Pith. sign in

REVIEW 3 major objections 5 minor 44 references

H-FedSN: Personalized Sparse Networks for Efficient and Accurate Hierarchical Federated Learning for IoT Applications

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read H-FedSN trains clients to learn only binary masks over frozen weights, cutting hierarchical federated learning communication by 58–238x while keeping accuracy.

desk verdict Plausible core idea but the headline communication reduction is an uplink-only artifact; the paper needs major revision before its quantitative claims hold up. read the letter →

arxiv 2412.06210 v2 pith:ZSR33M6R submitted 2024-12-09 cs.LG

Jiechao Gao , Yuangang Li , Yue Zhao , Brad Campbell This is my paper · ORCID
classification cs.LG
keywords HierarchicalFederatedLearningCommunicationEfficiencyIoTBayesianAggregationBinaryMasksSparseNetworksPersonalizedNon-IIDData
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 introduces H-FedSN, a hierarchical federated learning method for three-tier IoT networks (devices, edge servers, and cloud). Its central claim is that each client can obtain an accurate personalized model without ever transmitting weight updates: the device's randomly initialized weights stay frozen, and the client learns only a binary mask that selects which connections to keep. Shared mask layers are aggregated up the hierarchy with a Bayesian update, while private layers remain on the device for personalization. In experiments on MNIST and three IoT activity and gesture datasets under non-IID splits, H-FedSN reports communication reductions of 58 to 238 times compared with the HierFAVG baseline while matching or exceeding the accuracy of communication-focused baselines. If correct, this makes hierarchical federated learning practical on bandwidth-constrained IoT links.

What carries the argument

The load-bearing mechanism is a binary mask applied to frozen weights: the model is $w = m \odot w_{init}$ with $m \in \{0,1\}^d$ drawn from a Bernoulli distribution with learned parameter $\theta \in [0,1]^d$. Because only the mask is uploaded, each parameter costs 1 bit instead of 32, which is the source of the communication reduction; because the mask is partitioned into aggregated shared layers and local private layers, the method combines collaborative learning with per-client personalization. The Bayesian aggregation at edge and cloud servers treats each binary mask as a Bernoulli observation, cumulatively updates the parameters $\alpha$ and $\beta$ of a Beta distribution (with priors reset every 10 rounds), and uses the posterior mode $(\alpha-1)/(\alpha+\beta-2)$ as the next round's probability mask. This aggregation is intended to balance the influence of edges with different numbers of clients.

What would settle it

Run H-FedSN on one of the paper's datasets with the Bayesian aggregation at edge and cloud replaced by simple element-wise averaging of the uploaded binary masks, keeping mask training, frozen weights, and private layers unchanged; if accuracy does not drop, Bayesian aggregation is not load-bearing and the reported accuracy must be attributed to mask compression and personalization alone.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a hierarchical federated model can be built from a frozen random weight vector $w_{init}$ and per-client binary masks $m_k \sim \mathrm{Bernoulli}(\theta_k)$, giving the model $\dot{w}_k = m_k \odot w_{init}$. Local training updates a real-valued score mask $s_k$ through the sigmoid, and the straight-through estimator makes the non-differentiable Bernoulli sampling pass gradients back to the score. Masks are split into shared layers, which are aggregated at edge and cloud servers via Bayesian updates of Beta distribution parameters, and private layers that stay local, producing a personalized sparse network for each client. The paper reports that transmitting only these masks, at 1 bit per element, reduces communication cost by 58–238x compared to HierFAVG across four datasets under non-IID settings, with accuracy comparable to or better than personalized FL baselines.

Load-bearing premise

The central accuracy claim rests on the premise that a randomly initialized network with frozen weights can be made accurate for each client by learning only a binary mask, and that the Bayesian aggregation specifically contributes to this accuracy; if either fails, the method reduces to a communication-compression trick without the reported accuracy benefit.

