One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning: Design and Convergence Analysis
Open accessThe DOI matches its Crossref record; 35 of the 37 references it cites were found in a registry or an index.
Abstract
Federated edge learning (FEEL) is a popular framework for model training at an edge server using data distributed at edge devices (e.g., smart-phones and sensors) without compromising their privacy. In the FEEL framework, edge devices periodically transmit high-dimensional stochastic gradients to the edge server, where these gradients are aggregated and used to update a global model. When the edge devices share the same communication medium, the multiple access channel (MAC) from the devices to the edge server induces a communication bottleneck. To overcome this bottleneck, an efficient broadband analog transmission scheme has been recently proposed, featuring the aggregation of analog modulated gradients (or local models) via the waveform-superposition property of the wireless medium. However, the assumed linear analog modulation makes it difficult to deploy this technique in modern wireless systems that exclusively use digital modulation. To address this issue, we propose in this work a novel digital version of broadband over-the-air aggregation, called one-bit broadband digital aggregation (OBDA). The new scheme features one-bit gradient quantization followed by digital quadrature amplitude modulation (QAM) at edge devices and over-the-air majority-voting based decoding at edge server. We provide a comprehensive analysis of the effects of wireless channel hostilities (channel noise, fading, and channel estimation errors) on the convergence rate of the proposed FEEL scheme. The analysis shows that the hostilities slow down the convergence of the learning process by introducing a scaling factor and a bias term into the gradient norm. However, we show that all the negative effects vanish as the number of participating devices grows, but at a different rate for each type of channel hostility.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The text we found for this paper looks like another document, so the self-check wasn’t run.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474
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
A Survey on Mobile Edge Computing: The Communication Perspective
Mao, Y., You, C., Zhang, J., Huang, K., & Letaief, K. B. (2017) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2017.2745201 · Cited by 5,607 · cited 1×
“(Aggregated quantized local gradient estimates) g̃ = [ g̃[1] , g̃[2] , .”
Self-citationFederated 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
A Joint Learning and Communications Framework for Federated Learning over Wireless Networks
Chen, M., Yang, Z., Saad, W., Yin, C., Poor, H. V., & Cui, S. (2019) · arXiv · doi:10.48550/arxiv.1909.07972 · Cited by 1,547 · cited 1×
“In [21], a novel bandwidth allocation strategy is proposed for minimizing the model training loss of FEEL via convergence analysis accounting for packet transmission errors.”
MethodA Joint Learning and Communications Framework for Federated Learning Over Wireless Networks
Chen, M., Yang, Z., Saad, W., Yin, C., Poor, H. V., & Cui, S. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2020.3024629 · Cited by 1,547 · cited 1×
“In [21], a novel bandwidth allocation strategy is proposed for minimizing the model training loss of FEEL via convergence analysis accounting for packet transmission errors.”
MethodDeep 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 · cited 1×
“The last approach focuses on update compression by exploiting the sparsity of gradient updates [12], [13] and low-resolution gradient/model-parameter quantization [14]–[16].”
BackgroundFederated 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 4×
“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.”
BackgroundMethodInfluentialQSGD: Communication-Efficient SGD via Gradient Quantization and Encoding
Alistarh, D., Grubic, D., Li, J., Tomioka, R., & Vojnović, M. (2016) · London School of Economics and Political Science Research Online (London School of Economics and Political Science) · doi:10.48550/arxiv.1610.02132 · Cited by 859
Broadband Analog Aggregation for Low-Latency Federated Edge Learning
Zhu, G., Wang, Y., & Huang, K. (2020) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2019.2946245 · Cited by 761 · cited 8×
“A performance comparison between digital transmission using OFDMA, BAA developed in [6], and the proposed OBDA is presented in Fig.”
BackgroundMethodInfluentialSelf-citation
Research works printed but not linked: 2
- arunabha, 2010, Fundamentals of LTE
- 2010, Timing Advance (TA) in LTE
Cited by
Items 201–300 of 435
One-Bit Aggregation for Over-the-Air Federated Learning Against Byzantine Attacks
Miao, Y., Ni, W., & Tian, H. (2024) · IEEE Signal Processing Letters · doi:10.1109/lsp.2024.3384077 · Cited by 13
“Specifically, we model the impact of Byzantine attackers transmitting inverse gradient vectors, a scenario absent in [21].”
BackgroundMethodInfluentialReMAC: Digital Multiple Access Computing by Repeated Transmissions
Yan, X., Razavikia, S., & Fischione, C. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3565608 · Cited by 13
“One of the earliest attempts involves implementing signSGD for federated learning problems [29].”
FedMeld: A Model-Dispersal Federated Learning Framework for Space-Ground Integrated Networks
Chen, Q., Chen, X., & Huang, K. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3647858 · Cited by 13
Self-citationRandom 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
Yan, Z., Li, D., Zhang, Z., & He, J. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3277632 · Cited by 13
“The effect of uplink channel noise on model convergence was studied in [26] and [27].”
