Pith. sign in

REVIEW 2 major objections 4 minor 43 references

Cyclic, battery-aware scheduling can roughly halve the energy cost of energy-harvesting federated learning, and a selective variant that trains only on odd-numbered eligible chances does so without sacrificing accuracy.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

A cyclic scheduling framework for energy-harvesting federated learning that claims large energy savings, with a convergence analysis that appears internally inconsistent.

T0 review reviewed 2026-08-03 challenge →

load-bearing objection The FedBacys scheduling idea and energy experiments are plausible, but Theorem 1's proof has a participation-scaling error that invalidates the advertised convergence guarantee. the 2 major comments →

arxiv 2511.11949 v2 pith:DNQZJNHF submitted 2025-11-14 cs.LG cs.ITmath.IT

Computation-aware Energy-harvesting Federated Learning with Pipelined Cyclic Scheduling

Eunjeong Jeong , Nikolaos Pappas This is my paper
classification cs.LG cs.ITmath.IT
keywords federated learningenergy harvestingcyclic schedulingbattery-awareclient participationstationarity gapconvergence analysisenergy efficiency
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

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 aims to show that federated learning on energy-harvesting devices can be made practical when local training is explicitly modeled as a multi-slot, battery-draining operation. FedBacys groups clients by battery level and runs them through a rotating pipeline: each group's hub aggregates local updates and passes the model to the next group, so recharging in one group overlaps with computation in another. The paper proves a stationarity-gap bound for this cyclic schedule and argues that the pipeline's finite horizon avoids the exponential staleness factors that plague asynchronous FL analyses. Its FedBacys-Odd variant trains only on odd-numbered eligible chances, and the CIFAR-10 experiments show it reaches the same target accuracy as the full schedule while consuming roughly half the cumulative network energy at high charging probabilities.

Core claim

Under standard smoothness, bounded-variance, and gradient-divergence assumptions, FedBacys satisfies a stationarity-gap bound of the form min_t E||∇f(x_[t,0])||^2 = O( F/(BNTγη) + ρ²(ζ²+σ²)/(Nη²) + higher-order terms ), where η is a per-client upper bound on the probability of successful participation derived from the charging probability δ, local training cost κ, epoch length S, and battery capacity. The cyclic structure makes stale updates at most one epoch old, so the analysis avoids exponential staleness factors typical of asynchronous methods. Under the Polyak–Łojasiewicz condition the optimization error decays as O(1/T²) to a neighborhood set by noise and heterogeneity. Empirically, th

What carries the argument

The load-bearing object is the cyclic group pipeline: clients are split into G groups, each group is given R = floor(S/G) slots, and at the final slot of its window the group hub aggregates local updates and multicasts the intermediate model to the next group, with the last group relaying to the server. A deadline condition forces local training to start κ slots before the upload window, so clients train with a fresher reference model and avoid redundant work. The analysis compresses the causal, memory-dependent battery dynamics into a single bounded participation probability η (Lemma 3), and FedBacys-Odd adds an odd-chance counter that skips every other eligible training start, conserving κ

Load-bearing premise

The proof's central norm-splitting identity, Eq. (11), assumes the expected participation sum is Nη, but Eq. (5) defines q_i^t as 1/N for a participating client, which makes the expected sum equal to η, not Nη; if q_i is not redefined as an indicator, Theorem 1 does not follow from the given algebra.

