On Analog Gradient Descent Learning Over Multiple Access Fading Channels
Open accessThe DOI matches its Crossref record; 63 of the 64 references it cites were found in a registry or an index.
Abstract
We consider a distributed learning problem over multiple access channel (MAC) using a large wireless network. The computation is made by the network edge and is based on received data from a large number of distributed nodes which transmit over a noisy fading MAC. The objective function is a sum of the nodes' local loss functions. This problem has attracted a growing interest in distributed sensing systems, and more recently in federated learning. We develop a novel Gradient-Based Multiple Access (GBMA) algorithm to solve the distributed learning problem over MAC. Specifically, the nodes transmit an analog function of the local gradient using common shaping waveforms and the network edge receives a superposition of the analog transmitted signals used for updating the estimate. GBMA does not require power control or beamforming to cancel the fading effect as in other algorithms, and operates directly with noisy distorted gradients. We analyze the performance of GBMA theoretically, and prove that it can approach the convergence rate of the centralized gradient descent (GD) algorithm in large networks. Specifically, we establish a finite-sample bound of the error for both convex and strongly convex loss functions with Lipschitz gradient. Furthermore, we provide energy scaling laws for approaching the centralized convergence rate as the number of nodes increases. Finally, experimental results support the theoretical findings, and demonstrate strong performance of GBMA using synthetic and real data.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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 [174]
Citations
Cites
UCI Machine Learning Repository
Asuncion, A. (2007) · Medical Entomology and Zoology · Cited by 25,370
Goldsmith, A. (2005) · Cambridge University Press eBooks · doi:10.1017/cbo9780511841224 · Cited by 10,658
Mathematical Analysis of Random Noise
Rice, S. O. (1944) · Bell System Technical Journal · doi:10.1002/j.1538-7305.1944.tb00874.x · Cited by 5,480 · cited 1×
“If the channel has a fixed LOS component, the channel distribution can be well approximated by the Rician fading model [47].”
MethodMathematical Analysis of Random Noise
Rice, S. O. (1945) · Bell System Technical Journal · doi:10.1002/j.1538-7305.1945.tb00453.x · Cited by 2,898
Adaptive Federated Learning in Resource Constrained Edge Computing Systems
Wang, S., Tuor, T., Salonidis, T., Leung, K. K., Makaya, C., He, T., & Chan, K. (2019) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2019.2904348 · Cited by 2,232 · cited 2×
“In recent years, other stochastic gradient-descent (SGD) based methods were developed for federated learning [43], [44].”
BackgroundMethodRobust Stochastic Approximation Approach to Stochastic Programming
Nemirovski, A., Juditsky, A., Lan, G., & Shapiro, A. (2009) · SIAM Journal on Optimization · doi:10.1137/070704277 · Cited by 1,945
Federated Optimization: Distributed Machine Learning for On-Device Intelligence
Konečný, J., McMahan, H. B., Ramage, D., & Richtárik, P. (2016) · arXiv · doi:10.48550/arxiv.1610.02527 · Cited by 1,563
Pegasos: primal estimated sub-gradient solver for SVM
Shalev-Shwartz, S., Singer, Y., Srebro, N., & Cotter, A. (2011) · Mathematical Programming · doi:10.1007/s10107-010-0420-4 · Cited by 1,477
Federated Learning via Over-the-Air Computation
Yang, K., Jiang, T., Shi, Y., & Ding, Z. (2020) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2019.2961673 · Cited by 1,084 · cited 1×
“In [15], [16], multiple antennas were used at the receiver, where beamforming was used to maximize the number of devices scheduled for transmission.”
MethodBertin-Mahieux, T., Ellis, D. P. W., Whitman, B., & Lamere, P. (2011) · Columbia University · doi:10.7916/d8nz8j07 · Cited by 944 · cited 1×
“We used the dataset available by UCI Machine Learning Repository [58], extracted from the Million Song Dataset collaborative project between The Echo Nest and LabROSA [59].”
Method
Research works printed but not linked: 2
- 1998, Introductory lectures on convex programming volume i: Basic course, Lecture Notes Series, 3, 51
- bertin-mahieux, 0, Proc Int Conf Music Inf Retrieval
Cited by
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].”
