Pith. sign in

REVIEW 4 major objections 4 minor 54 references

Integrated user scheduling and beam steering in over-the-air federated learning for mobile IoT

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that a low-complexity greedy scheduling policy based on channel angles and channel strengths achieves lower over-the-air aggregation error and higher MNIST test accuracy than iterative joint optimization and existing…

desk verdict Plausible AirComp scheduling heuristic and solid simulations, but the theory misrepresents what the algorithm does and several proofs are invalid. read the letter →

arxiv 2508.00341 v1 pith:FQHNZJ65 submitted 2025-08-01 cs.DC

Shengheng Liu , Ningning Fu , Zhonghao Zhang , Yongming Huang , Tony Q. S. Quek This is my paper · ORCID
classification cs.DC
keywords over-the-aircomputationfederatedlearninguserschedulingreceivebeamformingdifference-of-convexlow-SNRmobileIoTMSEminimizationMNIST
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that user scheduling and receive beam steering in over-the-air federated learning can be decided jointly and cheaply, with a greedy channel-based policy outperforming heavier iterative optimization. The authors formulate the low-SNR aggregation problem as a max-min beamforming problem over which users to select and which receive direction to use, prove that users with smaller mutual channel angles and larger channel gains are preferable, and turn that insight into the low-complexity scheduling policy called Policy_greedy. In experiments, this policy achieves the lowest mean-square aggregation error and the highest MNIST test accuracy among all compared schemes, under both i.i.d. and non-i.i.d. data. If the claim holds, mobile IoT systems running over-the-air federated learning do not need the computational cost of iterative combinatorial optimization to obtain good aggregation.

What carries the argument

The carrying object is the max-min beamforming objective, written $\max_{m,S}\min_{k\in S} P|m^H h_k|^2$, which the MSE-minimization problem reduces to after the optimal transmit coefficients are substituted in. The argument runs through two channel-geometry facts: users whose channels have a smaller maximum pairwise angle raise the lower bound on the objective, and users with larger channel modulus strengthen it. These facts justify a greedy subset construction (Algorithm 3) that starts from one of the $G$ strongest channel users and repeatedly adds the unselected user minimizing the maximum inner-product deviation from the current subset, after which the beamformer is computed once with a difference-of-convex rank-one method.

What would settle it

A concrete check would be to construct a channel set in which the strongest-norm user lies far in angle from the other strong users, run the greedy policy and an exhaustive search over all subsets of the required size, and compare the resulting mean-square aggregation error: if exhaustive search materially beats the greedy subset, the paper's central scheduling criterion is incomplete.

Watch

Extended reading notes

Core claim

The central claim is that the receive beam steering vector and the set of scheduled users can be designed together to minimize the MSE of over-the-air aggregation, and that the optimality-relevant property of a user subset is captured by the maximum angle between its channel vectors and by their moduli. The paper proves that for users with equal-norm channels, the best achievable objective is lower bounded by $P\cos^2\alpha$ with $\alpha$ the largest pairwise channel angle, and that scaling channel vectors in the same direction only improves the objective; it then builds a greedy user selection policy on these two facts. The authors report that this policy, Policy_greedy, achieves the best MSE and the best test accuracy among the considered methods in both i.i.d. and non-i.i.d. MNIST experiments, making it the recommended low-complexity choice for low-SNR over-the-air federated learning.

Load-bearing premise

The policy recommendation rests on the assumption that the angle-and-modulus rules proved for idealized equal-norm and same-direction channel cases also govern the unequal-norm case that the greedy algorithm actually faces; the paper does not prove near-optimality of the sequential greedy selection in that general setting.

