Pith. sign in

REVIEW 4 major objections 5 minor 49 references

Optimal Batch-Size Control for Low-Latency Federated Learning with Device Heterogeneity

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper derives a closed-form rule that chooses global and per-device batch sizes to minimize end-to-end federated learning latency while preserving convergence accuracy.

desk verdict A solid systems paper that turns an intractable MINLP into a closed-form batch-size rule, with the caveat that 'optimal' holds for the fitted surrogate, not the true learning curve. read the letter →

arxiv 2507.15601 v2 pith:JJKQZSJ7 submitted 2025-07-21 cs.LG cs.SYeess.SY

Huiling Yang , Zhanwei Wang , Kaibin Huang This is my paper
classification cs.LGcs.SYeess.SY
keywords federatedlearningbatch-sizecontrolend-to-endlatencycommunication-computationtradeoffdeviceheterogeneityconvergenceanalysisslowfadingfast
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 show that the wall-clock time needed to train a federated model over heterogeneous wireless devices can be minimized by a closed-form batch-size rule, without solving the NP-hard mixed-integer program that the problem first appears to be. The argument rests on a communication-and-computation tradeoff: larger global batches sharpen gradient estimates and reduce the number of communication rounds, but they lengthen every round. The authors derive a round-batch scaling law from convergence analysis, fit its two parameters to real training data, and then prove that the optimal global batch size is the larger of a threshold and a simple closed-form point, with per-device batches given by an explicit linear allocation. Experiments on MNIST and CIFAR-10 report that this controller reaches the same accuracy thresholds faster than fixed, uniform, and exponentially growing batch-size schemes, with reported latency reductions of 26.7% and 37.5% over the no-local-adaptation baseline.

What carries the argument

The load-bearing object is the round-batch scaling law $N_{\epsilon}(B) = \hat{\alpha}/(\epsilon - \hat{\beta}/B)$, a two-parameter surrogate for the number of communication rounds required to reach a target accuracy as a function of the global batch size. It is fit by least squares to measured convergence-round data, and it converts the intractable mixed-integer problem into a unimodal one-dimensional minimization. The second mechanism is the latency equilibrium principle: for a fixed global batch size, per-round latency is minimized when every device's computation-plus-communication time is equal, so the optimal local batch size is the linear allocation that equalizes device latencies. Together these generate the continuous surrogate $\tilde{\psi}(B)$ whose unique minimizer is $B_{\epsilon}$, and the threshold $B_{\mathrm{th}}$ below which no equalizing allocation exists.

What would settle it

Run a dense sweep of global batch sizes on a fresh task, fit the two surrogate parameters using only a subset of low batch sizes, compute the closed-form $B^{*}$, and compare measured end-to-end latency at $B^{*}$ with latencies at neighboring values; if any neighbor is faster, or if the fitted law misses measured convergence rounds by more than the data noise near $B^{*}$, the optimality claim fails.

Watch

Extended reading notes

Core claim

Under slow-fading channels, the paper's Theorem 2 states that the optimal global batch size is $B^{*} = \max\{B_{\mathrm{th}}, \lfloor B_{\epsilon}\rceil_{\tilde{\psi}(\cdot)}\}$, where $B_{\epsilon} = \frac{\hat{\beta}}{\epsilon}\left(1+\sqrt{1+\frac{\hat{f}_{\Sigma}\epsilon}{H W \hat{\beta}}}\right)$ and the ceiling-floor notation means the integer neighbor of $B_{\epsilon}$ that gives the smaller surrogate latency. The matching device-level allocation is $b^{*}_{k,n} = \mathrm{round}\!\left(\frac{f_k}{H W}\left(\frac{H W B^{*} + \hat{f}_{\Sigma}}{f_{\Sigma}} - T^{\mathrm{cmm}}_k\right)\right)$, which drives every device to finish its round at the same time---the latency equilibrium principle that Lemma 1 proves optimal. The global formula balances the round saving from larger batches against the per-round computation they add, using the fitted scaling law $N_{\epsilon}(B) = \hat{\alpha}/(\epsilon - \hat{\beta}/B)$ as the surrogate for convergence speed. For fast fading, the same construction is evaluated at expected channel statistics and then raised to the instantaneous channel-aware threshold each round. The paper's central claim is that this two-part control is the latency minimizer under the fitted surrogate, and that the surrogate matches measured convergence behavior closely enough that the closed-form point lands on the empirical latency minimum.