MethodA Survey on Integrated Sensing, Communication, and Computation
Wen, D., Zhou, Y., Li, X., Shi, Y., Huang, K., & Letaief, K. B. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3521498 · Cited by 363
“Although transmitting model parameters reduce the communication overhead by a significant amount compared to transmitting raw data, scarce communication resources still limit the communication efficiency in wireless FL systems, which motivates the emergence of AirComp-based FL [22], [23], [78], [79].”
Over-the-Air Computing for Wireless Data Aggregation in Massive IoT
Zhu, G., Xu, J., Huang, K., & Cui, S. (2021) · IEEE Wireless Communications · doi:10.1109/mwc.011.2000467 · Cited by 235
Privacy for Free: Wireless Federated Learning via Uncoded Transmission With Adaptive Power Control
Liu, D., & Simeone, O. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2020.3036948 · Cited by 234
Over-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
“Additional related recent works on OTA FL, including [20], [22], [26], considered the distributed application of full gradient descent optimization over noisy channels.”
BackgroundMethodInfluentialSelf-citationFederated 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
“In [20], the authors developed a gradient-based multiple-access algorithm to directly deal with noise distorted gradients in FL systems.”
BackgroundWireless 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
Federated 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
Self-citationOptimized 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
Over-the-Air Computation for 6G: Foundations, Technologies, and Applications
Wang, Z., Zhao, Y., Zhou, Y., Shi, Y., Jiang, C., & Letaief, K. B. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3405448 · Cited by 167
“…frequency offsets in OFDM transmission induce phase noises to the transmitted signals, which may lead to the failure of coherent waveform superposition, an efficient protocol is proposed in [160] to address this issue by adding one more signaling round for phase error estimation and compensation.”
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
“b) Phase correction: phase correction (PC) is a preequalization strategy to achieve a coherent superposition by rotating the phase of the transmitted symbols based on the amount of the phase distortion due to the channel [10], [92], [94], [95], [96], [97].”
BackgroundGuo, H., Liu, A., & Lau, V. K. N. (2021) · IEEE Internet of Things Journal · doi:10.1109/jiot.2020.3002925 · Cited by 105
“The analog gradient aggregation (AGA) solution for FL is investigated in [17]–[21].”
BackgroundMethodInfluentialTowards 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 123
Amiri, M. M., Duman, T. M., Gunduz, D., Kulkarni, S. R., & Poor, H. V. P. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3065920 · Cited by 98
“MAC naturally provides the sum of the updates from the devices to the PS [1], [10]–[14], [17]–[19], [35]–[37].”
BackgroundConvergence 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 109
“FEEL over a broadband wireless fading MAC is studied in [14], where the devices have channel state information (CSI) to perform channel inversion, while [15] proposes analog transmission over the wireless fading MAC without any power control.”
BackgroundTime-Correlated Sparsification for Communication-Efficient Federated Learning
Özfatura, E., Özfatura, K., & Gündüz, D. (2021) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit45174.2021.9518221 · Cited by 81
Cao, X., Yang, B. F., Wang, K., Li, X., Yu, Z., Yuen, C., Zhang, Y., & Han, Z. (2024) · Proceedings of the IEEE · doi:10.1109/jproc.2024.3417332 · Cited by 84
“A large number of works applying ML techniques to the design of centralized-based OMA protocols have been studied [56]–[58], [66], [67], [69], [72], [74], [171], [174]–[176], [180], [181], [188], [190], [191], [248].”
BackgroundMethodWireless Federated Learning Over Resource-Constrained Networks: Digital Versus Analog Transmissions
Yao, J., Xu, W., Yang, Z., You, X., Bennis, M., & Poor, H. V. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3407822 · Cited by 73
“It was shown in [27] that the convergence rate of centralized learning remains approach-able with this analog approach without power control and beamforming.”
BackgroundOver-the-air federated learning: Status quo, open challenges, and future directions
Xiao, B., Yu, X., Ni, W., Wang, X., & Poor, H. V. (2024) · Fundamental Research · doi:10.1016/j.fmre.2024.01.011 · Cited by 73
“Based on the findings of Sery et al. [18] , this optimization problem can be translated to essentially construct an effi-cient channel with a large mean 𝜇 ℎ 𝐸 and a small variance 𝜎 2 ℎ 𝐸 .”
