradient method [174], quantized gradient method [175], sparsified gradient method [173], and operator splitting method [87].”
BackgroundMethodBeyond Transmitting Bits: Context, Semantics, and Task-Oriented Communications
Gunduz, D., Qin, Z., Aguerri, I. E., Dhillon, H. S., Yang, Z., Yener, A., Wong, K. K., & Chae, C.-B. (2023) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3223408 · Cited by 776
“This ‘over-the-air computation (OAC)’ approach has received significant interest in recent years thanks to its bandwidth efficiency [263], [264], [268],”
Semantic Communications for Future Internet: Fundamentals, Applications, and Challenges
Yang, W., Du, H., Liew, Z. Q., Lim, W. Y. B., Xiong, Z., Niyato, D., Chi, X., Shen, X., & Miao, C. (2023) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2022.3223224 · Cited by 670
“For example, in [180], gradient sparsification is adopted to compress the model at the transmitter by setting all but k elements with the highest magnitudes of entries to zero.”
Xu, W., Yang, Z., Ng, D. W. K., Levorato, M., Eldar, Y. C., & Debbah, M. (2023) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2023.3239189 · Cited by 660
“On the other hand, the work in [154] considered the”
Machine Learning at the Wireless Edge: Distributed Stochastic Gradient Descent Over-the-Air
Mohammadi Amiri, M., & Gunduz, D. (2020) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2020.2981904 · Cited by 548
“The digital and analog approaches proposed in this paper can be extended to more complicated channel models as it has been done in the follow up works [34]–[37].”
MethodSelf-citationDistributed Learning in Wireless Networks: Recent Progress and Future Challenges
Chen, M., Gunduz, D., Huang, K., Saad, W., Bennis, M., Feljan, A. V., & Poor, H. V. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118346 · Cited by 548
Communication-efficient edge ai: algorithms and systems
Shi, Y., Yang, K., Jiang, T., Zhang, J., & Letaief, K. B. (2020) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2020.3007787 · Cited by 488
“• Fast aggregation via over-the-air computation [21], [84], [85], [86] • Aggregation frequency control with limited bandwidth and computation resources [87], [88], [89] • Data reshuffling via index coding and pliable index coding for improving training performance [90], [91], [92] • Straggler…”
BackgroundWahab, O. A., Mourad, A., Otrok, H., & Taleb, T. (2021) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2021.3058573 · Cited by 487
“The authors of [90] propose to compress the computed gradient to a finite number of bits and communicate these bits to the parameter server.”
BackgroundXu, J., & Wang, H. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2020.3031503 · Cited by 466
“Recognizing the unique physical property of wireless transmission, [2], [20]–[22] propose analog model aggregation over the air, provided that a very stringent synchronization is available.”
BackgroundFederated Learning for Vehicular Internet of Things: Recent Advances and Open Issues
Du, Z., Wu, C., Yoshinaga, T., Yau, K.-L. A., Ji, Y., & Li, J. (2020) · IEEE Open Journal of the Computer Society · doi:10.1109/ojcs.2020.2992630 · Cited by 393
“Amiri and Gunduz [62] Three techniques to achieve efficient FL in a fading multiple access channel.”
MethodInfluentialHolistic Network Virtualization and Pervasive Network Intelligence for 6G
Shen, X., Gao, J., Wu, W., Li, M., Zhou, C., & Zhuang, W. (2022) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2021.3135829 · Cited by 392
“Over-theair computation based approaches are investigated in [194], [216], [217].”
MethodZhu, G., Du, Y., Gunduz, D., & Huang, K. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2020.3039309 · Cited by 376
“Existing works on over-the-air aggregation, as well as its application in the context of FEEL [6], [7], [17]–[19] consider analog modulation assuming that the transmitter can modulate the carrier waveform as desired, freely choosing the I/Q coefficients as arbitrary real number.”
BackgroundFederated Learning for 6G: Applications, Challenges, and Opportunities
Yang, Z., Chen, M., Wong, K.-K., Poor, H. V., & Cui, S. (2022) · Engineering · doi:10.1016/j.eng.2021.12.002 · Cited by 374
Federated Learning in Smart City Sensing: Challenges and Opportunities
Jiang, J. C., Kantarci, B., Oktug, S., & Soyata, T. (2020) · Sensors · doi:10.3390/s20216230 · Cited by 310
“Previous studies have tackled communication challenges regarding Federated Learning [176,177].”