Load-bearing premise

The entire optimality result depends on the fitted curve that predicts how many training rounds are needed for each batch size staying correct for all feasible batch sizes, especially the one the formula picks; if the curve bends differently outside the sampled points, the claimed optimum is not guaranteed for the real system.

Editorial extensions

If this is right

  • With slow-fading channels, the closed-form $B^{*}$ and $b^{*}_{k,n}$ compute in $O(K)$ arithmetic, replacing an NP-hard MINLP with a rule that can run in real time.
  • With fast-fading channels, setting $B^{*}_n = \max\{\bar{B}^{*}, B^{\mathrm{th}}_n\}$ each round keeps per-round latency at the straggler floor while using the extra samples to reduce the number of rounds, lowering E2E latency.
  • If bandwidth rises, the optimal global batch size falls; if device computation speed rises, the optimal global batch size grows, giving operators a direct knob for tuning the communication-computation tradeoff.
  • Reported results compare with fixed, uniform, and DBFL baselines, with the proposed adaptive control reaching the accuracy threshold with 26.7% (MNIST) and 37.5% (CIFAR-10) less E2E latency than the no-local-adaptation scheme.

Reading between the lines

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

  • The same surrogate-plus-equilibrium template should transfer to other round-level controls, such as gradient compression ratio, number of local epochs, or quantization bits, wherever a larger control value improves gradient quality but lengthens per-round time; the paper's communication-computation tradeoff is one instance of that family.
  • Because $\hat{\alpha}$ and $\hat{\beta}$ are fitted per task, a practical deployment would need a cheap online estimator of these two parameters from early training rounds rather than a full offline sweep; the paper does not develop this, but its own data-fitting step makes it a natural next step.
  • A natural extension is to treat the fast-fading rule as a stochastic online control problem and seek a regret guarantee relative to the clairvoyant channel-sequence optimum, which the paper does not provide.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes a communication-and-computation (C2)-aware batch-size control framework for synchronous federated learning over heterogeneous devices. The authors derive a non-convex convergence upper bound (Theorem 1), rearrange it into a round-batch scaling law N_epsilon(B) = alpha/(epsilon - beta/B) (Corollary 1), and fit alpha and beta by least squares to empirical convergence-round data on MNIST and CIFAR-10. They then formulate an E2E latency minimization problem, relax integrality, prove a latency-equilibrium lemma for per-device allocation, and derive a closed-form optimal global batch size (Theorem 2) for slow-fading channels. For fast-fading channels they present an adaptive protocol that recomputes a batch-size threshold from instantaneous communication latencies. Experiments compare the proposed scheme with fixed-batch baselines, DBFL, and a no-local-adaptation variant.

Significance. If the surrogate and the relaxations were fully justified, the paper would offer a practically attractive result: an O(K) closed-form batch-size control policy with an explicit C2 tradeoff and consistent experimental latency reductions. The convergence analysis follows a standard non-convex SGD template, and the latency-equilibrium lemma is sound for the continuous relaxation. The main contribution, however, is a surrogate-based optimization procedure rather than a proven optimal policy for the original MINLP: the optimality of Theorem 2 depends on an in-sample fitted functional form, and the integer-rounding gap is unquantified. These issues are load-bearing for the central claim, but they are addressable within the manuscript's scope.