BackgroundBlind Federated Learning via Over-the-Air q-QAM
Razavikia, S., da Silva, J. M. B., & Fischione, C. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3485117 · Cited by 13
“First, we need to present our assumptions on the loss function and the gradients, which are standard in the literature [19], [22], and [23].”
Robust federated learning for edge-intelligent networks
Gao, Z., Chen, X., & Shao, X. (2022) · Science China Information Sciences · doi:10.1007/s11432-020-3251-9 · Cited by 13
Hong, J., Park, S., & Choi, W. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3249252 · Cited by 13
“We consider the training of 6-layer convolutional neural network (CNN) which has been adopted in the relevant works [16], [26].”
BackgroundMethodOver-the-Air Computing for 6G - Turning Air into a Computer.
Zhu, G., Xu, J., & Huang, K. (2020) · Cited by 12
Self-citationDistributed Average Consensus via Noisy and Noncoherent Over-the-Air Aggregation
Yang, H., Chen, X., Huang, L., Dey, S., & Shi, L. (2024) · IEEE Transactions on Control of Network Systems · doi:10.1109/tcns.2024.3432827 · Cited by 12
“Under federated settings, over-the-air aggregation has been widely utilized and studied [22], [23], [24], [25], [26].”
Ai, Y., Chen, Q., Zhu, G., Wen, D., Jiang, H., Zeng, J., & Li, M. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3544318 · Cited by 12
“2) Communication Efficient FEL: The above works primarily focused on addressing the non-IID distribution issues without considering the communication efficient issue, which is essential in FEL mechanisms [17], [18], [19], [20], [21], [22].”
Self-citationJiang, W. P., Xiao, Z., Qian, L., Qin, S., Feng, G., & Wu, Y. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3568875 · Cited by 12
“In [49], [50], [51], the over-the-air computing methods were proposed to .”
Keçeci, C., Shaqfeh, M. O., Al‐Qahtani, F. S., Ismail, M., & Serpedin, E. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3262620 · Cited by 12
Ni, W., Liu, Y., Tian, H., Eldar, Y. C., & Huang, K. (2022) · International Conference on Communications · doi:10.1109/icc45855.2022.9838989 · Cited by 12
“Note that the above assumptions are standard in the convergence analysis of ML [3]–[5], [9]–[11].”
BackgroundSelf-citationPartitioned Edge Learning Over Fast Fading Channels
Jiang, Z., Wen, D., Liu, S., Zhu, G., & Yu, G. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3540281 · Cited by 11
“…caused by the frequent transmission of the high-dimensional AI model between the edge server and edge devices via resource allocation (e.g., [11], [12], [13], [14], [15]), device scheduling (e.g., [16], [17], [18], [19]), over-the-air computation (e.g., [20], [21], [22], [23], [24]), etc.”
Self-citationOver-the-Air ODE-Inspired Neural Network for Dual Task-Oriented Semantic Communications
Liu, M., An, J., Huang, C., & Yuen, C. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3569604 · Cited by 13
“…for fast global model aggregation via exploring the superposition property of the wireless MACs, including optimizing device selection and beamforming for enhanced participation [8], accelerating the convergence rate [9], [10], and improving both training accuracy and convergence speed [11].”
Over-the-Air Federated Learning With Joint Privacy-Accuracy Optimization
Feng, H., Wang, R., Liu, E., Ni, W., Niyato, D. T., & Jamalipour, A. (2025) · IEEE Transactions on Information Forensics and Security · doi:10.1109/tifs.2025.3620108 · Cited by 11
“This method was later extended to one-bit AirComp in [20].”
Quick and Reliable LoRa Data Aggregation Through Multi-Packet Reception
You, L., Tang, Z., Wang, P., Wang, Z., Dai, H., & Fu, L. (2023) · IEEE/ACM Transactions on Networking · doi:10.1109/tnet.2023.3323018 · Cited by 11
“Recently, references [32], [33], and [34] considers OFDM-based AirComp for federated edge learning.”
BackgroundPersonalized Over-The-Air Federated Learning with Personalized Reconfigurable Intelligent Surfaces
Mao, J., & Yener, A. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10446662 · Cited by 11
“The convergence upper bound is finite: as the channel estimation error is typically small, we adopt the Taylor expansion similar to [20].”
MethodFederated 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
“Unlike over-the-air FL schemes in [10]–[14] that depend on CSIT, FedCPU operates with constant power and does not rely on prior information.”
BackgroundMethodInfluentialCompute-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.”
InfluentialChen, Q., Cheng, H., Liang, Y., Zhu, G., Li, M., & Jiang, H. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3499375 · Cited by 11
“However, the limited computation, communication, and memory capabilities of edge devices are usually contradicted by the training model development trend [6], [7].”
