Federated Learning via Over-the-Air Computation
Open accessThe DOI matches its Crossref record; 61 of the 62 references it cites were found in a registry or an index.
Abstract
The stringent requirements for low-latency and privacy of the emerging high-stake applications with intelligent devices such as drones and smart vehicles make the cloud computing inapplicable in these scenarios. Instead, edge machine learning becomes increasingly attractive for performing training and inference directly at network edges without sending data to a centralized data center. This stimulates a nascent field termed as federated learning for training a machine learning model on computation, storage, energy and bandwidth limited mobile devices in a distributed manner. To preserve data privacy and address the issues of unbalanced and non-IID data points across different devices, the federated averaging algorithm has been proposed for global model aggregation by computing the weighted average of locally updated model at each selected device. However, the limited communication bandwidth becomes the main bottleneck for aggregating the locally computed updates. We thus propose a novel over-the-air computation based approach for fast global model aggregation via exploring the superposition property of a wireless multiple-access channel. This is achieved by joint device selection and beamforming design, which is modeled as a sparse and low-rank optimization problem to support efficient algorithms design. To achieve this goal, we provide a difference-of-convex-functions (DC) representation for the sparse and low-rank function to enhance sparsity and accurately detect the fixe…
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
“For instance, over-the-air federated learning (FL) provides a collaborative ML framework to train a global statistical model over wireless networks without accessing edge devices' private raw data [31].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [31]
In the citing paper
“Specifically, the transmitted signals at edge devices are first multiplied by the fading channels and then superposed over-the-air with additive channel noise, resulting in a noisy weighted sum of transmitted signals [31].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [31]
In the citing paper
“Specifically, for edge training systems, we need to jointly optimize the subcarrier and bandwidth allocation [88], [90], [241], transmit power and receive beamforming [31], [48], [178], passive beamforming at RIS [48], [178], device selection [31], [242] and activity detection [76], local updates computation [92], and global aggregation frequency control [130], thereby reducing the optimality gap”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [31]
In the citing paper
“For instance, the group sparsity can represent the combinatorial variables for edge devices selection in FL [31], edge devices activity detection [76], and inference tasks selection [53].”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [31]
In the citing paper
“Most of the resource allocation problems in edge AI need to solve a series of nonconvex optimization problems, e.g., nonconvex sparse optimization for device selection in wireless FL, nonconvex quadratic programming for transceiver design in over-the-air FL [31], lowrank matrix optimization for interference management in edge device distributed inference [52], and unit modulus constrained phase”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [31]
2 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
Hao, X.-B., & Zhang, G.-G. (2017) · Encycl. Semantic Comput. Robotic Intell. · doi:10.1142/s2425038416300184 · Cited by 86,401 · cited 1×
“THE astounding growth in data volume promotes widespread artificial intelligent applications such as image recognition and natural language processing [1], thanks to”
BackgroundLeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Boyd, S., & Vandenberghe, L. (2004) · doi:10.1017/cbo9780511804441 · Cited by 56,601
Vorontsova, E., Hildebrand, R., Gasnikov, A. V., & Stonyakin, F. (2010) · Computer Vision · doi:10.1007/978-1-4614-5257-7_6 · Cited by 39,270 · cited 2×
“The resulting problem is known as the sum-of-infeasibilities in the literature of optimization [39].”
Learning Multiple Layers of Features from Tiny Images
Krizhevsky, A. (2024) · doi:10.57702/zp44cu3g · Cited by 23,395
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 · cited 7×
“In particular, the federated averaging (FedAvg) algorithm [12] turns out to be a promising way to efficiently average the locally updated model at each device with unbalanced and non-IID data, thereby reducing the number of communication rounds between the center node and the end devices.”
InfluentialFederated Machine Learning: Concept and Applications
Yang, Q., Liu, Y., Chen, T., & Tong, Y. (2019) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3298981 · Cited by 6,322
Han, S., Mao, H., & Dally, W. J. (2015) · arXiv · doi:10.48550/arxiv.1510.00149 · Cited by 4,662 · cited 1×
“To address this issue, there is a growing body of recent works to reduce the storage overhead, time and power consumption in the inference process using the model compression methods via hardware and software co-design [7], [8].”
MethodTowards Federated Learning at Scale: System Design
Bonawitz, K., Eichner, H., Grieskamp, W., Huba, D., Ingerman, A., Ivanov, V., Kiddon, C., Konečný, J., Mazzocchi, S., McMahan, H. B., et al. (2019) · arXiv · doi:10.48550/arxiv.1902.01046 · Cited by 2,334 · cited 2×
“It has promising applications [14], [15] such as smart retail, smart healthcare, financial services, mobile content predictions, etc.”
BackgroundFederated Learning with Non-IID Data
Zhao, Y., Li, M., Lai, L., Suda, N., Civin, D., & Chandra, V. (2018) · arXiv · doi:10.48550/arxiv.1806.00582 · Cited by 2,077 · cited 1×
“, the data is generated by distinct distributions across different devices), imposes significant statistical challenges to fit a mode from the non-IID data [13], [17].”