Federated learning in a medical context: A systematic literature review
Pfitzner, B., Steckhan, N., & Arnrich, B. (2021) · ACM Transactions on Internet Technology · doi:10.1145/3412357 · Cited by 288
Edge Intelligence: Empowering Intelligence to the Edge of Network
Xu, D., Li, T., Li, Y., Su, X., Tarkoma, S., Jiang, T., Crowcroft, J., & Hui, P. (2021) · Proceedings of the IEEE · doi:10.1109/jproc.2021.3119950 · Cited by 272
What is Semantic Communication? A View on Conveying Meaning in the Era of Machine Intelligence
Lan, Q., Wen, D., Zhang, Z., Zeng, Q., Chen, X., Popovski, P., & Huang, K. (2021) · Journal of Communications and Information Networks · doi:10.23919/jcin.2021.9663101 · Cited by 255
Convergence of Update Aware Device Scheduling for Federated Learning at the Wireless Edge
Amiri, M. M., Gunduz, D., Kulkarni, S. R., & Poor, H. V. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3052681 · Cited by 232
“with sufficient quality and reliability [12], [16], [17], [19].”
BackgroundMethodInfluentialSelf-citationOver-the-Air Federated Learning From Heterogeneous Data
Sery, T., Shlezinger, N., Cohen, K., & Eldar, Y. (2021) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2021.3090323 · Cited by 219
“In the context of FL, which is a distributed machine learning setup, with OTA computations, the works [17], [18] considered scenarios where the model updates are sparse with an identical sparsity pattern, which is not likely to hold when the data is heterogeneous.”
BackgroundMethodInfluentialEdge Intelligence: Architectures, Challenges, and ApplicationsPreprint
Xu, D., Li, T., Li, Y., Su, X., Tarkoma, S., Jiang, T., Crowcroft, J., & Hui, P. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.12172 · Cited by 264
Gunduz, D., de Kerret, P., Sidiropoulos, N. D., Gesbert, D., Murthy, C. R., & van der Schaar, M. (2019) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2019.2933969 · Cited by 214
“This approach can also be extended to the scenario with fading [95], [96], in which case power control can be employed at the devices to align their transmissions at the same received power level.”
BackgroundApplications of federated learning in smart cities: recent advances, taxonomy, and open challenges
Zheng, Z., Zhou, Y., Sun, Y., Wang, Z., Liu, B., & Li, K. (2022) · Connection Science · doi:10.1080/09540091.2021.1936455 · Cited by 206
“In terms of bandwidth, the analogue solution is better than the digital solution (Amiri & Gündüz, 2020b).”
BackgroundFederated Learning via Intelligent Reflecting Surface
Wang, Z., Qiu, J., Zhou, Y., Shi, Y., Fu, L., Chen, W., & Letaief, K. B. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3099505 · Cited by 205
“For highdimensional model parameters, we can leverage specification and compressed sensing methods such as those proposed in [18], [19] to further reduce the model dimension, which is however out of the scope of this paper.”
MethodLo, S. K., Lu, Q., Wang, C., Paik, H.-Y., & Zhu, L. (2022) · ACM Computing Surveys · doi:10.1145/3450288 · Cited by 202
Wireless Federated Learning with Local Differential Privacy
Seif, M., Tandon, R., & Li, M. (2020) · 2020 IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit44484.2020.9174426 · Cited by 200
Wireless Communications for Collaborative Federated Learning
Chen, M., Poor, H. V., Saad, W., & Cui, S. (2020) · IEEE Communications Magazine · doi:10.1109/mcom.001.2000397 · Cited by 192
“Therefore, it is necessary to consider the optimization of wireless networks to improve FL performance, as pointed out in [7]–[9].”