Editorial extensions

If this is right

  • A three-tier FL system can reduce client-to-edge and edge-to-cloud traffic by one to two orders of magnitude compared with HierFAVG, which would make FL viable on low-bandwidth IoT links.
  • Communication cost becomes independent of the precision of model weights; increasing model capacity adds only 1 bit per new parameter to the uploaded masks.
  • Because private layers never leave the device, the local personalization component is never exposed to the server, an added privacy-adjacent property.
  • The periodic reset of the Beta priors gives the aggregation a built-in recency bias, which is claimed to prevent overfitting to historical masks and to handle imbalanced client counts across edges.

Reading between the lines

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

  • A testable extension is to vary the split between shared and private layers; the reported communication savings and accuracy will trace out a trade-off curve that the paper does not provide.
  • The same frozen-weight-mask formulation could be applied to two-tier FL or to other personalized-FL algorithms, potentially giving them the same compression benefit without changing their aggregation rules.
  • The Bayesian aggregation's reset period of 10 rounds is a hyperparameter; a sensitivity sweep would reveal whether the reported performance holds for shorter or longer windows.
  • An end-to-end accounting that includes the cloud's broadcast of the global probability mask would clarify the true savings in a real deployment, since the broadcast is not a binary mask.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes H-FedSN, a hierarchical federated learning method in which each client trains a binary mask over a frozen, randomly initialized network, with shared layers aggregated via Beta-Bernoulli Bayesian updates at edge and cloud levels and private layers kept local. Experiments on MNIST and three IoT datasets under four network configurations compare communication cost and accuracy against HierFAVG and several adapted FL baselines. The paper claims a 58–238x communication reduction relative to HierFAVG while achieving comparable or better accuracy.

Significance. If confirmed, the method could make hierarchical FL practical for bandwidth-limited IoT: exchanging 1-bit masks for shared layers is a structurally sound way to reduce uplink traffic, and the paper covers four datasets and four topologies, which is a useful empirical scope. The paper also ships the full algorithm descriptions in a reproducible form. However, the headline communication reduction is computed with an uplink-only metric, the accuracy comparison is confounded by a 10–35x difference in learning rates, and the Bayesian aggregation component is not ablated. These issues must be addressed before the central claims can be accepted.

major comments (3)
  1. [Section IV-B2, Algorithm 1 line 30] The communication-overhead metric in Section IV-B2 counts only client-to-edge and edge-to-cloud traffic, not the cloud-to-edge-to-client broadcast of the global probability mask theta_g,t performed in Algorithm 1 line 30 and described in Section III-C2. The reported 238.8x reduction for E2C5 MNIST is exactly s/(32d), where s is the number of shared parameters and d the total parameter count, confirming that only the 1-bit shared-mask uplink is charged. If the downstream broadcast of theta_g,t (32-bit floats per shared parameter) is counted, the end-to-end reduction is substantially smaller. The authors should either scope all claims to uplink-only traffic explicitly or include the downstream broadcast in the cost model and re-report the reduction factors, since the '58 to 238 times' figure is a central advertised result.
  2. [Section IV-B3] The experimental setting states that H-FedSN uses a learning rate of 0.01 for all datasets except WIDAR (0.035), while all baseline algorithms use a uniformly set learning rate of 0.001. This 10–35x difference is a major confound for the accuracy comparisons: a higher learning rate can fully explain an accuracy advantage, and the paper's claim of 'comparable or even improving' accuracy over baselines is therefore not reliably supported. The authors should tune the baselines per dataset (or use the same learning-rate schedule) and report the corresponding accuracy numbers, ideally with multiple seeds.
  3. [Section III-C and Section IV-C] The paper attributes accuracy improvements to Bayesian aggregation at edge and cloud levels, but never ablates this component. A minimal ablation would replace the Bayesian update (Algorithms 2 and 3) with a simple element-wise mean of the submitted binary masks, keeping all other parts (mask training, private layers, communication) identical. Without such a comparison, the reader cannot determine whether the reported accuracy comes from the Bayesian aggregation, from mask training, or from the private layers; this makes the design justification for a central algorithmic contribution unsupported.
minor comments (5)
  1. [Algorithm 3, line 2] The reset step in the cloud aggregation uses 'alpha_e,t-1 = beta_e,t-1 = lambda_0', but this should presumably be the global parameters alpha_g,t-1 and beta_g,t-1.
  2. [Algorithms 2 and 3] The predicate ResPriors(t) is never defined; the text says the priors are reset 'at the start of every 10th training round', which is ambiguous (t mod 10 = 0 or t mod 10 = 1?). Please specify the exact condition and use the same convention at edge and cloud.
  3. [Section III-C1] The formula for beta_e,t is typeset ambiguously as 'beta_e,t = beta_e,t-1 + |Ke| · 1 - sum ms_k,t'; please use explicit element-wise notation with the all-ones vector and parentheses, e.g., beta_e,t = beta_e,t-1 + |Ke|*1 - sum_{k in Ke} ms_k,t.
  4. [Section IV-B1] The model is described as a '4-layer CNN' but the text lists two convolutional blocks (four convolutional layers) plus three fully-connected layers; please clarify the layer count and specify the convolutional kernel sizes and pooling configurations.
  5. [Table I] The WIDAR dataset is listed with dimension 22 x 20 x 20; please verify this against the source and clarify what each axis represents, since the preceding description says '22 x 20 x 20, covering the time axis and x-y velocity features'.