Background
Research works printed but not linked: 2
- rockafellar, 2015, Convex Analysis
- tao, 1997, Convex analysis approach to DC programming: Theory, algorithms and applications, Acta Math Vietn, 22, 289
Cited by
Items 501–600 of 999
Qin, F., Xu, S., Li, C., Xu, Y., & Yang, L. (2024) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2024.3395453 · Cited by 13
“Extensive study has been conducted about the facilitation of FL within wireless networks [4], [5], [6], [7], [8], [9], [10], [11].”
BackgroundReconfigurable Intelligent Surface Empowered Federated Edge Learning With Statistical CSI
Li, H., Wang, R., Wu, J. H., Zhang, W., & Soto, I. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3334250 · Cited by 12
“The specific operating procedures are similar to [10] and [37], which are omitted here due to space limitation.”
BackgroundEmpowering 6G maritime communications with distributed intelligence and over-the-air model sharing
Zetas, M., Spantideas, S., Giannopoulos, A., Νομικός, Ν., & Trakadas, P. T. (2024) · Frontiers in Communications and Networks · doi:10.3389/frcmn.2023.1280602 · Cited by 14
Over-the-Air Federated Averaging With Limited Power and Privacy Budgets
Yan, N., Wang, K., Pan, C., Chai, K. K., Shu, F., & Wang, J. (2023) · IEEE Transactions on Communications · doi:10.1109/tcomm.2023.3338738 · Cited by 13
“On the other hand, over-the-air FL (OTA-FL) [1], [11], [16], [17], [18], [19], [20], [21], [22], [23], [24] is promising to alleviate the above-mentioned communication bottleneck and reduce the communication and computation latency.”
BackgroundGraph Neural Networks Over the Air for Decentralized Tasks in Wireless Networks
Gao, Z., & Gündüz, D. (2025) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2025.3534685 · Cited by 18
“For example, it is used in federated learning for gradient transmissions among devices and the server [51], [52], [53], or for collaborative inference across an ensemble of models at the edge [54].”
Three-Dimensional Cooperative Positioning for Internet of Things Provenance
Bi, S., Cui, J., Ni, W., Jiang, Y. F., Yu, S., & Wang, X. (2022) · IEEE Internet of Things Journal · doi:10.1109/jiot.2022.3170564 · Cited by 11
“, problem (7) without (7f), is increasingly unlikely [28].”
BackgroundBeamforming Design for Integrated Sensing and SWIPT System
Zeng, X., Xing, L., Wu, Y., & Shi, Y. (2022) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc54779.2022.9977463 · Cited by 13
“And for the computation of the subgradient of h, there is a helpful proposition as follow that have been proved in [24].”
Self-citationDouble-RIS Assisted Over-the-Air Computation
Li, J., Fu, M., Zhou, Y., & Shi, Y. (2021) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps52748.2021.9682077 · Cited by 10
“To quantify the computation error, a common design metric that has been widely adopted in AirComp is the mean-squared error (MSE) between the estimated function and the target function [2]-[5].”
BackgroundMethodInfluentialSelf-citationOptimal Receive Beamforming for Over-the-Air Computation
Fang, W., Zou, Y., Zhu, H., Shi, Y., & Zhou, Y. (2021) · IEEE International Workshop on Signal Processing Advances in Wireless Communications (SPAWC) · doi:10.1109/spawc51858.2021.9593251 · Cited by 12
“In particular, by exploiting the superposition property of multiple access channels (MACs), wireless data aggregation can be achieved in one transmission interval by allowing all IoT devices to transmit concurrently over the same radio channel [5].”
BackgroundSelf-citationAirMixML: Over-the-Air Data Mixup for Inherently Privacy-Preserving Edge Machine Learning
Koda, Y., Park, J., Bennis, M., Vepakomma, P., & Raskar, R. (2021) · Globecom · doi:10.1109/globecom46510.2021.9685232 · Cited by 10
“Channel superposition of analog modulation is proposed to achieve scalability [1], [4], [12], [13] and DP [2]–[4].”
BackgroundGradient Estimation for Federated Learning over Massive MIMO Communication SystemsPreprint
Jeon, Y.-S., Amiri, M., Li, J., & Poor, H. (2020) · arXiv.org · doi:10.48550/arxiv.2003.08059 · Cited by 10
“For example, in [23], a joint optimization framework for device selection and receive beamforming is proposed to maximize the number of selected devices while guaranteeing the target mean square error (MSE) requirement.”
Differentially Quantized Gradient Methods
Lin, C.-Y., Kostina, V., & Hassibi, B. (2022) · IEEE Transactions on Information Theory · doi:10.1109/tit.2022.3171173 · Cited by 21
Over-the-Air Inference Through Analog Computation Over Multi-Hop MIMO Networks
Bian, C., Hua, M., & Gündüz, D. (2025) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2025.3602209 · Cited by 8
“For instance, linear operations such as summations and polynomials can be carried out by exploring the signal-superposition property of the wireless medium, which have been widely applied in federated learning [3], [4].”
Jiang, 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 .”
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
“Index Terms —Federated learning, latency minimization, communication-computation tradeoff, batch-size adaptation.”