BackgroundDesign and Analysis of Uplink and Downlink Communications for Federated Learning
Zheng, S., Shen, C., & Chen, X. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2020.3041388 · Cited by 191
Federated Learning for Channel Estimation in Conventional and RIS-Assisted Massive MIMO
Elbir, A. M., & Çöleri, S. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3128392 · Cited by 190
“Therefore, the corrupted model parameters and gradients at the tth iteration are given as [24, 41]”
BackgroundMethodInfluentialAge-Based Scheduling Policy for Federated Learning in Mobile Edge Networks
Yang, H. H., Arafa, A., Quek, T. Q. S., & Vincent Poor, H. (2020) · ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp40776.2020.9053740 · Cited by 189
“This salient feature of on-device training brings along great advantages of eliminating the large communication overheads as well as preserving data privacy, and hence making FL particularly relevant for mobile applications [6–10].”
BackgroundFederated Machine Learning for Intelligent IoT via Reconfigurable Intelligent Surface
Yang, K., Shi, Y., Zhou, Y., Yang, Z., Fu, L., & Chen, W. (2020) · IEEE Network · doi:10.1109/mnet.011.2000045 · Cited by 184
“Over-the-air computation (AirComp) provides a novel simultaneous access technique to support fast local model aggregation for federated machine learning via exploiting the signal superposition property of multi-access channels [6]– [8].”
MethodOpportunities of Federated Learning in Connected, Cooperative, and Automated Industrial Systems
Savazzi, S., Nicoli, M., Bennis, M., Kianoush, S., & Barbieri, L. (2021) · IEEE Communications Magazine · doi:10.1109/mcom.001.2000200 · Cited by 181
“More recently, analog FL and hybrid analog-digital implementations have been proposed for fast and synchronous model averaging [4].”
MethodFederated Learning: A signal processing perspective
Gafni, T., Shlezinger, N., Cohen, K., Eldar, Y. C., & Poor, H. V. (2022) · IEEE Signal Processing Magazine · doi:10.1109/msp.2021.3125282 · Cited by 176
“This may be exploited by applying random compressive linear precoding, which can be applied in fading channels without requiring the users to have channel state information, having the receiver recover the aggregated model via compressed sensing mechanisms [16].”
BackgroundOptimized Power Control Design for Over-the-Air Federated Edge Learning
Cao, X., Zhu, G., Xu, J., Wang, Z., & Cui, S. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126060 · Cited by 173
On Analog Gradient Descent Learning Over Multiple Access Fading Channels
Sery, T., & Cohen, K. (2020) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2020.2989580 · Cited by 172
“In the case of fading channels [10], each node uses power control to cancel the channel effect at the receiver, where nodes that experience deep fading do not transmit.”
BackgroundFederated Learning Over Wireless IoT Networks With Optimized Communication and Resources
Chen, H., Huang, S., Zhang, D., Xiao, M., Skoglund, M., & Poor, H. V. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3151193 · Cited by 202
“Typical approaches in this direction range from: 1) compressing the uploaded gradients via coding [12], [13], quantization [14], [15], and sparsification [16], [17]; 2) limiting the model sharing by only updating clients with significant training improvement [18], [19]; or 3) accelerating the training process by adopt…”
Federated Learning With Non-IID Data in Wireless Networks
Zhao, Z., Feng, C., Hong, W., Jiang, J., Jia, C., Quek, T. Q. S., & Peng, M. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3108197 · Cited by 165
Liu, H., Yuan, X., & Zhang, Y.-J. A. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3086116 · Cited by 165
“The authors in [11], [12] studied over-the-air model aggregation with gradient sparsification and compression in Gaussian channels and wireless fading channels.”
BackgroundA Survey on Over-the-Air Computation
Şahin, A., & Yang, R. (2023) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2023.3264649 · Cited by 162
“For example, in [180] and [77], for distributed learning, the EDs sparsify their gradient estimates before they project them into a lower”
BackgroundShen, M., Wang, X., Park, J. H., Yi, Y., & Che, W. (2022) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2022.3222491 · Cited by 162
“Consequently, system performance could be deteriorated by packet dropouts [6], time delays [7], fading channels [8], and so on.”