Self-citationToward Optimal Rate-Delay Tradeoff for Computation Over Multiple Access Channel
Chen, L., Zhao, N., Chen, Y., Yu, F. R., & Wei, G. (2021) · IEEE Transactions on Communications · doi:10.1109/tcomm.2021.3068493 · Cited by 11
DCT-Based Air Interface Design for Function Computation
Martinez-Gost, M., Pérez-Neira, A. I., & Lagunas, M. A. (2023) · IEEE Open Journal of Signal Processing · doi:10.1109/ojsp.2023.3241584 · Cited by 11
On the Design of Communication Efficient Federated Learning over Wireless Networks.Preprint
Jin, R., He, X., & Dai, H. (2020) · arXiv (Cornell University) · Cited by 11
One-Bit Over-the-Air Aggregation for Communication-Efficient Federated Edge Learning
Zhu, G., Du, Y., Gündüz, D., & Huang, K. (2020) · Globecom · doi:10.1109/globecom42002.2020.9322334 · Cited by 10
“Proof: See Appendix A in the extended version [13].”
BackgroundSelf-citationBest Effort Voting Power Control for Byzantine-resilient Federated Learning Over the Air
Fan, X., Wang, Y., Huo, Y., & Tian, Z. (2022) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops53468.2022.9814495 · Cited by 10
“(8) Next, we provide two transmit power allocation schemes for normal local workers: the existing channel-inversion (CI) transmission [13], [16] and our proposed best effort voting (BEV) scheme.”
InfluentialPrivacy-Enhanced Over-the-Air Federated Learning via Client-Driven Power Balancing
Kim, B., Seo, H., & Choi, W. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3605478 · Cited by 9
Optimal Batch-Size Control for Low-Latency Federated Learning With Device Heterogeneity
Yang, H., Wang, Z., & Huang, K. (2026) · IEEE Transactions on Communications · doi:10.1109/tcomm.2026.3666674 · Cited by 9
“The convergence analysis builds upon the following assumptions, which are widely adopted in non-convex optimization for FL systems [18], [26], [35], [36], [37].”
Self-citationShahabodini, S., Mansoori, M., Abouei, J., & Plataniotis, K. N. (2023) · Transactions on Emerging Telecommunications Technologies · doi:10.1002/ett.4926 · Cited by 9
“For local dataset collection, CL is used which means all users must transmit their sample pairs to the BS. ) 20 : This model also uses a fully connected DNN to handle the channel estimation problem.”
BackgroundInfluentialQiao, L., Gao, Z., Li, Z., & Gündüz, D. (2023) · IEEE International Symposium on Information Theory (ISIT) · doi:10.1109/isit54713.2023.10206560 · Cited by 9
Integrating Sensing, Communication, and Computation in the Sky
Tang, Y., Zhu, G., Xu, W., Cheung, M. H., Lok, T.-M., & Cui, S. R. (2024) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp48485.2024.10446598 · Cited by 9
“We consider general smooth convex learning problems with the following commonly adopted assumptions [13–16]: 1) Smoothness: Each local loss function F k ( w ) is Lips-chitz continuous with constant L : ∀ w i and w j , 3) Unbiasedness and bounded variance: The mean and variance of local gradient g k…”
BackgroundSelf-citationAsynchronous Federated Learning Over Non-IID Data via Over-the-Air Computation
Ma, Q., Song, X., Zhou, J., Wang, H., Liao, Y., Liu, J., & Xu, H. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3641928 · Cited by 9
“To this end, over-the-air computation (AirComp) [13], a new analog non-orthogonal multiple access (NOMA) technique, is motivated to break through the limitation of communication resources, and reduce transmission delay for large-scale FL [14], [15], [16], [17], [18], [19].”
RIS-Assisted Over-the-Air Adaptive Federated Learning with Noisy Downlink
Mao, J., & Yener, A. (2023) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops57953.2023.10283543 · Cited by 8
“To bound the statistical error and channel estimation error types, we use the Taylor expansion to analyze the channel estimation error similar to [22]: h i t ĥt = 1 1+ ∆t ht = 1− ∆ i t ht +O(( i t ht )(2)).”
MethodSemi-Asynchronous Over-the-Air Federated Learning Over Heterogeneous Edge Devices
Kou, Z., Ji, Y., Yang, D., Zhang, S., & Zhong, X. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3466556 · Cited by 8
Frequency Modulation Aggregation for Federated Learning
Martinez-Gost, M., Pérez-Neira, A. I., & Lagunas, M. A. (2023) · Globecom · doi:10.1109/globecom54140.2023.10437413 · Cited by 8
Model Compression by Count Sketch for Over-the-Air Stateless Federated Learning
Ahn, J.-H., Kim, G.-Y., Kim, D. H., & You, C. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3376771 · Cited by 8
“Priorstudies [14], [15], [16], [17], [18], [19], [20], [21], [22] have established that AirComp FL surpasses the performance of FL based on traditional digital communication techniques (digital FL).”
