Pith. sign in

REVIEW 4 major objections 4 minor 40 references

Efficient Split Federated Learning for Large Language Models over Communication Networks

T0 review · 4 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Split federated learning with LoRA matches centralized fine-tuning accuracy on GPT-2 while cutting training latency by up to 60%.

desk verdict Solid systems paper with a genuinely new joint optimization, but the rank-selection result leans on an unvalidated offline convergence curve that needs to be made explicit or the reported latency gains from rank selection cannot be trusted. read the letter →

arxiv 2504.14667 v2 pith:6EA4AWZC submitted 2025-04-20 cs.LG cs.NI

Kai Zhao , Zhaohui Yang , Ye Hu , Mingzhe Chen , Chen Zhu , Zhaoyang Zhang This is my paper · ORCID
classification cs.LGcs.NI
keywords splitfederatedlearningLoRAlargelanguagemodelsedgenetworksresourceallocationtraininglatencyminimizationmodelsplittingcommunication-computationtradeoff
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 establishes that split federated learning can be made practical for fine-tuning large language models on resource-constrained edge devices when combined with low-rank adaptation (LoRA). It builds a total-training-delay model that couples communication and computation through two knobs: where the model is split between clients and a main server, and what LoRA rank is used. The paper then solves the resulting mixed-integer optimization by alternating a greedy subchannel assignment, a convex power-allocation step, and exhaustive searches over split layer and rank. Simulations on GPT-2 with the E2E dataset show the framework reaches final test perplexity within 0.001 of centralized LoRA fine-tuning while sharply reducing client-side computation, and the optimized resource allocation cuts total training latency by up to 60% relative to a random-allocation baseline.

What carries the argument

The load-bearing object is the delay identity in Eq. (17): total training delay is the product of the number of global rounds $E(r)$ and the sum of per-round time $I T_{\text{local}}$ plus the slowest adapter upload, with $T_{\text{local}}$ itself the sum of bottleneck client forward/upload time, server forward/backward time, and bottleneck client backward time. Every term is an explicit function of the split vector $\mu$ (which layers live on the client side), the LoRA rank $r$ (which controls trainable parameters, upload size, and convergence speed), subchannel assignment, and transmit power. This single expression turns 'make fine-tuning fast' into a mixed-integer nonconvex program, and the paper's block-coordinate-descent algorithm attacks it by fixing three variable blocks at a time: greedy allocation for subchannels, a log-convexified convex solve for power, and exhaustive search over the small integer sets of split layers and candidate ranks.

What would settle it

Run SflLLM on a different downstream task, or with a non-IID client data split, after fitting $E(r)$ on E2E; then record the actual number of global rounds to the target loss for each candidate rank. If the observed $E(r)$ curve differs enough to change which rank minimizes total training delay, the reported latency reductions do not generalize.

Watch

Extended reading notes

Core claim

The central claim is that SflLLM, a split federated learning framework for LLMs, achieves accuracy essentially equal to centralized fine-tuning while making client-side training feasible on low-resource devices. The vehicle is a delay objective, $T = E(r)(I T_{\text{local}} + \max_k T_k^f)$, where $E(r)$ is the number of global rounds needed to reach a target loss, $I$ the local steps per round, $T_{\text{local}}$ the slowest client's forward-plus-upload time plus server forward and backward time plus slowest client's backward time, and $T_k^f$ the time to upload client LoRA adapters to a federated server. Minimizing this objective chooses subchannel assignment, transmit power, the layer at which the model is split, and the LoRA rank. The reported result is that this joint choice reduces training latency by up to 60% versus random allocation and, in the tested regime, rank selection contributes more to the latency saving than communication tuning does.

Load-bearing premise

The delay formula assumes that the fitted curve $E(r)$, giving the number of training rounds needed to reach the target accuracy for each LoRA rank, transfers from an offline representative dataset to the actual deployment data, target accuracy, and channel conditions without re-measurement.

