Pith. sign in

REVIEW 3 major objections 6 minor 38 references

FedLBW: A Loss-Based Weighting Strategy for Federated Learning on Non-IID Data in Wireless Networks

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read FedLBW claims that weighting client updates by inverse validation loss on a small server proxy set, instead of by dataset size, raises accuracy and convergence under non-IID data and client dropout.

desk verdict A simple, honestly evaluated proxy-based aggregation rule with a real gap in its convergence proof—worth refereeing once the sampling argument is fixed. read the letter →

arxiv 2608.07007 v1 pith:BABZCZMF submitted 2026-08-07 cs.AI cs.DCcs.ETcs.LG

Majid Kundroo , Tinku Singh , Taehong Kim This is my paper · ORCID
classification cs.AIcs.DCcs.ETcs.LG
keywords FederatedlearningNon-IIDdataLoss-basedaggregationServerproxydatasetClientdropoutWirelessnetworksConvergenceanalysis
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 argues that in federated learning the standard practice of weighting client updates by dataset size is the wrong signal under non-IID data and wireless dropouts. It proposes FedLBW, which weights each returned model by the inverse of its validation loss on a small, class-balanced proxy dataset held by the server, so low-loss models get more influence. The authors claim this substitution yields higher accuracy and faster convergence, with up to 7.6% higher accuracy on CIFAR-10 at the most skewed Dirichlet setting, and keeps accuracy stable when up to half of clients drop out. The reason to care is that the scheme replaces an unavailable global data distribution with a cheap, server-side quality signal during aggregation.

What carries the argument

The load-bearing mechanism is the inverse-loss weighting rule at Eq. (4): replace the data-size coefficients $n_c/\sum_j n_j$ in FedAvg with coefficients $\beta_c = (1/L_c^{t+1})/\sum_{j \in S_t} (1/L_j^{t+1})$. Here $L_c^{t+1}$ is the average cross-entropy loss of client $c$'s updated model on the server's small class-balanced proxy set $V$. The inversion converts lower validation loss into higher aggregation weight while keeping the update a convex combination, and the proxy set lets the server evaluate every participating client on a common fixed reference distribution. The convergence analysis leans on Assumption 5, which bounds how far these loss-based weights deviate from uniform and how concentrated their squared values are, because that concentration factor $\rho$ controls both the stochastic-noise term and the weighting-bias term in Theorem 4.1.

What would settle it

Compute the Spearman correlation between client rankings induced by proxy validation loss and rankings induced by true held-out accuracy across many rounds; if that correlation is near zero or negative, FedLBW's weights systematically favor worse models, and the experiment should show FedLBW matching or falling below FedAvg. Table 5 already shows the correlation dropping to 0.42 for an SVHN proxy, so the sharp check is a stronger proxy mismatch where the ranking is reversed and FedLBW's accuracy falls below its data-size-weighted baseline.

Watch

Extended reading notes

Core claim

The central claim is that the aggregation step in federated learning should weight client updates by the inverse of each client model's validation loss on a small server-side proxy dataset, rather than by the number of local samples. Formally, the global update is $w^{t+1} = \sum_{c \in S_t} \beta_c w_c^{t+1}$ with $\beta_c \propto 1/L_c^{t+1}$ and $L_c^{t+1}$ the average cross-entropy loss of client $c$'s returned model on the proxy set. The paper reports that FedLBW outperforms FedAvg, FedAvgM, FedProx, FedNova, FedLAW, and FedDkw across FashionMNIST, CIFAR-10, and CIFAR-100, with the largest gains at extreme non-IID skew, and that it retains most of its accuracy under client dropout while FedAvg degrades sharply. It also contributes a nonconvex convergence bound that separates optimization error, stochastic noise, client heterogeneity, and loss-weight concentration, and argues that bounded loss ratios give FedLBW a smaller concentration factor, and thus less bias, than FedAvg.

Load-bearing premise

The load-bearing premise is that validation losses computed on the server's small proxy dataset rank clients by their true generalization quality; if the proxy's ranking is wrong, FedLBW will up-weight the wrong clients and its advantage over FedAvg collapses.