BackgroundReliable Majority Vote Computation With Complementary Sequences for UAV Waypoint Flight Control
Şahin, A., & Wang, X. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3393759 · Cited by 8
“Note that a similar assumption is also made in [11], [12], and [24], where the time-variation in the channel is captured by the realizations of h k,i in (8).”
BackgroundFundamental Limits of Over-the-Air Optimization: Are Analog Schemes Optimal?
Jha, S. K., Mayekar, P., & Tyagi, H. (2022) · IEEE Journal on Selected Areas in Information Theory · doi:10.1109/jsait.2022.3177582 · Cited by 8
“In [23], authors perform one-bit quantization of subgradients similar”
BackgroundMethodResultInfluentialChirp-Based Majority Vote Computation for Federated Edge Learning and Distributed Localization
Hoque, S. S. M., & Şahin, A. (2023) · IEEE Open Journal of the Communications Society · doi:10.1109/ojcoms.2023.3271204 · Cited by 8
“The scheme of [3], [4], [8] experience high PMEPR as they are based on OFDM transmission.”
BackgroundMethodInfluentialTowards 6G Intelligence: The Role of Generative AI in Future Wireless NetworksPreprint
Mohsin, M. A., Ahmad, J., Nawaz, M. H., & Jamshed, M. A. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2508.19495 · Cited by 8
Digitalizing Over-the-Air Computation via the Novel Complement Coded Modulation
Wang, Z., Yao, J., Xu, W., Shi, W., & Huang, K. (2026) · IEEE Communications Letters · doi:10.1109/lcomm.2025.3650626 · Cited by 7
“For example, the authors of [8] proposed a binary phase shift keying (BPSK)-based one-bit digital aggregation scheme for majority voting, where the voting result was determined by the sign of the received signal.”
Self-citationBiased Federated Learning Under Wireless Heterogeneity
Abrar, M. F. U., & Michelusi, N. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2026.3678173 · Cited by 7
“…propagation conditions, and is assumed constant throughout FL runtime and known to the PS. 2 We emphasize that, unlike existing works [7], [13]– [15], [18] assuming identical average path loss across de-1 By the strong convexity of F and the optimality condition ∇ F ( w ∗ ) = 0 , it holds that…”
Joint Beamforming Design for Multifunctional RIS-Aided Over-the-Air Federated Learning
Zhang, X., Tian, H., Ni, W., & Yang, Z. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3548100 · Cited by 7
“The former assumes that the gradient noise, defined by the random discrepancy between the local gradient estimate and the global full gradient, has zero expectation, which typically holds when local data are independently and identically distributed (i.i.d.) [31].”
Model Pruning for Efficient Over-the-Air Federated Learning in Tactical Networks
Khan, F. M. A., Abou-Zeid, H., & Hassan, S. A. (2023) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops57953.2023.10283773 · Cited by 7
“As modern communication equipment uses digital modulation [15], we deploy orthogonal frequency division multiplexing (OFDM) by dividing the available bandwidth B into M orthogonal sub-channels which has also been utilized in [16, 17].”
MethodLiang, Y., Chen, Q., Zhu, G., & Jiang, H. (2023) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops57953.2023.10283720 · Cited by 7
“Besides, the authors in [11] implemented the over-the-air computation (AirComp) technique into the FEEL to accelerate the model aggregation.”
MethodSelf-citationPinching-antenna-enabled Federated Learning: Tail Latency, Participation, and Convergence AnalysisPreprint
Lin, Y., Chen, Z., & Ding, Z. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.23315 · Cited by 7
“For example, Zhu et al. [18] proposed a one-bit OTA-FL scheme where users quantize gradients to 1-bit and transmit over a common channel with digital modulation, using a majority-vote decoding at the server.”
One-Bit Byzantine-Tolerant Distributed Learning via Over-the-Air Computation
Yang, Y., Wu, Y., Jiang, Y., & Shi, Y. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3326438 · Cited by 6
“According to the additive structure of (4), the Majority Vote decoding scheme completely fits the category of AirComp [30], [31], [34], [39], which can be leveraged to support a communication-efficient Byzantine-tolerant distributed learning system.”
BackgroundMethodInfluentialWireless Federated k-Means Clustering with Non-coherent Over-the-Air Computation
Şahin, A. (2023) · IEEE Military Communications Conference (MILCOM) · doi:10.1109/milcom58377.2023.10356362 · Cited by 6
Over-the-Air Computation Enabled Semi-Asynchronous Wireless Federated Learning
Zheng, Z., Deng, Y., Yi, W., Shin, H., & Nallanathan, A. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3582727 · Cited by 6
“These affine constraints are introduced to maintain the stability of the model parameter scale during training, while also simplifying the convergence analysis.”