Editorial extensions

If this is right

  • A client that only holds the front part of a transformer can fine-tune an LLM to roughly the same final perplexity as centralized LoRA training, with the tested deviation below 0.001.
  • Increasing LoRA rank accelerates convergence but raises per-step compute and upload traffic, so the delay-minimizing rank is an interior value that the exhaustive search finds rather than a fixed hyperparameter.
  • The joint allocation shifts its emphasis with the bottleneck: when bandwidth is scarce, subchannel and power optimization dominate the gain; when client compute is weak, split-point selection dominates.
  • The complete optimization loop has complexity $O(\tau_{\max} K M^2)$, making it implementable in a network controller with modest iteration counts.

Reading between the lines

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

  • The paper leaves rank-dependent convergence as an empirically fitted curve; a closed-form bound on $E(r)$ as a function of rank and data heterogeneity would let the same delay model be re-optimized online during training instead of only offline.
  • Because the delay model neglects downlink broadcasts and activation-gradient downloads, its predicted optimal rank and split point could shift in regimes where those costs are large; adding downlink terms is a direct testable extension.
  • Allowing each client its own split layer, rather than one global split point, would likely further reduce straggler losses in heterogeneous networks, at the price of a larger search space the paper does not explore.
  • The reported results are at GPT-2 scale on a single dataset; whether the accuracy and latency tradeoffs persist for larger open-weight LLMs or other tasks is not established and could be tested with the same framework.
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 proposes SflLLM, a split federated learning framework that combines model splitting with LoRA for fine-tuning large language models on resource-constrained edge networks. A joint optimization problem is formulated to minimize total training delay by jointly selecting subchannel allocation, transmit power, model split point, and LoRA rank. The authors propose a block coordinate descent algorithm with a greedy subchannel allocation, a convex reformulation of the power control subproblem, and exhaustive search for split point and rank. Simulations on the E2E dataset with GPT-2 variants show comparable accuracy to centralized LoRA and reduced training latency relative to several baselines.

Significance. If the delay model and optimization are validated, the framework addresses a practically important problem: distributed fine-tuning of LLMs over heterogeneous wireless edge networks. The paper contains several sound and useful components: the convex reformulation of the power control subproblem (Eqs. (22)-(24)) appears internally consistent, the complexity analysis in Section VI.D is explicit, and the accuracy comparison to centralized LoRA in Table IV (PPL deviation within 0.001) is a clean sanity check. The main significance of the claimed result, however, rests on the adaptive LoRA rank selection, and that claim is currently built on an empirically fitted E(r) curve whose transferability is not established.