Editorial extensions

If this is right

  • In non-IID federated learning, switching aggregation weights from dataset size to inverse validation loss should improve final global accuracy, with the largest measured gain of 7.6% over FedAvg on CIFAR-10 at Dirichlet $\alpha=0.1$.
  • Convergence should be faster in early rounds because the aggregation direction is dominated by low-loss clients; the reported convergence curves show smoother and steeper accuracy increases in the first rounds.
  • Under client dropout, FedLBW should stay near its no-dropout accuracy because surviving reliable clients keep high weight; at a dropout probability of 0.5 on CIFAR-10 it preserves about 90% of its baseline performance while FedAvg loses about 28%.
  • The convergence bound implies that the method's advantage is governed by the weight concentration factor $\rho$: with tightly bounded validation-loss ratios, FedLBW's $\rho$ can be smaller than FedAvg's, reducing the bias term $G^2(\rho-1)$.

Reading between the lines

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

  • Editorial inference: FedLBW effectively turns the server's proxy dataset into a learned similarity measure, so the method's ceiling is set by proxy representativeness; one natural extension is to update or diversify the proxy online when client data drift is detected.
  • Editorial inference: the inverse-loss rule is a deterministic performance-weighted ensemble, so combining it with client sampling that is proportional to $\beta_c$ rather than to $n_c$ would directly align the partial-participation analysis with the algorithm actually run.
  • Editorial inference: a testable extension is temporal smoothing or clipping of the loss-based weights, which the paper mentions only as an enforcement mechanism for the theory but does not evaluate as a practical stabilizer.
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

3 major / 6 minor

Summary. The paper introduces FedLBW, a federated aggregation rule that replaces dataset-size weights with weights inversely proportional to each client model's validation loss on a server-held proxy set. The authors motivate the method for wireless settings with non-IID data and dropouts, provide a convergence analysis in Section 4, and report experiments on FashionMNIST, CIFAR-10, and CIFAR-100 showing 1.5-7.7 percentage point accuracy gains over FedAvg and consistent gains over FedLAW, FedDkw, and other baselines, plus a dropout-robustness study and a proxy-shift sensitivity study.

Significance. The empirical study is broad and the core idea is simple and plausible: if the proxy ranking is informative, inverse-loss weighting can suppress poorly generalizing clients. The paper's strengths are the range of datasets, the three-seed repetitions, the comparison against six baselines, and the honest proxy-robustness experiments in Table 5. However, the theoretical contribution is not sound in its current form: the main convergence bound relies on a biased-sampling step and on unverified concentration assumptions, and the theorem's objective does not match the algorithm's actual round-dependent weighting. The empirical claims may survive a revision, but the paper's advertised theoretical guarantee should not be taken as established.