Bayesian AirComp with Sign-Alignment Precoding for Wireless Federated Learning
Park, C., Lee, S., & Lee, N. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685854 · Cited by 6
“The OBDA system has features of the truncated channel-inversion precoding and majority-votingbased aggregation at the server [8].”
MethodInfluentialLeveraging Channel Noise for Sampling and Privacy via Quantized Federated Langevin Monte Carlo
Zhang, Y., Liu, D., & Simeone, O. (2022) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51304.2022.9833991 · Cited by 6
Joint Power Control for Over-the-Air Computation in Multicarrier Wireless System
Xie, X. N., Hua, C., & Hong, J. (2022) · Globecom · doi:10.1109/globecom48099.2022.10000853 · Cited by 6
“In [9], [10], AirComp was extended for model aggregation in the distributed edge learning system, which is shown to reduce the communication latency without significant loss of the learning accuracy [11].”
MethodWang, Q., Cui, Y., Li, C., Zou, J., & Xiong, H. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3244084 · Cited by 6
“Recently, distributed learning (or federated learning) has been actively investigated [2], [3], [4], [5], [6].”
BackgroundRatioFi: Unlocking the Potential of WiFi CSI
Zhang, D., Cai, Z., Zhu, G., Li, H., Li, X., Shi, Q., & Shen, C. (2023) · International Conference on Ubiquitous Communication (Ucom) · doi:10.1109/ucom59132.2023.10257637 · Cited by 6
Self-citationHeterogeneous Resources in Infrastructures of the Edge Network Paradigm: A Comprehensive Review
Alsaffar, Q. S., & Ayed, L. J. B. (2024) · Karbala International Journal of Modern Science · doi:10.33640/2405-609x.3360 · Cited by 6
Over-the-Air Computing in OFDM Systems
Evgenidis, N. G., Tegos, S. A., Diamantoulakis, P. D., & Karagiannidis, G. K. (2024) · IEEE Communications Letters · doi:10.1109/lcomm.2024.3446174 · Cited by 6
“The distributed nature of OTA computing has also attracted considerable attention as a means to enable distributed learning techniques, such as federated learning (FL) [7].”
BackgroundChen, X., Wang, A., Deng, X., & Gui, J. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3462722 · Cited by 6
“A more aggressive quantization method is to compress gradients to three possible values [7], or two possible values [8], [9], i.e., one bit.”
Federated Edge Learning via Reconfigurable Intelligent Surface with One-Bit Quantization
Li, H., Wang, R., Wu, J. H., & Zhang, W. (2022) · Globecom · doi:10.1109/globecom48099.2022.10001550 · Cited by 6
“We assume that the global gradient v̄ can be perfectly transmitted to the devices due to the enough transmit power available and full downlink bandwidth at the BS, as in [3], [5].”
BackgroundTinyFL: A Lightweight Federated Learning Method with Efficient Memory-and-Communication
Cheng, H. L., Chen, Q., Liang, Y., & Zhu, G. (2023) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps58843.2023.10464633 · Cited by 6
“…the application of memory-and-communication intensive FL faces severe challenges due to the limitation of memory resources at edge devices and the wireless communication bandwidths at the network [2], [3], which would lead to low model training efficiency and high communication overhead.”
BackgroundSelf-citationSparse Training for Federated Learning With Regularized Error Correction
Greidi, R., & Cohen, K. (2024) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2024.3463505 · Cited by 6
“Addressing these issues is a dynamic and evolving field with numerous prior works spanning various areas, including weight pruning [17]–[19], over-the-air methods [13]–[16], [20], and model or gradient compression [21]– [30].”
BackgroundOver-the-Air Federated Learning: Rethinking Edge AI Through Signal ProcessingPreprint
Azimi‐Abarghouyi, S. M., Fischione, C., & Huang, K. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.03719 · Cited by 6
“…since been widely adopted in subsequent AirFL studies, including [26], [30]–[32], where it has been extended to support digital modulations through one-bit aggregation [26], large-scale cellular deployments [30], hierarchical multi-cluster networks [32], and principal-component-based learning [31].”
Self-citationOut-of-Air Computation: Enabling Structured Function Extraction from Wireless SuperpositionPreprint
Azimi‐Abarghouyi, S. M. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2604.04312 · Cited by 6
“To overcome these limitations, a new paradigm has recently emerged that seeks to realize AirComp using digital communication techniques, often referred to as digital Air-Comp [16]–[26].”
Turning Channel Noise into an Accelerator for Over-the-Air Principal Component Analysis
Zhang, Z., Wang, R., Li, T., Sun, L., Lau, V. K. N., & Huang, K. (2021) · 2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51858.2021.9593170 · Cited by 5
Self-citationBiased Over-the-Air Federated Learning under Wireless Heterogeneity
Abrar, M. F. U., & Michelusi, N. (2024) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops59551.2024.10615488 · Cited by 5
“We highlight here that the existing schemes [7]–[9], [16] assume the same average path loss across devices, yielding uniform device participation, p m = 1 N , ∀ m ∈ [ N ] , so that ( ˜ P) and (P) become equivalent.”
