Pith. sign in

REVIEW 4 major objections 6 minor 32 references

Explainable Semantic Federated Learning Enabled Industrial Edge Network for Fire Surveillance

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read This paper aims to establish that a semantic federated learning framework, XSFL, can make industrial fire surveillance spectrally efficient, privacy-preserving, and explainable by freezing per-device weights chosen by Fisher information…

desk verdict A coherent SC+FL+explainability system paper whose delay advantage rests on a shaky compute model and whose accuracy claims lack experimental support. read the letter →

arxiv 2412.19979 v1 pith:SIDJBIVL submitted 2024-12-28 cs.LG cs.CRcs.ITmath.IT

classification cs.LGcs.CRcs.ITmath.IT
keywords semanticcommunicationfederatedlearningfiresurveillanceexplainableAImodelfreezingFisherinformationindustrialInternetofThingsedgecomputing
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 aims to establish that fire surveillance over an industrial edge network can be made spectrally efficient, privacy-preserving, and explainable in one framework. The framework, eXplainable Semantic Federated Learning (XSFL), lets IIoT cameras transmit only compact semantic symbols instead of raw frames, and trains the semantic encoder-decoder cooperatively at the edge so raw data never leaves the devices. To handle heterogeneous devices, the Adaptive Client Training (ACT) strategy freezes a per-device fraction of model weights chosen by a Fisher-information importance score, which is reported to cut training delay while preserving or improving accuracy. An explainable semantic communication (ESC) mechanism then produces heatmaps that align extracted semantics with fire and smoke regions in the input image. On the CDD fire dataset, the paper reports that XSFL beats FedAvg, STC, FTTQ, and FedPAQ on accuracy, precision, specificity, F1-score, and recall, and that it gives the shortest average training delay per device.

What carries the argument

The load-bearing mechanism is the Adaptive Client Training (ACT) strategy. For device $n$, ACT clusters devices by local data volume, sets a trainable proportion $\zeta_n = \mathrm{Mean}(C_n^i)/D_{\max}$, and freezes the $1-\zeta_n$ fraction of global weights with the smallest values of the Fisher-weighted quadratic $I_{n}^{g,t} = (w_{g,t} - w_{n,t})^\top \mathrm{diag}(I_n(w_{n,t}))(w_{g,t} - w_{n,t})$, using the Fisher information matrix as a Hessian approximation. Freezing low-importance weights is what reduces local computation per device while allowing the remaining weights to adapt to local data. The second mechanism is the Explainable SC (ESC) mechanism, which computes gradients of semantic features with respect to the last convolutional feature map, weights them with second-order coefficients, and applies leaky ReLU to form a localization heatmap; this is what connects a prediction back to image regions.

What would settle it

A decisive experiment is to run XSFL on the CDD dataset with the same frozen-weight fractions, but choose the frozen weights by a random index set instead of by the Fisher-information score. If random freezing reproduces XSFL's reported accuracy, precision, F1, and delay, then the importance-based selection rule is not what carries the result; if random freezing clearly underperforms, the Fisher criterion is doing the claimed work.

Watch

Extended reading notes

Core claim

The central discovery is that federated training of a semantic communication model can be adapted to heterogeneous IIoT devices by freezing different weight subsets per device, and that this simultaneously improves the accuracy-delay trade-off and preserves explainability. Concretely, the paper claims XSFL achieves the best values on all five classification indicators on the CDD fire dataset, and that its average training delay per device is the shortest among the compared schemes. The reported reason is that ACT's weight freezing, guided by the Fisher information of each local model, lets limited-resource devices update fewer parameters while well-resourced devices train more, improving local adaptation and global generalization. The ESC heatmaps, the paper argues, show that the extracted semantics correspond to visible fire and smoke regions, making the predictions interpretable.

Load-bearing premise

The paper's accuracy advantage rests on the premise that freezing the $1-\zeta_n$ fraction of weights with the smallest Fisher-information-based $I_{n}^{g,t}$ values improves local adaptation and global generalization without degrading accuracy; the premise is asserted through the Hessian approximation, and no ablation varies the freezing criterion.

Editorial extensions

If this is right

  • Fire-surveillance devices could send semantic symbols rather than raw frames, reducing uplink bandwidth while retaining classification accuracy.
  • The maximum per-round delay in synchronous federated learning shrinks because slower devices train fewer weights, so more devices can stay within the delay budget.
  • The ESC heatmaps give rescue crews a human-checkable link between a fire alert and the image evidence, which is the kind of transparency needed for life-critical decisions.
  • The reported optimization objective, combining global loss and weighted delay, gives a single number for comparing federated schemes on the accuracy-latency trade-off.