Circularity Check

2 steps flagged · score 6.0 of 10

Headline communication-reduction claim is forced by an uplink-only cost metric; the 'Bayesian aggregation' is a running average in disguise.

  1. self definitional [Section IV-B2 'Performance Metrics'; Algorithm 1 line 30; Section III-C2]
    "Communication Overhead: We assessed the volume of data transmitted during the training process between clients and edge servers, as well as from edge servers to the cloud. ... Broadcast θg,t to all edges e ∈ E, and all clients ... the cloud broadcasts the probability mask θg,t to all clients through the edge servers."

    The cost metric counts only the client-to-edge and edge-to-cloud uplinks, while H-FedSN's own Algorithm 1 requires a full-precision probability mask θg,t ∈ [0,1]^s to be broadcast from cloud to edges and clients every round, with no binarization or compression stated. The reported 238.8× MNIST reduction is exactly the uplink bit-ratio 32d/s (full 32-bit weights versus 1-bit masks), so the headline figure follows from the metric definition rather than from an end-to-end accounting. If the downstream θg broadcasts are charged even once per edge and client, the end-to-end ratio drops by an order of magnitude; the central quantitative claim is therefore true by construction only under an asymmetric definition of 'communication overhead.'

  2. renaming known result [Section III-C1 'Edge Server Bayesian Aggregation'; Algorithm 2]
    "This approach marks a significant departure from traditional federated learning methods, such as FedAvg, which primarily rely on simple averaging ... αe,t = αe,t−1 + Σ ms_k,t; βe,t = βe,t−1 + |Ke|·1 − Σ ms_k,t ... θe,t = (αe,t−1)/(αe,t+βe,t−2)."

    With λ0 = 1 and reset priors every 10 rounds, the Beta-Bernoulli posterior-mode formula simplifies elementwise to θe,t = (1/|Ke|) Σ_{k∈Ke} ms_k,t, i.e., the empirical mean of the uploaded binary masks. The paper presents this as 'Bayesian aggregation' and as a significant departure from FedAvg's simple averaging, but its own update equations reduce exactly to a (sliding) average of the mask bits. The Bayesian label is a renaming of a known result rather than a distinct aggregation mechanism.