InfluentialCombating Interference for Over-the-Air Federated Learning: A Statistical Approach via RIS
Shi, W., Yao, J., Xu, W., Xu, J., You, X., Eldar, Y. C., & Zhao, C. (2025) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2025.3536023 · Cited by 73
“Numerical analyses in [5], [14] show that the derived upper bound is not strictly tight and can exhibit notable deviation from the true loss function, posing challenges in ensuring optimal convergence performance.”
Over-the-Air Decentralized Federated Learning
Shi, Y., Zhou, Y., & Shi, Y. (2021) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit45174.2021.9517780 · Cited by 67
Revisiting Analog Over-the-Air Machine Learning: The Blessing and Curse of Interference
Yang, H. H., Chen, Z., Quek, T. Q. S., & Poor, H. V. (2021) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2021.3139231 · Cited by 62
Privacy Amplification for Federated Learning via User Sampling and Wireless Aggregation
Seif, M., Chang, W.-T., & Tandon, R. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118408 · Cited by 60
Wireless for Machine Learning: A Survey
Hellström, H., da Silva, J. M. B., Amiri, M. M., Chen, M., Fodor, V., Poor, H. V., & Fischione, C. (2022) · Foundations and Trends® in Signal Processing · doi:10.1561/2000000114 · Cited by 59
Wireless for Machine Learning: A Survey
Hellström, H., da Silva, J. M. B., Amiri, M. M., Chen, M., Fodor, V., Poor, H. V., & Fischione, C. (2022) · doi:10.1561/9781638280071 · Cited by 58
Over-the-Air Federated Learning and Optimization
Zhu, J., Shi, Y., Zhou, Y., Jiang, C., Chen, W., & Letaief, K. B. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3352280 · Cited by 57
“As in most of the existing literature on AirComp [27], [56], [60], we assume that perfect channel state information (CSI) is available at each edge device and the edge sever.”
BackgroundMethodInfluentialOver-the-Air Aggregation for Federated Learning: Waveform Superposition and Prototype Validation
Guo, H., Zhu, Y., Ma, H., Lau, V. K. N., Huang, K., Li, X., Nong, H., & Zhou, M. (2021) · Journal of Communications and Information Networks · doi:10.23919/jcin.2021.9663107 · Cited by 56
“The authors in [26], [27] prove that the training may converge even without transmit power control and user scheduling by designing a dynamic learning rate.”
Communication-Efficient Federated Edge Learning via Optimal Probabilistic Device Scheduling
Zhang, M., Zhu, G., Wang, S., Jiang, J., Liao, Q., Zhong, C., & Cui, S. R. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3166941 · Cited by 55
“Practical AirComp solutions have been developed for various system setting, including narrow band system [10], [11], broadband system [12], multi-antenna system [13], [14], misaligned system [15], system with digital modulation [16], system without CSI support [17] and the system with heterogeneity…”
BackgroundFederated Learning over Wireless Networks: A Band-limited Coordinated Descent Approach
Zhang, J., Li, N., & Dedeoğlu, M. (2021) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom42981.2021.9488818 · Cited by 53
Differentially Private Federated Learning via Reconfigurable Intelligent Surface
Yang, Y., Zhou, Y., Wu, Y., & Shi, Y. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3168066 · Cited by 53
Communication Efficient Federated Learning over Multiple Access Channels
Chang, W.-T., & Tandon, R. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2001.08737 · Cited by 43
Privacy-Enhanced Decentralized Federated Learning at Dynamic Edge
Chen, S., Wang, Y., Yu, D., Ren, J., Xu, C., & Zheng, Y. (2023) · IEEE Transactions on Computers · doi:10.1109/tc.2023.3239542 · Cited by 48
“In analog schemes [27], [28], [33], [34], [35], nodes transmit data directly over the air simultaneously.”
BackgroundDecentralized SGD with Over-the-Air Computation
Özfatura, E., Rini, S., & Gündüz, D. (2020) · Globecom · doi:10.1109/globecom42002.2020.9322286 · Cited by 48
“Indeed, in the context of PS-based FL, it has been shown that, rather than trying to minimize interference, the signal superposition property of the wireless medium can be harnessed to achieve over-the-air computation (OAC) [4], [5], [10]–[12], which in turn, can reduce the communication requirements and increase the…”
BackgroundDistributed Intelligence in Wireless Networks
Liu, X., Yu, J., Liu, Y., Gao, Y., Mahmoodi, T., Lambotharan, S., & Tsang, D. H. K. (2023) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2023.3265425 · Cited by 48
“To realize efficient uplink model aggregation in FL, an analog AirComp was proposed as a communication and computation co-design approach by exploiting the additive nature of the wireless multiple access channels [161], [174].”