Reading between the lines

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

  • A test not run in the paper: compare ACT's Fisher-based freezing against random freezing of the same fraction on the same dataset; if accuracy ties, the Fisher criterion itself is not the driver of the reported gain.
  • The explainability evaluation in the paper is qualitative; a quantitative faithfulness score for the heatmaps, such as deleting highlighted regions and measuring the accuracy drop, would let future work compare ESC against other explanation methods.
  • If the reported five-metric dominance holds under varying channel noise, the same framework could extend to other disaster-surveillance tasks such as smoke, flood, or industrial anomaly alerts, but that transfer is not established here.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes an Industrial Edge Semantic Network (IESN) for fire surveillance, in which IIoT devices transmit only semantic features to an edge server. To train the semantic communication (SC) model while preserving privacy, the authors introduce XSFL (eXplainable Semantic Federated Learning), which combines an Adaptive Client Training (ACT) strategy and an Explainable SC (ESC) mechanism. ACT clusters devices by local data volume and freezes a device-specific fraction of global model weights selected via a Fisher-information-based importance measure, with the aim of reducing local training delay and improving accuracy. ESC uses a leakyReLU-based activation mapping to produce heatmaps that visualize which input regions drive the extracted semantics. The paper reports experiments on the Comprehensive Disaster Dataset (CDD) with ten simulated IIoT devices, comparing XSFL against FedAvg, STC, FTTQ, and FedPAQ, and claims that XSFL achieves the best accuracy/precision/specificity/F1/recall and the shortest average training delay, hence the best accuracy-delay trade-off.

Significance. If the central claims were substantiated, XSFL would be a useful contribution: it addresses a real problem (privacy-preserving, resource-adaptive, explainable semantic communication for industrial edge fire surveillance) and combines ideas from federated learning, Fisher-information-based weight freezing, and saliency-style explanation in a novel way. The paper also gives a clear system architecture and a reasonably organized set of algorithms. However, the evidence provided is currently too weak to support the claimed dominance over the baselines. The performance claim rests on a single dataset, ten simulated devices, no error bars or significance tests, no code, and no ablation of the ACT freezing criterion. The delay advantage is derived from an assumption that freezing parameters reduces per-sample computation cost linearly, which is not justified for deep neural networks. The explainability evaluation is qualitative only. These gaps are load-bearing because the paper's main selling points are exactly the accuracy-delay trade-off and the explainability mechanism.