full rationale

The accuracy claims are empirical comparisons against external baselines (HierFAVG, FedPer, FedRS, FedCAMS, TOPK), and the local mask-gradient update is a standard straight-through estimator; neither reduces to the paper's own assumptions. The self-citations ([33], [34]) appear only in the related-work discussion of personalized FL and do not carry the argument, so they do not raise the score by themselves. However, the headline communication-reduction claim is not self-contained: the 'communication overhead' metric counts only the two uplink legs, while the required continuous-valued global mask broadcast is excluded. The 238.8× MNIST figure is exactly the 32-bit-to-1-bit uplink ratio, so the central quantitative result is partly an artifact of the metric definition. In addition, the Beta-Bernoulli aggregation with λ0=1 simplifies to the sample mean of the masks, making the claimed departure from simple averaging a renaming. These two issues make the central communication-efficiency claim partially circular by construction, warranting a score of 6 rather than 0-2.

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

The method depends on four hand-chosen hyperparameters (lambda_0, reset period, learning rate, private-layer split) and on domain assumptions inherited from prior work on mask training, the straight-through estimator, and non-IID partitioning. No new entities are introduced. The central claim is therefore an engineering combination of existing components plus tuning choices, not a derivation.

free parameters (4)
  • Initial Beta hyperparameter lambda_0 = 1
    Set at edge and cloud servers to define the prior; chosen without justification, though a value of 1 is a standard uniform prior.
  • Beta prior reset period = 10 rounds
    Priors are reset every 10 training rounds to avoid overfitting; this interval is chosen by hand and is not ablated.
  • Learning rate for H-FedSN = 0.01 (0.035 for WIDAR)
    H-FedSN's learning rate differs from the baseline rate of 0.001; this is an experimental tuning choice that affects the accuracy comparison.
  • Number of private layers = last 3 layers
    The last three layers of the 4-layer CNN are designated private; the choice of split point is not studied.
assumptions (4)
  • standard math Beta-Bernoulli conjugacy gives a closed-form posterior update
    Used in Section III-C for edge and cloud aggregation; a standard result.
  • domain assumption Straight-through estimator provides usable gradients through Bernoulli sampling
    Adopted from [39] in Section III-B; standard in binary network training.
  • domain assumption Mask training on frozen random weights can produce accurate models
    Based on [38] (supermask); the central premise for H-FedSN's accuracy.
  • domain assumption The non-IID label-imbalance partitioning approximates IoT data heterogeneity
    Uses the partition strategy from [25]; accepted in the FL literature.

how reviews work

0 comments
Cite this review

Pith. "Pith review of H-FedSN: Personalized Sparse Networks for Efficient and Accurate Hierarchical Federated Learning for IoT Applications." pith.science (2026). https://pith.science/paper/ZSR33M6R

@misc{pith2026241206210,
  author       = {Pith},
  title        = {Pith review of: H-FedSN: Personalized Sparse Networks for Efficient and Accurate Hierarchical Federated Learning for IoT Applications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZSR33M6R}},
  note         = {Machine review of arXiv:2412.06210}
}
read the original abstract

The proliferation of Internet of Things (IoT) has increased interest in federated learning (FL) for privacy-preserving distributed data utilization. However, traditional two-tier FL architectures inadequately adapt to multi-tier IoT environments. While Hierarchical Federated Learning (HFL) improves practicality in multi-tier IoT environments by multi-layer aggregation, it still faces challenges in communication efficiency and accuracy due to high data transfer volumes, data heterogeneity, and imbalanced device distribution, struggling to meet the low-latency and high-accuracy model training requirements of practical IoT scenarios. To overcome these limitations, we propose H-FedSN, an innovative approach for practical IoT environments. H-FedSN introduces a binary mask mechanism with shared and personalized layers to reduce communication overhead by creating a sparse network while keeping original weights frozen. To address data heterogeneity and imbalanced device distribution, we integrate personalized layers for local data adaptation and apply Bayesian aggregation with cumulative Beta distribution updates at edge and cloud levels, effectively balancing contributions from diverse client groups. Evaluations on three real-world IoT datasets and MNIST under non-IID settings demonstrate that H-FedSN significantly reduces communication costs by 58 to 238 times compared to HierFAVG while achieving high accuracy, making it highly effective for practical IoT applications in hierarchical federated learning scenarios.