Editorial extensions

If this is right

  • In low-SNR mobile IoT settings, Policy_greedy is the recommended scheduling strategy because it matches or beats the iterative method while replacing repeated $O(N_r^6)$ beamformer solves with a one-shot $O(K^2)$ subset search plus a single DC beamformer computation.
  • Larger user pools lower per-round MSE for the proposed policies but not for random selection, so user diversity helps only when scheduling actively picks aligned, high-gain users.
  • Increasing the number of scheduled users $S$ raises aggregation error, making participation count a tunable trade-off against communication-side model quality; more receive antennas relax that trade.
  • Lower MSE and higher test accuracy go together in the reported MNIST experiments, so aggregation-error reduction is itself a learning-performance strategy in low-SNR over-the-air federated learning.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension the paper leaves implicit is applying the same angle-and-modulus criterion per subcarrier in frequency-selective channels, which would broaden the policy to wideband AirComp and could be tested without changing the core optimizer.
  • Because problem (9) is reduced to maximum clique, the greedy policy's suboptimality gap likely grows with the total user count; measuring that gap statistically for large $K$ would indicate when the greedy parameter $G$ must be enlarged.
  • The MSE-to-accuracy link is demonstrated on MNIST only; testing the policy on tasks with larger models and stronger data heterogeneity would show how much of the accuracy gain transfers beyond this benchmark.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper considers over-the-air federated learning (AirComp FL) in low-SNR mobile IoT, where only a subset of users can transmit in each round. It formulates an MSE-minimization problem with fixed selection size and per-user power constraints, reduces it to a max-min beamforming alignment problem (problem (9)), and proposes two solution approaches: an iterative DC-based alternating method (Algorithms 1-2) and a low-complexity channel-based greedy scheduling policy (Policy and Policy_greedy, Algorithm 3). The empirical section reports MSE, CDF, runtime, and MNIST accuracy comparisons against several baselines, claiming that Policy_greedy achieves the best MSE and test accuracy in both i.i.d. and non-i.i.d. settings.

Significance. If the claims were fully supported, the paper would provide a practically useful low-complexity scheduling heuristic for AirComp FL, together with a DC-based beamforming solver. The problem formulation is clean, the transmit coefficient follows prior work [42] without parameter fitting, and the simulation study is reasonably broad: it includes MSE versus K, CDFs over 50 runs, complexity/runtime comparisons, and i.i.d./non-i.i.d. MNIST tasks. The main weaknesses are in the supporting theory: the NP-hardness proof in Appendix A is not a valid reduction, the linear-convergence claim in Theorem 4.1 is a generic performance-estimation argument not tied to the proposed algorithm, and the validation of the greedy proxy in Theorem 5.1 contains a false equality and does not cover the actual selection rule of Algorithm 3. These issues are load-bearing for the paper's stated contributions, though they appear repairable in a revision.