major comments (3)
  1. [Section 4.3, Lemma 4.2] Lemma 4.2 does not follow from Assumptions 2-4 and Eq. (4). With uniform client sampling and loss-based weights beta_t, the conditional expectation of u_t = sum_{c in S_t} beta_c grad F_c(w_t) - grad F(w_t) contains a deterministic bias. In the special case beta_c = 1/|S_t|, E[u_t | w_t] = (1/C) sum_c grad F_c(w_t) - sum_c (n_c/N) grad F_c(w_t), which is not zero unless client sizes are equal or sampling is proportional to n_c. Table 1 shows n_c ranges from 14 to 1409 on CIFAR-10 at alpha=0.1, exactly the regime highlighted in the paper. The first term in the bound controls only the variance around the unweighted client mean; the second term is a pure weighting-bias term unrelated to the n_c mismatch. Therefore the stated bound is false, and Theorem 4.1's convergence to a stationary point of F in Eq. (2) is not established for the setup defined in Section 3.2. Repairing this requires either sampling clients with probability proportional to n_c or proving a new bound that explicitly controls the weighted-vs-unweighted gradient mismatch.
  2. [Section 4.2, Assumption 5; Section 4.4, Corollary 4.2] Assumption 5 is doing the entire work of the claimed theoretical advantage, but it is neither derived from the proxy-loss mechanism nor empirically verified. The theorem's method-specific bias term is G^2(rho-1), and Corollary 4.2 asserts FedLBW achieves smaller rho than FedAvg provided Lmin/Lmax >= delta. No result in the paper establishes this loss-ratio condition, and Section 6.4 shows the proxy ranking can be weakly aligned with client quality (Spearman rho=0.4219 for SVHN), so the condition cannot be taken for granted. The authors should either derive delta (or rho) from properties of the validation loss and proxy set, or report measured values of rho and B_max in the experiments; otherwise the comparison with FedAvg is circular.
  3. [Section 4.4, Theorem 4.1 vs Section 3.2, Eq. (5)] The convergence analysis studies the static data-size-weighted objective F in Eq. (2), but Algorithm 1 implements the round-dependent objective in Eq. (5), whose coefficients beta_t depend on the current validation losses. The proof bounds descent on F while an update with beta_t != n_c/N is applied; even if Lemma 4.2 were repaired, the theorem would still be about a different optimization problem. The paper should state explicitly which objective FedLBW is claimed to converge to, and the analysis should handle the time-varying nature of the loss-based weights.
minor comments (6)
  1. [Section 3.3, Algorithm 1] Section 3.3 refers to 'the aggregation step line 10', but in Algorithm 1 the aggregation happens on line 9 and the ClientUpdate function starts at line 11; please correct the cross-reference.
  2. [Figures 2-4] The y-axis labels read 'Accuracy (%)' while the tick values are 0.0-0.9; please make the units consistent by either plotting percentages (e.g., 20-80) or changing the label to 'Accuracy'.
  3. [Section 6.4, Table 5] The symbol rho is used both for the Spearman rank correlation coefficient and for the weight concentration factor in Section 4.1; please use a different symbol for one of these quantities.
  4. [Table 1] The 'Training Samples per Client' statistics are useful, but the header layout and the 'Data Distribution' entries make the table hard to parse; consider a clearer multi-row layout with a separate row for each dataset and alpha value.
  5. [Section 6.1, Table 2] The reported gains over FedAvg are arithmetic differences of means; given the reported standard deviations (e.g., 58.63 +/- 3.02 vs 50.97 +/- 3.55 at CIFAR-10 alpha=0.1), a significance test or a statement about overlap would strengthen the comparison.
  6. [Section 3.5] The complexity O(|V||S_t|) for the server-side loss computation should be stated as a per-round cost and should clarify that it assumes one forward pass per client model over the proxy set; otherwise it may be read as an end-to-end training cost.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FedLBW's empirical claims rest on out-of-sample test accuracy and no fitted parameter is relabeled as a prediction; the Lemma 4.2 gap is a correctness concern, not a circular reduction.

full rationale

FedLBW's derivation chain is not equivalent to its inputs. The inverse-validation-loss weights in Eq. (4) are computed from server-side proxy losses, and the reported accuracies in Tables 2-4 are final test accuracies, so the empirical claim is not obtained by fitting test labels. No constant or parameter is fit to a subset and then predicted on a closely related quantity. The self-citations ([4], [14], [24], [34]) are background, hyperparameter, and software-tool citations and are not load-bearing; the proxy-data motivation is supported by external references ([10], [11], [12], [32]). The convergence analysis has a genuine formal gap: Lemma 4.2 bounds u_t by sampling variance plus a weight-concentration term but omits the deterministic bias between the uniformly sampled client average and the data-size-weighted objective F of Eq. (2), so Theorem 4.1's guarantee is unsupported in the unequal-n_c regime of Table 1. This is an omitted-proof or correctness flaw, not a circular step, because the theorem's conclusion is not assumed as an input. Assumption 5(b) is best read as defining rho for normalized weights (such a rho always exists in [1, |S_t|]) rather than as asserting concentration; Corollary 4.2 is explicitly conditional on the loss-ratio bound delta. Section 6.4 openly reports degraded proxy ranking under SVHN (rho=0.42), a robustness limitation that does not make the out-of-sample evaluation circular. Overall, no load-bearing step reduces by construction to its own input.