What would settle it

Recompute Eq. (11) under the paper's own definition q_i^t = 1/N for participants: the target BNη∇f requires E[Σ_i q_i^t] = Nη, whereas the definition gives E[Σ_i q_i^t] = η, so the identity claimed in step (a) fails unless q_i is silently redefined. Separately, rerun the Table I setting (CIFAR-10, N=100, S=30, κ=20, δ=1.0, G=5): if FedBacys-Odd's total energy is not close to 524,850 battery units with final accuracy near FedBacys, the empirical energy claim is falsified.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the bounds hold, battery-aware cyclic scheduling can replace random client selection in EHFL without losing convergence guarantees: performance is controlled by a finite-horizon staleness term rather than exponential asynchrony factors.
  • The odd-chance rule implies that deliberately skipping a large fraction of training opportunities can preserve accuracy, because conserved battery energy is spent later on fresher updates rather than wasted on stale ones.
  • The condition 1 - F_Binomial(κ-1; S, δ) ≥ 1/(6√N) gives a concrete design rule: choose the epoch length S long enough that a client starting with an empty battery can harvest κ units by its upload deadline with probability at least 1/(6√N).
  • The reported energy savings come from scheduling alone, not from compression or quantization, so systems that already use compression could expect additional savings on top of those techniques.
  • Under the PL condition, the convergence error decays as O(1/T²) down to a floor set by client heterogeneity and participation variability, identifying stable participation and balanced scheduling as the levers that shrink that floor.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Extension: the odd-chance rule is a generic 'skip until fresher' principle, so the same counter-based selection could be applied to communication-constrained FL, where the scarce resource is uplink budget rather than battery energy.
  • Extension: the bound's dependence on η suggests a testable prediction: as the charging probability δ decreases, performance should degrade through the η² term in the floor, and the paper's threshold condition marks where that floor starts to dominate; experiments below that threshold would test the theory.
  • Extension: the deadline-cramming rule could be made per-client by allowing heterogeneous κ values, letting slower or larger-model devices participate without forcing a single global epoch length S; this is a direct generalization of the scheduling mechanism.
  • Extension: because the error floor contains heterogeneity terms ζ² and β̃², grouping clients by data distribution in addition to battery level could shrink the floor; the paper lists semantics-aware adaptation as future work but does not test this combination.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. The paper proposes FedBacys and FedBacys-Odd, energy-harvesting federated learning algorithms that organize clients into cyclically scheduled groups with pipelined intra-group aggregation. The main claimed contributions are (i) a convergence bound on the stationarity gap (Theorem 1) and a PL-based rate (Theorem 2) under a multi-slot local training model, and (ii) an experimental demonstration that the schemes, especially FedBacys-Odd, reduce cumulative energy consumption relative to several FL/EHFL baselines. The scheduling idea is interesting and the experimental setup is fairly broad, but the central convergence proof contains a scaling inconsistency in the participation weights that invalidates the advertised theorems as stated.

Significance. If the convergence claims were valid, the paper would make a useful contribution to EHFL by addressing the under-modeled cost of multi-slot local training and by showing that cyclic, battery-aware scheduling can preserve convergence while reducing energy. The experimental study, which includes several baselines and a range of charging probabilities, provides suggestive evidence that the proposed algorithms are energy-efficient in practice. However, the theoretical guarantee is the primary scientific claim of the paper, and the proof of Theorem 1 is not correct as written. No code or machine-checked proofs are provided, so the experiments are not independently reproducible from the manuscript alone. The significance of the paper therefore rests on a convergence result that is currently unsupported.

major comments (2)
  1. [Appendix B-A, Eq. (11); §V Eq. (5)] The proof of Theorem 1 mis-scales the expected participant sum. Eq. (5) defines q_i^t = 1/N for participating clients and 0 otherwise. Lemma 3 bounds each client's participation probability by η≤1, so E[Σ_i q_i^t] = (1/N)Σ_i P(i participates) ≤ η. Eq. (11) nevertheless centers the decomposition on BNη∇f(x_{t,0}) and divides by 2BNη, i.e. it treats the expected weighted sum as Nη. The justification given ('expectation of the size of the participant subset ... bounded above by Nη') confuses the number of participants with the weighted sum Σ q_i. This changes the descent coefficient by a factor of N and propagates into Eqs. (14)-(15). Concretely, full participation gives Σ_i q_i^t = 1, so consistency with Eq. (11)'s target would require Nη=1, i.e. η=1/N, contradicting Lemma 4's η ≥ 1/(6√N) for N>36. Thus Theorem 1, and consequently Theorem 2, are not established.
  2. [Appendix C, Lemma 3; §V] Lemma 3 is advertised as giving a 'significantly tighter' upper bound η on participation probability, but its proof only derives the trivial bound η≤1. Since η is the sole parameter encoding battery dynamics in the convergence bound, the claimed finite-horizon staleness bound and the stated dependence of Theorem 1 on system parameters are unsupported. Additionally, the 'Impact of the odd-chances-only policy' paragraph asserts without proof that η decays from 1 to 1/2 as T→∞; if this is a modeling assumption rather than a theorem, it should be labeled as such and its effect on the bound quantified. As written, η acts as an arbitrary free parameter in the proof, not as a derived system quantity.
minor comments (4)
  1. [§V, Eq. (5)] The notation q_i^t is used both as an aggregation weight (1/N) and, in the text, as a participation probability. This conflation is likely the source of the N-scaling error in the proof; please rename or explicitly distinguish the weight from the participation indicator.
  2. [Algorithm 1] The pseudo-code for FedBacys-Odd is hard to follow: the LOCALTRAIN call appears in both the general condition branch and the odd-chance branch, making it unclear when a client actually trains. A cleaner separation of the two policies would help.
  3. [§VI, Table I and Fig. 3] The energy and accuracy results appear to be from single runs. Reporting multiple seeds with error bars would strengthen the empirical claims, especially for the 'roughly half energy' comparison.
  4. [Appendix B-A] The symbol E_t is used for both the conditional expectation E[·|F_t] and the battery level E_s^{(i)} in Section III. Please use distinct notations to avoid confusion.