major comments (4)
  1. [Section VI.C, Eq. (17), P4 (Eq. (26))] The total delay T in Eq. (17) multiplies per-round latency by E(r), the number of global rounds required to reach target accuracy, and subproblem P4 in Eq. (26) selects the LoRA rank by minimizing this expression over r. Section VI.C states only that E(r) 'can be estimated offline through pretraining on a representative dataset,' but the paper gives no fitted form, no table of E(r) values, and no transferability analysis across data distributions, non-IID partitions, target accuracy levels, or deployment channel conditions. Because E(r) is the only rank-dependent convergence term in the objective, the selected rank and the reported latency reductions (especially the comparison with baseline d in Section VII.C) are calibrated to this assumed curve. If E(r) does not transfer to the actual deployment, P4 can select a non-optimal rank and the stated latency improvement is an artifact of the assumed function rather than a property of the proposed optimization. This is a load-bearing gap for the paper's central claim and requires either a theoretical characterization of E(r), a robust formulation, or a validation study across deployment conditions.
  2. [Section I (contribution bullet), Section VI.C, Section VIII] The second contribution bullet in Section I claims that the paper 'theoretically analyze[s] how the rank of the LoRA module affects training overhead and convergence speed.' However, Section VI.C does not provide any theoretical convergence analysis; it instead relies on an offline empirical estimate of E(r). Section VIII explicitly defers 'analyzing the effect of rank selection on convergence through theoretical derivations' to future work. This is an internal inconsistency between the claimed contribution and the delivered content. The authors should either supply the theoretical analysis or revise the contribution statement and the optimization formulation to present E(r) as an empirical input whose estimation and validation are part of the method.
  3. [Section VII.C, baseline d; Section VI.C] The evaluation protocol for E(r) in the baselines is unspecified. Baseline d is described as 'random rank selection' with the proposed subchannel, power, and split allocation. For the latency computation of this baseline, it is not stated whether the randomly chosen rank is assigned the same fitted E(r) curve, an independent convergence measurement, or some other value. If the fitted E(r) is used for randomly chosen ranks as well, then the comparison between the proposed rank selection and baseline d circularly favors the fitted rank; if a different protocol is used, it should be described explicitly. The same ambiguity applies to the other baselines when they involve rank selection. A clear and consistent protocol for assigning E(r) to all compared schemes is needed before the latency comparisons can be interpreted.
  4. [Section VI.D, Algorithm 3] The paper asserts that the proposed BCD algorithm 'reliably converges to a stable and effective solution within a finite number of iterations, regardless of initialization,' while also stating that formal convergence guarantees are precluded. No convergence plots of the BCD objective across iterations, no initialization sensitivity study, and no comparison of different initializations are presented in Section VII. As the algorithm is a heuristic, the empirical convergence claim is currently unsupported by the exhibited results. The authors should add a convergence study (e.g., objective value versus iteration count for several random initializations) or temper the claim to match the evidence.
minor comments (4)
  1. [Section VII.C, Figs. 6-8] The captions of Figs. 6, 7, and 8 are identical to the caption of Fig. 4 ('The effect of LoRA rank on the number of steps required to reach the target loss value') and do not describe the actual plotted quantities (latency versus clients' compute power, main server compute capability, and transmit power, respectively). These captions should be corrected.
  2. [Table III] Table III contains a long sequence of corrupted glyph strings ('/uni00000013/uni00000015/...') in the row labels and some numeric entries, making part of the table unreadable. The table should be regenerated with a proper font encoding.
  3. [Section V.A, bullet list] Two bullets in the decision-variable list are both labeled 'r': the first defines the binary subchannel allocation variables r_i,s_k and r_i,f_k, and the second defines the integer LoRA rank r. Using the same symbol for both is confusing; the subchannel variables should be renamed (e.g., a_i,s_k and a_i,f_k) or the rank should be denoted differently.
  4. [Section I, first paragraph] There are several typographical errors in the introduction, including 'has has significantly advanced,' 'lnternet of things,' and 'A appropriately designed.' These should be corrected in a copyediting pass.

Circularity Check

1 steps flagged · score 5.0 of 10

Rank-selection latency gains are forced by the offline-fitted E(r) curve in Eq. (17)/P4, while the rest of the resource allocation is independent.

  1. fitted input called prediction [Section VI.C (P4, Eq. (26)) with Section V.B (Eq. (17))]
    "The total training delay across all global rounds is given by: T(r^s,r^f,p^s,p^f,μ,r)=E(r)(IT_local+max_k{T_f^k}), (17) where E(r) as a function of the rank r denotes the number of global iterations required to achieve the desired accuracy. ... P4: min_r eT (26) s.t. C7, C8, C9, C10, where the number of global training rounds E(r) can be estimated offline through pretraining on a representative dataset."

    P4 selects the LoRA rank by minimizing eT, and eT is exactly Eq. (17) with E(r) as a multiplicative factor. E(r) is not derived from theory; it is an empirically estimated input ('estimated offline through pretraining on a representative dataset'), and the paper's own Fig. 4 is the steps-to-target-loss curve used for this purpose. The reported latency reduction of the proposed rank selection over random rank (Baseline d) is then measured with the same T objective. Consequently, the optimal rank and the measured latency gain are forced by the fitted E(r) curve: if a different E(r) were used, P4 could select a different rank and the claimed gains would change.