Assumptions & free parameters 3 free parameters · 7 assumptions · 0 invented entities

The central numerical claims are empirical and do not require fitted constants. The theory, however, relies on unproven concentration of loss-based weights (Assumption 5) and on an implicit equal-data-size or proportional-sampling assumption in Lemma 4.2. The proxy dataset size is a hand-set design choice. No new physical or mathematical entities are introduced.

free parameters (3)
  • Server proxy dataset size per class = 100 samples/class for FashionMNIST and CIFAR-10; 10 samples/class for CIFAR-100
    Hand-set design choice following FedLAW. It controls the variance of the validation-loss estimates that feed Eq. (4), but it is not fitted to the target accuracy.
  • Weight concentration factor rho
    Introduced in Assumption 5(b). The convergence bound becomes favorable when rho is small, but Algorithm 1 contains no mechanism that guarantees any particular rho, so this is an assumed constant rather than a measured or fitted one.
  • Uniform-weight deviation bound B_max
    Introduced in Assumption 5(a) to control the weighting bias term in the convergence bound. No procedure in the paper estimates or enforces this bound.
assumptions (7)
  • standard math Each local objective F_c is L_f-smooth (Assumption 1).
    Standard smoothness assumption used in the descent lemma of Theorem 4.1.
  • standard math Mini-batch gradients are unbiased with bounded variance and bounded gradient norm (Assumption 2).
    Used in Lemmas 4.1 and 4.3. It is not verified for the CNN and ResNet models used in the experiments.
  • domain assumption Client heterogeneity is bounded by zeta^2 (Assumption 3).
    Standard bounded-heterogeneity condition. It cannot be checked from the Dirichlet partition parameters alone.
  • domain assumption Clients in S_t are sampled uniformly at random without replacement, independent of mini-batch sampling (Assumption 4).
    Matches Algorithm 1 line 4, but this sampling rule conflicts with the data-size-weighted global objective when client dataset sizes are unequal.
  • ad hoc to paper Loss-based weights satisfy bounded deviation and concentration bounds (Assumption 5).
    This is the load-bearing assumption of the convergence analysis. The theorem's advantage over FedAvg relies on rho being small, but no training or clipping mechanism in Algorithm 1 enforces the bound.
  • domain assumption Proxy validation loss on the server's small dataset ranks clients by true generalization quality on the global task.
    Needed for Eq. (4) to up-weight better clients. Table 5 shows that the ranking quality degrades under domain shift, with Spearman rho dropping to 0.42 for SVHN.
  • ad hoc to paper Uniform sampling yields unbiased estimates of the data-size-weighted global gradient.
    Lemma 4.2's finite-population variance formula implicitly requires this. With unequal n_c as reported in Table 1, the uniform sample mean is biased by the weight imbalance across clients.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FedLBW: A Loss-Based Weighting Strategy for Federated Learning on Non-IID Data in Wireless Networks." pith.science (2026). https://pith.science/paper/BABZCZMF

@misc{pith2026260807007,
  author       = {Pith},
  title        = {Pith review of: FedLBW: A Loss-Based Weighting Strategy for Federated Learning on Non-IID Data in Wireless Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BABZCZMF}},
  note         = {Machine review of arXiv:2608.07007}
}
read the original abstract

Federated Learning (FL) enables collaborative machine learning (ML) across distributed clients while preserving privacy. However, efficient model convergence in FL remains challenging, especially in wireless networks where non-independent and identically distributed (non-IID) data and frequent client dropouts are common. Traditional FL algorithms, such as FedAvg, rely solely on dataset size to weight client updates. This introduces biases towards clients with larger datasets and makes the process sensitive to non-IID data, outliers, and client dropouts. To address these challenges, we propose Federated Learning with Loss-Based Weighting (FedLBW), a novel aggregation method that assigns each client's update a weight proportional to the inverse of its validation loss, computed using a small proxy dataset on the server, rather than its dataset size. This ensures that lower-loss models exert greater influence during aggregation, prioritizing the most reliable updates and boosting overall performance. Through extensive experiments across multiple datasets, including FashionMNIST (CNN), CIFAR-10 (ResNet-18), and CIFAR-100 (ResNet-34), we demonstrate that FedLBW achieves higher accuracy and faster convergence compared to baseline algorithms such as FedAvg, FedAvgM, FedProx, FedNova, FedLAW and FedDkw, with notable improvements of up to 7.6 % higher accuracy on CIFAR-10 in extreme non-IID cases. Moreover, FedLBW showcases exceptional resilience to increasing dropout probabilities, consistently maintaining significantly higher accuracy even in challenging conditions. These results establish FedLBW as an effective and resilient solution for FL in wireless network environments, offering marked improvements in model accuracy, convergence speed, and robustness to non-IID data and client dropouts.