major comments (4)
  1. [§IV-A, Eq. (18); §IV-C, Theorem 2] The central optimality claim is established only for the surrogate psi-tilde(B) in (31), whose parameters hat-alpha and hat-beta are least-squares fitted to the same tasks used later for validation. The functional form N_epsilon(B) = alpha/(epsilon - beta/B) is a rearrangement of the Theorem 1 upper bound, but the actual parameter values are free, and Fig. 2 reports only an in-sample fit: no residuals, confidence intervals, or hold-out checks are given. Because B_epsilon in (55) is proportional to the pole hat-beta/epsilon, errors in the fit or in the assumed functional form translate directly into a biased optimum, and the pole lies in the steep region of the fitted curve for the reported MNIST parameters. The manuscript should either provide a robustness bound on B* under a bounded surrogate error or validate the scaling law out-of-sample over a grid that contains the predicted B*, with error bars.
  2. [§IV-C, Eq. (34)] Rounding each b*_{k,n} to the nearest integer can violate the global batch-sum constraint (20c), because sum_k round(x_k) need not equal B*. The claim that the induced performance gap is negligible rests on experimental observation rather than a mathematical bound, and Theorem 2's proof does not address the integer MINLP formulation of Problem (20). A feasibility-correction step or an explicit approximation bound is needed before the closed-form allocation can be called optimal for the original problem.
  3. [§V-B, Algorithm 1] The adaptive protocol changes the global batch size B_n in every round, whereas Theorem 1 and Corollary 1 assume a fixed global batch size B_n = B for all rounds. Consequently, Eqs. (21), (31), and the optimality of B*_n = max(bar-B*, Bth_n) in (39) are not justified by the convergence analysis. The fast-fading extension may be a reasonable heuristic, but the introduction and conclusion present it as part of the optimal-control result; the paper should clearly label it as heuristic or provide a convergence analysis that covers time-varying batch sizes.
  4. [§IV-C, Eqs. (30)-(33)] Theorem 2 optimizes the continuous relaxation psi-tilde(B), not the exact integer objective psi(B) in (30) with its ceiling operation. Fig. 4 shows a close match for two displayed scenarios, but no general bound is given for the gap between the relaxed minimizer and the true integer minimizer of Problem (20). Since the original problem is MINLP and claimed to be NP-hard, the closed form requires either an exactness proof or a quantified integrality gap before it can be stated as the solution of Problem (20).
minor comments (5)
  1. [§IV-C, Eq. (24) and surrounding text] The definition of Bth uses a floor operator in (24), while the continuous threshold tilde-Bth in (31) is real-valued; the text should state explicitly whether the piecewise switch in Fig. 3 and the condition B > Bth refer to the integer threshold or its continuous relaxation, otherwise the two are easily conflated.
  2. [Appendix B, proof of Lemma 1] In the contradiction argument, the perturbed allocation tilde-b'_ki = tilde-b*_ki +/- delta may leave the feasible interval [1, B-K+1]; adding a short feasibility check would make the proof complete.
  3. [§VI-A, experimental settings] The convergence tolerance epsilon = 0.5 is set alongside validation-accuracy thresholds epsilon_A = 95% and 58%, but the relationship between the gradient-norm threshold epsilon in the analysis and the validation-accuracy proxy epsilon_A is never stated; this connection should be made explicit.
  4. [§II-B, Eq. (5)] The symbol W denotes per-sample computational workload, which is natural, but the same letter is reused for the model dimension in the text around Eq. (2); using distinct symbols for workload and model dimension would avoid confusion.
  5. [§VI-C, Fig. 8] The DBFL baseline is described as having a 'stiff incremental factor', but the paper does not specify how the DBFL parameters are chosen in the experiments; a sentence describing the configuration would improve reproducibility.

Circularity Check

1 steps flagged · score 4.0 of 10

The closed-form optimum in Theorem 2 minimizes the E2E latency objective only under the least-squares fitted round-batch law; the Fig. 7 'validation' uses the same in-sample measurements that produced the fit, so the headline optimality is conditional on a fitted surrogate rather than an independently predicted empirical quantity.

  1. fitted input called prediction [Section IV-A (Eq. 18) and Section VI-B (Fig. 7)]
    "Building on the collected data samples {B(m), N(m) εA }, we obtain the estimates of α and β in (17), denoted as ˆα and ˆβ, respectively... ˆα, ˆβ = arg min ... (18) ... Fig. 7 ... E2E latency vs. global batch size by empirical measurements, theoretical predictions from the surrogate model in (31), and the analytically derived optimum point via (33). ..."

    α̂ and β̂ are least-squares fitted to measured convergence rounds N^{(m)}_{εA}(B^{(m)}) for the same task and accuracy threshold (Eq. 18). The E2E objective ψ(B)=Nε(B)·τmin(B) in Eqs. (30)-(31) is then built from this fitted law Nε(B)=α̂/(ε-β̂/B), and Theorem 2's B* in Eq. (33) is the minimizer of that fitted objective. The Fig. 7 'empirical measurements' are the same Nε data used in the fit (same datasets, same thresholds), so the agreement shown is in-sample consistency, not an independent prediction. Because Bε is proportional to the surrogate pole β̂/ε, small errors in α̂/β̂ or in the assumed 1/(ε-β/B) shape directly determine the claimed optimum; no hold-out or residual analysis quantifies this.