Circularity Check

0 steps flagged

No significant circularity: the convergence analysis is derived from stated assumptions and the energy-efficiency claims are empirical, not fitted predictions.

full rationale

The central convergence claim (Theorem 1) is not a restatement of its inputs: it is derived from L-smoothness, unbiased stochastic gradients, bounded gradient divergence, the participation-bound parameter η, and a drift bound (Lemma 1). η is introduced as an upper bound on per-user participation probability obtained from system parameters (Lemma 3), not fitted to the final stationarity gap; the later statement that η behaves as a decaying coefficient toward 1/2 for FedBacys-Odd is an additional modeling assertion, not a parameter chosen to force the bound. The energy-efficiency results (Table I, Fig. 3) are experimental comparisons, not predictions derived from the fitted values. The self-citation [1] identifies the origin of FedBacys, but the proof and experiments in this manuscript do not rely on [1] for the load-bearing convergence argument. There are mathematical concerns in the proof—for instance, Eq. (5) defines q_i^t = 1/N for participants while Eq. (11) invokes Nη as the expected participant sum, which appears to be a scaling inconsistency—but an internal proof error of that kind is a soundness/correctness issue, not circularity, since the claimed stationarity bound does not reduce to its own assumptions by construction. No circular step can be exhibited from the paper's own equations.

Axiom & Free-Parameter Ledger

2 free parameters · 4 axioms · 0 invented entities

The proof relies on standard FL assumptions plus two paper-specific moves: a participation bound η that collapses to 1, and an implicit redefinition of q_i as an indicator rather than 1/N. The latter is a load-bearing algebraic step that is not justified.

free parameters (2)
  • η (participation upper bound) = not fitted; claimed determined by S,κ,δ,E_max; Lemma 3 shows it equals 1
    The convergence bound's quality depends on η being a tight bound less than 1, but the paper's own Lemma 3 proves the universal upper bound is 1. For FedBacys-Odd, the authors assert η decays to 1/2 without a proof. This functions as a free parameter in Theorem 1.
  • batch size B and local epochs = not stated in experiments
    The algorithm and convergence bound depend on B (number of local steps), but the experimental section never specifies the mini-batch size or number of local epochs, leaving an unstated degree of freedom.
axioms (4)
  • standard math Smoothness, unbiased gradients, bounded variance, bounded gradient divergence (Assumptions 1-3)
    Standard in FL convergence proofs; acceptable as background.
  • domain assumption Intra-group and inter-group data heterogeneity bounds (Assumption 4)
    Assumes bounded gradient deviation within and across groups, which is plausible but unverified and not tied to measurable data properties.
  • standard math Polyak-Łojasiewicz condition (Assumption 5)
    Used for Theorem 2; not guaranteed for deep networks like the CNN in experiments.
  • ad hoc to paper The expected participant sum is Nη, i.e., each participant contributes weight 1
    This is the algebra in Eq. (11)-(14); it conflicts with the stated definition q_i^t=1/N in Eq. (5). The proof depends on this identity to factor the stationarity gap.

reviewed 2026-08-03 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Computation-aware Energy-harvesting Federated Learning with Pipelined Cyclic Scheduling." pith.science (2026). https://pith.science/paper/DNQZJNHF