Figures

Figures reproduced from arXiv: 2412.06210 by the authors.

Figure 1
Figure 1. H-FedSN: Personalized and structured sparse models [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of H-FedSN framework [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. H-FedSN Local Training in Client. 1) Clients receive the global probability mask θg,t−1 from the cloud server, then combine θg,t−1 with θ p k,t−1 to generate θk,t. Each client then inputs this probability mask into the inverse Sigmoid function Sigmoid−1 (·), obtaining the score mask sk,t = Sigmoid−1 (θk,t). 2) Clients transform back to θk,t = Sigmoid(sk,t), then use Bernoulli sampling to extract binary masks mk,t fr… view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Edge Server Bayesian Aggregation At Round [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Cloud Server Bayesian Aggregation At Round [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: WIDAR Dataset NoN-IID Data Distribution Across 5 [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Communication cost for different algorithms across [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Inference accuracy for different algorithms across [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Communication cost for different algorithms across [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: Inference accuracy for different algorithms across [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 13
Figure 13. Figure 13: Communication cost for different algorithms across [PITH_FULL_IMAGE:figures/full_fig_p010_13.png]
Figure 14
Figure 14. Figure 14: Inference accuracy for different algorithms across [PITH_FULL_IMAGE:figures/full_fig_p010_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 29 canonical work pages

  1. [1]

    Srda: Mobile sensing based fluid overload detection for end stage kidney disease patients using sensor relation dual autoencoder,

    M. Tang, J. Gao, G. Dong, C. Yang, B. Campbell, B. Bowman, J. M. Zoellner, E. Abdel-Rahman, and M. Boukhechba, “Srda: Mobile sensing based fluid overload detection for end stage kidney disease patients using sensor relation dual autoencoder,” in Conference on Health, Inference, and Learning. PMLR, 2023, pp. 133–146

    work page 2023

  2. [2]

    Pfdrl: Personalized federated deep reinforcement learning for residen- tial energy management,

    J. Gao, W. Wang, F. Nikseresht, V . Govinda Rajan, and B. Campbell, “Pfdrl: Personalized federated deep reinforcement learning for residen- tial energy management,” in Proceedings of the 52nd International Conference on Parallel Processing, 2023, pp. 402–411

    work page 2023

  3. [3]

    Federated machine learning: Concept and applications,

    Q. Yang, Y . Liu, T. Chen, and Y . Tong, “Federated machine learning: Concept and applications,” ACM Transactions on Intelligent Systems and Technology (TIST), vol. 10, no. 2, pp. 1–19, 2019

    2019

  4. [4]

    Communication-efficient learning of deep networks from decentralized data,

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

    2017

  5. [5]

    Federated optimization in heterogeneous networks,

    T. Li, A. K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V . Smith, “Federated optimization in heterogeneous networks,” Proceedings of Machine learning and systems , vol. 2, pp. 429–450, 2020

    2020

  6. [6]

    Scaffold: Stochastic controlled averaging for federated learn- ing,

    S. P. Karimireddy, S. Kale, M. Mohri, S. Reddi, S. Stich, and A. T. Suresh, “Scaffold: Stochastic controlled averaging for federated learn- ing,” in International conference on machine learning . PMLR, 2020, pp. 5132–5143

    2020

  7. [7]

    Understanding the smart city domain: A literature review,

    L. G. Anthopoulos, “Understanding the smart city domain: A literature review,” Transforming city governments for successful smart cities , pp. 9–21, 2015

    work page 2015

  8. [8]