BackgroundLearning Rate Optimization for Federated Learning Exploiting Over-the-Air Computation
Xu, C., Liu, S., Yang, Z., Huang, Y., & Wong, K.-K. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118402 · Cited by 159
“Specifically, a gradient sparsification and random linear projection based approach is proposed and the reduced data was transmitted via AirComp to address the bandwidth and power limitations, which outperformed its digital counterpart [23], [24].”
MethodGradient Statistics Aware Power Control for Over-the-Air Federated Learning
Zhang, N., & Tao, M. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3065748 · Cited by 178
“By exploiting the sparsity pattern in gradient vectors, the work [21] projects the gradient estimate in each device into a lowdimensional vector and transmits only the important gradient entries while accumulating the error from previous iterations.”
BackgroundMethodDynamic Scheduling for Over-the-Air Federated Edge Learning With Energy Constraints
Sun, Y., Zhou, S., Niu, Z., & Gündüz, D. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126078 · Cited by 174
Integrated Sensing, Communication, and Computation Over-the-Air: MIMO Beamforming Design
Li, X., Liu, F., Zhou, Z., Zhu, G., Wang, S., Huang, K., & Gong, Y. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3233795 · Cited by 168
“The existing literature have investigated the implementation of AirComp in FEEL from different perspectives, including communication-learning tradeoff [16], devices scheduling [17], update compression [18], beamforming design [19], hyper-parameters control [20], learning rate control [21], power control [22], digital…”
BackgroundMethodMulti-Stage Hybrid Federated Learning Over Large-Scale D2D-Enabled Fog Networks
Hosseinalipour, S., Azam, S. S., Brinton, C. G., Michelusi, N., Aggarwal, V., Love, D. J., & Dai, H. (2022) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2022.3143495 · Cited by 162
“Related Work Researchers have considered the effects of limited and imperfect communication capabilities in wireless networks – such as channel fading, packet loss, and limited bandwidth – on the operation of federated learning [12]–[15].”
BackgroundMethodInfluentialChallenges, Applications and Design Aspects of Federated Learning: A Survey
Rahman, K. M. J., Ahmed, F., Akhter, N., Hasan, M., Amin, R., Aziz, K. E., Islam, A. K. M. M., Mukta, M. S. H., & Islam, A. K. M. N. (2021) · IEEE Access · doi:10.1109/access.2021.3111118 · Cited by 147
Federated Learning for Computationally Constrained Heterogeneous Devices: A Survey
Pfeiffer, K., Rapp, M., Khalili, R., & Henkel, J. (2023) · ACM Computing Surveys · doi:10.1145/3596907 · Cited by 173
“Most of the current research focuses on making the transmission more eicient by using compression and quantization schemes [3, 28, 54, 72].”
BackgroundSemantic Communication Meets Edge Intelligence
Yang, W., Liew, Z. Q., Lim, W. Y. B., Xiong, Z., Niyato, D. T., Chi, X., Cao, X., & Letaief, K. B. (2022) · IEEE Wireless Communications · doi:10.1109/mwc.004.2200050 · Cited by 141
“For example, in [13], gradient sparsification is adopted to compress the model at the transmitter by setting all but k elements with the highest magnitudes of entries to zero.”
MethodCommunication-efficient distributed learning: An overview
Cao, X., Başar, T., Diggavi, S., Eldar, Y. C., Letaief, K. B., Poor, H. V., & Zhang, J. (2023) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2023.3242710 · Cited by 139
Federated Learning Over Wireless Device-to-Device Networks: Algorithms and Convergence Analysis
Xing, H., Simeone, O., & Bi, S. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118400 · Cited by 160
“At the physical layer, over-theair computation (AirComp) was investigated in [15], [24], [25], [28], [29] as a promising solution to support simultaneous transmissions by leveraging the waveform superposition property of the wireless medium.”
MethodFederated learning: Applications, challenges and future directions
Bharati, S., Mondal, M. R. H., Podder, P., & Prasath, V. B. S. (2022) · International Journal of Hybrid Intelligent Systems · doi:10.3233/his-220006 · Cited by 194
“Using a remote parameter server and data from each device’s own data collection, the authors of [60] investigated the use of FL in edge computing and wireless networks.”
An Incentive Mechanism for Cross-Silo Federated Learning: A Public Goods Perspective
Tang, M., & Wong, V. W. S. (2021) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom42981.2021.9488705 · Cited by 131
“Some other works considered resource optimization problem for FL, e.g., [13], [14].”