full rationale

Most parts of the paper are self-contained engineering modeling: the delay expressions (8)-(15) are standard, P1-P3 are solved by greedy, convex, and exhaustive methods, and the SflLLM accuracy result is empirically compared against centralized LoRA baselines. There is no load-bearing self-citation chain; the only self-citation (ref. [17]) supports a general remark about joint computation-communication optimization. However, the adaptive rank-selection contribution rests on the empirical function E(r) inserted into Eq. (17). Section I claims a theoretical analysis of rank-versus-convergence, but Section VIII defers such analysis to future work ('Future research should focus on analyzing the effect of rank selection on convergence through theoretical derivations'), confirming that E(r) is a fitted input rather than a derived result. Since P4 minimizes T, the rank choice and the latency reductions relative to random rank are consequences of the assumed E(r) rather than an independently validated prediction. This makes the rank-selection claim partially circular, while the resource-allocation and accuracy claims retain independent content.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The delay model rests on empirical workload-scaling assumptions and on the offline-estimated convergence function E(r). No new physical entities or mediators are introduced. The assumptions are stated in the text, which is good, but they are load-bearing for the central latency-reduction claim.

free parameters (1)
  • E(r), global training rounds to target accuracy = not tabulated; empirical curves in Figs. 3-4
    Estimated offline through pretraining on a representative dataset and inserted into the total delay objective (17). Rank selection in subproblem P4 minimizes against this fitted function.
assumptions (3)
  • domain assumption E(r) estimated on a representative dataset remains valid for the deployment scenario and for all clients
    Used in total delay T = E(r)(I T_local + max Tf_k). If E(r) shifts with data heterogeneity or deployment conditions, the rank selection is miscalibrated. Entered in Section VI.C.
  • domain assumption Downlink transmission and server aggregation latencies are negligible
    Section V.A states that activation-gradient downloads and model broadcasting are negligible because servers transmit at high power and data sizes are small. This removes two phases from the delay model.
  • domain assumption Computation workload scales linearly with LoRA rank and backward pass costs twice the forward pass
    Used in the Phi and Delta-Phi workload definitions in Eqs. (8)-(13) and in Table III. The linear scaling in rank and the 2x backward factor are not validated against real profiling on GPT-2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient Split Federated Learning for Large Language Models over Communication Networks." pith.science (2026). https://pith.science/paper/6EA4AWZC

@misc{pith2026250414667,
  author       = {Pith},
  title        = {Pith review of: Efficient Split Federated Learning for Large Language Models over Communication Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6EA4AWZC}},
  note         = {Machine review of arXiv:2504.14667}
}
read the original abstract

Fine-tuning pre-trained large language models (LLMs) in a distributed manner poses significant challenges on resource-constrained edge networks. To address this challenge, we propose SflLLM, a novel framework that integrates split federated learning with parameter-efficient fine-tuning techniques. By leveraging model splitting and low-rank adaptation (LoRA), SflLLM reduces the computational burden on edge devices. Furthermore, the introduction of a federated server facilitates parallel training and enhances data privacy. To accommodate heterogeneous communication conditions and diverse computational capabilities of edge devices, as well as the impact of LoRA rank selection on model convergence and training cost, we formulate a joint optimization problem of both communication and computation resource. The formulated problem jointly optimizes subchannel allocation, power control, model splitting point selection, and LoRA rank configuration, aimed at minimizing total training delay. An iterative optimization algorithm is proposed to solve this problem efficiently. Specifically, a greedy heuristic is employed for subchannel allocation, the power control subproblem is reformulated as a convex optimization problem using auxiliary variables, and an exhaustive search is adopted for optimal split position and rank selection. Simulation results demonstrate that the proposed SflLLM framework achieves comparable model accuracy while significantly reducing client-side computational requirements. Furthermore, the proposed resource allocation scheme and adaptive LoRA rank selection strategy notably reduce the training latency compared to conventional approaches.

Figures

Figures reproduced from arXiv: 2504.14667 by the authors.