BackgroundInfluentialOn Designing Modulation for Over-the-Air Computation — Part II: Pyramid Sampling
Razavikia, S., & Fischione, C. (2026) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2026.3742995 · Cited by 5
“Then, we tailored the distance among the induced constellation points D C p r i , r j q to the distribution of the noise, which yields robust constellation diagrams for Gaussian, Laplace, and heavy-tailed noise distributions.”
On Designing Modulation for Over-the-Air Computation — Part I: Noise-Aware Design
Razavikia, S., & Fischione, C. (2026) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2026.3742996 · Cited by 5
“To compensate for the fading effects in (3), we use channel inversion power control [46], [47], where we adjust the transmit power as the inverse of the channel, i.e., p k “ h ˚ k {| h k | 2 for k P r K s .”
Split Federated Learning for Resource-Constrained Edge Computing Networks
Wang, X., Song, S., Zhang, Z., Hou, X., Li, Z., Xing, T., & Zhang, X.-P. (2025) · IEEE Transactions on Consumer Electronics · doi:10.1109/tce.2025.3626460 · Cited by 5
“In [27], the waveform super-position properties of broadband channels were leveraged to reduce transmission latency, while the effects of channel fading and imperfect channel state information on FL convergence were systematically investigated.”
VecComp: Vector Computing via MIMO Digital Over-the-Air Computation
Razavikia, S., da Silva, J. M. B., & Fischione, C. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2026.3688265 · Cited by 5
“Towards devising a digital aggregation method, one-bit broadband digital aggregation, using simple binary phase shift keying modulation, has been introduced in [5] with its extension to non-coherent system model studied in [15] and [17].”
Optimising Decision Making on Communication Systems: The Federated Learning Approach
Stergiou, K. D., Psannis, K. E., Roumeliotis, Μ., Kokkonis, G., & Ishibashi, Y. (2021) · IEEE International Conference on Information, Communication and Networks (ICICN) · doi:10.1109/icicn52636.2021.9674002 · Cited by 5
Chirp-Based Over-the-Air Computation for Long-Range Federated Edge Learning
Hoque, S. S. M., Adeli, M. H., & Şahin, A. (2022) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc54779.2022.9978134 · Cited by 5
“Similar to OBDA [7], CSC-MV is compatible with digital modulation.”
MethodInfluentialCo-Phase Over-the-Air Aggregation for Multi-Server Federated Learning with Randomized Transmissions
Choi, J. (2023) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2023-spring57618.2023.10199823 · Cited by 5
“For digital OTA, one-bit quantization in [10] or randomized quantization as in [17] [9] can be considered.”
BackgroundMethodInfluentialJin, R., He, X., Zhong, C., Zhang, Z., Quek, T. Q. S., & Dai, H. (2024) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2024.3454986 · Cited by 5
Multi-Task Federated Edge Learning (MTFeeL) With SignSGD
Mahara, S. S., Shruti, M., & Bharath, B. N. (2022) · National Conference on Communications · doi:10.1109/ncc55593.2022.9806778 · Cited by 5
“An extension of this called one-bit broadband digital aggregation (OBDA) is proposed in [14].”
BackgroundChirp-Based Over-the-Air Computation for Power-Efficient Distributed Localization
Hoque, S. S. M., & Şahin, A. (2023) · IEEE International Black Sea Conference on Communications and Networking · doi:10.1109/blackseacom58138.2023.10299785 · Cited by 5
“In [8], the authors propose an equivalent digital framework named one-bit broadband digital aggregation (OBDA), where distributed training based on majority vote (MV) is considered.”
BackgroundOver-the-Air Federated Edge Learning with Error-Feedback One-Bit Quantization and Power Control
Liu, Y., Liu, D., Jin, R., Zhu, G., & Shi, Q. (2024) · IEEE International Conference on Computer and Communications (ICCC) · doi:10.1109/iccc62609.2024.10942260 · Cited by 5
“In another word, EFOBDA achieves better performance than OBDA, since the latter is worse than the scheme without quantization as shown in [29].”
InfluentialSelf-citationCoherence-Aware Over-the-Air Distributed Learning Under Heterogeneous Link Impairments
Karbalayghareh, M., Love, D. J., & Brinton, C. G. (2026) · IEEE Journal on Selected Areas in Information Theory · doi:10.1109/jsait.2026.3672794 · Cited by 4
“While digital FL emphasizes e ffi cient scheduling and bandwidth allocation 1 Throughout, we classify devices by channel coherence conditions.”