Figures

Figures reproduced from arXiv: 2608.07007 by the authors.

Figure 1
Figure 1. Overview of the proposed Federated Learning with Loss-Based Weighting [PITH_FULL_IMAGE:figures/full_fig_p008_1.png] view at source ↗
Figure 2
Figure 2. Global accuracy over rounds for the FashionMNIST dataset using a CNN model [PITH_FULL_IMAGE:figures/full_fig_p023_2.png] view at source ↗
Figure 3
Figure 3. Global accuracy over rounds for the CIFAR-10 dataset using a ResNet-18 ar [PITH_FULL_IMAGE:figures/full_fig_p024_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Global accuracy over rounds for the CIFAR-100 dataset using a ResNet-34 [PITH_FULL_IMAGE:figures/full_fig_p024_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 24 canonical work pages

  1. [1]

    Konečný, H

    J. Konečný, H. B. McMahan, F. X. Yu, P. Richtárik, A. T. Suresh, D.Bacon, FederatedLearning: StrategiesforImprovingCommunication Efficiency, CoRR (2016) 1–10arXiv:1610.05492. URLhttp://arxiv.org/abs/1610.05492

    arXiv 2016

  2. [2]

    T. Wang, Z. Zheng, F. Lin, Federated learning framework based on trimmed mean aggregation rules, Expert Systems with Applications 270 (2025) 126354. doi:10.1016/j.eswa.2024.126354

    arXiv 2025

  3. [3]

    Z. Chen, W. Yi, H. Shin, A. Nallanathan, Adaptive semi-asynchronous federated learning over wireless networks, IEEE Transactions on Com- munications (2024) 1–1doi:10.1109/TCOMM.2024.3425635

    arXiv 2024

  4. [4]

    A. W. Mamond, M. Kundroo, S.-e. Yoo, S. Kim, T. Kim, Fldqn: Co- operative multi-agent federated reinforcement learning for solving travel time minimization problems in dynamic environments using sumo sim- ulation, Sensors 25 (3) (2025). doi:10.3390/s25030911. URLhttps://www.mdpi.com/1424-8220/25/3/911

    work page doi:10.3390/s25030911 2025

  5. [5]

    McMahan, E

    B. McMahan, E. Moore, D. Ramage, S. Hampson, B. A. y Arcas, Communication-efficient learning of deep networks from decentralized data, in: Artificial intelligence and statistics, PMLR, 2017, pp. 1273– 1282

    2017

  6. [6]

    Y. Hu, H. Ren, C. Hu, J. Deng, X. Xie, An element-wise weights aggre- gation method for federated learning, in: 2023 IEEE International Con- ferenceonDataMiningWorkshops(ICDMW),IEEE,2023, pp.188–196

    work page 2023

  7. [7]

    Annunziata, M

    D. Annunziata, M. Canzaniello, M. Savoia, S. Cuomo, F. Piccialli, Benchmarking federated learning on high-performance computing: Ag- gregation methods and their impact, in: 2024 32nd Euromicro Interna- tional Conference on Parallel, Distributed and Network-Based Process- ing (PDP), 2024, pp. 207–214. doi:10.1109/PDP62718.2024.00036

    work page arXiv 2024

  8. [8]

    Y. Sun, Y. Mao, J. Zhang, Mimic: Combating client dropouts in feder- ated learning by mimicking central updates, IEEE Transactions on Mo- bile Computing 23 (7) (2024) 7572–7584. doi:10.1109/tmc.2023.3338021. URLhttp://dx.doi.org/10.1109/TMC.2023.3338021 32

    arXiv 2024

Show all 38 references
  1. [9]

    Ghasemi, A

    A. Ghasemi, A. Keshavarzi, A. M. Abdelmoniem, O. R. Nejati, T. De- rikvand, Edge intelligence for intelligent transport systems: Approaches, challenges, and future directions, Expert Systems with Applications 280 (2025) 127273. doi:https://doi.org/10.1016/j.eswa.2025.127273. U...