MethodOver-the-Air Federated Learning via Second-Order Optimization
Yang, P., Jiang, Y., Wang, T., Zhou, Y., Shi, Y., & Jones, C. N. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3185156 · Cited by 41
“Since only one global parameter vector needs to be broadcasted, the total communication cost of this step is negligible [25], [27], [31], [47].”
BackgroundMethodOver-The-Air Federated Learning Over Scalable Cell-Free Massive MIMO
Sifaou, H., & Li, G. Y. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3315962 · Cited by 41
“Significant efforts have been recently dedicated for the design and performance analysis of FL over wireless networks [4], [5], [6], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], referred to as federated edge earning (FEEL).”
BackgroundOver-the-Air Federated Learning with Joint Adaptive Computation and Power Control
Yang, H., Qiu, P., Jia, L., & Yener, A. (2022) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit50566.2022.9834342 · Cited by 38
Federated Learning Over-the-Air by Retransmissions
Hellström, H., Fodor, V., & Fischione, C. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3268742 · Cited by 38
“The first baseline (max power) is based on the scheme proposed in [20].”
BackgroundMethodResultInfluentialLi, 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
“, analog gradient methods [11], quantization schemes [12], and gradient sparsification approaches [13].”
Device Scheduling in Over-the-Air Federated Learning Via Matching Pursuit
Bereyhi, A., Vagollari, A., Asaad, S., Müller, R. R., Gerstacker, W., & Poor, H. V. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3284376 · Cited by 37
“In [62], the authors propose an algorithm for OTA-FL via AirComp, considering a risk-”
MethodAccelerated Gradient Descent Learning Over Multiple Access Fading Channels
Paul, R., Friedman, Y., & Cohen, K. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118410 · Cited by 36
Self-citationDistributed Learning Over Markovian Fading Channels for Stable Spectrum Access
Gafni, T., & Cohen, K. (2022) · IEEE Access · doi:10.1109/access.2022.3171666 · Cited by 35
“Model-free learning strategies were developed in [36], [37] for orthogonal channels, compact models [38], and multiple access channel strategies were developed in [39], [40].”
MethodSelf-citationTemporal-Structure-Assisted Gradient Aggregation for Over-the-Air Federated Edge Learning
Fan, D., Yuan, X., & Zhang, Y. A. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118348 · Cited by 34
COTAF: Convergent Over-the-Air Federated Learning
Sery, T., Shlezinger, N., Cohen, K., & Eldar, Y. C. (2020) · Globecom · doi:10.1109/globecom42002.2020.9322580 · Cited by 34
“As in [15], [17], [20], [21], we utilize analog signalling, namely, each vector x nt consists of continuous-amplitude quantities, rather than a set of discrete symbols or bits, as common in digital communications.”
BackgroundMethodInfluentialSelf-citationFast 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].”
BackgroundMethodInfluentialOptimizing Over-the-Air Computation in IRS-Aided C-RAN Systems
Yu, D., Park, S.-H., Simeone, O., & Shamai, S. S. (2020) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc48557.2020.9154243 · Cited by 33
Client Selection for Generalization in Accelerated Federated Learning: A Multi-Armed Bandit Approach
Ben-Ami, D., Cohen, K., & Zhao, Q. Q. (2025) · IEEE Access · doi:10.1109/access.2025.3543441 · Cited by 33
“Ashortversion of this paper was accepted for presentation in the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2023 [1].”
Self-citationOver-the-Air Federated Learning with Retransmissions
Hellström, H., Fodor, V., & Fischione, C. (2021) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51858.2021.9593119 · Cited by 32
FedOComp: Two-Timescale Online Gradient Compression for Over-the-Air Federated Learning
Xue, Y., Su, L., & Lau, V. K. N. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3165268 · Cited by 32
“Specifically, analog pulse amplitude modulation (PAM) directly modulates the amplitude of the uncoded transmitted vectors [30], while digital modulation can also be applied after the proper quantization of the transmitted vector s tk , as in [11].”