BackgroundFederated Learning in Multi-RIS-Aided Systems
Ni, W., Liu, Y., Yang, Z., Tian, H., & Shen, X. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3130444 · Cited by 134
Transmission Power Control for Over-the-Air Federated Averaging at Network Edge
Cao, X., Zhu, G., Xu, J., & Cui, S. R. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3143217 · Cited by 136
Energy-Aware Analog Aggregation for Federated Learning with Redundant Data
Sun, Y., Zhou, S., & Gündüz, D. (2020) · International Conference on Communications · doi:10.1109/icc40277.2020.9148853 · Cited by 129
“The communication burden of FL mainly comes from the global model aggregation, which can be reduced by efficient scheduling and resource allocation [10]–[14], gradient quantization and sparsification [14]–[16], or via analog aggregation [15]–[17].”
BackgroundInfluentialAbkenar, F. S., Ramezani, P., Iranmanesh, S., Murali, S., Chulerttiyawong, D., Wan, X., Jamalipour, A., & Raad, R. (2022) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2022.3211462 · Cited by 131
“The replacement of fixed ECNs with mobile ECNs can improve the efficiency and performance of the corresponding networks because the mobile ECNs provide complementary resources at the closest distance to the TNs [33], [38].”
BackgroundDynamic Aggregation for Heterogeneous Quantization in Federated Learning
Chen, S., Shen, C., Zhang, L., & Tang, Y. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3076613 · Cited by 132
“The authors of [6] study the quantization and protocol design for uplink and downlink communications of FL. Bandwidth-limited fading medium access control (MAC) between the clients and server is investigated in [26], which also proposes new methods for distributed SGD over a shared wireless channel.”
BackgroundMethodReputation-Based Federated Learning for Secure Wireless Networks
Song, Z., Sun, H., Yang, H. H., Wang, X., Zhang, Y., & Quek, T. Q. S. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3079104 · Cited by 122
Semantic Communication: A Survey on Research Landscape, Challenges, and Future Directions
Getu, T. M., Kaddoum, G., & Bennis, M. (2024) · Proceedings of the IEEE · doi:10.1109/jproc.2024.3520707 · Cited by 110
Decentralized Federated Learning via SGD over Wireless D2D Networks
Xing, H., Simeone, O., & Bi, S. (2020) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc48557.2020.9154332 · Cited by 122
Liu, P., Zhu, G., Wang, S., Jiang, W., Luo, W., Poor, H. V., & Cui, S. R. (2022) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2022.3226836 · Cited by 109
“While interesting, the works in [6], [7], [9], [10], [11], [12], [13], [14] all took the learning error as the objective function.”
BackgroundInterference Management for Over-the-Air Federated Learning in Multi-Cell Wireless Networks
Wang, Z., Zhou, Y., Shi, Y., & Zhuang, W. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3180799 · Cited by 108
“Note that, when the dimension of the model/gradient parameter is quite large, the compression techniques can be leveraged to enable the parameters to be transmitted in one coherence block [22], [39].”
BackgroundAn Energy and Carbon Footprint Analysis of Distributed and Federated Learning
Savazzi, S., Rampa, V., Kianoush, S., & Bennis, M. (2022) · IEEE Transactions on Green Communications and Networking · doi:10.1109/tgcn.2022.3186439 · Cited by 133
“In contrast to digital implementations, analog FL designs [9], [17] get over the restrictions of time scheduled”
BackgroundGuo, H., Liu, A., & Lau, V. K. N. (2020) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3002925 · Cited by 123
“The existing works on the AGA solution suggest updating the model by the SGD algorithm with full gradient and a constant learning rate [16], [17] or the ADAM optimizer [18]–[20].”
BackgroundMethodApplications and Techniques for Fast Machine Learning in Science
Deiana, A. M., Tran, N. V., Agar, J., Blott, M., Guglielmo, G. D., Duarte, J. M., Harris, P. C., Hauck, S., Liu, M., Neubauer, M. S., et al. (2022) · Frontiers in Big Data · doi:10.3389/fdata.2022.787421 · Cited by 125
Zhao, T., Li, F., & He, L. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3170900 · Cited by 113
“Keeping the data at IIoT devices not only protects the privacy but also mitigates the network congestion [6]–[9].”