full rationale

The paper's convergence analysis (Theorem 1) and the per-round latency equilibrium allocation (Lemma 1) are derived independently and are not circular: Theorem 1 is a standard SGD upper bound with explicit assumptions, and Lemma 1 is a min-max latency argument. The circularity concern is confined to how the round-batch law Nε(B)=α/(ε-β/B) is used. Corollary 1 obtains this functional form by rearranging the Theorem 1 upper bound, but α and β are then treated as unknown and least-squares fitted to measured convergence rounds from the same MNIST/CIFAR tasks (Eq. 18). Theorem 2 and the adaptive algorithms then optimize an E2E objective whose Nε(B) factor is exactly this fitted law, and Fig. 7 presents the resulting curve as 'theoretical predictions' validated against the same fitted data. Thus the central 'optimal batch size' claim is statistically forced by the fitted input rather than being an independent prediction of an unmeasured quantity. The paper is transparent that the parameters are data-driven, and no load-bearing self-citation (e.g., ref. [43] for continuous relaxation) is involved, so this is partial circularity rather than a fully self-referential derivation.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The analytical core is a standard non-convex SGD convergence bound (Assumptions 1 to 4) plus a min-max latency equalization lemma. The non-standard load-bearing input is the empirical fit of the round-batch scaling law and the continuous relaxation with integer rounding, which are not derived from first principles.

free parameters (2)
  • alpha (convergence prefactor) = 34.5 (MNIST), 30.0 (CIFAR-10)
    Coefficient in N_epsilon(B)=alpha/(epsilon-beta/B); fitted by least squares in Eq. 18 using validation-accuracy thresholds; depends on task, model, and learning rate.
  • beta (batch-size sensitivity) = 23.2 (MNIST), 123.3 (CIFAR-10)
    Coefficient controlling how much global batch size reduces convergence rounds; fitted to data; directly enters B_epsilon and the optimal B*.
assumptions (6)
  • standard math The loss function is L-smooth (Assumption 1).
    Section III-A, Assumption 1; used in the descent lemma for Theorem 1.
  • standard math The loss function is lower bounded (Assumption 2).
    Section III-A, Assumption 2; used to telescope the potential function in the proof of Theorem 1.
  • domain assumption Local mini-batch gradients are unbiased estimates of the global gradient (Assumption 3).
    Section III-A, Assumption 3; requires each device's data distribution to match the global distribution, which fails for non-i.i.d. FL; experiments partition data i.i.d.
  • domain assumption Gradient variance is bounded by sigma^2 / b (Assumption 4).
    Section III-A, Assumption 4; the inverse-batch-size variance scaling produces the beta/B term in the convergence law.
  • ad hoc to paper The fitted surrogate law N_epsilon(B)=alpha/(epsilon-beta/B) holds for all feasible B, including unsampled values.
    After fitting alpha and beta on sampled B values, the paper treats the law as exact when forming psi(B) in Eq. 30 and when deriving B*.
  • ad hoc to paper Rounding the continuous relaxed batch allocations to integers has negligible effect on optimality.
    The footnote after Eq. 27 and Eq. 34 assert the gap is negligible without a formal bound.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimal Batch-Size Control for Low-Latency Federated Learning with Device Heterogeneity." pith.science (2026). https://pith.science/paper/JJKQZSJ7

@misc{pith2026250715601,
  author       = {Pith},
  title        = {Pith review of: Optimal Batch-Size Control for Low-Latency Federated Learning with Device Heterogeneity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JJKQZSJ7}},
  note         = {Machine review of arXiv:2507.15601}
}
abstract

Federated learning (FL) has emerged as a popular approach for collaborative machine learning in sixth-generation (6G) networks, primarily due to its privacy-preserving capabilities. The deployment of FL algorithms is expected to empower a wide range of Internet-of-Things (IoT) applications, e.g., autonomous driving, augmented reality, and healthcare. The mission-critical and time-sensitive nature of these applications necessitates the design of low-latency FL frameworks that guarantee high learning performance. In practice, achieving low-latency FL faces two challenges: the overhead of computing and transmitting high-dimensional model updates, and the heterogeneity in communication-and-computation (C$^2$) capabilities across devices. To address these challenges, we propose a novel C$^2$-aware framework for optimal batch-size control that minimizes end-to-end (E2E) learning latency while ensuring convergence. The framework is designed to balance a fundamental C$^2$ tradeoff as revealed through convergence analysis. Specifically, increasing batch sizes improves the accuracy of gradient estimation in FL and thus reduces the number of communication rounds required for convergence, but results in higher per-round latency, and vice versa. The associated problem of latency minimization is intractable; however, we solve it by designing an accurate and tractable surrogate for convergence speed, with parameters fitted to real data. This approach yields two batch-size control strategies tailored to scenarios with slow and fast fading, while also accommodating device heterogeneity. Extensive experiments using real datasets demonstrate that the proposed strategies outperform conventional batch-size adaptation schemes that do not consider the C$^2$ tradeoff or device heterogeneity.