The Analysis and Optimization of Volatile Clients in Over-the-Air Federated Learning
Shi, F., Lin, W., Wang, X., Li, K., & Zomaya, A. Y. (2024) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2024.3427709 · Cited by 12
“In AirComp FL, although not constrained by bandwidth, existing research has demonstrated that both the quality and quantity of participating clients significantly influence model training outcomes [18], [19], [20].”
BackgroundMethodInfluentialHu, Z., Yan, J., & Zhang, Y. A. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3446669 · Cited by 6
“With over-the-air model aggregation, local updates are aligned at the PS through transmit power control [3], [4], [5].”
Privacy-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
Reconfigurable Intelligent Surfaces-Aided Federated Learning in Over-the-Air Computation
Kim, M., & Park, D. (2024) · IEEE Wireless Communications Letters · doi:10.1109/lwc.2024.3399828 · Cited by 18
“However, the FL performance improves by employing device selection strategies that exclude devices that increase aggregation errors [10], [11].”
BackgroundMethodInfluentialEfficient and Secure Aggregation Framework for Federated-Learning-Based Spectrum Sharing
Li, W., Chen, G., Zhang, X., Wang, N., Ouyang, D., & Chen, C. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3357575 · Cited by 9
“For example, Yang et al. [22] introduced an over-the-air computation-based approach for fast global model aggregation, where the device selection and beamforming design are modeled as a sparse and low-rank optimization problem.”
MethodLightweight Differential Frameworks for CSI Feedback in Time-Varying Massive MIMO Systems
Zhang, Y., Zhang, X., & Liu, Y. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3345938 · Cited by 9
“It is being deployed for various purposes, including spectral efficiency optimization [18] and energy efficiency enhancement [19].”
BackgroundJoint 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
“3 The design of a common λt among the N mobile devices at each iteration t was studied in [18], [19], [21], [22], [24]-[27], and is outside the scope of this paper.”
BackgroundMethodMiso: Misalignment Allowed Optimization for Multiantenna Over-the-Air Computation
Tang, S., Zhang, C., Li, J., & Obana, S. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3292882 · Cited by 11
“The proposed Miso method (Algorithm 2) is compared with Algorithm 1, the antenna selection (AntSel) method [16] and the zero-forcing (ZF) method [14].”
BackgroundMethodInfluentialFeng, T.-Y., Xu, J., Zhou, Z., & Luo, Y. (2023) · Journal of Cases on Information Technology · doi:10.4018/jcit.333858 · Cited by 12
“This makes it difficult for banks to effectively assess their credit situation, resulting in credit losses and high credit risk (He et al., 2022; Yang et al., 2020; Sattler et al., 2020).”
BackgroundPower 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
“…of noise 𝜎 𝑛 range from − 80dBm to − 68dBm. (1) Minimum Mean Squared Error (MMSE): In each communication Specifically { 𝑎 } are set based on [14, 19, 32] by zero forcing as as (25), and the same SCA method can be used to design f 𝑡 (2) Greedy Spatial Device Selection (GSDS)[22]: Joint…”
MethodLi, J.-M., Wang, Z.-Y., Zhang, Y., Zhu, P.-C., Wang, D., & You, X.-H. (2023) · IEEE Systems Journal · doi:10.1109/jsyst.2022.3168022 · Cited by 9
“Specifically, for the PSD matrix F , the constraint (26d) has the following equivalent form [26]: Denote nuclear norm and spectral norm of matrix F as F (cid:4) ∗ = (cid”
BackgroundBeamforming 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
“Unlike the existing works [20], [21], [22], we formulate our design problem assuming the channel states between the server and the devices can change over communication rounds, and the device selection and beamforming solutions are computed in each round based on the current channel state…”
BackgroundMethodResultInfluentialLearning-based Multi-Objective Resource Allocation for Over-the-Air Federated Learning
Tu, X., & Zhu, K. (2022) · Globecom · doi:10.1109/globecom48099.2022.10000875 · Cited by 9
“In [3], the authors jointly designed the receive beamforming device selection to maximize the quantity of participating devices with the mean squared error (MSE) constraint for the improvement of model training.”
BackgroundFederated and Meta learning over Non-Wireless and Wireless Networks: A Tutorial
Liu, X., Deng, Y.-S., Nallanathan, A., & Bennis, M. (2022) · arXiv.org · doi:10.48550/arxiv.2210.13111 · Cited by 9
“…is defined as Motivated by [162], given the receiver beamforming vector m , the MSE is minimized by using the zero-forcing transmitter Based on [161], authors in [163] considered broadband analog aggregation (BAA) for OAC model aggregation to further maximize the number of scheduled devices…”
BackgroundMethodOver-the-Air Split Learning with MIMO-Based Neural Network and Constellation-Based Activation
Yang, Y.-Z., Zhang, Z.-Y., & Yang, Z.-H. (2022) · International Workshop on Machine Learning for Signal Processing · doi:10.1109/mlsp55214.2022.9943518 · Cited by 9
“OAC has been applied to the typical federated learning (FL) scheme [4], where OAC is used to calculate the weighted sum of the parameters on different edge devices in FL.”