BackgroundCHARLES: Channel-Quality-Adaptive Over-the-Air Federated Learning over Wireless Networks
Mao, J., Yang, H., Qiu, P., Liu, J., & Yener, A. (2022) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51304.2022.9833989 · Cited by 32
Mathew, C., & Asha, D. (2024) · Journal of Internet Services and Information Security · doi:10.58346/jisis.2024.i3.021 · Cited by 32
Byzantine-Resilient Federated Machine Learning via Over-the-Air Computation
Huang, S., Zhou, Y., Wang, T., & Shi, Y. (2021) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops50388.2021.9473694 · Cited by 31
“As in most existing works on AirComp [8], [15], [43], [44],”
Over-the-Air Computation in Correlated Channels
Frey, M., Bjelaković, I., & Stańczak, S. (2021) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2021.3106115 · Cited by 30
Soret, B., Nguyen, L. D., Seeger, J., Bröring, A., Issaid, C. B., Samarakoon, S., Elgabli, A., Kulkarni, V., Bennis, M., & Popovski, P. (2021) · IEEE Transactions on Green Communications and Networking · doi:10.1109/tgcn.2021.3138792 · Cited by 29
Non-Coherent Over-the-Air Decentralized Gradient Descent
Michelusi, N. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3460690 · Cited by 28
“The paper [38] focuses on over-the-air computation in the FL setting.”
BackgroundCommunication-Efficient Split Learning Based on Analog Communication and Over the Air Aggregation
Krouka, M., Elgabli, A., Issaid, C. B., & Bennis, M. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685045 · Cited by 28
“A novel Gradient-Based Multiple Access (GBMA) algorithm was proposed to solve the distributed learning problem over MAC in [11], where the nodes use common shaping waveforms to transmit an analog function of the local gradient.”
MethodFederated learning with stochastic quantization
Li, Y., Li, W., & Xue, Z. (2022) · International Journal of Intelligent Systems · doi:10.1002/int.23056 · Cited by 27
Robust Federated Learning via Over-the-Air Computation
Sifaou, H., & Li, G. Y. (2022) · Machine learning for signal processing ... · doi:10.1109/mlsp55214.2022.9943401 · Cited by 27
“As in [18, 19, 21, 24, 25], we assume that perfect channel state information (CSI) is available at the clients and the CS.”
BackgroundWireless Federated Langevin Monte Carlo: Repurposing Channel Noise for Bayesian Sampling and Privacy
Liu, D., & Simeone, O. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3215663 · Cited by 26
“Each block contains m channel uses, allowing the uncoded transmission of a gradient vector via non-orthogonal multiple access (NOMA) as in [6] and [20].”
MethodCommunication-Efficient and Distributed Learning Over Wireless Networks: Principles and Applications
Park, J., Samarakoon, S., Elgabli, A., Kim, J., Bennis, M., Kim, S.-L., & Debbah, M. (2021) · Proceedings of the IEEE · doi:10.1109/jproc.2021.3055679 · Cited by 25
“For this reason, it is common to allow transmissions only when the channel gains exceed a certain threshold [109]–[111].”
BackgroundDistributed Constrained Online Convex Optimization Over Multiple Access Fading Channels
Cao, X., & Başar, T. (2022) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2022.3185897 · Cited by 25
Multi-Slot Over-the-Air Computation in Fading Channels
Tang, S., Popovski, P., Zhang, C., & Obana, S. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3245304 · Cited by 25
“The authors in [24] investigated direct model update based on the noisy distorted gradient.”
BackgroundAnomaly Search With Multiple Plays Under Delay and Switching Costs
Lambez, T., & Cohen, K. (2021) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2021.3136810 · Cited by 24
“Modern application areas of search problems with limited sensing resources include narrowband spectrum scanning [27], event detection by a fusion center that communicates with sensors using narrowband transmissions [28]–[30], and sensor visual search studied recently by neuroscientists [31].”
BackgroundSelf-citationMultiple Parallel Federated Learning via Over-the-Air Computation
Shi, G., Guo, S., Ye, J., Saeed, N., & Dang, S. (2022) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2022.3194821 · Cited by 24
“The tractable FL convergence analysis of full gradient descent optimization was done in [34], [35].”