    2025

  2. [10]

    Z. Li, T. Lin, X. Shang, C. Wu, Revisiting weighted aggregation in federated learning with neural networks, in: International Conference on Machine Learning, PMLR, 2023, pp. 19767–19788

    2023

  3. [11]

    H. Yang, P. Qiu, P. Khanduri, M. Fang, J. Liu, Understanding server- assisted federated learning in the presence of incomplete client partici- pation, in: R. Salakhutdinov, Z. Kolter, K. Heller, A. Weller, N. Oliver, J. Scarlett, F. Berkenkamp (Eds.), Proceedings of the 41st I...

    2024

  4. [12]

    Zawad, X

    S. Zawad, X. Ma, J. Yi, C. Li, M. Zhang, L. Yang, F. Yan, Y. He, Fed- cust: Offloading hyperparameter customization for federated learning, Perform. Eval. 167 (C) (Apr. 2025). doi:10.1016/j.peva.2024.102450. URLhttps://doi.org/10.1016/j.peva.2024.102450

    2025

  5. [13]

    T.-M. H. Hsu, H. Qi, M. Brown, Measuring the effects of non- identical data distribution for federated visual classification (2019). arXiv:1909.06335. URLhttps://arxiv.org/abs/1909.06335

    2019 arXiv

  6. [14]

    Kundroo, T

    M. Kundroo, T. Kim, Demystifying impact of key hyper-parameters in federated learning: A case study on cifar-10 and fashionmnist, IEEE Access 12 (2024) 120570–120583. doi:10.1109/ACCESS.2024.3450894

    2024

  7. [15]

    Y. Zhao, M. Li, L. Lai, N. Suda, D. Civin, V. Chandra, Federated learning with non-iid data, arXiv preprint arXiv:1806.00582 (2018)

    2018 arXiv

  8. [16]

    T. Li, A. K. Sahu, A. Talwalkar, V. Smith, Federated learning: Chal- lenges, methods, and future directions, IEEE signal processing magazine 37 (3) (2020) 50–60. 33

    2020

  9. [17]

    Q. Li, Z. Wen, Z. Wu, S. Hu, N. Wang, Y. Li, X. Liu, B. He, A survey on federatedlearningsystems: Vision, hypeandrealityfordataprivacyand protection, IEEE Transactions on Knowledge & Data Engineering 35 (04) (2023) 3347–3366. doi:10.1109/TKDE.2021.3124599

    2023

  10. [18]

    Smith, C.-K

    V. Smith, C.-K. Chiang, M. Sanjabi, A. S. Talwalkar, Federated multi- task learning, Advances in neural information processing systems 30 (2017)

    2017

  11. [19]

    Reddi, Z

    S. Reddi, Z. Charles, M. Zaheer, Z. Garrett, K. Rush, J. Konečný, S. Kumar, H. B. McMahan, Adaptive federated optimization (2021). arXiv:2003.00295. URLhttps://arxiv.org/abs/2003.00295

    2021 arXiv

  12. [20]

    5442–5444

    J.C.Duchi, P.L.Bartlett, M.J.Wainwright, Randomizedsmoothingfor (parallel) stochastic optimization, in: 2012 IEEE 51st IEEE Conference on Decision and Control (CDC), IEEE, 2012, pp. 5442–5444

    2012

  13. [21]

    Zaheer, S

    M. Zaheer, S. Reddi, D. Sachan, S. Kale, S. Kumar, Adaptive methods for nonconvex optimization, in: S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, R. Garnett (Eds.), Advances in Neural Information Processing Systems, Vol. 31, Curran Associates, Inc., 2018, ...

    2018

  14. [22]

    D. P. Kingma, J. Ba, Adam: A method for stochastic optimization, arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  15. [23]

    S. P. Karimireddy, S. Kale, M. Mohri, S. J. Reddi, S. U. Stich, A. T. Suresh, SCAFFOLD: Stochastic Controlled Averaging for Federated Learning, in: 37thInternationalConferenceonMachineLearning, ICML 2020, Vol. PartF16814, 2020, pp. 5088–5099. arXiv:1910.06378

    2020 arXiv

  16. [24]

    Kundroo, T

    M. Kundroo, T. Kim, Efficient federated learning with adaptive client- side hyper-parameter optimization, in: 2023 IEEE 43rd International Conference on Distributed Computing Systems (ICDCS), 2023, pp. 973–