Non-Convex Over-the-Air Heterogeneous Federated Learning: A Bias–Variance Trade-off
Abrar, M. F. U., & Michelusi, N. (2026) · International Conference on Communications · doi:10.1109/icc59461.2026.11587568 · Cited by 4
“Critically, unlike existing works [5]–[7], [12] assuming identical average path loss across devices ( Λ m =Λ n , ∀ m, n ∈ [ N ] ), we consider a heterogeneous wireless environment where devices experience distinct path losses and study uplink transmission of local gradients via over-the-air (OTA)…”
InfluentialConvergence Analysis of Over-the-Air FL with Compression and Power Control via Clipping
Wen, H., Xing, H., & Simeone, O. (2023) · Globecom · doi:10.1109/globecom54140.2023.10437291 · Cited by 4
Semantic Communications for Joint Image Recovery and Classification
Lyu, Z., Zhu, G., Xu, J., Ai, B., & Cui, S. R. (2023) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps58843.2023.10464706 · Cited by 4
“However, the proposed frameworks in above works [4]–[6] only focused on data recovery, which fail to serve on-growing needs of executing AI tasks in future wireless networks [7].”
BackgroundSelf-citationOn Differential Privacy for Wireless Federated Learning with Non-coherent Aggregation
Seif, M., Şahin, A., Poor, H. V., & Goldsmith, A. J. (2023) · Globecom · doi:10.1109/globecom54140.2023.10437931 · Cited by 4
“Similarly, [7] proposes one-bit broadband digital aggregation (OBDA) with TCI.”
BackgroundResultWaveform Design for Over-the-Air Computing
Evgenidis, N. G., Mitsiou, N. A., Tegos, S. A., Diamantoulakis, P. D., Sarigiannidis, P. G., Rekanos, I. T., & Karagiannidis, G. K. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3613838 · Cited by 4
“In addition to improving OTA computing, some other technologies can be improved by it.”
LoLaFL: Low-Latency Federated Learning via Forward-Only Propagation
Zhang, J., Huang, J., & Huang, K. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3625961 · Cited by 4
“For the second problem, techniques like federated transfer learning can be utilized for model initialization and speed up the convergence [34].”
Self-citationDistributionally Robust Over-the-Air Computation in Presence of Channel Uncertainties
Zhang, H., Tang, X., Zhang, R., Turlykozhayeva, D., Ussipov, N., & Han, Z. (2024) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc57260.2024.10571139 · Cited by 4
“In [10], the authors design an elaborate federated edge learning scheme, which features over-the-air majority-vote based gradient-decoding at the edge server.”
BackgroundAnalog MIMO Communication for One-Shot Distributed Principal Component Analysis
Chen, X., Larsson, E. G., & Huang, K. (2022) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2022.3182484 · Cited by 4
“For example, we consider the DPCA traffic latency-critical and make no assumptions on uplink/downlink reciprocity (that could otherwise facilitate synchronization for AirAggregate purposes).”
BackgroundMethodInfluentialSelf-citationOne-Bit Distributed Mean Estimation with Unknown Variance
Kumar, R., & Vatedka, S. (2025) · Transactions on machine learning research · doi:10.48550/arxiv.2501.18502 · Cited by 4
AI-in-the-Loop Sensing and Communication Joint Design for Edge IntelligencePreprint
Cai, Z., Cao, X., Chen, X., Cui, Y., Zhu, G., Huang, K., & Cui, S. R. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2502.10203 · Cited by 4
Self-citationCommunication-Efficient Multi-Server Federated Learning via Over-the-Air Computation
Han, R., Ma, J., Bai, L., Choi, J., & Zhang, W. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3573600 · Cited by 3
“Restrictions apply. the channel state information (CSI) [19], [20], [21], others do not require CSI [22].”
Liang, H., Wen, H., Wu, K., Letaief, K. B., & Xing, H. (2026) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2026.3740652 · Cited by 3
“Moreover, for an edge server equipped with a multi-antenna base station (BS), Yang et selection and beamforming optimization subject to aggregation error requirements [3].”
Error Analysis for Over-the-Air Federated Learning under Misaligned and Time-Varying Channels
Ma, X., Zehtabi, S., Kim, T., & Brinton, C. G. (2025) · Globecom · doi:10.1109/globecom59602.2025.11432115 · Cited by 3
“In [4], it has been verified that AirComp can provide strong noise tolerance and reduced latency compared to conventional orthogonal multiple access (OMA) methods.”
Nguyen, D. Q., & Kim, T. (2023) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf59524.2023.10476780 · Cited by 3
Ultra-Reliable and Low-Latency Wireless Hierarchical Federated Learning: Performance Analysis
Zhang, H., Xu, P., & Dai, B. (2024) · Entropy · doi:10.3390/e26100827 · Cited by 3
“We assume that the channel matrix elements follow an i.i.d. distribution as CN ( 0, 1 ) [5,6,17].”