@misc{pith2026251111949,
  author       = {Pith},
  title        = {Pith review of: Computation-aware Energy-harvesting Federated Learning with Pipelined Cyclic Scheduling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DNQZJNHF}},
  note         = {Machine review of arXiv:2511.11949}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Federated learning (FL) is a powerful paradigm for distributed learning, but increasing model complexity leads to significant energy consumption from client-side computations for local training. This challenge is critical in energy-harvesting FL (EHFL) systems, where the participation availability of each device fluctuates because of limited energy. To address this, we propose PipeCycle, a battery-aware distributed learning framework that organizes clients into pipelined cyclic groups. When a group completes its intra-group aggregation, its aggregated model is relayed directly to a newly formed group as a reference for local training, allowing multiple groups to coexist in the pipeline while overlapping client recharging periods with active training in other pipeline stages. We provide a convergence analysis of PipeCycle under a realistic energy consumption model in which local training spans multiple time slots, and show that the cyclic structure of the pipeline imposes a finite-horizon staleness bound that avoids the exponential factors typical of asynchronous FL analyses. Numerical experiments across both IID and non-IID data and various battery charging probabilities show that PipeCycle reaches a target accuracy with substantially lower cumulative energy than existing FL baselines, particularly under severe label skew where competing cyclic schemes collapse to near-chance accuracy.

Figures

Figures reproduced from arXiv: 2511.11949 by Eunjeong Jeong, Nikolaos Pappas.

Figure 1
Figure 1. Figure 1: A schematic view of energy-harvesting FL based on proba [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: An example of a FedBacys network with N = 10 users assigned into G = 3 groups. Within each group, (1) each user trans￾mits local updates to the group’s hub if they satisfy the conditions to send local updates; (2) the hub aggregates the received updates; (3) the hub sends the updated intra-group model to the next group by multicast. The final group’s hub sends the aggregated model to the server, which is s… view at source ↗
Figure 3
Figure 3. Figure 3: Test accuracies from operating different FL algorithms with [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Energy consumption of FedBacys and FedBacys-Odd with [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Test accuracy of FedBacys and FedBacys-Odd with respect [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

43 extracted references · 4 linked inside Pith

  1. [1]

    Battery-aware cyclic scheduling in energy- harvesting federated learning,

    E. Jeong and N. Pappas, “Battery-aware cyclic scheduling in energy- harvesting federated learning,” in2025 IEEE 26th International Work- shop on Signal Processing and Artificial Intelligence for Wireless Communications (SPAWC), pp. 1–5, 2025

    2025

  2. [2]

    Federated optimization: Distributed machine learning for on-device intelligence,

    J. Kone ˇcn`y, H. B. McMahan, D. Ramage, and P. Richt ´arik, “Federated optimization: Distributed machine learning for on-device intelligence,” arXiv preprint arXiv:1610.02527, 2016

    Pith/arXiv arXiv 2016

  3. [3]

    Prabakeran, T

    S. Prabakeran, T. Sethukarasi, and V . Indumathi,Federated Learn- ing Frameworks for Energy-Efficient AI in Distributed Data Centres, p. 399–426. IGI Global Scientific Publishing, Apr. 2025

    2025

  4. [4]

    Green federated learning,

    A. Yousefpour, S. Guo, A. Shenoy, S. Ghosh, P. Stock, K. Maeng, S.- W. Kr ¨uger, M. Rabbat, C.-J. Wu, and I. Mironov, “Green federated learning,” inFederated Learning and Analytics in Practice: Algorithms, Systems, Applications, and Opportunities, 2023

    2023

  5. [5]

    Green federated learning: A new era of green aware AI,

    D. Thakur, A. Guzzo, G. Fortino, and F. Piccialli, “Green federated learning: A new era of green aware AI,”ACM Comput. Surv., vol. 57, Mar. 2025

    2025

  6. [6]

    Energy-efficient federated learning with resource allocation for green IoT edge intelligence in B5G,

    A. Salh, R. Ngah, L. Audah, K. S. Kim, Q. Abdullah, Y . M. Al-Moliki, K. A. Aljaloud, and H. N. Talib, “Energy-efficient federated learning with resource allocation for green IoT edge intelligence in B5G,”IEEE Access, vol. 11, pp. 16353–16367, 2023. 8

    2023

  7. [7]

    FedZero: Leveraging renewable excess energy in federated learning,

    P. Wiesner, R. Khalili, D. Grinwald, P. Agrawal, L. Thamsen, and O. Kao, “FedZero: Leveraging renewable excess energy in federated learning,” inProceedings of the 15th ACM International Conference on Future and Sustainable Energy Systems, e-Energy ’24, (New York, NY , USA), p. 373–385, Association for Computing Machinery, 2024