BackgroundFederated Edge Learning With Misaligned Over-the-Air Computation
Shao, Y., Gündüz, D., & Liew, S. C. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3125798 · Cited by 127
“Prior works on OAC, with the exception of [17], considered only the perfectly aligned case [10], [11], [13], [14], [20], where there is neither channel-gain misalignment nor time misalignment, which we refer to as the aligned OAC.”
BackgroundAmiri, M. M., Duman, T. M., Gündüz, D., Kulkarni, S. R., & Poor, H. V. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3065920 · Cited by 115
“Several studies have incorporated over-the-air computation into FEEL utilizing the superposition property of the wireless multiple access channel (MAC) for reliable transmission from the devices to the PS, where the MAC naturally provides the sum of the updates from the devices to the PS [1], [9]–[13], [16]–[18], [34]…”
MethodInfluentialSelf-citationDecentralized Wireless Federated Learning With Differential Privacy
Chen, S., Yu, D., Zou, Y., Yu, J., & Cheng, X. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3145010 · Cited by 100
Federated Learning Over Noisy Channels: Convergence Analysis and Design Examples
Wei, X., & Shen, C. (2022) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2022.3140788 · Cited by 115
Towards efficient communications in federated learning: A contemporary survey
Zhao, Z., Mao, Y., Liu, Y., Song, L., Ouyang, Y., Chen, X., & Ding, W. (2023) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2022.12.053 · Cited by 121
Convergence of Federated Learning Over a Noisy Downlink
Amiri, M. M., Gündüz, D., Kulkarni, S. R., & Poor, H. V. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3103874 · Cited by 108
“The power-limited wireless devices then transmit their local model updates to the PS over a bandwidth-limited fading MAC in an analog fashion, whose superiority over digital transmission for the uplink has been shown in the literature [12], [13], [16].”
BackgroundMethodInfluentialSelf-citationTask-Oriented Over-the-Air Computation for Multi-Device Edge AI
Wen, D., Jiao, X., Liu, P., Zhu, G., Shi, Y., & Huang, K. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3294703 · Cited by 108
“…AirComp has been a promising technique widely exploited in federated edge learning for communication efficiency enhancement (see, e.g., [38], [39], [40], [41], [42]) and in many other crtical tasks such as distributed localization and wireless control systems (see, e.g., [43] and references…”
Salau, B. A., Rawal, A., & Rawat, D. B. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3170449 · Cited by 106
“For instance federated learning has been applied to handle scalability between the radio resources block and number of devices in a network, network overload and power allocation [47], [87], [46].”
Semi-Synchronous Federated Learning Protocol With Dynamic Aggregation in Internet of Vehicles
Liang, F., Yang, Q., Liu, R., Wang, J., Sato, K., & Guo, J. B. (2022) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2022.3148872 · Cited by 92
“In FL, the computing nodes train local models by using their own data, respectively, and then upload the local model instead of data, to a logical centralized parameter server that aggregates all local models into a global model [10].”
MethodFederated Learning and Its Role in the Privacy Preservation of IoT Devices
Alam, T., & Gupta, R. (2022) · Future Internet · doi:10.3390/fi14090246 · Cited by 92
“This paper addresses the use of FL in smart cities, including communications, healthcare, and the Internet of Things [85,89,90].”
BackgroundA Compressive Sensing Approach for Federated Learning Over Massive MIMO Communication Systems
Jeon, Y., Amiri, M. M., Li, J., & Poor, H. V. (2020) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2020.3038407 · Cited by 91
“More generally, if all the local training data sets have an identical distribution, the wireless devices may have similar sparsity patterns [20], [21]; in this case, the transmitted signals constructed from (2) may not be sparse.”
BackgroundSelf-citationFederated Learning Over Energy Harvesting Wireless Networks
Hamdi, R., Chen, M., Said, A. B., Qaraqe, M., & Poor, H. V. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2021.3089054 · Cited by 84
“The impact of wireless channel randomness on the FL performance was studied in [9] where the authors proposed a novel efficient transmission method based on gradient sparsification.”