MethodReconfigurable Intelligent Surface-Assisted Wireless Federated Learning With Imperfect Aggregation
Sun, P., Liu, E., Ni, W., Wang, R., Xing, Z., Li, B., & Jamalipour, A. (2024) · IEEE Transactions on Communications · doi:10.1109/tcomm.2024.3450605 · Cited by 24
“Furthermore, algorithms developed based on these assumptions are prevalent in the literature due to their tractability, e.g., [14], [34], and [35], and experimentally validated on ML models with non-convex loss functions [35], [36].”
Random Orthogonalization for Federated Learning in Massive MIMO Systems
Wei, X., Shen, C., Yang, J., & Poor, H. V. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3302335 · Cited by 22
“The assumption of full CSIT is relaxed in [32] by only using the phase information of each individual channel.”
MethodAnalog Compression and Communication for Federated Learning over Wireless MAC
Abdi, A., Saidutta, Y. M., & Fekri, F. (2020) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc48557.2020.9154309 · Cited by 22
Over-the-Air Federated Learning via Weighted Aggregation
Azimi‐Abarghouyi, S. M., & Tassiulas, L. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3463754 · Cited by 21
“There is no prior knowledge of local data statistics of the devices, consistent with [9]–[19], [23]–[30].”
BackgroundMethodResultInfluentialDong, Y., Wang, L., Wang, J., Hu, X., Zhang, H., Yu, F. R., & Leung, V. C. M. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3329375 · Cited by 20
“In this vein, the analog WFL algorithms were proposed to adapt the information exchange to the AWGN channels [13], [26] and fading channels [27], [28], [29], [30], [31], [32], [33], [34], [35].”
BackgroundMethodInfluentialPrivate Over-the-Air Federated Learning at Band-Limited Edge
Tao, Y., Chen, S., Zhang, C., Wang, D., Yu, D., Cheng, X., & Dressler, F. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3411295 · Cited by 20
“As a result, federated learning based on AirComp, i.e., over-the-air federated learning (OTA-FL), has recently gained considerable interest in edge intelligence research [13], [14], [15], [16], [17], [18].”
BackgroundOptimal Adaptive Power Control for Over-The-Air Federated Edge Learning Under Fading Channels
Yu, X., Xiao, B., Ni, W., & Wang, X. (2023) · IEEE Transactions on Communications · doi:10.1109/tcomm.2023.3283790 · Cited by 19
“• Fixed-power allocation [5]: The edge devices scale the gradient with a fixed factor ρn,t = √ P0,∀n, at any round.”
BackgroundUnleashing Edgeless Federated Learning With Analog Transmissions
Yang, H. H., Chen, Z., & Quek, T. Q. S. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3352405 · Cited by 19
“As a result, the system not only scales easily but also attains significant energy saving [11] (as demonstrated in [18], the energy cost can be reduced by two orders of magnitude compared to digital communication-based methods).”
BackgroundMethodInfluentialRiemannian Low-Rank Model Compression for Federated Learning With Over-the-Air Aggregation
Xue, Y., & Lau, V. K. N. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3284381 · Cited by 18
“For the downlink broadcast from the central server to the local devices, we assume the transmission is error-free since the server has higher transmit power with forward error correction compared to the uplink access [19, 20].”
BackgroundMethodInfluentialOver-the-Air Federated Learning Exploiting Channel Perturbation
Hamidi, S. M., Mehrabi, M., Khandani, A. K., & Gündüz, D. (2022) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51304.2022.9833972 · Cited by 17
Over-the-Air Statistical Estimation
Lee, C.-Z., Barnes, L. P., & Özgür, A. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118412 · Cited by 15
Federated Learning from Heterogeneous Data via Controlled Air Aggregation with Bayesian Estimation
Gafni, T., Cohen, K., & Eldar, Y. C. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3351469 · Cited by 15
“Additional related works on OTA FL, including [22], [23], [24], [25], considered the distributed application of full gradient descent optimization over noisy channels.”
BackgroundSelf-citationBeamforming and Device Selection Design in Federated Learning With Over-the-Air Aggregation
Kalarde, F. M., Dong, M., Liang, B., Ahmed, Y., & Cheng, H. T. (2024) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2024.3372893 · Cited by 15
“Such over-the-air computation has attracted growing interest due to its advantages of efficient utilization of bandwidth and reduced communication latency over the conventional approach of orthogonal multiple access [4], [5], [6], [7], [8], [9], [10].”