major comments (4)
  1. [Sec. 5, Algorithm 3] The text preceding Algorithm 3 states that 'the selected users subset S is obtained with the largest obj value among the G considered objective values,' and the paper attributes the empirical advantage of Policy_greedy to this selection. However, Algorithm 3 never computes the objective of problem (9) or the equivalent MSE objective of (7). The quantity tmp computed at lines 5, 12, and 18 is max_i min_s |h_s^H h_i|, a pairwise-alignment proxy, and lines 20-22 select the subset with the largest stored value of this proxy. Theorems 5.1 and 5.2 do not establish that this proxy is order-equivalent to the true max-min beamforming objective: Theorem 5.1 is an equal-norm lower bound for a fixed subset, and Theorem 5.2 is a same-direction scaling monotonicity. Neither covers the unequal-norm candidate ranking executed by Algorithm 3. The reported MSE and accuracy advantages of Policy_greedy are therefore not backed by the selection rule as described. The authors should either modify Algorithm 3 to evaluate the true objective of (9) for each candidate subset, prove that the proxy is order-equivalent to (or bounds) the true objective, or present the method explicitly as an unproven heuristic and temper the corresponding claims.
  2. [Sec. 5, Theorem 5.1 proof, Eq. (19)] In the proof of Theorem 5.1, the equality min_{i in S} |h_p^H h_i| = cos alpha in Eq. (19) is generally false. Since alpha is the maximum angle between any two channels in S, every channel satisfies |h_p^H h_i| >= cos alpha, and the minimum is usually strictly larger than cos alpha; the member h_p itself has inner product 1 with itself. The contradiction only needs the inequality min_i |h_p^H h_i| >= cos alpha, so the theorem's conclusion is likely repairable, but the proof as written is incorrect. Please correct this step and state the valid inequality explicitly.
  3. [Sec. 4.1, Theorem 4.1 and Appendix B] Theorem 4.1 claims linear convergence of the DC algorithm, but the proof in Appendix B is a generic performance-estimation argument for a difference-of-convex decomposition with abstract functions f1, f2 and parameters mu1, mu2, L1, L2. The proof never identifies these functions with the iterates of Algorithm 1 or with problem (14), and it assumes the PL inequality for a single-user quadratic f(m)=|m^H h_k|^2, whereas Algorithm 1 solves the max-min problem (10) over all k in S. The PL property for the max-min objective is never verified. As written, the result does not establish linear convergence of the proposed algorithm. Please either provide a proof specific to the algorithm's update rule or state a corrected convergence result with explicitly verified assumptions.
  4. [Appendix A] The claimed NP-hardness proof of problem (9) is not a valid polynomial-time reduction from Maximum Clique. It asserts without proof that problem (9) 'can be transformed' to selecting a subset maximizing pairwise inner products, but Theorems 5.1 and 5.2 provide only bounds, not an equivalence between problem (9) and the pairwise-inner-product objective. The proof then says the threshold tau is adjusted iteratively until |S| = S, which means the construction is not a single polynomial-time mapping, and no correctness or termination proof is provided. The reduction also fails to map clique-size decision instances to fixed-S instances of problem (9). The heuristic methods do not depend on this claim, but the formal NP-hardness statement should be removed or replaced by a correct reduction.
minor comments (4)
  1. [Sec. 3.1, Eqs. (1)-(2)] The desired signal y_des in (2) is the sum over the selected subset S, while the global model update in (1) averages over all K users. Since only S users participate, please clarify the normalization (or state that the scaling factor eta absorbs the 1/|S| factor), as this affects the interpretation of the aggregation error.
  2. [Sec. 6.6] The paragraph on learning performance says 'only the remaining seven methods were included' but then lists six methods ('Iterative', 'Policy', 'Policy_greedy', 'Subgradient', 'Iterative reweighted', and 'RL') and reports seven average MSE/sigma^2 values. Please correct the enumeration and the mapping of values to methods.
  3. [Algorithm 2, line 6] The line 'Select the user with the largest S projection values' should read 'Select the users with the largest S projection values', since a subset of size S is being chosen.
  4. [Table 1 and Sec. 5] There are small typographical errors: 'parametcers' should be 'parameters' in Table 1, and 'lager modulus' should be 'larger modulus' in Remark 5.2.1. Please proofread the final text.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the MSE and accuracy results are self-contained simulations, and the transmit coefficient is lifted from external prior work, not from fitted outputs; the noted Algorithm 3/proof problems are correctness defects rather than circular reasoning.

full rationale

The derivation chain is essentially self-contained with respect to circularity. The optimal transmit coefficient b_k is taken from external [42] and substituted into the MSE problem; the resulting max-min beamforming problem (9) and the DC-based Algorithm 1 are standard reformulations, not predictions that assume their own conclusions. All reported MSE and MNIST accuracy numbers come from direct simulations of the proposed and baseline algorithms, and no parameter is fitted to those outputs, so no fitted input is renamed as a prediction. The self-cited work [39] appears only as a "Random beamforming" comparison baseline and is therefore not load-bearing. The low-complexity policy rests on Theorems 5.1 and 5.2, which are geometric statements about fixed subsets rather than circular uses of the target optimum. Two genuine defects should be flagged, but they are not circularity: (i) the prose before Algorithm 3 says "the selected users subset S is obtained with the largest obj value among the G considered objective values," whereas Algorithm 3's stored val/tmp is the max-min pairwise inner product rather than the problem-(9) objective, so the stated selection rule and the code differ; and (ii) Eq. (19) writes min_i |m^H h_i| = cos alpha for m = h_p, while by definition this minimum is generally strictly larger than cos alpha (the needed inequality still holds). These are internal-validity and proof-correctness concerns, not instances where a claimed prediction is equivalent by construction to an input or to a self-citation chain.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities. The main free parameters are algorithm hyperparameters (G, S) and experiment settings (P, learning rates). The key domain assumptions are the i.i.d. Rayleigh channel model and the unit-variance analog symbol abstraction. The DC algorithm's convergence and rank-one recovery are assumed, not rigorously established.