major comments (4)
  1. [Section IV and Eq. (11)] The claim that XSFL has the shortest training delay is not established. Section IV states that the computation time complexity of XSFL is O(T G Davg |\hat w^{n,tr}_{g,t}|), i.e., linear in the number of trainable parameters, but Eq. (11) defines the local training delay as D_n Φ_n(w_n,t) G / f_n,t, where Φ_n(w_n,t) is the cost of a forward-backward pass of the full local model. In a deep neural network, freezing a weight does not remove its forward multiplication or necessarily remove it from backward gradient propagation, so the actual computation savings are typically much smaller than the frozen fraction. Since Section IV itself notes that |\hat w^n_{g,t}| = |w_{n,t}|, the communication complexity is unchanged, and the only claimed advantage over the baselines is this uncertain computation saving. The paper needs wall-clock training-time measurements or FLOP counts, not just a complexity expression, to support the Section V-C claim of shortest average delay and the optimization-objective comparison in Fig. 5.
  2. [Section V-A and V-B] The accuracy-dominance claim is supported only by one dataset (CDD), ten simulated IIoT devices, and radar plots without error bars, confidence intervals, or significance tests. The simulation settings omit essential details: the SC model architecture, the number of communication rounds T, the local epochs G, the learning rate and optimizer, the number of clusters C in ACT, the actual ζ_n values per device, the LeakyReLU slope α in Eq. (25), and the way the Fisher information matrix in Eq. (18) is computed. The paper's own explanation for the accuracy gain is speculative ("We speculate the excellent performance of XSFL is attributed to the ACT strategy"). Without ablations that vary the freezing criterion, the freezing fraction, and the cluster count, the claim that XSFL achieves the best performance in all indicators is not demonstrated.
  3. [Section III-B, Eqs. (18)-(21)] The ACT strategy's core premise is not validated. The Fisher information matrix I_n(w_n,t) is introduced without a definition, and Eq. (21) claims that I_n^g,t approximates (1/2)(w_g,t - w_n,t)^T H_n,t (w_g,t - w_n,t) at w_n,t = w_g,t, but the relationship between the quadratic form in Eq. (18) and the Hessian of the local loss is not derived. More importantly, the paper provides no evidence that freezing the weights with the smallest values of I_n^g,t improves local adaptation or global generalization; an ablation comparing Fisher-based freezing with random freezing or magnitude-based freezing is needed. As written, the only delay benefit of ACT is an artifact of the computation model in Eq. (11), so the accuracy benefit is an empirical claim that currently lacks supporting experiments.
  4. [Section III-C, Algorithm 3] The statement that ESC "can be used ... without any pre-training, hence there is no extra computation cost" is internally inconsistent with Eqs. (22)-(24), which require computing first, second, and third derivatives of the semantic features with respect to feature-map activations. These gradient computations have real computational cost and may be expensive on resource-constrained IIoT devices. The paper should either remove the no-extra-cost claim or provide a complexity analysis and empirical timing for the ESC heatmap generation.
minor comments (6)
  1. [Section III-B, Eq. (17)] The notation Mean(C_i^n) in Eq. (17) is ambiguous: C_i^n is defined as a set, but it is not clear whether it contains data volumes D_n or device indices; the text should specify that the mean is taken over the data volumes of devices in the cluster.
  2. [Section V-B, Fig. 3] Radar plots without numerical values overlaid are hard to read; please provide a table with exact values and standard deviations across multiple random seeds.
  3. [Section V-C, Figs. 4 and 5] The figures lack clear axis labels and units; specify what "training delay" measures (wall-clock time or simulated CPU cycles) and what the optimization-objective weights are in Fig. 5 beyond the stated coefficient 1e-3.
  4. [Section II-D, Eq. (11)] All devices are set to the same CPU frequency f_n,t = 2 GHz in the experiments, so the heterogeneity in computational resources that motivates ACT is not actually simulated; please vary f_n,t and D_n to demonstrate the claimed adaptivity.
  5. [Section III-C, Eq. (25)] The summation index in Eq. (25) uses k both as the kernel index and as the upper limit; please use a different variable, e.g., sum over e=1 to K, and make the definition of K explicit.
  6. [Section V-D] The explainability evaluation is purely qualitative; consider adding a quantitative metric such as insertion/deletion scores or pointing-game accuracy to substantiate the claim that ESC "accurately marks" fire and smoke regions.

Circularity Check

1 steps flagged · score 6.0 of 10

The delay-dominance claim reduces to the paper's own linear-in-trainable-parameters complexity model, while the accuracy claim remains an independent empirical comparison.

  1. self definitional [Section IV (Complexity Analysis); see also Section III-B and Eq. (11)]
    "Simultaneously, the number of trainable local parameters in the IIoT devices is reduced, leading to a corresponding decrease in the FL training delay [21], [22]. ... In XSFL, partial parameters are frozen, namely, these parameters are not trained, hence the time complexity of computation is O(T GDavg| ˆwn,tr g,t |), where ˆwn,tr g,t means the trained weights of ˆwn g,t. Due to | ˆwn,tr g,t | < | ˆwn g,t| and the weights freezing is carried on the ES, hence O(T GDavg| ˆwn,tr g,t |) < O(T GDavg|wg,t|)."

    Eq. (11) defines local training delay as dL_n(t)=D_n Φ_n(w_n,t) G / f_n,t, where Φ_n(w_n,t) is the CPU cycles for a forward-backward pass of the full local model w_n,t. Section IV then asserts that freezing 1−ζ_n of the weights makes the computation complexity O(T G Davg |^w^{n,tr}_{g,t}|), i.e., linear only in the trainable-parameter count. This is not derived from Eq. (11); it defines delay to be proportional to the number of trainable parameters. With that definition, the conclusion that XSFL has the shortest average training delay (Section V-C, Fig. 4) is immediate from the ACT freezing rule (Eqs. 19-20) and is not an empirical or theoretical prediction.