Figure 1
Figure 1. Illustration of the proposed SflLLM framework over wireless networks. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Illustration of SflLLM training procedure. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 4
Figure 4. Steps required to achieve target validation loss under varying LoRA [PITH_FULL_IMAGE:figures/full_fig_p011_4.png] view at source ↗
Figures from the paper (3 more)
Figure 3
Figure 3. Figure 3: Validation loss on the E2E dataset with GPT2-S and GPT2-M for [PITH_FULL_IMAGE:figures/full_fig_p011_3.png]
Figure 5
Figure 5. Figure 5: Total latency versus total bandwidth of each client. [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: The effect of LoRA rank on the number of steps required to reach [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 6 canonical work pages

  1. [1]

    Gpt-4 technical report,

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat et al. , “Gpt-4 technical report,” arXiv preprint arXiv:2303.08774 , 2023

    arXiv 2023

  2. [2]

    Palm: Scal- ing language modeling with pathways,

    A. Chowdhery, S. Narang, J. Devlin, M. Bosma, G. Mishra, A. Roberts, P. Barham, H. W. Chung, C. Sutton, S. Gehrmann et al., “Palm: Scal- ing language modeling with pathways,” Journal of Machine Learning Research, vol. 24, no. 240, pp. 1–113, 2023

    2023

  3. [3]

    Scaling laws for neural language models,

    J. Kaplan, S. McCandlish, T. Henighan, T. B. Brown, B. Chess, R. Child, S. Gray, A. Radford, J. Wu, and D. Amodei, “Scaling laws for neural language models,” arXiv preprint arXiv:2001.08361 , 2020

    arXiv 2001

  4. [4]

    Large language models in medicine,

    A. J. Thirunavukarasu, D. S. J. Ting, K. Elangovan, L. Gutierrez, T. F. Tan, and D. S. W. Ting, “Large language models in medicine,” Nature medicine, vol. 29, no. 8, pp. 1930–1940, 2023

    1930

  5. [5]

    Bloomberggpt: A large language model for finance,

    S. Wu, O. Irsoy, S. Lu, V . Dabravolski, M. Dredze, S. Gehrmann, P. Kambadur, D. Rosenberg, and G. Mann, “Bloomberggpt: A large language model for finance,” arXiv preprint arXiv:2303.17564 , 2023

    arXiv 2023

  6. [6]

    Fate- llm: A industrial grade federated learning framework for large language models,

    T. Fan, Y . Kang, G. Ma, W. Chen, W. Wei, L. Fan, and Q. Yang, “Fate- llm: A industrial grade federated learning framework for large language models,” arXiv preprint arXiv:2310.10049 , 2023

    arXiv 2023

  7. [7]

    Federatedscope-llm: A comprehensive package for fine-tuning large language models in federated learning,

    W. Kuang, B. Qian, Z. Li, D. Chen, D. Gao, X. Pan, Y . Xie, Y . Li, B. Ding, and J. Zhou, “Federatedscope-llm: A comprehensive package for fine-tuning large language models in federated learning,” in Proceed- ings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2024, pp. 5260–5271

    2024

  8. [8]

    Openfedllm: Training large language models on decentralized private data via federated learning,

    R. Ye, W. Wang, J. Chai, D. Li, Z. Li, Y . Xu, Y . Du, Y . Wang, and S. Chen, “Openfedllm: Training large language models on decentralized private data via federated learning,” in Proceedings of the 30th ACM SIGKDD conference on knowledge discovery and data mining , 2024, pp. 6137–6147

    2024

Show all 40 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]

    Split learning for health: Distributed deep learning without sharing raw patient data,

    P. Vepakomma, O. Gupta, T. Swedish, and R. Raskar, “Split learning for health: Distributed deep learning without sharing raw patient data,” arXiv preprint arXiv:1812.00564 , 2018

    2018 arXiv

  3. [11]

    Efficient parallel split learning over resource-constrained wireless edge networks,

    Z. Lin, G. Zhu, Y . Deng, X. Chen, Y . Gao, K. Huang, and Y . Fang, “Efficient parallel split learning over resource-constrained wireless edge networks,” IEEE Transactions on Mobile Computing , vol. 23, no. 10, pp. 9224–9239, 2024

    2024

  4. [12]

    Split learning in 6g edge networks,

    Z. Lin, G. Qu, X. Chen, and K. Huang, “Split learning in 6g edge networks,” IEEE Wireless Communications, 2024

    2024

  5. [13]

    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,” in Proceedings of the AAAI conference on artificial intelligence , vol. 36, no. 8, 2022, pp. 8485– 8493

    2022

  6. [14]