    Smart farming: An overview,

    A. L. Virk, M. A. Noor, S. Fiaz, S. Hussain, H. Hussain, M. Rehman, M. Ahsan, and W. Ma, “Smart farming: An overview,” Smart Village Technology, 2020

    work page 2020

Show all 44 references
  1. [9]

    Client-edge-cloud hierarchical federated learning,

    L. Liu, J. Zhang, S. Song, and K. B. Letaief, “Client-edge-cloud hierarchical federated learning,” in ICC 2020-2020 IEEE international conference on communications (ICC) . IEEE, 2020, pp. 1–6

    2020

  2. [10]

    Federated learning with extreme label skew: A data extension approach,

    S. A. Tijani, X. Ma, R. Zhang, F. Jiang, and R. Doss, “Federated learning with extreme label skew: A data extension approach,” in 2021 International Joint Conference on Neural Networks (IJCNN) , 2021, pp. 1–8

    2021

  3. [11]

    Federated learning technology in serial topology for iot networks,

    J. Sheng, J. Xiong, and B. Liu, “Federated learning technology in serial topology for iot networks,” 2023 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB) , pp. 1–5, 2023

    2023

  4. [12]

    Artificial intelligence-aided digital twin design: A systematic review,

    N. Hao, Y . Li, K. Liu, S. Liu, Y . Lu, B. Xu, C. Li, J. Chen, L. Yue, T. Fu et al. , “Artificial intelligence-aided digital twin design: A systematic review,” 2024

    2024

  5. [13]

    FedBCGD: Communication-efficient accelerated block coordinate gradient descent for federated learning,

    J. Liu, F. Shang, Y . Liu, H. Liu, Y . Li, and Y . Gong, “FedBCGD: Communication-efficient accelerated block coordinate gradient descent for federated learning,” in ACM Multimedia 2024 , 2024. [Online]. Available: https://openreview.net/forum?id=lAFO0SUjXD

    2024

  6. [14]

    A model parameter update strategy for enhanced asynchronous federated learning algorithm,

    Y . Sun, X. Gao, Y . Shen, J. Xie, J. Yang, and N. Si, “A model parameter update strategy for enhanced asynchronous federated learning algorithm,” in 2023 IEEE 3rd International Conference on Computer Systems (ICCS), 2023, pp. 9–15

    2023

  7. [15]

    Personalized federated learning: A meta-learning approach,

    A. Fallah, A. Mokhtari, and A. Ozdaglar, “Personalized federated learning: A meta-learning approach,” arXiv preprint arXiv:2002.07948 , 2020

    2002 arXiv

  8. [16]

    Ternary compression for communication-efficient federated learning,

    J. Xu, W. Du, Y . Jin, W. He, and R. Cheng, “Ternary compression for communication-efficient federated learning,” IEEE Transactions on Neural Networks and Learning Systems , vol. 33, no. 3, pp. 1162–1176, 2022

    2022

  9. [17]

    Communication-efficient adaptive federated learning,

    Y . Wang, L. Lin, and J. Chen, “Communication-efficient adaptive federated learning,” ArXiv, vol. abs/2205.02719, 2022

    2022 arXiv

  10. [18]

    Fedrs: Federated learning with restricted softmax for label distribution non-iid data,

    X.-C. Li and D.-C. Zhan, “Fedrs: Federated learning with restricted softmax for label distribution non-iid data,” in Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , ser. KDD ’21. New York, NY , USA: Association for Computing Machinery, 2021...