Figures

Figures reproduced from arXiv: 2507.15601 by the authors.

Figure 1
Figure 1. FL system with synchronous model updates. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Convergence round vs. global batch size relationship, validated [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Per-round latency comparison of proposed allocation method with [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Two C2 tradeoff regimes through comparison of E2E latency ψ(B) and its continuous relaxation ψ˜(B) under different bandwidth configurations (CNN model, MNIST dataset). computing device as B grows. These results demonstrate the superior latency performance of the propos…
Figure 5
Figure 5. Figure 5: Adaptive batch-size control mechanisms. T cmm k receive proportionally larger batches, avoiding resource waste when waiting for stragglers. Since both B∗ and {b ∗ k,n} are closed-form expressions requiring only simple arithmetic across the K devices, the total computat…
Figure 8
Figure 8. Figure 8: Learning performance comparison of different batch control schemes [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 7
Figure 7. Figure 7: The illustration of the C2 tradeoff: E2E latency vs. global batch size by empirical measurements, theoretical predictions from the surrogate model in (31), and the analytically derived optimum point via (33). We conduct H = 5 local updates per round for a CNN on MNIST …
Figure 10
Figure 10. Figure 10: Optimal global batch size vs. computation speed scaling factor [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

49 extracted references · 38 canonical work pages

  1. [1]

    A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,

    W. Saad, M. Bennis, and M. Chen, “A vision of 6G wireless systems: Applications, trends, technologies, and open research problems,” IEEE Netw., vol. 34, no. 3, pp. 134–142, 2019

    work page 2019

  2. [2]

    Integrated sensing and edge AI: Realizing intelligent perception in 6G,

    Z. Liu, X. Chen, H. Wu, Z. Wang, X. Chen, D. Niyato, and K. Huang, “Integrated sensing and edge AI: Realizing intelligent perception in 6G,” arXiv preprint arXiv:2501.06726 , 2025

    arXiv 2025

  3. [3]

    Space–ground fluid AI for 6G edge intelligence,

    Q. Chen, Z. Wang, X. Chen, J. Wen, D. Zhou, S. Ji, M. Sheng, and K. Huang, “Space–ground fluid AI for 6G edge intelligence,” Eng

    work page

  4. [4]

    Communication-efficient learning of deep networks from decentralized data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Proc. Int. Conf. Artif. Intell. Statist. (AISTATS) . PMLR, 2017, pp. 1273–1282

    work page 2017

  5. [5]

    Federated machine learning: Concept and applications,

    Q. Yang, Y . Liu, T. Chen, and Y . Tong, “Federated machine learning: Concept and applications,” ACM Trans. Intell. Syst. Technol. , vol. 10, no. 2, pp. 1–19, 2019

    2019

  6. [6]

    Advances and open problems in federated learning,

    P. Kairouz, H. B. McMahan, B. Avent, A. Bellet, M. Bennis, A. N. Bhagoji, K. Bonawitz, Z. Charles, G. Cormode, R. Cummings et al. , “Advances and open problems in federated learning,” Found. Trends Mach. Learn., vol. 14, no. 1–2, pp. 1–210, 2021. 14

    work page 2021

  7. [7]

    Fedhome: Cloud-edge based personalized federated learning for in-home health monitoring,

    Q. Wu, X. Chen, Z. Zhou, and J. Zhang, “Fedhome: Cloud-edge based personalized federated learning for in-home health monitoring,” IEEE Trans. Mobile Comput. , vol. 21, no. 8, pp. 2818–2832, 2020

    work page 2020

  8. [8]

    GeFL: Gradient encryption- aided privacy preserved federated learning for autonomous vehicles,