free parameters (4)
  • G (greedy candidate count) = 5 (default)
    Number of top channel-gain users considered as starting points in Algorithm 3. Chosen by hand; sensitivity sweep in Section 6.5 shows diminishing returns as G grows.
  • S (selected user subset size) = 10 (default); varied 5, 10, 20
    Constraint (5c) fixes the number of selected users. Experiment parameter, not fitted to data.
  • P (max transmit power) = 0 dBm
    Transmit power constraint (5d) in simulations (Section 6).
  • Learning rates = 0.1 (user), 0.5 (aggregator)
    FL training hyperparameters in Section 6.6.
assumptions (5)
  • domain assumption Channel vectors follow i.i.d. complex Gaussian distribution h_k ~ CN(0,I) with block fading constant during transmission.
    Section 3.1. Standard Rayleigh fading model; may not capture low-SNR mobile IoT blockages and tunnels motivating the paper.
  • domain assumption Local model update symbols are normalized to unit variance: E[s_k s_k^H] = I, and the model is transmitted as one analog symbol, ignoring model dimension and frequency selectivity.
    Section 3.1. Common AirComp abstraction; treats the full model vector as a single scalar-like symbol.
  • standard math The optimal transmit coefficient is b_k = h_k^H m / (sqrt(eta) ||m^H h_k||^2), taken from [42].
    Used in Section 3.2 to reduce the problem; borrowed from prior work without proof.
  • ad hoc to paper The DC approximation (14) with v as the leading eigenvector of M_{t-1} converges to a rank-one solution, and the bisection feasibility check correctly identifies feasibility.
    Algorithm 1 depends on this; not proved, only asserted.
  • ad hoc to paper The objective f(m) = |m^H h_k|^2 satisfies the Polyak-Lojasiewicz inequality, yielding linear convergence of the DC method.
    Theorem 4.1 and Appendix B; the condition is assumed, not verified for Algorithm 1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Integrated user scheduling and beam steering in over-the-air federated learning for mobile IoT." pith.science (2026). https://pith.science/paper/FQHNZJ65

@misc{pith2026250800341,
  author       = {Pith},
  title        = {Pith review of: Integrated user scheduling and beam steering in over-the-air federated learning for mobile IoT},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FQHNZJ65}},
  note         = {Machine review of arXiv:2508.00341}
}
read the original abstract

The rising popularity of Internet of things (IoT) has spurred technological advancements in mobile internet and interconnected systems. While offering flexible connectivity and intelligent applications across various domains, IoT service providers must gather vast amounts of sensitive data from users, which nonetheless concomitantly raises concerns about privacy breaches. Federated learning (FL) has emerged as a promising decentralized training paradigm to tackle this challenge. This work focuses on enhancing the aggregation efficiency of distributed local models by introducing over-the-air computation into the FL framework. Due to radio resource scarcity in large-scale networks, only a subset of users can participate in each training round. This highlights the need for effective user scheduling and model transmission strategies to optimize communication efficiency and inference accuracy. To address this, we propose an integrated approach to user scheduling and receive beam steering, subject to constraints on the number of selected users and transmit power. Leveraging the difference-of-convex technique, we decompose the primal non-convex optimization problem into two sub-problems, yielding an iterative solution. While effective, the computational load of the iterative method hampers its practical implementation. To overcome this, we further propose a low-complexity user scheduling policy based on characteristic analysis of the wireless channel to directly determine the user subset without iteration. Extensive experiments validate the superiority of the proposed method in terms of aggregation error and learning performance over existing approaches.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 54 canonical work pages

  1. [42]

    Kai Yang, Tao Jiang, Yuanming Shi, and Zhi Ding. 2020. Federated learning via over-the-air computation. IEEE Trans. Wireless Commun. 19, 3 (Mar. 2020), 2022–2035. ACM Trans. Internet Technol., Vol. 20, No. 12, Article 333. Publication date: December 2025. 333:22 Liu et al

    work page 2020

  2. [1]