MethodJoint Optimization for Federated Learning Over the Air
Fan, X., Wang, Y., Huo, Y., & Tian, Z. (2022) · International Conference on Communications · doi:10.1109/icc45855.2022.9838269 · Cited by 11
“Although optimization issues are considered in [9], [10], the optimization is conducted on communication side alone, without an underlying connection to FL. Noticeably, while the optimization in existing works boils down to maximizing the number of selected workers, our theoretical results indicate…”
ResultOver-The-Air Federated Learning under Byzantine AttacksPreprint
Sifaou, H., & Li, G. Y. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2205.02949 · Cited by 8
“As in [21], [22], [24], [27], [28], we assume that perfect channel state information (CSI) is available at the clients and the CS.”
BackgroundMethodFederated learning-based aerial image segmentation for collision-free movement and landing
Chhikara, P., Tekchandani, R., Kumar, N., & Tanwar, S. (2021) · ACM MobiCom Workshop on Drone Assisted Wireless Communications for 5G and Beyond (DroneCom) · doi:10.1145/3477090.3481051 · Cited by 12
“While in the air, the drone gathers data in the form of images and offers FL for data interpretation and inferring tasks like trajectory devising and cooperative target identification [17, 19].”
FL-MAC-RDP: Federated Learning over Multiple Access Channels with Rényi Differential Privacy
Wu, S., Yu, M., Ahmed, M. A. M., Qian, Y., & Tao, Y.-H. (2021) · International Journal of Theoretical Physics · doi:10.1007/s10773-021-04867-0 · Cited by 9
Zhang, D. (2021) · Journal of Ambient Intelligence and Humanized Computing · doi:10.1007/s12652-021-03272-z · Cited by 12
Zhu, Y., Xu, J., Xie, Y., Jiang, J., Yang, X.-Z., & Li, Z.-B. (2021) · International Conference on Communication, Computing & Security · doi:10.1109/icccs52626.2021.9449265 · Cited by 9
“…for different application scenarios, such as appropriately reducing communication frequency to ensure learning efficiency, or adding a logical judgment after aggregation to determine the quality of the local computation received to improve the robustness of the federated learning system [9], [10].”
BackgroundChannel-Driven Monte Carlo Sampling for Bayesian Distributed Learning in Wireless Data Centers
Liu, D., & Simeone, O. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3118406 · Cited by 14
Computation Over Multi-Access Channels: Multi-Hop Implementation and Resource Allocation
Wu, F., Chen, L., Zhao, N., Chen, Y., Yu, F. R., & Wei, G. (2020) · IEEE Transactions on Communications · doi:10.1109/tcomm.2020.3034934 · Cited by 16
“Besides, in federated learning, to obtain the average learning model from the local models of distributed nodes, CoMAC is deployed to significantly reduce the aggregation latency [4].”
MethodOne-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
“The first works in the literature that studied FEEL taking into account the physical layer resource constraints focus on over-the-air aggregation [5]–[8].”
BackgroundMethodAge of Aggregated Information: Timely Status Update with Over-The-Air Computation
Li, J., Zhou, Y., Chen, H. H., & Shi, Y. (2020) · Globecom · doi:10.1109/globecom42002.2020.9322298 · Cited by 11
“…AirComp has been applied in different applications, e.g., to achieve fast high-mobility multi-modal sensing in MIMO networks [15] and to accelerate the convergence of federated learning [16], how to leverage AirComp for boosting the data freshness in IoT networks still remains an open problem.”
MethodSelf-citationMultigroup Multicast Transmission via Intelligent Reflecting Surface
Ding, L., An, Q., Shi, Y., & Zhou, Y. (2020) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2020-fall49728.2020.9348503 · Cited by 8
“According to [18], the rank-one constraint of a PSD matrix can be equivalently written as the zero difference between the trace norm and the spectral norm.”
BackgroundSelf-citationJamming Attacks Detection and Ejection in Over the Air Computation Concentrated Solar Power Systems
Adhikary, D., Plusquellic, J., & Tsiropoulou, E. E. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3606344 · Cited by 9
“The problem of optimizing the device selection and beam-forming design is studied in [11] aiming at maximizing the number of participating devices while minimizing the model aggregation errors in order to improve the communication efficiency and the learning performance.”
Pinching Antennas Meet AI in Next-Generation Wireless Networks
Fang, F., Ding, Z.-G., Leung, V. C. M., & Hanzo, L. (2025) · arXiv.org · doi:10.48550/arxiv.2511.07442 · Cited by 8
“To address this challenge, AirComp has emerged as a powerful technique for wireless data aggregation in edge learning systems by exploiting the superposition property of multi-access wireless channels [12].”
Self-citationRIS-Assisted Multi-Cell Over-the-Air Computation
Xiao, Y., Tegos, S. A., Huang, S., Diamantoulakis, P. D., Tyrovolas, D., Ma, Z., Karagiannidis, G. K., & Fan, P. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3560708 · Cited by 9
“…in OTA computation reduces the need for extensive back-and-forth data transmission, thereby decreasing latency and bandwidth requirements for data aggregation, which in turn enhances network efficiency and enables efficient wireless model aggregation across multiple domains [6], [7], [8], [9].”