    Lora: Low-rank adaptation of large language models

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chen et al., “Lora: Low-rank adaptation of large language models.” ICLR, vol. 1, no. 2, p. 3, 2022

    2022

  7. [15]

    S-lora: Serving thousands of concurrent lora adapters,

    Y . Sheng, S. Cao, D. Li, C. Hooper, N. Lee, S. Yang, C. Chou, B. Zhu, L. Zheng, K. Keutzer et al. , “S-lora: Serving thousands of concurrent lora adapters,” arXiv preprint arXiv:2311.03285 , 2023

    2023 arXiv

  8. [16]

    Splitlora: A split parameter-efficient fine-tuning framework for large language models,

    Z. Lin, X. Hu, Y . Zhang, Z. Chen, Z. Fang, X. Chen, A. Li, P. Vepakomma, and Y . Gao, “Splitlora: A split parameter-efficient fine-tuning framework for large language models,” arXiv preprint arXiv:2407.00952, 2024

    2024 arXiv

  9. [17]

    Compression ratio allocation for probabilistic semantic communication with RSMA,

    Z. Zhao, Z. Yang, Y . Hu, C. Zhu, M. Shikh-Bahaei, W. Xu, Z. Zhang, and K. Huang, “Compression ratio allocation for probabilistic semantic communication with RSMA,” IEEE Trans. Commun. , pp. 1–1, 2025

    2025

  10. [18]

    Distributed learning in wireless networks: Recent progress and future challenges,

    M. Chen, D. G ¨und¨uz, K. Huang, W. Saad, M. Bennis, A. V . Feljan, and H. V . Poor, “Distributed learning in wireless networks: Recent progress and future challenges,” IEEE Journal on Selected Areas in Communications, vol. 39, no. 12, pp. 3579–3605, 2021

    2021

  11. [19]

    Speeding up distributed machine learning using codes,

    K. Lee, M. Lam, R. Pedarsani, D. Papailiopoulos, and K. Ramchandran, “Speeding up distributed machine learning using codes,” IEEE Trans- actions on Information Theory , vol. 64, no. 3, pp. 1514–1529, 2017

    2017

  12. [20]

    Toward energy- efficient federated learning over 5g+ mobile devices,

    D. Shi, L. Li, R. Chen, P. Prakash, M. Pan, and Y . Fang, “Toward energy- efficient federated learning over 5g+ mobile devices,” IEEE Wireless Communications, vol. 29, no. 5, pp. 44–51, 2022

    2022

  13. [21]

    Service delay minimization for federated learning over mobile devices,

    R. Chen, D. Shi, X. Qin, D. Liu, M. Pan, and S. Cui, “Service delay minimization for federated learning over mobile devices,” IEEE Journal on Selected Areas in Communications , vol. 41, no. 4, pp. 990–1006, 2023

    2023

  14. [22]

    Toward an automated auction framework for wireless federated learning services market,

    Y . Jiao, P. Wang, D. Niyato, B. Lin, and D. I. Kim, “Toward an automated auction framework for wireless federated learning services market,” IEEE Transactions on Mobile Computing , vol. 20, no. 10, pp. 3034–3048, 2020

    2020

  15. [23]

    Federated learning over multi- hop wireless networks with in-network aggregation,

    X. Chen, G. Zhu, Y . Deng, and Y . Fang, “Federated learning over multi- hop wireless networks with in-network aggregation,” IEEE Transactions on Wireless Communications, vol. 21, no. 6, pp. 4622–4634, 2022

    2022

  16. [24]

    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 Transactions on Wireless Communications , vol. 20, no. 2, pp. 1188– 1200, 2020

    2020

  17. [25]

    Split learning over wireless networks: Parallel design and resource management,

    W. Wu, M. Li, K. Qu, C. Zhou, X. Shen, W. Zhuang, X. Li, and W. Shi, “Split learning over wireless networks: Parallel design and resource management,” IEEE Journal on Selected Areas in Communications , vol. 41, no. 4, pp. 1051–1066, 2023

    2023

  18. [26]

    Adaptsfl: Adaptive split federated learning in resource-constrained edge networks,