    R. Parekh, N. Patel, R. Gupta, N. K. Jadav, S. Tanwar, A. Alharbi, A. Tolba, B.-C. Neagu, and M. S. Raboaca, “GeFL: Gradient encryption- aided privacy preserved federated learning for autonomous vehicles,” IEEE Access, vol. 11, pp. 1825–1839, 2023

    work page 2023

Show all 49 references
  1. [9]

    Federated learning: Strategies for improving communication efficiency,

    J. Kone ˇcn`y, H. B. McMahan, F. X. Yu, P. Richt ´arik, A. T. Suresh, and D. Bacon, “Federated learning: Strategies for improving communication efficiency,” arXiv preprint arXiv:1610.05492 , 2016

    2016 arXiv

  2. [10]

    Model pruning enables efficient federated learning on edge devices,

    Y . Jiang, S. Wang, V . Valls, B. J. Ko, W.-H. Lee, K. K. Leung, and L. Tassiulas, “Model pruning enables efficient federated learning on edge devices,” IEEE Trans. Neural Netw. Learn. Syst. , vol. 34, no. 12, pp. 10 374–10 386, 2022

    2022

  3. [11]

    Deploying federated learning in large-scale cellular networks: Spatial convergence analysis,

    Z. Lin, X. Li, V . K. Lau, Y . Gong, and K. Huang, “Deploying federated learning in large-scale cellular networks: Spatial convergence analysis,” IEEE Trans. Wireless Commun. , vol. 21, no. 3, pp. 1542–1556, 2021

    2021

  4. [12]

    UVeQFed: Universal vector quantization for federated learning,

    N. Shlezinger, M. Chen, Y . C. Eldar, H. V . Poor, and S. Cui, “UVeQFed: Universal vector quantization for federated learning,”IEEE Trans. Signal Process., vol. 69, pp. 500–514, 2020

    2020

  5. [13]

    Communication-efficient federated learning with dual-side low-rank compression,

    Z. Qiao, X. Yu, J. Zhang, and K. B. Letaief, “Communication-efficient federated learning with dual-side low-rank compression,” arXiv preprint arXiv:2104.12416, 2021

    2021 arXiv

  6. [14]

    Splitfed: When federated learning meets split learning,

    C. Thapa, P. C. M. Arachchige, S. Camtepe, and L. Sun, “Splitfed: When federated learning meets split learning,” inProc. AAAI Conf. Artif. Intell., vol. 36, no. 8, 2022, pp. 8485–8493

    2022

  7. [15]

    Broadband analog aggregation for low-latency federated edge learning,

    G. Zhu, Y . Wang, and K. Huang, “Broadband analog aggregation for low-latency federated edge learning,” IEEE Trans. Wireless Commun. , vol. 19, no. 1, pp. 491–506, 2019

    2019

  8. [16]

    Federated learning via over- the-air computation,

    K. Yang, T. Jiang, Y . Shi, and Z. Ding, “Federated learning via over- the-air computation,” IEEE Trans. Wireless Commun., vol. 19, no. 3, pp. 2022–2035, 2020

    2022

  9. [17]

    Spectrum breathing: Protecting over-the-air federated learning against interference,

    Z. Wang, K. Huang, and Y . C. Eldar, “Spectrum breathing: Protecting over-the-air federated learning against interference,” IEEE Trans. Wire- less Commun., 2024

    2024

  10. [18]

    Airbreath sensing: Protecting over-the-air distributed sensing against interference,

    Z. Wang, M. Cui, H. Yang, Q. Zeng, M. Sheng, and K. Huang, “Airbreath sensing: Protecting over-the-air distributed sensing against interference,” arXiv preprint arXiv:2508.11267 , 2025

    2025

  11. [19]

    Federated learning over wireless fading channels,

    M. M. Amiri and D. G ¨und¨uz, “Federated learning over wireless fading channels,” IEEE Trans. Wireless Commun. , vol. 19, no. 5, pp. 3546– 3557, 2020

    2020

  12. [20]

    Accurate, large minibatch SGD: Training ImageNet in 1 hour,

    P. Goyal, P. Doll ´ar, R. Girshick, P. Noordhuis, L. Wesolowski, A. Kyrola, A. Tulloch, Y . Jia, and K. He, “Accurate, large minibatch SGD: Training ImageNet in 1 hour,” arXiv preprint arXiv:1706.02677 , 2017

    2017 arXiv

  13. [21]

    To talk or to work: Dynamic batch sizes assisted time efficient federated learning over future mobile edge devices,