    2024

  8. [8]

    FedBiF: Communication-efficient federated learning via bits freezing,

    S. Li, Q. Li, H. Wang, R. Li, J. Lin, and W. Zhong, “FedBiF: Communication-efficient federated learning via bits freezing,”IEEE Transactions on Parallel and Distributed Systems, vol. 36, no. 12, pp. 2668–2678, 2025

    2025

  9. [9]

    FAST: Enhancing federated learning through adaptive data sampling and local training,

    Z. Wang, H. Xu, Y . Xu, Z. Jiang, J. Liu, and S. Chen, “FAST: Enhancing federated learning through adaptive data sampling and local training,” IEEE Transactions on Parallel and Distributed Systems, vol. 35, no. 2, pp. 221–236, 2024

    2024

  10. [10]

    FedALT: Federated fine- tuning through adaptive local training with rest-of-world LoRA,

    J. Bian, L. Wang, L. Zhang, and J. Xu, “FedALT: Federated fine- tuning through adaptive local training with rest-of-world LoRA,”AAAI Conference on Artificial Intelligence (AAAI), 2026

    2026

  11. [11]

    Energy, scalability, data and security in massive IoT: Current landscape and future directions,

    I. Cheikh, S. Roy, E. Sabir, and R. Aouami, “Energy, scalability, data and security in massive IoT: Current landscape and future directions,” arXiv preprint arXiv:2505.03036, 2025

    Pith/arXiv arXiv 2025

  12. [12]

    Optimizing resource allocation in UA V-assisted ultra-dense networks for enhanced performance and security,

    P.-G. Ye, J. Zheng, X. Ren, J. Huang, Z. Zhang, Y . Pang, and G. Kou, “Optimizing resource allocation in UA V-assisted ultra-dense networks for enhanced performance and security,”Information Sciences, vol. 679, p. 120788, 2024

    2024

  13. [13]

    Federated learning over energy harvesting wireless networks,

    R. Hamdi, M. Chen, A. B. Said, M. Qaraqe, and H. V . Poor, “Federated learning over energy harvesting wireless networks,”IEEE Internet of Things Journal, vol. 9, no. 1, pp. 92–103, 2022

    2022

  14. [14]

    Online optimization for over-the-air federated learning with energy harvest- ing,

    Q. An, Y . Zhou, Z. Wang, H. Shan, Y . Shi, and M. Bennis, “Online optimization for over-the-air federated learning with energy harvest- ing,”IEEE Transactions on Wireless Communications, vol. 23, no. 7, pp. 7291–7306, 2024

    2024

  15. [15]

    Federated learning with energy harvesting devices,

    L. Zeng, D. Wen, G. Zhu, C. You, Q. Chen, and Y . Shi, “Federated learning with energy harvesting devices,”IEEE Transactions on Green Communications and Networking, vol. 8, no. 1, pp. 190–204, 2024

    2024

  16. [16]

    Fast federated learning in the presence of arbitrary device unavailability,

    X. Gu, K. Huang, J. Zhang, and L. Huang, “Fast federated learning in the presence of arbitrary device unavailability,” inAdvances in Neural Information Processing Systems, vol. 34, pp. 12052–12064, Curran Associates, Inc., 2021

    2021

  17. [17]

    On the convergence of federated averaging with cyclic client participation,

    Y . J. Cho, P. Sharma, G. Joshi, Z. Xu, S. Kale, and T. Zhang, “On the convergence of federated averaging with cyclic client participation,” in Proceedings of the 40th International Conference on Machine Learning, vol. 202 ofProceedings of Machine Learning Research, pp. 5677–5721, PMLR, 23–29 Jul 2023

    2023

  18. [18]

    Adaptive biased user scheduling for heterogeneous wireless federate learning network,

    C. Wu, Y . Ren, D. K. So, and J. Tang, “Adaptive biased user scheduling for heterogeneous wireless federate learning network,”arXiv preprint arXiv:2505.05231, 2025

    Pith/arXiv arXiv 2025

  19. [19]

    A framework for sustainable federated learning,

    B. G ¨uler and A. Yener, “A framework for sustainable federated learning,” in2021 19th International Symposium on Modeling and Optimization in Mobile, Ad hoc, and Wireless Networks (WiOpt), pp. 1–8, 2021