Improving communication performance of Federated Learning: A networking perspective
Amadeo, M., Campolo, C., Ruggeri, G., & Molinaro, A. (2025) · Computer Networks · doi:10.1016/j.comnet.2025.111353 · Cited by 12
Weighted Average Consensus Algorithms in Distributed and Federated Learning
Tedeschini, B. C., Savazzi, S., & Nicoli, M. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3528982 · Cited by 13
“Additionally, Over-The-Air (OTA) computations and communications [25], [26], [27] require efficient techniques for data transmission and model aggregation while minimizing latency and bandwidth.”
Over-the-Air FEEL With Integrated Sensing: Joint Scheduling and Beamforming Design
Asaad, S., Wang, P., & Tabassum, H. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3529509 · Cited by 5
“The initial work [5] focused on studying the joint design of beamforming and device scheduling policy in MIMO settings.”
Wang, J., Liu, Y., Liang, B., & Dong, M. (2025) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom55648.2025.11044570 · Cited by 9
“Various solutions have been proposed in the literature, including model truncation [10], device selection [11], aggregation error minimization [12], and truncated channel inversion [13].”
Federated Low-Rank Adaptation for Large Language Model Fine-Tuning Over Wireless Networks
Wang, Z., Zhou, Y., Shi, Y.-M., & Letaief, K. (2024) · Global Communications Conference · doi:10.1109/globecom52923.2024.10901572 · Cited by 8
“To facilitate an efficient high-dimensional model exchange under limited radio resources, many of existing approaches have been investigated from the perspective of device scheduling [12], power allocation [13], beamforming design [14], and bandwidth optimization [15].”
Self-citationSimulafed: an enhanced federated simulated environment for privacy and security in health
Rivas, J. M., Fernandez‐Basso, C., Morcillo-Jiménez, R., Paños-Basterra, J., Ruiz, M. D., & Martín-Bautista, M. J. (2024) · Computing · doi:10.1007/s00607-024-01364-0 · Cited by 9
“It is particularly useful in scenarios where data privacy is a concern, such as in healthcare [7], finance and banking [8], telecommunications [9], manufacturing and industry 4.0 [10], SmartCities and IoT [11], Retail [12], and Energy [13].”
InfluentialOver-the-Air Computation in Cell-Free Massive MIMO Systems
Chen, C., Björnson, E., & Fischione, C. (2024) · IEEE International Workshop on Computer Aided Modeling and Design of Communication Links and Networks · doi:10.1109/camad62243.2024.10942978 · Cited by 8
“Future Internet-of-Things (IoT) are expected to aggregate enormous amounts of data or computation results transmitted from a large number of edge devices [1], [2].”
Wen, H., Xing, H., & Simeone, O. (2025) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2025.3640114 · Cited by 9
“A popular class of methods has explored the use of over-the-air computing [20], an approach known as over-the-air FL (AirFL) [18], [19], [20], to overcome communication bottlenecks on the shared radio channel for conventional FL. Channel impairments may not necessarily be deleterious for…”
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.”
BackgroundDistributed 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
“For example, for the agents performing federated learning, the required information can be the weighted summation of their neighbors’ gradients [22].”
Deep Reinforcement Learning‐Based Multireconfigurable Intelligent Surface for MEC Offloading
Qu, L., Huang, A., Pan, J., Dai, C., Garg, S., & Hassan, M. M. (2024) · International Journal of Intelligent Systems · doi:10.1155/2024/2960447 · Cited by 11
Device Scheduling for Relay-Assisted Over-the-Air Aggregation in Federated Learning
Zhang, F., Chen, J., Wang, K., & Chen, W. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3346920 · Cited by 11
“[11] and [12] utilize over-the-air computation (AirComp) to design a FL framework that integrates communication and computation.”
MethodOver-the-Air Computation Empowered Federated Learning: A Joint Uplink-Downlink Design
Zhang, D., Xiao, M., & Skoglund, M. (2023) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2023-fall60731.2023.10333467 · Cited by 7
“In the next section, we will jointly optimize each edge node transmit and receive equalization coefficients along with the edge server forwarding matrix to minimize the maximum MSE among the K edge nodes in each training round, aiming to improve the performance of the considered wireless FL system.”
BackgroundMethodInfluentialFedDAD: Solving the Islanding Problem of SAR Image Aircraft Detection Data
Jia, Z., Zheng, H., Wang, R., & Zhou, W. (2023) · Remote Sensing · doi:10.3390/rs15143620 · Cited by 10
“There have been many improved experiments based on FedAvg [43,44], but the Federated Distribution Average Deviation algorithm (FedDAD) proposed in this paper is orthogonal to these methods, and thus FedAvg is used as the baseline for comparison.”
MethodOver-the-Air Federated Learning Over MIMO Channels: A Sparse-Coded Multiplexing Approach
Zhong, C., & Yuan, X. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3567598 · Cited by 7
“Some recent studies have focused on MIMO beamforming design for OA-FL [21], [22], [23].”