    D. Shi, L. Li, M. Wu, M. Shu, R. Yu, M. Pan, and Z. Han, “To talk or to work: Dynamic batch sizes assisted time efficient federated learning over future mobile edge devices,” IEEE Trans. Wireless Commun. , vol. 21, no. 12, pp. 11 038–11 050, 2022

    2022

  14. [22]

    ARM Cortex-M7 Processor Datasheet,

    ARM Holdings, “ARM Cortex-M7 Processor Datasheet,” Online, Oct. 2023, available: https://www.arm.com/products/silicon-ip-cpu/cortex-m/ cortex-m7, Accessed: Oct. 10, 2023

    2023

  15. [23]

    Apple A18 Pro Chip Specifications,

    Apple Inc., “Apple A18 Pro Chip Specifications,” Online, Oct. 2024, available: https://nanoreview.net/en/soc/apple-a18-pro, Accessed: Oct. 10, 2024

    2024

  16. [24]

    Asynchronous federated optimization,

    C. Xie, S. Koyejo, and I. Gupta, “Asynchronous federated optimization,” arXiv preprint arXiv:1903.03934 , 2019

    1903 arXiv

  17. [25]

    Asynchronous federated learning over wireless communication networks,

    Z. Wang, Z. Zhang, Y . Tian, Q. Yang, H. Shan, W. Wang, and T. Q. Quek, “Asynchronous federated learning over wireless communication networks,” IEEE Trans. Wireless Commun. , vol. 21, no. 9, pp. 6961– 6978, 2022

    2022

  18. [26]

    Asynchronous federated learning on heterogeneous devices: A survey,

    C. Xu, Y . Qu, Y . Xiang, and L. Gao, “Asynchronous federated learning on heterogeneous devices: A survey,” Comput. Sci. Rev. , vol. 50, p. 100595, 2023

    2023

  19. [27]

    Revisiting distributed synchronous SGD,

    J. Chen, X. Pan, R. Monga, S. Bengio, and R. Jozefowicz, “Revisiting distributed synchronous SGD,” arXiv preprint arXiv:1604.00981 , 2016

    2016 arXiv

  20. [28]

    Bandwidth allocation for multiple federated learning services in wireless edge networks,

    J. Xu, H. Wang, and L. Chen, “Bandwidth allocation for multiple federated learning services in wireless edge networks,” IEEE Trans. Wireless Commun., vol. 21, no. 4, pp. 2534–2546, 2021

    2021

  21. [29]

    Client selection and bandwidth allocation in wireless federated learning networks: A long-term perspective,

    J. Xu and H. Wang, “Client selection and bandwidth allocation in wireless federated learning networks: A long-term perspective,” IEEE Trans. Wireless Commun., vol. 20, no. 2, pp. 1188–1200, 2020

    2020

  22. [30]

    Joint device schedul- ing and resource allocation for latency constrained wireless federated learning,

    W. Shi, S. Zhou, Z. Niu, M. Jiang, and L. Geng, “Joint device schedul- ing and resource allocation for latency constrained wireless federated learning,” IEEE Trans. Wireless Commun., vol. 20, no. 1, pp. 453–467, 2020

    2020

  23. [31]

    Wirelessly powered federated edge learning: Optimal tradeoffs between convergence and power transfer,

    Q. Zeng, Y . Du, and K. Huang, “Wirelessly powered federated edge learning: Optimal tradeoffs between convergence and power transfer,” IEEE Trans. Wireless Commun. , vol. 21, no. 1, pp. 680–695, 2021

    2021

  24. [32]

    Adaptive batch size for federated learning in resource-constrained edge computing,

    Z. Ma, Y . Xu, H. Xu, Z. Meng, L. Huang, and Y . Xue, “Adaptive batch size for federated learning in resource-constrained edge computing,” IEEE Trans. Mobile Comput. , vol. 22, no. 1, pp. 37–53, 2021

    2021

  25. [33]

    AMBLE: Adjusting mini-batch and local epoch for federated learning with heterogeneous devices,

    J. Park, D. Yoon, S. Yeo, and S. Oh, “AMBLE: Adjusting mini-batch and local epoch for federated learning with heterogeneous devices,” J. Parallel Distrib. Comput., vol. 170, pp. 13–23, 2022

    2022

  26. [34]

    Optimal batch allocation for wireless federated learning,

    J. Song and S.-W. Jeon, “Optimal batch allocation for wireless federated learning,” IEEE Internet Things J. , 2024

    2024

  27. [35]