    2021

  11. [19]

    Federated learning with personalization layers,

    M. G. Arivazhagan, V . Aggarwal, A. K. Singh, and S. Choud- hary, “Federated learning with personalization layers,” arXiv preprint arXiv:1912.00818, 2019

    1912 arXiv

  12. [20]

    Smartphone and smartwatch-based biometrics using activities of daily living,

    G. M. Weiss, K. Yoneda, and T. Hayajneh, “Smartphone and smartwatch-based biometrics using activities of daily living,” Ieee Ac- cess, vol. 7, pp. 133 190–133 202, 2019

    2019

  13. [21]

    Widar 3.0: Wifi-based activity recognition dataset,

    Z. Yang, Y . Zhang, G. Zhang, Y . Zheng, and G. Chi, “Widar 3.0: Wifi-based activity recognition dataset,” 2020. [Online]. Available: https://dx.doi.org/10.21227/7znf-qp86

    2020 doi

  14. [22]

    Zero-effort cross-domain gesture recognition with wi-fi,

    Y . Zheng, Y . Zhang, K. Qian, G. Zhang, Y . Liu, C. Wu, and Z. Yang, “Zero-effort cross-domain gesture recognition with wi-fi,” in Proceed- ings of the 17th annual international conference on mobile systems, applications, and services , 2019, pp. 313–325

    2019

  15. [23]

    Design considerations for the wisdm smart phone-based sensor mining architecture,

    J. W. Lockhart, G. M. Weiss, J. C. Xue, S. T. Gallagher, A. B. Grosner, and T. T. Pulickal, “Design considerations for the wisdm smart phone-based sensor mining architecture,” in Proceedings of the Fifth International Workshop on Knowledge Discovery from Sensor Data , 2011, pp. 25–33

    2011

  16. [24]

    Gradient-based learning applied to document recognition,

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

    1998

  17. [25]

    Federated learning on non-iid data silos: An experimental study,

    Q. Li, Y . Diao, Q. Chen, and B. He, “Federated learning on non-iid data silos: An experimental study,” in 2022 IEEE 38th international conference on data engineering (ICDE) . IEEE, 2022, pp. 965–978

    2022

  18. [26]

    Sparse communication for distributed gradient descent,

    A. F. Aji and K. Heafield, “Sparse communication for distributed gradient descent,” arXiv preprint arXiv:1704.05021 , 2017

    2017 arXiv

  19. [27]

    Industrial automation using iot,

    R. A. Sekar, T. Prabakaran, A. Sudhakar, and R. S. Kumar, “Industrial automation using iot,” in AIP Conference Proceedings, vol. 2393, no. 1. AIP Publishing, 2022

    2022

  20. [28]

    A review on iot healthcare monitoring applications and a vision for transforming sensor data into real-time clinical feedback,

    H. H. Nguyen, F. Mirza, M. A. Naeem, and M. Nguyen, “A review on iot healthcare monitoring applications and a vision for transforming sensor data into real-time clinical feedback,” in 2017 IEEE 21st Interna- tional Conference on Computer Supported Cooperative Work in Design (C...

    2017

  21. [29]

    A survey on federated learning for resource-constrained iot devices,

    A. Imteaj, U. Thakker, S. Wang, J. Li, and M. Amini, “A survey on federated learning for resource-constrained iot devices,” IEEE Internet of Things Journal , vol. 9, pp. 1–24, 2021

    2021

  22. [30]

    Fedscr: Structure-based communi- cation reduction for federated learning,

    X. Wu, X. Yao, and C.-L. Wang, “Fedscr: Structure-based communi- cation reduction for federated learning,” IEEE Transactions on Parallel and Distributed Systems , vol. 32, no. 7, pp. 1565–1577, 2021

    2021

  23. [31]

    Fast federated learning by balancing communication trade-offs,

    M. Nori, S. Yun, and I. Kim, “Fast federated learning by balancing communication trade-offs,” IEEE Transactions on Communications , vol. 69, pp. 5168–5182, 2021

    2021

  24. [32]

    Toward communication-learning trade- off for federated learning at the network edge,

    J. ji Ren, W. Ni, and H. Tian, “Toward communication-learning trade- off for federated learning at the network edge,” IEEE Communications Letters, vol. 26, pp. 1858–1862, 2022

    2022

  25. [33]