full rationale

The central accuracy claim (Section V-B, Fig. 3) is an empirical comparison on the external CDD dataset against FedAvg, STC, FTTQ, and FedPAQ; it is not derived from the ACT equations, so it is not circular. The ESC mechanism is a leakyReLU-weighted Grad-CAM variant (Eqs. 22-25) and, while incremental and cited to Grad-CAM, is not used to derive the performance claims. I find no load-bearing self-citation chain: citations [3], [8], and [15] to the authors' prior work are background, and the Fisher/Hessian approximation is attributed to the external reference [20]. The one circular step is the delay-dominance claim: Eq. (11) defines local delay via CPU cycles for a full-model forward-backward pass, but Section IV redefines computation complexity as O(T G Davg |trainable weights|). Under that definition, freezing 1−ζ_n weights trivially guarantees the shortest delay shown in Fig. 4, so the accuracy-delay dominance is partially built into the complexity model rather than tested. Score 6 reflects this partial, construction-level circularity; the accuracy component remains independent.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities; its contributions are algorithmic (ACT, ESC) and architectural (IESN). The ledger lists one hand-set coefficient and several unreported hyperparameters, plus five load-bearing assumptions, indicating a heuristic system that relies on unvalidated design choices.

free parameters (6)
  • Number of clusters C in ACT k-medoids
    Eqs. (15)-(16) require C, but Section V does not state the value used; it determines which devices share a freezing proportion.
  • Local training epochs G
    Algorithm 1 takes G as input, but no value is reported; it directly affects training delay and accuracy.
  • Communication rounds T
    Algorithm 1 takes T as input, but no value is reported; convergence and delay results depend on it.
  • SC model architecture and optimizer hyperparameters
    Section II-A says convolutional or transformer layers can be used, but the actual network, learning rate, batch size, and optimizer are not specified; these dominate the reported accuracy.
  • LeakyReLU slope alpha in Eq. (25)
    alpha is constrained to (0,1) but the value is not given; it affects the ESC heatmaps.
  • Delay weighting coefficient in optimization objective = 1e-3
    Section V-C sets this coefficient to balance delay and global loss; conclusions in Fig. 5 depend on this hand-chosen value.
assumptions (5)
  • domain assumption The Fisher information matrix approximates the Hessian and Eq. (21) is equal to the loss function near the global model.
    Section III-B4 invokes [20] under unspecified regularity conditions; this underpins ACT's freezing choices.
  • domain assumption The physical channel can be simulated by neural networks so gradients backpropagate end-to-end.
    Section II-A cites [14], [15]; this is standard in semantic communication but assumes the simulated channel is faithful to real wireless behavior.
  • domain assumption Computation and communication time scale linearly with model size.
    Section IV relies on this proportionality to conclude that XSFL reduces training delay; it is an approximation that may not hold for all hardware.
  • ad hoc to paper Freezing the selected least-important weights improves local adaptation and global generalization.
    Section V-B says the performance gain is speculated to come from ACT; no theorem or ablation establishes this claim.
  • ad hoc to paper The ESC heatmap visually captures the true relationship between semantics and input monitoring data.
    Section V-D shows only example heatmaps; no quantitative evaluation or user study supports the claimed explainability.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Explainable Semantic Federated Learning Enabled Industrial Edge Network for Fire Surveillance." pith.science (2026). https://pith.science/paper/SIDJBIVL

@misc{pith2026241219979,
  author       = {Pith},
  title        = {Pith review of: Explainable Semantic Federated Learning Enabled Industrial Edge Network for Fire Surveillance},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SIDJBIVL}},
  note         = {Machine review of arXiv:2412.19979}
}
read the original abstract

In fire surveillance, Industrial Internet of Things (IIoT) devices require transmitting large monitoring data frequently, which leads to huge consumption of spectrum resources. Hence, we propose an Industrial Edge Semantic Network (IESN) to allow IIoT devices to send warnings through Semantic communication (SC). Thus, we should consider (1) Data privacy and security. (2) SC model adaptation for heterogeneous devices. (3) Explainability of semantics. Therefore, first, we present an eXplainable Semantic Federated Learning (XSFL) to train the SC model, thus ensuring data privacy and security. Then, we present an Adaptive Client Training (ACT) strategy to provide a specific SC model for each device according to its Fisher information matrix, thus overcoming the heterogeneity. Next, an Explainable SC (ESC) mechanism is designed, which introduces a leakyReLU-based activation mapping to explain the relationship between the extracted semantics and monitoring data. Finally, simulation results demonstrate the effectiveness of XSFL.

Figures

Figures reproduced from arXiv: 2412.19979 by the authors.