MethodCommunication-Efficient Federated Learning via Quantized Compressed Sensing
Oh, Y., Lee, N., Jeon, Y., & Poor, H. V. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3201207 · Cited by 96
“Fortunately, this information loss can be partially compensated by accumulating the dropped gradient entries and then by adding these entries in the next iteration [16], [17].”
Radio Resource Management for Cellular-Connected UAV: A Learning Approach
Li, Y., & Aghvami, A. H. (2023) · IEEE Transactions on Communications · doi:10.1109/tcomm.2023.3262826 · Cited by 82
Resource Constrained Vehicular Edge Federated Learning With Highly Mobile Connected Vehicles
Pervej, M. F., Jin, R., & Dai, H. (2023) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2023.3273700 · Cited by 91
“The energy budget, required CPU cycle to process per bit data, per unit energy cost and φ̄vs are randomly chosen from [20, 30] Joules, [20, 30], [5, 10] units and [10, 20] units, respectively.”
BackgroundMethodA survey: Distributed Machine Learning for 5G and beyond
Nassef, O., Sun, W., Purmehdi, H., Tatipamula, M., & Mahmoodi, T. (2022) · Computer Networks · doi:10.1016/j.comnet.2022.108820 · Cited by 94
Joint Optimization of Communications and Federated Learning Over the Air
Fan, X., Wang, Y., Huo, Y., & Tian, Z. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3130111 · Cited by 86
A Comprehensive Survey on Communication-Efficient Federated Learning in Mobile Edge Environments
Jia, N., Qu, Z., Ye, B., Wang, Y., Hu, S., & Guo, S. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3535957 · Cited by 105
“The sparse gradients are projected into a low-dimensional vector to reduce the bandwidth usage [70].”
Wang, S., Hosseinalipour, S., Gorlatova, M. A., Brinton, C. G., & Chiang, M. (2022) · IEEE Transactions on Network and Service Management · doi:10.1109/tnsm.2022.3216326 · Cited by 76
“Distributed ML over wireless networks: Recent literature concerning ML by wireless networks has shifted towards federated learning [14], and is mostly focused on studying the convergence and behavior of federated learning over wireless networks [15], [16], [17], [18], [19], [20].”
BackgroundComputation Scheduling for Distributed Machine Learning With Straggling Workers
Amiri, M. M., & Gündüz, D. (2019) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2019.2952051 · Cited by 80
Self-citationUpdate Aware Device Scheduling for Federated Learning at the Wireless Edge
Amiri, M. M., Gündüz, D., Kulkarni, S. R., & Poor, H. V. (2020) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit44484.2020.9173960 · Cited by 71
“This generalizes the approach studied in [11], where only a single device is scheduled based on the channel conditions.”
BackgroundSelf-citationConvergence Analysis and System Design for Federated Learning Over Wireless Networks
Wan, S., Lu, J., Fan, P., Shao, Y., Peng, C., & Letaief, K. B. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118351 · Cited by 71
Communicate to Learn at the Edge
Gündüz, D., Kurka, D. B., Jankowski, M., Amiri, M. M., Özfatura, E., & Sreekumar, S. (2020) · IEEE Communications Magazine · doi:10.1109/mcom.001.2000394 · Cited by 76
“Hence, rather than transmitting individual updates in an orthogonal fashion, signal superposition property of the wireless medium can be exploited to directly convey the sum of the local parameters through over-the-air computation [13], [14].”
BackgroundMethodSelf-citation1-Bit Compressive Sensing for Efficient Federated Learning Over the Air
Fan, X., Wang, Y., Huo, Y., & Tian, Z. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3209190 · Cited by 80
“The power allocation scheme in [21], [22] scales the power of the vectors containing the gradient information of different devices to satisfy the average power constraint.”
BackgroundMethodInfluentialAdaptive Transmission Scheduling in Wireless Networks for Asynchronous Federated Learning
Lee, H.-S., & Lee, J.-W. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118353 · Cited by 76
“…existing works focus on reducing the communication costs of a single local model transmission of each edge device; analog model aggregation methods using the inherent characteristic of a wireless medium [3], [4], a gradient reusing method [16], and local gradient compressing methods [17], [18].”