Secure Over-the-Air Computation Using Zero-Forced Artificial Noise
Maßny, L., & Wachter-Zeh, A. (2023) · IEEE Information Theory Workshop (ITW) · doi:10.1109/itw55543.2023.10161677 · Cited by 8
Analysis of Error Feedback in Federated Non-Convex Optimization with Biased CompressionPreprint
Li, X., & Li, P. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2211.14292 · Cited by 31
Federated-Learning-Driven Radio Access Networks
Foukalas, F. T. (2022) · IEEE Wireless Communications · doi:10.1109/mwc.102.2100113 · Cited by 10
“It is obvious that in case of two users, for , UE1 and UE2, which need to be allocated with a chunk of resources dynamically, the resource allocation for high bandwidth.”
BackgroundKeçeci, 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
Federated Stochastic Gradient Descent Begets Self-Induced Momentum
Yang, H. H., Liu, Z., Fu, Y., Quek, T. Q. S., & Poor, H. V. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746995 · Cited by 8
“, the server can select many clients for parameter updates in each iteration, that leads to an increase in N which in turns reduces β and results in a faster convergence rate, and vice versa; and (iii) this result also provides further evidence to the claim that having more clients participate in each round of FL trai…”
ResultJoint Communication-Learning Design for RIS-Assisted Federated Learning
Liu, H., Yuan, X., & Zhang, Y. A. (2021) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops50388.2021.9473672 · Cited by 11
“In contrast to the existing work that considers the communication objective [6], [7] or a rough proxy of the learning performance [4], we develop a unified framework that tracks the impact of device selection and communication error on the FL training loss under mild assumptions on the learning…”
BackgroundMethodChen, X., Liu, A., & Zhao, M. (2020) · IEEE Communications Letters · doi:10.1109/lcomm.2020.2999718 · Cited by 12
“Thereby, it induces huge power consumption and CSI signaling overhead for MIMO, making FDC particular inadequate in many applications such as high-mobility multi-modal sensing (HMS) and federated learning [6]–[8].”
BackgroundAnalog Computing With Hybrid Couplers and Phase Shifters
Nerini, M., Liu, X., & Clerckx, B. (2026) · IEEE Transactions on Microwave Theory and Techniques · doi:10.1109/tmtt.2026.3721382 · Cited by 9
“The operation of computing the product Wu over a wireless channel by directly observing the received signal v is known as over-the-air computing [20], and has interesting applications to federated learning [21].”
Cao, X., Wen, D., Bi, S., Cui, Y., Zhu, G., Hu, H., & Eldar, Y. C. (2026) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2026.3674960 · Cited by 6
“To improve learning performance, extensive research was explored ranging from device selection [17], power control optimization [18], [19], interference mitigation [20], to differential privacy [21].”
FRACTAL: Data-Aware Clustering and Communication Optimization for Decentralized Federated Learning
Ma, Q., Liu, J., Xu, H., Jia, Q., & Xie, R. (2024) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2024.3403381 · Cited by 12
“Another approach to improve wireless communication effi-ciency for FL model aggregation is over-the-air computation (AirComp) [33], [34], [35].”
Dynamic Client Selection for Over-the-Air Federated Learning Network
Shi, F., Lin, W.-W., Peng, C.-D., Zhong, C., Wang, D., & Cheng, M. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3523947 · Cited by 7
“For example, Yang et al. [31] used experiments to prove that selecting more clients participating in the model training can make a better training performance.”
Heterogeneity-Aware Client Sampling for Optimal and Efficient Federated LearningPreprint
Weng, S., Ren, C., Xiao, M., & Skoglund, M. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.11304 · Cited by 5
“The rationale behind the proposed method is to design a biased client sampling scheme to counteract the biased optimization dynamics arising from heterogeneous communication q m such that the non-vanishing term χ 2 ω ∥ Ω ( r ) = 0 in each round, the FL process will converge to the pursued global…”
Federated Learning Meets Fluid Antenna: Towards Robust and Scalable Edge Intelligence
Park, S., & Seo, H. (2025) · doi:10.48550/arxiv.2503.03054 · Cited by 7
“Among these advancements, AirComp has emerged as a key technology for alleviating communication bottlenecks in FL by enabling simultaneous model update transmissions [13]–[20].”
Zhao, Y., Xu, M., Wang, P., & Niyato, D. (2025) · doi:10.48550/arxiv.2503.00011 · Cited by 7
“By allowing all devices to transmit simultaneously, OTA achieves high spectral efficiency and accelerates training convergence [19], [20], although robust designs must account for imperfect channel state information, synchronization requirements, and noise-induced aggregation errors [20], [21].”
Wu, M., Liwang, M., Su, Y., Li, L., Hosseinalipour, S., Wang, X., Dai, H., & Jiao, Z. (2025) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2025.3624561 · Cited by 9
“Further, [21], [22] aim to reduce the required global aggregation rounds.”
An Online Federated Learning Framework via Over-the-Air Computing for UAV Object Detection
Li, Y., & Liu, Y. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3461650 · Cited by 12
“The authors in [12] introduced an AirComp-based FL strategyaimingforfastmodelaggregation,leveragingthesuperposition propertyinherentinawirelessmultiple-accesschannel.”