    2023

  17. [25]

    J. Wang, Q. Liu, H. Liang, G. Joshi, H. V. Poor, A novel framework for the analysis and design of heterogeneous federated learning, IEEE Transactions on Signal Processing 69 (2021) 5234–5249

    2021

  18. [26]

    H.Wang, M.Yurochkin, Y.Sun, D.S.Papailiopoulos, Y.Khazaeni, Fed- erated learning with matched averaging, in: 8th International Confer- ence on Learning Representations, ICLR 2020, Addis Ababa, Ethiopia, April 26-30, 2020, OpenReview.net, 2020, pp. 1–16. URLhttps://openreview.n...

    2020

  19. [27]

    Pillutla, S

    K. Pillutla, S. M. Kakade, Z. Harchaoui, Robust aggregation for feder- ated learning, IEEE Transactions on Signal Processing 70 (2022) 1142– 1154

    2022

  20. [28]

    Zhang, Y

    J. Zhang, Y. Li, D. Wu, Y. Zhao, S. Palaiahnakote, Sffl: Self- aware fairness federated learning framework for heterogeneous data distributions, Expert Systems with Applications 269 (2025) 126418. doi:https://doi.org/10.1016/j.eswa.2025.126418. URLhttps : / / www . sciencedire...

    2025

  21. [29]

    R. Wang, J. Wang, X. Li, J. Lai, F. Zhang, X. Pei, M. K. Khan, Cesa: Communication efficient secure aggregation scheme via sparse graph in federated learning, Journal of Network and Computer Applications 231 (2024) 103997. doi:https://doi.org/10.1016/j.jnca.2024.103997. URLhtt...

    2024

  22. [30]

    Y. Wang, B. Kantarci, Reputation-enabled federated learning model aggregation in mobile platforms, in: ICC 2021 - IEEE International Conference on Communications, 2021, pp. 1–6. doi:10.1109/ICC42927.2021.9500928

    2021

  23. [31]

    J. Geng, Y. Hou, X. Tao, J. Wang, B. Luo, Adaptive federated learning in heterogeneous wireless networks with independent sampling, in: ICC 2024-IEEE International Conference on Communications, IEEE, 2024, pp. 944–949

    2024

  24. [32]

    Scott, A

    J. Scott, A. Cahill, Improved modelling of federated datasets using mixtures-of-Dirichlet-multinomials, in: R. Salakhutdinov, Z. Kolter, 35 K. Heller, A. Weller, N. Oliver, J. Scarlett, F. Berkenkamp (Eds.), Pro- ceedings of the 41st International Conference on Machine Learnin...

    2024

  25. [33]

    X. Li, K. Huang, W. Yang, S. Wang, Z. Zhang, On the Convergence of FedAvg on Non-IID Data, arXiv preprint arXiv:1907.02189 (2019) 1– 26arXiv:1907.02189. URLhttp://arxiv.org/abs/1907.02189

    2019 arXiv

  26. [34]

    Kundroo, G

    M. Kundroo, G. Haider, N. Khoa, A. W. Mamond, T. Kim, Fedeasy : Federated learning with ease, SoftwareX 31 (2025) 102276. doi:https://doi.org/10.1016/j.softx.2025.102276

    2025

  27. [35]

    B. Li, S. Chen, K. Yu, Feddkw – federated learning with dynamic kullback–leibler-divergence weight, ACM Trans. Asian Low-Resour. Lang. Inf. Process. (Apr. 2023). doi:10.1145/3594779. URLhttps://doi.org/10.1145/3594779

    2023 doi

  28. [36]

    Spearman, The proof and measurement of association between two things, The American journal of psychology 100 (3/4) (1987) 441–471

    C. Spearman, The proof and measurement of association between two things, The American journal of psychology 100 (3/4) (1987) 441–471

    1987

  29. [37]

    Heusel, H

    M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler, S.Hochreiter, Gans trained by a two time-scale update rule converge to a local nash equi- librium, Advances in neural information processing systems 30 (2017). 36

    2017

  30. [974]

    doi:10.1109/ICDCS57875.2023.00103. 34

    2023

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.