    2021

  20. [20]

    Energy-efficient federated edge learning with streaming data: A lyapunov optimization approach,

    C.-H. Hu, Z. Chen, and E. G. Larsson, “Energy-efficient federated edge learning with streaming data: A lyapunov optimization approach,”IEEE Transactions on Communications, vol. 73, no. 2, pp. 1142–1156, 2025

    2025

  21. [21]

    Over-the-air computation enabled semi-asynchronous wireless federated learning,

    Z. Zheng, Y . Deng, W. Yi, H. Shin, and A. Nallanathan, “Over-the-air computation enabled semi-asynchronous wireless federated learning,” IEEE Transactions on Communications, vol. 73, no. 10, pp. 8919–8936, 2025

    2025

  22. [22]

    Power of redundancy: Surplus client scheduling for federated learning against user uncertainties,

    Y . Li, F. Li, L. Chen, L. Zhu, P. Zhou, and Y . Wang, “Power of redundancy: Surplus client scheduling for federated learning against user uncertainties,”IEEE Transactions on Mobile Computing, vol. 22, no. 9, pp. 5449–5462, 2023

    2023

  23. [23]

    Joint client selection and bandwidth allocation algorithm for federated learning,

    H. Ko, J. Lee, S. Seo, S. Pack, and V . C. M. Leung, “Joint client selection and bandwidth allocation algorithm for federated learning,” IEEE Transactions on Mobile Computing, vol. 22, no. 6, pp. 3380–3390, 2023

    2023

  24. [24]

    Age of information based client selection for wireless federated learning with diversified learning capabilities,

    L. Dong, Y . Zhou, L. Liu, Y . Qi, and Y . Zhang, “Age of information based client selection for wireless federated learning with diversified learning capabilities,”IEEE Transactions on Mobile Computing, vol. 23, no. 12, pp. 14934–14945, 2024

    2024

  25. [25]

    An incentive auction for heterogeneous client selection in federated learning,

    J. Pang, J. Yu, R. Zhou, and J. C. Lui, “An incentive auction for heterogeneous client selection in federated learning,”IEEE Transactions on Mobile Computing, vol. 22, no. 10, pp. 5733–5750, 2023

    2023

  26. [26]

    Knowledge-aided federated learning for energy-limited wireless networks,

    Z. Chen, W. Yi, Y . Liu, and A. Nallanathan, “Knowledge-aided federated learning for energy-limited wireless networks,”IEEE Transactions on Communications, vol. 71, no. 6, pp. 3368–3386, 2023

    2023

  27. [27]

    Federated learning under intermittent client availability and time-varying communication con- straints,

    M. Ribero, H. Vikalo, and G. de Veciana, “Federated learning under intermittent client availability and time-varying communication con- straints,”IEEE Journal of Selected Topics in Signal Processing, vol. 17, no. 1, pp. 98–111, 2023

    2023

  28. [28]

    On federated learning with energy har- vesting clients,

    C. Shen, J. Yang, and J. Xu, “On federated learning with energy har- vesting clients,” inICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 8657–8661, 2022

    2022

  29. [29]

    Update estimation and scheduling for over-the-air federated learning with energy harvesting devices,

    F. Bagci, B. Tegin, M. Kazemi, and T. M. Duman, “Update estimation and scheduling for over-the-air federated learning with energy harvesting devices,” inWorkshop on Integrated, Intelligent and Ubiquitous Con- nectivity for 6G and Beyond (IEEE ICC 2025 – IEEE International Conference on Communications), 2025

    2025

  30. [30]

    Federated learning- distillation alternation for resource-constrained IoT,

    R. V . da Silva, O. L. A. L ´opez, and R. D. Souza, “Federated learning- distillation alternation for resource-constrained IoT,”arXiv preprint arXiv:2505.20456, 2025

    Pith/arXiv arXiv 2025

  31. [31]

    Any- costFL: Efficient on-demand federated learning over heterogeneous edge devices,

    P. Li, G. Cheng, X. Huang, J. Kang, R. Yu, Y . Wu, and M. Pan, “Any- costFL: Efficient on-demand federated learning over heterogeneous edge devices,” inIEEE INFOCOM 2023 - IEEE Conference on Computer Communications, pp. 1–10, 2023