Feature-Based Federated Transfer Learning: Communication Efficiency, Robustness and Privacy
Wang, F., Gursoy, M. C., & Velipasalar, S. (2024) · IEEE Transactions on Machine Learning in Communications and Networking · doi:10.1109/tmlcn.2024.3408131 · Cited by 12
“…[50], federated distillation [51], [52], [53], pruning [54], [55] or partially trainable network [56], [57], and over-the-air computation [58], [59], [60]) still consider the transmission of gradient updates and can achieve a relatively limited reduction in payload and experience degradation…”
BackgroundSCALA: Split Federated Learning With Concatenated Activations and Logit Adjustments
Yang, J., & Liu, Y. (2025) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2025.3649075 · Cited by 8
“Since many clients possess limited compute, battery life, or network connectivity, it is infeasible for every client to take part in each communication round [39], [40].”
Federated Learning in Heterogeneous Networks With Unreliable Communication
Zheng, P., Zhu, Y. L., Hu, Y., Zhang, Z. J., & Schmeink, A. (2023) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2023.3311824 · Cited by 51
“Moreover, existing client selection strategies based on a combinatorial optimization problem [30], [31], [32], [33] including our earlier work [1] are NP-hard and are not scalable with the dimension of the number of clients in an FL cross-device setting.”
BackgroundMethodCollaborative Learning of Different Types of Healthcare Data From Heterogeneous IoT Devices
Zhao, L., Cai, L., & Lu, W.-S. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3307675 · Cited by 8
“Extensive research has focused on developing advanced FL algorithms to enhance learning performance, with emphasis on privacy preservation [11], [12] and learning efficiency [13].”
BackgroundDistributed Clustering for Cooperative Multi-Task Learning Networks
Li, J., Wang, W., Abbas, W., & Koutsoukos, X. (2023) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2023.3276854 · Cited by 7
Fang, W., Han, D.-J., & Brinton, C. G. (2025) · IEEE Transactions on Networking · doi:10.1109/ton.2025.3538661 · Cited by 14
“The authors of [30], [31], and [32] proposed to utilize AirComp to accelerate the convergence of FedAvg in single-cell networks.”
Chen, D., Zhu, Y., Wang, D., Wang, H., Xie, J., Zhang, X.-P., & Han, Z. (2023) · IEEE Transactions on Intelligent Vehicles · doi:10.1109/tiv.2022.3224655 · Cited by 7
“Obviously, when performing federated analytics, the results will be better if more vehicles with high love of variety unity are included [27].”
BackgroundDevice Scheduling for Over-the-Air Federated Learning with Differential Privacy
Yan, N., Wang, K., Pan, C., & Chai, K. K. (2023) · International Conference on Communications · doi:10.1109/icc45041.2023.10278671 · Cited by 6
“Therefore, we propose a device scheduling scheme for the aligned DP-OTA-FL, referred to as S-DPOTAFL.”
BackgroundInfluentialMatching Pursuit Based Scheduling for Over-the-Air Federated Learning
Bereyhi, A., Vagollari, A., Asaad, S., Müller, R. R., Gerstacker, W., & Poor, H. (2022) · arXiv.org · doi:10.48550/arxiv.2206.06679 · Cited by 7
“Consequently, various sub-optimal algorithms for scheduling are proposed in the literature [38] and [45]–[47].”
BackgroundMethodResultInfluentialJoint Coverage and Resource Allocation for Federated Learning in UAV-Enabled Networks
Yahya, M., & Maghsudi, S. (2022) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc51071.2022.9771880 · Cited by 7
“The work in [5] exploits the signal superposition property of multi-access channels to allocate the limited bandwidth to maximize the number of participants in the model update to accelerate the FL convergence.”
BackgroundWireless Federated Learning over MIMO Networks: Joint Device Scheduling and Beamforming Design
Huang, S., Zhang, P., Mao, Y., Lian, L., Wu, Y., & Shi, Y. (2022) · IEEE International Conference on Communications workshops · doi:10.1109/iccworkshops53468.2022.9814684 · Cited by 8
“In analog FL systems, over-the-air computation (AirComp) technology is broadly leveraged to imple- ment the efficient concurrent transmission of locally computed updates by exploiting the superposition property of wireless multiple access channels (MAC) [10]–[12].”
BackgroundMethodSelf-citationBayesian 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
“Over-the-air computation (AirComp)-based FL has been recently proposed as a communication-bandwidth efficient aggregation method [7]–[12].”
BackgroundMethodMAFAT: Memory-Aware Fusing and Tiling of Neural Networks for Accelerated Edge Inference
Farley, J., & Gerstlauer, A. (2023) · IFIP advances in information and communication technology · doi:10.1007/978-3-031-34214-1_7 · Cited by 5
Vertical federated learning based on DFP and BFGSPreprint
Song, W., & Shen, X. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2101.09428 · Cited by 6
“A series of works on horizontal federated learning has been proposed [6], [7], each client has a part of the sample, but has all the data attributes.”