    Z. Lin, G. Qu, W. Wei, X. Chen, and K. K. Leung, “Adaptsfl: Adaptive split federated learning in resource-constrained edge networks,” arXiv preprint arXiv:2403.13101, 2024

    2024 arXiv

  19. [27]

    Wireless distributed learning: A new hybrid split and federated learning approach,

    X. Liu, Y . Deng, and T. Mahmoodi, “Wireless distributed learning: A new hybrid split and federated learning approach,” IEEE Transactions on Wireless Communications, vol. 22, no. 4, pp. 2650–2665, 2022

    2022

  20. [28]

    A joint communication and learning framework for hierarchical split federated learning,

    L. U. Khan, M. Guizani, A. Al-Fuqaha, C. S. Hong, D. Niyato, and Z. Han, “A joint communication and learning framework for hierarchical split federated learning,” IEEE Internet of Things Journal, vol. 11, no. 1, pp. 268–282, 2023

    2023

  21. [29]

    Ringsfl: An adaptive split federated learning towards taming client heterogeneity,

    J. Shen, N. Cheng, X. Wang, F. Lyu, W. Xu, Z. Liu, K. Aldubaikhy, and X. Shen, “Ringsfl: An adaptive split federated learning towards taming client heterogeneity,” IEEE Transactions on Mobile Computing, vol. 23, no. 5, pp. 5462–5478, 2023

    2023

  22. [30]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

    2016

  23. [31]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556 , 2014

    2014 arXiv

  24. [32]

    Imagenet classification with deep convolutional neural networks,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” Advances in neural informa- tion processing systems , vol. 25, 2012

    2012

  25. [33]

    Low-parameter federated learning with large language models,

    J. Jiang, H. Jiang, Y . Ma, X. Liu, and C. Fan, “Low-parameter federated learning with large language models,” in International Conference on Web Information Systems and Applications . Springer, 2024, pp. 319– 330

    2024

  26. [34]

    Fedcollm: A parameter-efficient federated co-tuning framework for large and small language models,

    T. Fan, Y . Kang, G. Ma, L. Fan, K. Chen, and Q. Yang, “Fedcollm: A parameter-efficient federated co-tuning framework for large and small language models,” arXiv preprint arXiv:2411.11707 , 2024

    2024 arXiv

  27. [35]

    Accelerating split federated learning over wireless communication networks,

    C. Xu, J. Li, Y . Liu, Y . Ling, and M. Wen, “Accelerating split federated learning over wireless communication networks,” IEEE Transactions on Wireless Communications, vol. 23, no. 6, pp. 5587–5599, 2023

    2023

  28. [36]

    Esfl: Ef- ficient split federated learning over resource-constrained heterogeneous wireless devices,

    G. Zhu, Y . Deng, X. Chen, H. Zhang, Y . Fang, and T. F. Wong, “Esfl: Ef- ficient split federated learning over resource-constrained heterogeneous wireless devices,” IEEE Internet of Things Journal , 2024

    2024

  29. [37]

    Applications of second-order cone programming,

    M. S. Lobo, L. Vandenberghe, S. Boyd, and H. Lebret, “Applications of second-order cone programming,” Linear algebra and its applications , vol. 284, no. 1-3, pp. 193–228, 1998

    1998

  30. [38]

    Edge and central cloud computing: A perfect pairing for high energy efficiency and low-latency,

    X. Hu, L. Wang, K.-K. Wong, M. Tao, Y . Zhang, and Z. Zheng, “Edge and central cloud computing: A perfect pairing for high energy efficiency and low-latency,” IEEE Transactions on Wireless Communications , vol. 19, no. 2, pp. 1070–1083, 2019

    2019

  31. [39]

    The e2e dataset: New challenges for end-to-end generation,

    J. Novikova, O. Du ˇsek, and V . Rieser, “The e2e dataset: New challenges for end-to-end generation,” arXiv preprint arXiv:1706.09254 , 2017

    2017 arXiv

  32. [40]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever et al., “Language models are unsupervised multitask learners,” OpenAI blog , vol. 1, no. 8, p. 9, 2019

    2019

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.