    Accelerating DNN training in wireless federated edge learning systems,

    J. Ren, G. Yu, and G. Ding, “Accelerating DNN training in wireless federated edge learning systems,” IEEE J. Sel. Areas Commun. , vol. 39, no. 1, pp. 219–232, 2020

    2020

  28. [36]

    Adaptive batchsize selection and gradient compression for wireless federated learning,

    S. Liu, G. Yu, R. Yin, J. Yuan, and F. Qu, “Adaptive batchsize selection and gradient compression for wireless federated learning,” in Proc. IEEE Glob. Commun. Conf. (GLOBECOM) . IEEE, 2020, pp. 1–6

    2020

  29. [37]

    DYNAMITE: Dynamic interplay of mini-batch size and aggregation frequency for federated learning with static and streaming datasets,

    W. Liu, X. Zhang, J. Duan, C. Joe-Wong, Z. Zhou, and X. Chen, “DYNAMITE: Dynamic interplay of mini-batch size and aggregation frequency for federated learning with static and streaming datasets,” IEEE Trans. Mobile Comput. , vol. 23, no. 7, pp. 7664–7679, 2023

    2023

  30. [38]

    On the convergence properties of a K-step aver- aging stochastic gradient descent algorithm for nonconvex optimization,

    F. Zhou and G. Cong, “On the convergence properties of a K-step aver- aging stochastic gradient descent algorithm for nonconvex optimization,” arXiv preprint arXiv:1708.01012 , 2017

    2017 arXiv

  31. [39]

    Parallel restarted SGD with faster con- vergence and less communication: Demystifying why model averaging works for deep learning,

    H. Yu, S. Yang, and S. Zhu, “Parallel restarted SGD with faster con- vergence and less communication: Demystifying why model averaging works for deep learning,” in Proc. AAAI Conf. Artif. Intell. , vol. 33, no. 01, 2019, pp. 5693–5700

    2019

  32. [40]

    One-bit over-the-air aggregation for communication-efficient federated edge learning: Design and convergence analysis,

    G. Zhu, Y . Du, D. Gunduz, and K. Huang, “One-bit over-the-air aggregation for communication-efficient federated edge learning: Design and convergence analysis,” IEEE Trans. Wireless Commun. , vol. 20, no. 3, pp. 2120–2135, 2021

    2021

  33. [41]

    A method for the solution of certain non-linear problems in least squares,

    K. Levenberg, “A method for the solution of certain non-linear problems in least squares,” Quart. Appl. Math. , vol. 2, no. 2, pp. 164–168, 1944

    1944

  34. [42]

    M. R. Garey and D. S. Johnson, Computers and intractability . wh freeman New York, 2002, vol. 29

    2002

  35. [43]

    Ultra- low-latency edge inference for distributed sensing,

    Z. Wang, A. E. Kalør, Y . Zhou, P. Popovski, and K. Huang, “Ultra- low-latency edge inference for distributed sensing,” arXiv preprint arXiv:2407.13360, 2024

    2024 arXiv

  36. [44]

    Accessed: Oct

    IBM ILOG CPLEX Optimization Studio , IBM Corporation, Armonk, NY , USA, 2024, available: https://www.ibm.com/products/ilog-cplex- optimization-studio. Accessed: Oct. 10, 2024

    2024

  37. [45]

    Accessed: Oct

    Gurobi Optimizer Reference Manual, Gurobi Optimization, LLC, Hous- ton, TX, USA, 2024, available: https://www.gurobi.com/documentation/. Accessed: Oct. 10, 2024

    2024

  38. [46]

    Gradient-based learning applied to document recognition,

    Y . LeCun, L. Bottou, Y . Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proc. IEEE, vol. 86, no. 11, pp. 2278– 2324, 1998

    1998

  39. [47]

    Learning multiple layers of features from tiny images,

    A. Krizhevsky, “Learning multiple layers of features from tiny images,” University of Toronto, Technical Report, 2009

    2009

  40. [48]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , Las Vegas, NV , USA, 2016, pp. 770–778

    2016

  41. [49]

    Revisiting outage for edge inference systems,

    Z. Wang, Q. Zeng, H. Zheng, and K. Huang, “Revisiting outage for edge inference systems,” arXiv preprint arXiv:2504.03686 , 2025

    2025

Pith tools

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