BackgroundSignal Processing Optimization for Federated Learning over Multi-User MIMO Uplink Channel
Huh, M., Yu, D., & Park, S.-H. (2021) · International Conference on Information Networking · doi:10.1109/icoin50884.2021.9333891 · Cited by 8
“The design and analysis of the AirComp technique were studied in [4]–[6] in various aspects.”
MethodReconfigurable Intelligent Surface Assisted Non-Orthogonal Unicast and Broadcast Transmission
An, Q., Shi, Y.-M., & Zhou, Y. (2020) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2020-spring48590.2020.9128547 · Cited by 7
“In detail, as W 0 is a PSD matrix and Tr( W 0 ) > 0 , the rank-one constraint, i.e., rank( W 0 ) = 1 , can be expressed as the difference between the trace norm and the nuclear norm, given by where , and σ i ( W 0 ) is the i -th singular value of matrix W 0 [17].”
Self-citationLarge-Scale Beamforming for Massive MIMO via Randomized Sketching
Choi, H., Jiang, T., Shi, Y., Liu, X., Zhou, Y., & Letaief, K. B. (2021) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2021.3071543 · Cited by 7
“For example, in wireless federated systems [39], the high dimensional local model parameters are required to be aggregated at an edge server, which leads to a communication bottleneck.”
BackgroundSelf-citationAn improved blockchain-based multi-region Federated Learning framework for crop disease diagnosis
Qin, Y., Xu, C., Zhou, Q., Zhang, L., & Zhang, Y. (2025) · Computers & Electrical Engineering · doi:10.1016/j.compeleceng.2025.110181 · Cited by 7
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).”
BackgroundLow-Complexity OTFS-Based Over-the-Air Computation Design for Time-Varying Channels
Huang, X., Hellström, H., & Fischione, C. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3521982 · Cited by 5
“In [8], the authors investigated AirComp for fast global model aggregation in federated learning (FL), where the authors assumed a constant channel vector between a fusion center (FC) and each device.”
Robust Beamforming for Reconfigurable Intelligent Surface-Assisted Multi-Cell Downlink Transmissions
Gao, H., Yang, X., Chen, N., Chen, S., Yang, Y., & Yuen, C. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3346906 · Cited by 7
“Inparticular,therank-onecon-straintcanbefirstlytransformedintothefollowingformby usingthepropertiesofthenorm[35], where (cid:2) W l,k ∗ and (cid:2) W l,k 2 denote the trace norm and spectral norm of the matrix W l,k .”
BackgroundValue of Information and Timing-aware Scheduling for Federated Learning
Khan, M. A., Yang, H. H., Chen, Z., Iera, A., & Παππάς, Ν. (2023) · IEEE Conference on Standards for Communications and Networking · doi:10.1109/cscn60443.2023.10453168 · Cited by 7
“These protocols range from reducing transmission delay [9] and maximizing spectrum usage [10] to cleverly alternating between choosing UEs with different channel conditions [11].”
BackgroundA Random Access Scheme for Federated Learning Over Massive MIMO Systems
Han, H., Zhao, J., & Zhou, X. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3278256 · Cited by 8
“Therefore, massive MIMO systems are suitable for supporting the wireless FL, which has been studied in [13], [14], [15], [16], [17], [18], [19], and [20].”
BackgroundMethodInfluentialAnalog-digital Scheduling for Federated Learning: A Communication-Efficient Approach
Abrar, M. F. U., & Michelusi, N. (2023) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf59524.2023.10477031 · Cited by 5
“To overcome this challenge and ensure scalable device participation, several works [9], [13], [14] have proposed analog over-the-air (OTA) computation for FL-based solutions.”
BackgroundMethodTIFeD: a Tiny Integer-based Federated learning algorithm with Direct feedback alignment
Colombo, L., Falcetta, A., & Roveri, M. (2023) · International Conference on AI-ML Systems (AIMLSystems) · doi:10.1145/3639856.3639867 · Cited by 8
“Other works in the field, such as [4, 6, 18], try to optimize the FedAvg algorithm either by improving its efficiency or by using more complex and sophisticated aggregation functions, but without addressing the issue of TinyML-specific design.”
Metadata
- Authors
- Yang, K., Jiang, T., Shi, Y., & Ding, Z.
- Affiliations
- Chinese Academy of Sciences (CN)
- ShanghaiTech University (CN)
- University of California, Davis (US)
- Venue
- IEEE Transactions on Wireless Communications
- Published
- 2020-03
- Volume
- 19
- Issue
- 3
- Pages
- 2022-2035
- ISSN
- 1536-1276 (ISSN-L)
- 1558-2248
- OpenAlex
- W2999074226
- Cited by
- 1,084
- References
- 62
- Citations by year
- 2026: 86 · 2025: 154 · 2024: 176 · 2023: 190 · 2022: 189 · 2021: 190 · 2020: 80
- FWCI
- 63.79
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Wireless Communications (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (61601290)
- National Science Foundation (CNS-1702752, ECCS-1711823)
- 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 › Engineering › Electrical and Electronic Engineering › Advanced Wireless Communication Technologies
- Keywords
- federated learning, over-the-air computation, edge machine learning, model aggregation, device selection, beamforming design, difference of convex functions, DC algorithm, non-IID data, data privacy
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.