    2023

  32. [32]

    Federated learning under periodic client participation and heterogeneous data: A new communication-efficient algorithm and analysis,

    M. Crawshaw and M. Liu, “Federated learning under periodic client participation and heterogeneous data: A new communication-efficient algorithm and analysis,” inAdvances in Neural Information Processing Systems, vol. 37, pp. 8240–8299, Curran Associates, Inc., 2024

    2024

  33. [33]

    Communication-efficient local SGD with age-based worker selection,

    F. Zhu, J. Zhang, and X. Wang, “Communication-efficient local SGD with age-based worker selection,”The Journal of Supercomputing, vol. 79, no. 12, pp. 13794–13816, 2023

    2023

  34. [34]

    Clustered federated learning: Model-agnostic distributed multitask optimization under privacy con- straints,

    F. Sattler, K.-R. M ¨uller, and W. Samek, “Clustered federated learning: Model-agnostic distributed multitask optimization under privacy con- straints,”IEEE Transactions on Neural Networks and Learning Systems, vol. 32, no. 8, pp. 3710–3722, 2021

    2021

  35. [35]

    Dynamic clustering in federated learning,

    Y . Kim, E. A. Hakim, J. Haraldson, H. Eriksson, J. M. B. da Silva, and C. Fischione, “Dynamic clustering in federated learning,” inICC 2021 - IEEE International Conference on Communications, pp. 1–6, 2021

    2021

  36. [36]

    Federated learning with soft cluster- ing,

    C. Li, G. Li, and P. K. Varshney, “Federated learning with soft cluster- ing,”IEEE Internet of Things Journal, vol. 9, no. 10, pp. 7773–7782, 2022

    2022

  37. [37]

    An efficient framework for clustered federated learning,

    A. Ghosh, J. Chung, D. Yin, and K. Ramchandran, “An efficient framework for clustered federated learning,”IEEE Transactions on Information Theory, vol. 68, no. 12, pp. 8076–8091, 2022

    2022

  38. [38]

    Active client selection for clustered federated learning,

    H. Huang, W. Shi, Y . Feng, C. Niu, G. Cheng, J. Huang, and Z. Liu, “Active client selection for clustered federated learning,”IEEE Trans- actions on Neural Networks and Learning Systems, vol. 35, no. 11, pp. 16424–16438, 2024

    2024

  39. [39]

    Energy harvesting sensor nodes: Survey and implications,

    S. Sudevalayam and P. Kulkarni, “Energy harvesting sensor nodes: Survey and implications,”IEEE Communications Surveys & Tutorials, vol. 13, no. 3, pp. 443–461, 2011

    2011

  40. [40]

    Communication-Efficient Learning of Deep Networks from Decentral- ized Data,

    B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y. Arcas, “Communication-Efficient Learning of Deep Networks from Decentral- ized Data,” inProceedings of the 20th International Conference on Artificial Intelligence and Statistics, vol. 54 ofProceedings of Machine Learning Research, pp. 1273–1282, PMLR, 20–22 Apr 2017

    2017

  41. [41]

    FedSeq: A hybrid federated learning framework based on sequential in-cluster training,

    Z. Chen, D. Li, R. Ni, J. Zhu, and S. Zhang, “FedSeq: A hybrid federated learning framework based on sequential in-cluster training,” IEEE Systems Journal, vol. 17, no. 3, pp. 4038–4049, 2023

    2023

  42. [42]

    A unified analysis of federated learning with arbi- trary client participation,

    S. Wang and M. Ji, “A unified analysis of federated learning with arbi- trary client participation,” inAdvances in Neural Information Processing Systems, vol. 35, pp. 19124–19137, Curran Associates, Inc., 2022. APPENDIXA PRELIMINARIES A. Notations Based on Alg. 1, the update of global models between two groups with indexg+ 1andgwithin the same epochtcan b...

    2022

  43. [43]

    This inequality is equivalent toh(t ∗)≥0, which is mutually inclusive toh(t)≥0for allt >0(and automaticallyh 1(t)≥h 2(t)for allt >0holds). Lemma 3.(Upper bound of participation probability at each epoch) OnceS, κ, δ, Emax are determined, the upper bound of the successful participation probability of any useri∈ U, denoted asη, is uniquely determined and is...

This paper was first reviewed by deepseek-v4-flash on August 3, 2026.