BackgroundFederated Learning with Downlink Device Selection
Amiri, M. M., Kulkarni, S. R., & Poor, H. V. (2021) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51858.2021.9593132 · Cited by 16
Harnessing wireless channels for scalable and privacy-preserving federated learning
Elgabli, A., Park, J., Issaid, C. B., & Bennis, M. (2021) · IEEE Transactions on Communications · doi:10.1109/tcomm.2021.3078783 · Cited by 14
Decentralized Over-the-Air Federated Learning by Second-Order Optimization Method
Yang, P., Jiang, Y., Wen, D., Wang, T., Jones, C. N., & Shi, Y. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3327610 · Cited by 14
“In order to further mitigate communication overhead and improve spectral efficiency in this step, we adopt the over-the-air computation (AirComp) technique [48], [49] that is widely used in wireless FL studies [39], [40], [41], [42], [43], [44].”
BackgroundMethodPrivacy Amplification for Federated Learning via User Sampling and Wireless Aggregation
Seif, M., Chang, W.-T., & Tandon, R. (2021) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit45174.2021.9518031 · Cited by 14
Random Orthogonalization for Federated Learning in Massive MIMO Systems
Wei, X., Shen, C., Yang, J., & Poor, H. V. (2022) · International Conference on Communications · doi:10.1109/icc45855.2022.9838988 · Cited by 13
“The assumption of full CSIT is relaxed in [33] by only using the phase information of each individual channel.”
MethodOnline Model Updating with Analog Aggregation in Wireless Edge Learning
Wang, J., Dong, M., Liang, B., Boudreau, G., & Abou-Zeid, H. (2022) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom48880.2022.9796860 · Cited by 13
“This regularization scheme was adopted in [22]-[24].”
BackgroundHong, J., Park, S., & Choi, W. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3249252 · Cited by 13
“…strategies have explored the channel inversion power control to cancel out the effect of channel fading, and they have shown to achieve a convergence to the optimal model parameters under some constraints on the training model and the simplified channel model without pathloss [18], [20], [21].”
BackgroundPersonalizing Federated Learning with Over-The-Air Computations
Chen, Z., Li, Z., Yang, H. H., & Quek, T. Q. S. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10095533 · Cited by 13
“Nevertheless, A-OTA computations inevitably introduce the random channel fading and interference into the aggregated gradients, leading to performance degradations such as the slower convergence and instability [5,10].”
BackgroundOver-the-Air Computing for 6G - Turning Air into a Computer.
Zhu, G., Xu, J., & Huang, K. (2020) · Cited by 12
Joint Online Optimization of Model Training and Analog Aggregation for Wireless Edge Learning
Wang, J., Liang, B., Dong, M., Boudreau, G., & Abou-Zeid, H. (2023) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2023.3318474 · Cited by 12
“Restrictions apply. focused on analog gradient aggregation in FL [21], [22], [23], [24], [25], [26], [27].”
BackgroundMethodInfluentialOver-The-Air Computation in Correlated Channels
Frey, M., Bjelakovic, I., & Stanczak, S. (2021) · IEEE Information Theory Workshop (ITW) · doi:10.1109/itw46852.2021.9457577 · Cited by 12
Latency Minimization for Wireless Federated Learning With Heterogeneous Local Model Updates
Zhu, J., Shi, Y., Fu, M., Zhou, Y., Wu, Y., & Fu, L. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3285937 · Cited by 12
“Therefore, each edge device can be assumed to receive the current shared global model θ ( t + 1 , 0 ) without distortion [61], [62], [63].”
BackgroundA Sequential Gradient-Based Multiple Access for Distributed Learning over Fading Channels
Sery, T., & Cohen, K. (2019) · Allerton Conference on Communication, Control, and Computing · doi:10.1109/allerton.2019.8919883 · Cited by 12
“In the extended version of this paper [44] we address these issues.”
BackgroundMethodInfluentialSelf-citationOver-the-Air Multitask Federated Learning via Cost-Effective Hybrid Beamforming
Yao, J., Xu, W., Zhu, G., Yang, Z., Huang, K., & Niyato, D. T. (2025) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2025.3589658 · Cited by 11
“Therefore, applying AirComp to FL training process can deeply integrate uplink gradient transmission and model aggregation, paving a way for communication-efficient FL over resource-constrained wireless networks [10], [11], [12].”