    3GPP TR 45.820: Cellular System Support for Ultra-Low Complexity and Low Throughput Internet of Things (Release 13)

    2015. 3GPP TR 45.820: Cellular System Support for Ultra-Low Complexity and Low Throughput Internet of Things (Release 13). Technical Report TR 45.820. 3rd Generation Partnership Project (3GPP). https://portal.3gpp.org/desktopmodules/ Specifications/SpecificationDetails.aspx?specificationId=2702 Version 13.1.0

    work page 2015

  3. [2]

    Hadi Abbaszadehpeivasti, Etienne de Klerk, and Moslem Zamani. 2024. On the rate of convergence of the difference- of-convex algorithm (DCA). J. Optim. Theory Appl. 202, 1 (2024), 475–496

    work page 2024

  4. [3]

    Mohammad Mohammadi Amiri and Deniz Gündüz. 2020. Federated learning over wireless fading channels. IEEE Trans. Wireless Commun. 19, 5 (May 2020), 3546–3557

    work page 2020

  5. [4]

    Jiang Bian, Abdullah Al Arafat, Haoyi Xiong, Jing Li, Li Li, Hongyang Chen, Jun Wang, Dejing Dou, and Zhishan Guo

    work page

  6. [5]

    Yekta Said Can and Cem Ersoy. 2021. Privacy-preserving federated deep learning for wearable IoT-based biomedical monitoring. ACM Trans. Internet Technol. 21, 1, Article 21 (Jan. 2021), 17 pages

    work page 2021

  7. [6]

    Xiaowen Cao, Zhonghao Lyu, Guangxu Zhu, Jie Xu, Lexi Xu, and Shuguang Cui. 2024. An overview on over-the-air federated edge learning. IEEE Wireless Commun. 31, 3 (June 2024), 202–210

    work page 2024

  8. [7]

    Mingzhe Chen, Zhaohui Yang, Walid Saad, Changchuan Yin, H Vincent Poor, and Shuguang Cui. 2020. A joint learning and communications framework for federated learning over wireless networks. IEEE Trans. Wireless Commun. 20, 1 (Jan. 2020), 269–283

    work page 2020

Show all 54 references
  1. [8]

    Shuxiao Chen, Qinqing Zheng, Qi Long, and Weijie J. Su. 2024. Minimax estimation for personalized federated learning: An alternative between FedAvg and local training? J. Mach. Learn. Res. 24, 1 (Mar. 2024), 59 pages

    2024

  2. [9]

    Ingrid Daubechies, Ronald DeVore, Massimo Fornasier, and C Sinan Güntürk. 2010. Iteratively reweighted least squares minimization for sparse recovery. Commun. Pure Appl. Math. 63, 1 (2010), 1–38

    2010

  3. [10]

    Jun Du, Bingqing Jiang, Chunxiao Jiang, Yuanming Shi, and Zhu Han. 2023. Gradient and channel aware dynamic scheduling for over-the-air computation in federated edge learning systems. IEEE J. Sel. Areas Commun. 41, 4 (Apr. 2023), 1035–1050

    2023

  4. [11]

    Liang Gao, Huazhu Fu, Li Li, Yingwen Chen, Ming Xu, and Cheng-Zhong Xu. 2022. FedDC: Federated Learning with Non-IID Data via Local Drift Decoupling and Correction. In Proc. IEEE Comput. Soc. Conf. Comput. Vision Pattern Recognit. (CVPR). 10102–10111

    2022

  5. [12]

    Benjamin Grimmer. 2019. Convergence Rates for Deterministic and Stochastic Subgradient Methods without Lipschitz Continuity. SIAM J. Optim. 29, 2 (2019), 1350–1365

    2019

  6. [13]

    Gregory Gutin and Royal Holloway. 2004. 5.3 Independent sets and cliques. Handbook of graph theory (2004), 389–402

    2004

  7. [14]