BackgroundMethodCommunication-efficient distributed AI strategies for the IoT edge
Mwase, C., Jin, Y., Westerlund, T., Tenhunen, H., & Zou, Z. (2022) · Future Generation Computer Systems · doi:10.1016/j.future.2022.01.013 · Cited by 74
Massive Digital Over-the-Air Computation for Communication-Efficient Federated Edge Learning
Qiao, L., Gao, Z., Mashhadi, M. B., & Gündüz, D. (2024) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2024.3431572 · Cited by 81
“One of these emerging frameworks is federated edge learning (FEEL), where a central server coordinates multiple AIoT devices to train an ML model using their respective local datasets and computing resources [10].”
BackgroundMethodResource Allocation for Latency-Aware Federated Learning in Industrial Internet of Things
Gao, W., Zhao, Z., Min, G., Ni, Q., & Jiang, Y. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2021.3073642 · Cited by 68
“Considering the fading channels, Mohammad et al. [24] put forward an analog scheme to exploit gradient sparsification and achieve more efficient use of limited channel bandwidth.”
BackgroundOn the Convergence of Decentralized Federated Learning Under Imperfect Information Sharing
Chellapandi, V. P., Upadhyay, A., Hashemi, A., & Żak, S. H. (2023) · IEEE Control Systems Letters · doi:10.1109/lcsys.2023.3290470 · Cited by 38
“Other works including [17], [18], [19], [20], [21] study the impact of noise in server-assisted FL.”
BackgroundLi, L., Yang, L., Guo, X., Shi, Y., Wang, H., Chen, W., & Letaief, K. B. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118431 · Cited by 37
“Moreover, resource allocation becomes critical to improve the learning convergence rate and model prediction accuracy in wireless FL, including the transmission power control [13]–[16], model aggregation beamforming [17], [18], and device scheduling [18]–[21].”
Fast Convergence Algorithm for Analog Federated Learning
Xia, S., Zhu, J., Yang, Y., Zhou, Y., Shi, Y., & Chen, W. (2021) · ICC 2021 - IEEE International Conference on Communications · doi:10.1109/icc42927.2021.9500875 · Cited by 33
“Hence, the power constraint of the edge server is not as strict as the edge devices, and the downlink communication is assumed to be error-free [14], [15], [19].”
BackgroundFederated Learning-Based Cell-Free Massive MIMO System for Privacy-Preserving
Zhang, J., Zhang, J., Ng, D. W. K., & Ai, B. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3225812 · Cited by 30
“In particular, multiple UEs are allowed to jointly train a global ML model without having to exchange raw data among them or transfer their data to the CPU [12].”
Background
Metadata
- Authors
- Amiri, M. M., & Gunduz, D.
- Affiliations
- Princeton University (US)
- Venue
- IEEE Transactions on Wireless Communications
- Published
- 2020-05
- Volume
- 19
- Issue
- 5
- Pages
- 3546-3557
- ISSN
- 1536-1276 (ISSN-L)
- 1558-2248
- OpenAlex
- W2963647784
- Cited by
- 625
- References
- 44
- Citations by year
- 2026: 47 · 2025: 66 · 2024: 93 · 2023: 111 · 2022: 129 · 2021: 126 · 2020: 47
- FWCI
- 40.56
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Wireless Communications (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- IRIS UNIMORE (University of Modena and Reggio Emilia) (submittedVersion)
- IRIS UNIMORE (University of Modena and Reggio Emilia) (submittedVersion, cc-by)
- Spiral (Imperial College London) (submittedVersion)
- arXiv (Cornell University)
- Funders
- European Research Council (ERC) Starting Grant BEACON (677854)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed Sensor Networks and Detection Algorithms
- Physical Sciences › Computer Science › Artificial Intelligence › Stochastic Gradient Optimization Techniques
- Keywords
- federated learning, wireless edge computing, fading multiple access channel, distributed stochastic gradient descent, over-the-air computation, gradient quantization, gradient sparsification, power allocation, bandwidth-limited channels, non-IID data, imperfect channel state information, parameter server
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.