BackgroundScalable and Reliable Over-the-Air Federated Edge Learning
Egger, M., Hofmeister, C., Kaya, C., Bitar, R., & Wachter-Zeh, A. (2024) · Globecom · doi:10.1109/globecom52923.2024.10901264 · Cited by 3
“In a limited bandwidth regime where n k , each element of the gradient can be quantized to a single bit, and quadrature amplitude modulation (QAM) can be used for digital modulation with majority-vote-based decoding at the federator [11].”
Air-FedGA: A Grouping Asynchronous Federated Learning Mechanism Exploiting Over-The-Air Computation
Ma, Q., Zhou, J., Hou, X., Liu, J., Xu, H., Miao, J., & Jia, Q. (2025) · IEEE International Parallel and Distributed Processing Symposium (IPDPS) · doi:10.1109/ipdps64566.2025.00051 · Cited by 3
“They further expand their work by using one-bit quantization at workers, followed by modulation and majority-voting-based decoding at the parameter server, to reduce the communication overhead [15].”
Jha, N. K., Guo, H., & Lau, V. K. N. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3338046 · Cited by 3
“Single bit-quantisation [41] developed for federated learning in deep learning is also not applicable for general OTA applications.”
BackgroundImproving Wireless Federated Learning via Joint Downlink-Uplink Beamforming Over Analog Transmission
Zhang, C., Dong, M., Liang, B., Afana, A., & Ahmed, Y. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3595700 · Cited by 3
“Besides the analog over-the-air aggregation designs mentioned above, a few recent works have considered digital over-the-air aggregation [27]–[29], where coding schemes with scalar or vector quantization schemes are proposed.”
Fundamental limits of over-the-air optimization: Are analog schemes optimal?
Jha, S. K., Mayekar, P., & Tyagi, H. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685770 · Cited by 3
“In [23], authors perform one-bit quantization of subgradients similar”
BackgroundMethodResultInfluentialDynamic Clustering and Power Control for Two-Tier Over-the-Air Federated Learning
Guo, W. M., Huang, C. Z., Qin, X., Yang, L. X., & Zhang, W. (2022) · IEEE/CIC International Conference on Communications in China (ICCC) · doi:10.1109/iccc55456.2022.9880666 · Cited by 3
“However, the distortion caused by the additive noise and channel fading of the wireless channels degrades the FL performance, and existing works to address this issue exhibit similar solutions that they discard the devices with relatively weak channel gains [5], [6], which results in insufficient training data exploit…”
BackgroundOver-the-Air Federated Learning in MIMO Cloud Radio Access Networks
Ma, H., Yuan, X., & Ding, Z. J. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3549499 · Cited by 3
Private Federated Learning with Dynamic Power Control via Non-Coherent Over-the-Air ComputationPreprint
Zhang, A., Guo, S., & Liu, S. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2308.02881 · Cited by 3
“Under the dynamic power control method, we derive that stochasticity-induced error in the wireless fading channel for all i is bounded as: where is defined as the gradient-signal-to-data-noise ratio (Zhu et al., 2021).”
Metadata
- Authors
- Zhu, G., Du, Y., Gunduz, D., & Huang, K.
- Affiliations
- Shenzhen Research Institute of Big Data (CN)
- University of Hong Kong (HK)
- Venue
- IEEE Transactions on Wireless Communications
- Published
- 2021-03
- Volume
- 20
- Issue
- 3
- Pages
- 2120-2135
- ISSN
- 1536-1276 (ISSN-L)
- 1558-2248
- OpenAlex
- W3108590825
- Cited by
- 376
- References
- 37
- Citations by year
- 2026: 30 · 2025: 45 · 2024: 61 · 2023: 93 · 2022: 68 · 2021: 58
- FWCI
- 23.57
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Transactions on Wireless Communications (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- IRIS UNIMORE (University of Modena and Reggio Emilia) (submittedVersion, cc-by)
- Spiral (Imperial College London) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (62001310)
- National Key Research and Development Program of China (2018YFB1800800)
- Guangdong Province Key Area Research and Development Program (2018B030338001)
- Shenzhen Peacock Plan (KQTD2015033114415450)
- European Research Council (ERC) through the Starting Grant BEACON (677854)
- U.K. EPSRC (EP/T023600/1)
- Guangdong Basic and Applied Basic Research Foundation (2019B1515130003)
- Hong Kong Research Grants Council (17208319, 17209917)
- Innovation and Technology Fund (GHP/016/18GD)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- Physical Sciences › Computer Science › Computer Networks and Communications › IoT and Edge/Fog Computing
- Physical Sciences › Computer Science › Computer Networks and Communications › Distributed Sensor Networks and Detection Algorithms
- Keywords
- federated edge learning, over-the-air aggregation, 1-bit quantization, convergence analysis, channel estimation error, fading, channel noise, communication bottleneck, edge devices
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.