    Farzin Haddadpour, Mohammad Mahdi Kamani, Aryan Mokhtari, and Mehrdad Mahdavi. 2021. Federated learning with compression: Unified analysis and sharp guarantees. In International Conference on Artificial Intelligence and Statistics. PMLR, 2350–2358

    2021

  8. [15]

    Jenny Hamer, Mehryar Mohri, and Ananda Theertha Suresh. 2020. FedBoost: A Communication-Efficient Algorithm for Federated Learning. In Proc. 37th Int. Conf. Machin. Learn. (ICML) , Vol. 119. 3973–3983

    2020

  9. [16]

    Henrik Hellström, Viktoria Fodor, and Carlo Fischione. 2023. Federated learning over-the-air by retransmissions. IEEE Trans. Wireless Commun. 22, 12 (Dec. 2023), 9143–9156. ACM Trans. Internet Technol., Vol. 20, No. 12, Article 333. Publication date: December 2025. Integrated ...

    2023

  10. [17]

    Lee Swindlehurst, and Daeyoung Park

    Minsik Kim, A. Lee Swindlehurst, and Daeyoung Park. 2023. Beamforming vector design and device selection in over-the-air federated learning. IEEE Trans. Wireless Commun. 22, 11 (Nov. 2023), 7464–7477

    2023

  11. [18]

    Hongtao Li, Longyao Ran, Shengyao Chen, Zhiyong Cheng, and Feng Xi. 2024. Antenna Selection and Receive Beamforming for Multi-Functional Sparse Linear Array via Consensus ADMM. IEEE Trans. Veh. Technol. 73, 7 (July 2024), 10435–10450

    2024

  12. [19]

    Yang Li, Yuan Wu, Yuxiao Song, Liping Qian, and Weijia Jia. 2023. Dynamic user-scheduling and power allocation for SWIPT aided federated learning: A deep learning approach. IEEE Trans. Mob. Comput. 22, 12 (Dec. 2023), 6956–6969

    2023

  13. [20]

    Zhenyi Lin, Xiaoyang Li, Vincent K. N. Lau, Yi Gong, and Kaibin Huang. 2022. Deploying federated learning in large-scale cellular networks: Spatial convergence analysis. IEEE Trans. Wireless Commun. (Mar. 2022)

    2022

  14. [21]

    Su Liu, Jiong Yu, Xiaoheng Deng, and Shaohua Wan. 2022. FedCPF: An efficient-communication federated learning approach for vehicular edge computing in 6G communication networks. IEEE Trans. Intell. Transp. Syst. 23, 2 (Feb. 2022), 1616–1629

    2022

  15. [22]

    Canyi Lu, Jinhui Tang, Shuicheng Yan, and Zhouchen Lin. 2016. Nonconvex Nonsmooth Low Rank Minimization via Iteratively Reweighted Nuclear Norm. IEEE Trans. Image Process. 25, 2 (Feb. 2016), 829–839

    2016

  16. [23]

    Sidiropoulos, Paul Tseng, and Shuzhong Zhang

    Zhi-Quan Luo, Nicholas D. Sidiropoulos, Paul Tseng, and Shuzhong Zhang. 2007. Approximation Bounds for Quadratic Optimization with Homogeneous Quadratic Constraints. SIAM J. Optim. 18, 1 (Feb. 2007), 1–28

    2007

  17. [24]

    Common Sense Media. 2022. Kids and the Metaverse: What parents, policymakers, and companies need to know. https://www.commonsensemedia.org/sites/default/files/featured-content/files/metaverse-white-paper.pdf

    2022

  18. [25]

    Mohammad Mohammadi Amiri and Deniz Gündüz. 2020. Machine learning at the wireless edge: Distributed stochastic gradient descent over-the-air. IEEE Trans. Signal Process. 68 (Mar. 2020), 2155–2169

    2020

  19. [26]

    Truc Nguyen and My T. Thai. 2023. Preserving privacy and security in federated learning. IEEE/ACM Trans. Netw. 32, 1 (Aug. 2023), 833–843

    2023

  20. [27]

    Jing Qiao, Shikun Shen, Shuzhen Chen, Xiao Zhang, Tian Lan, Xiuzhen Cheng, and Dongxiao Yu. 2023. Communication resources limited decentralized learning with privacy guarantee through over-the-air computation. In Proc. 24th Int. Symp. Mobile Ad Hoc Networking Comput. (MobiHoc ...