    Fedmetamed: Federated meta-learning for person- alized medication in distributed healthcare systems,

    J. Gao and Y . Li, “Fedmetamed: Federated meta-learning for person- alized medication in distributed healthcare systems,” arXiv preprint arXiv:2412.03851, 2024

    2024 arXiv

  26. [34]

    Fed-ldr: Federated local data-infused graph creation with node-centric model refinement,

    J. Gao, Y . Li, and S. F. Ahmed, “Fed-ldr: Federated local data-infused graph creation with node-centric model refinement,” arXiv preprint arXiv:2411.04936, 2024

    2024 arXiv

  27. [35]

    Client scheduling and resource management for efficient training in heterogeneous iot-edge federated learning,

    Y . Cui, K. Cao, G. Cao, M. Qiu, and T. Wei, “Client scheduling and resource management for efficient training in heterogeneous iot-edge federated learning,” IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems , vol. 41, no. 8, pp. 2407–2420, 2022

    2022

  28. [36]

    Federated learning with hierarchical clustering of local updates to improve training on non-iid data,

    C. Briggs, Z. Fan, and P. Andras, “Federated learning with hierarchical clustering of local updates to improve training on non-iid data,” in 2020 International Joint Conference on Neural Networks (IJCNN) , 2020, pp. 1–9

    2020

  29. [37]

    Towards fast and accurate federated learning with non-iid data for cloud- based iot applications,

    T. Liu, J. Ding, T. Wang, M. Pan, and M. Chen, “Towards fast and accurate federated learning with non-iid data for cloud- based iot applications,” Journal of Circuits, Systems and Computers , vol. 31, no. 13, p. 2250235, 2022. [Online]. Available: https: //doi.org/10.1142/S021...

    2022 doi

  30. [38]

    Deconstructing lottery tickets: Zeros, signs, and the supermask,

    H. Zhou, J. Lan, R. Liu, and J. Yosinski, “Deconstructing lottery tickets: Zeros, signs, and the supermask,” Advances in neural information processing systems, vol. 32, 2019

    2019

  31. [39]

    Estimating or propagating gradients through stochastic neurons for conditional computation,

    Y . Bengio, N. L ´eonard, and A. Courville, “Estimating or propagating gradients through stochastic neurons for conditional computation,” arXiv preprint arXiv:1308.3432, 2013

    2013 arXiv

  32. [40]

    Bayesian signsgd optimizer for federated learning,

    P. A. Ferreira, P. N. da Silva, V . Gottin, R. Stelling, and T. Calmon, “Bayesian signsgd optimizer for federated learning,” Advances in Neural Information Processing Systems , vol. 34, 2021

    2021

  33. [41]

    Activity recog- nition from accelerometer data,

    N. Ravi, N. Dandekar, P. Mysore, and M. L. Littman, “Activity recog- nition from accelerometer data,” in Aaai, vol. 5, no. 2005. Pittsburgh, PA, 2005, pp. 1541–1546

    2005

  34. [42]

    Transition-aware human activity recognition using smartphones,

    J.-L. Reyes-Ortiz, L. Oneto, A. Sam `a, X. Parra, and D. Anguita, “Transition-aware human activity recognition using smartphones,” Neu- rocomputing, vol. 171, pp. 754–767, 2016

    2016

  35. [43]

    Human activity recognition with smart- phone sensors using deep learning neural networks,

    C. A. Ronao and S.-B. Cho, “Human activity recognition with smart- phone sensors using deep learning neural networks,” Expert systems with applications, vol. 59, pp. 235–244, 2016

    2016

  36. [44]

    Evaluation of the efficacy of iot deployment on petro-retail operations,

    S. Purohit et al. , “Evaluation of the efficacy of iot deployment on petro-retail operations,” Turkish Journal of Computer and Mathematics Education (TURCOMAT), vol. 12, no. 4, pp. 356–363, 2021

    2021

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.