Communication and Energy Efficient Wireless Federated Learning With Intrinsic Privacy
Zhang, Z., Guo, Y., Fang, Y. M., & Gong, Y. (2023) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2023.3341788 · Cited by 11
Federated Learning via Lattice Joint Source-Channel Coding
Azimi‐Abarghouyi, S. M., & Varshney, L. R. (2024) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit57864.2024.10619502 · Cited by 11
“Lastly, approaches that adjust for the channel at the transmitter can run into channel estimation errors, causing the values to be multiplied by unpredictable gains when received [16].”
BackgroundCompute-Update Federated Learning: A Lattice Coding Approach
Azimi‐Abarghouyi, S. M., & Varshney, L. R. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3491993 · Cited by 11
“It differes from [2], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [32] which use predefined weights in FL, such as equal or proportional to local dataset sizes.”
InfluentialAdaptive Federated Learning Over the Air
Wang, C., Chen, Z., Παππάς, Ν., Yang, H. H., Quek, T. Q. S., & Poor, H. V. (2025) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2025.3585002 · Cited by 10
“…sum of clients’ uploaded parameters, rather than requesting the precise value of each client’s parameter, a line of recent studies [8], [14], [15], [16], [27], [28], [29] proposed integrating OTA computations with FL, arriving at a low-latency multi-access edge learning scheme (which is commonly…”
InfluentialPower Minimization in Federated Learning with Over-the-air Aggregation and Receiver Beamforming
Kalarde, F. M., Liang, B., Dong, M., Ahmed, Y., & Cheng, H. T. (2023) · ACM International Conference on Modeling, Analysis and Simulation of Wireless and Mobile Systems (MSWiM) · doi:10.1145/3616388.3617534 · Cited by 10
Zeroth-Order Over-the-Air Federated Large Model Tuning Over Edge Networks
Chen, Z., Yang, H. H., Li, Z., Park, J., & Quek, T. Q. S. (2025) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2025.3585344 · Cited by 10
“Although these methods could effectively reduce the overall communication overhead, but still require computation of the full or partial first-order (FO) gradients for iterative model update [7], with excessive computational cost for model tuning.”
Compressive Sensing-Based Power Allocation Optimization for Energy Harvesting IoT Nodes
Zhang, J., Xie, G., Han, G., Yu, Z. L., Gu, Z., & Li, Y. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3131159 · Cited by 10
Zenati, D., Maimon, T., & Cohen, K. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3528813 · Cited by 9
“…spectral resources and manage data transmissions effectively, has become a critical challenge in modern communication networks, such as 5G and beyond wireless and mobile networks [1]–[4], cognitive radio networks [2], [5]–[8], distributed [9], [10] and federated learning systems [11], [12].”
BackgroundSelf-citationFree Privacy Protection for Wireless Federated Learning: Enjoy It or Suffer From It?
Li, W., Lv, T., Zhao, X., Yuan, X., & Ni, W. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3581042 · Cited by 9
“Operating in the analog domain, OTA-FL allows the participating clients to upload their models within the same resource block [17], [25], [26].”
Metadata
- Authors
- Sery, T., & Cohen, K.
- Affiliations
- Ben-Gurion University of the Negev (IL)
- Venue
- IEEE Transactions on Signal Processing
- Published
- 2020
- Volume
- 68
- Pages
- 2897-2911
- ISSN
- 1053-587X (ISSN-L)
- 1941-0476
- OpenAlex
- W2969343398
- Cited by
- 172
- References
- 64
- Citations by year
- 2026: 8 · 2025: 24 · 2024: 29 · 2023: 29 · 2022: 28 · 2021: 37 · 2020: 15
- FWCI
- 12.81
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Signal Processing (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- United States-Israel Binational Science Foundation (2017723)
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed Sensor Networks and Detection Algorithms
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Artificial Intelligence › Stochastic Gradient Optimization Techniques
- Keywords
- multiple access channel, fading channels, distributed learning, federated learning, noisy gradients, convergence rate, energy scaling law, finite-sample error bounds, large-scale wireless networks
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.