    2023

  21. [28]

    Tomer Sery, Nir Shlezinger, Kobi Cohen, and Yonina C. Eldar. 2020. COTAF: Convergent over-the-air federated learning. In Proc. IEEE Glob. Commun. Conf. (GLOBECOM) . 1–6

    2020

  22. [29]

    Suhail Mohmad Shah and Vincent K. N. Lau. 2023. Model compression for communication efficient federated learning. IEEE Trans. Neural Networks Learn. Syst. 34, 9 (Sept. 2023), 5937–5951

    2023

  23. [30]

    W. Shi, S. Zhou, and Z. Niu. 2020. Device scheduling with fast convergence for wireless federated learning. In Proc. IEEE Int. Conf. Commun. (ICC) . 1–6

    2020

  24. [31]

    Sebastian U. Stich. 2019. Local SGD converges fast and communicates little. In Proc. 7th Int. Conf. Learn. Represent. (ICLR). 1–17

    2019

  25. [32]

    Mainul Haque, Li Chen, Fei Xu, and Xu Yuan

    Abeda Sultana, Md. Mainul Haque, Li Chen, Fei Xu, and Xu Yuan. 2022. Eiffel: Efficient and fair scheduling in adaptive federated learning. IEEE Trans. Parallel Distrib. Syst. 33, 12 (Dec. 2022), 4282–4294

    2022

  26. [33]

    Juncheng Wang, Ben Liang, Min Dong, Gary Boudreau, and Hatem Abou-Zeid. 2024. Joint online optimization of model training and analog aggregation for wireless edge learning. IEEE/ACM Trans. Networking 32, 2 (Apr. 2024), 1212–1228

    2024

  27. [34]

    Tian Wang, Md Zakirul Alam Bhuiyan, Guojun Wang, Lianyong Qi, Jie Wu, and Thaier Hayajneh. 2020. Preserving balance between privacy and data integrity in edge-assisted Internet of things. IEEE Internet Things J. 7, 4 (Apr. 2020), 2679–2689

    2020

  28. [35]

    Wenqi Wei, Ka-Ho Chow, Yanzhao Wu, and Ling Liu. 2024. Demystifying data poisoning attacks in distributed learning as a service. IEEE Trans. Serv. Comput. 17, 1 (Jan. 2024), 237–250

    2024

  29. [36]

    Xueyu Wu, Xin Yao, and Cho-Li Wang. 2021. FedSCR: Structure-based communication reduction for federated learning. IEEE Trans. Parallel Distrib. Syst. 32, 7 (July 2021), 1565–1577

    2021

  30. [37]

    Wenchao Xia, Tony QS Quek, Kun Guo, Wanli Wen, Howard H Yang, and Hongbo Zhu. 2020. Multi-armed bandit-based client scheduling for federated learning. IEEE Trans. Wireless Commun. 19, 11 (Nov. 2020), 7108–7123

    2020

  31. [38]

    Quek, Jun Zhang, and Hongbo Zhu

    Wenchao Xia, Wanli Wen, Kai-Kit Wong, Tony Q.S. Quek, Jun Zhang, and Hongbo Zhu. 2021. Federated-learning-based client scheduling for low-latency wireless communications. IEEE Wireless Commun. 28, 2 (Apr. 2021), 32–38

    2021

  32. [39]

    Chunmei Xu, Shengheng Liu, Yongming Huang, Bjorn Ottersten, and Dusit Niyato. 2024. Random aggregate beam- forming for over-the-air federated learning in large-scale networks. arXiv preprint arXiv:2403.18946 (2024)

    2024 arXiv

  33. [40]

    Chunmei Xu, Shengheng Liu, Zhaohui Yang, Yongming Huang, and Kai-Kit Wong. 2021. Learning rate optimization for federated learning exploiting over-the-air computation. IEEE J. Sel. Areas Commun. 39, 12 (Dec. 2021), 3742–3756

    2021

  34. [41]

    Jinlai Xu and Balaji Palanisamy. 2021. Optimized contract-based model for resource allocation in federated geo- distributed clouds. IEEE Trans. Serv. Comput. 14, 2 (Mar. 2021), 530–543

    2021

  35. [43]

    Kai Yang, Yuanming Shi, Yong Zhou, Zhanpeng Yang, Liqun Fu, and Wei Chen. 2020. Federated machine learning for intelligent IoT via reconfigurable intelligent surface. IEEE Network 34, 5 (Sept./Oct. 2020), 16–22

    2020

  36. [44]

    Naifu Zhang and Meixia Tao. 2021. Gradient statistics aware power control for over-the-air federated learning. IEEE Trans. Wireless Commun. 20, 8 (Aug. 2021), 5115–5128

    2021

  37. [45]

    Qinnan Zhang, Zehui Xiong, Jianming Zhu, Sheng Gao, and Wanting Yang. 2024. A privacy-preserving auction mechanism for learning model as an NFT in blockchain-driven metaverse. ACM Trans. Multimedia Comput. Commun. Appl. 20, 7 (Mar. 2024), 24 pages

    2024

  38. [46]

    Tinghao Zhang, Kwok-Yan Lam, Jun Zhao, and Jie Feng. 2025. Joint Device Scheduling and Bandwidth Allocation for Federated Learning Over Wireless Networks. IEEE Trans. Wireless Commun. 24, 1 (Jan. 2025), 3–18

    2025

  39. [47]

    Yang, Debin Liu, Wanli Lu, Chenlu Zhu, and Yiheng Ruan

    Ruonan Zhao, Laurence T. Yang, Debin Liu, Wanli Lu, Chenlu Zhu, and Yiheng Ruan. 2023. Tensor-empowered LSTM for communication-efficient and privacy-enhanced cognitive federated learning in intelligent transportation systems. ACM Trans. Multimedia Comput. Commun. Appl. 20, 2 (...

    2023

  40. [48]

    Zhongyuan Zhao, Chenyuan Feng, Wei Hong, Jiamo Jiang, Chao Jia, Tony Q. S. Quek, and Mugen Peng. 2022. Federated learning with non-IID data in wireless networks. IEEE Trans. Wireless Commun. 21, 3 (Sept. 2022), 1927–1942

    2022

  41. [49]

    Yuhao Zhou, Qing Ye, and Jiancheng Lv. 2022. Communication-efficient federated learning with compensated overlap- FedAvg. IEEE Trans. Parallel Distrib. Syst. 33, 1 (Jan. 2022), 192–205

    2022

  42. [50]

    Guangxu Zhu, Yuqing Du, Deniz Gündüz, and Kaibin Huang. 2021. One-bit over-the-air aggregation for communication- efficient federated edge learning: Design and convergence analysis. IEEE Trans. Wireless Commun. 20, 3 (Aug. 2021), 2120–2135

    2021

  43. [51]

    G. Zhu, Y. Wang, and K. Huang. 2020. Broadband analog aggregation for low-latency federated edge learning. IEEE Trans. Wireless Commun. 19, 1 (Jan. 2020), 491–506

    2020

  44. [52]

    Guangxu Zhu, Jie Xu, Kaibin Huang, and Shuguang Cui. 2021. Over-the-air computing for wireless data aggregation in massive IoT. IEEE Wireless Commun. 28, 4 (Aug. 2021), 57–65

    2021

  45. [53]

    Alphan Şahin and Rui Yang. 2023. A survey on over-the-air computation. IEEE Commun. Surv. Tutorials 25, 3 (Thirdquarter 2023), 1877–1908. A NP-HARDNESS PROOF OF PROBLEM (9) We establish the NP-hardness proof of Problem (9) via a reduction from the maximum clique problem, which...

    2023

  46. [2022]

    IEEE Internet Things J

    Machine learning in real-time Internet of things (IoT) systems: A survey. IEEE Internet Things J. 9, 11 (June 2022), 8364–8386

    2022

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.