Figure 1
Figure 1. The illustration of IESN for fire surveillance. (a) [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The illustration of the XSFL framework. (a) ACT strategy; (b) ESC mechanism. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. The average training delay of each IIoT device using [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (2 more)
Figure 5
Figure 5. Figure 5: The values of the optimization objective for different [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: The explainable results of XSFL [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 25 canonical work pages

  1. [1]

    An iot based efficient fire supervision monitoring and alerting system,

    B. Prabha, “An iot based efficient fire supervision monitoring and alerting system,” in 2019 Third International conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud)(I-SMAC) . IEEE, 2019, pp. 414–419

  2. [2]

    Real-time wsn based early flood detection and control monitoring system,

    T. M. Thekkil and N. Prabakaran, “Real-time wsn based early flood detection and control monitoring system,” in 2017 International Con- ference on Intelligent Computing, Instrumentation and Control Tech- nologies (ICICICT). IEEE, 2017, pp. 1709–1713

  3. [3]

    AI driven heterogeneous mec system with uav assistance for dynamic environment: Challenges and solutions,

    F. Jiang, K. Wang, L. Dong, C. Pan, W. Xu, and K. Yang, “AI driven heterogeneous mec system with uav assistance for dynamic environment: Challenges and solutions,” IEEE Network, vol. 35, no. 1, pp. 400–408, 2020

  4. [4]

    Efficient fire detection for uncertain surveillance environment,

    K. Muhammad, S. Khan, M. Elhoseny, S. H. Ahmed, and S. W. Baik, “Efficient fire detection for uncertain surveillance environment,” IEEE Transactions on Industrial Informatics , vol. 15, no. 5, pp. 3113–3122, 2019

  5. [5]

    Privacy-preserving efficient fire detection system for indoor surveillance,

    A. Jain and A. Srivastava, “Privacy-preserving efficient fire detection system for indoor surveillance,” IEEE Transactions on Industrial Infor- matics, vol. 18, no. 5, pp. 3043–3054, 2022

  6. [6]

    Design of an iot approach for security surveillance system for industrial process monitoring using raspberry-pi,

    G. Manikandan, D. Karunkuzhali, D. Geetha, and V . Kavitha, “Design of an iot approach for security surveillance system for industrial process monitoring using raspberry-pi,” in AIP Conference Proceedings , vol. 2519, no. 1. AIP Publishing LLC, 2022, p. 030024

  7. [7]

    A novel method for fire detection and alarming notifi- cation using iot technology,

    S. Banka, B. Kanchanapalli, B. Annam, A. Akula, S. Thirukkovalluru, and S. Jarpula, “A novel method for fire detection and alarming notifi- cation using iot technology,” in 2023 Second International Conference on Electronics and Renewable Systems (ICEARS) , 2023, pp. 560–565

  8. [8]

    Large ai model-based semantic communications,

    F. Jiang, Y . Peng, L. Dong, K. Wang, K. Yang, C. Pan, and X. You, “Large ai model-based semantic communications,” IEEE Wireless Com- munications, vol. 31, no. 3, pp. 68–75, 2024

Show all 32 references
  1. [9]

    Privacy-preserved credit data sharing integrating blockchain and federated learning for industrial 4.0,

    F. Yang, Y . Qiao, M. Z. Abedin, and C. Huang, “Privacy-preserved credit data sharing integrating blockchain and federated learning for industrial 4.0,” IEEE Transactions on Industrial Informatics , vol. 18, no. 12, pp. 8755–8764, 2022

  2. [10]

    Natural disaster monitoring with wireless sensor networks: A case study of data- intensive applications upon low-cost scalable systems,

    D. Chen, Z. Liu, L. Wang, M. Dou, J. Chen, and H. Li, “Natural disaster monitoring with wireless sensor networks: A case study of data- intensive applications upon low-cost scalable systems,” Mobile Networks and Applications, vol. 18, no. 5, pp. 651–663, 2013

  3. [11]

    Federated learning,

    Q. Yang, Y . Liu, Y . Cheng, Y . Kang, T. Chen, and H. Yu, “Federated learning,” Synthesis Lectures on Artificial Intelligence and Machine Learning, vol. 13, no. 3, pp. 1–207, 2019

  4. [12]

    The need for uncertainty quantification in machine-assisted medical decision making,

    E. Begoli, T. Bhattacharya, and D. Kusnezov, “The need for uncertainty quantification in machine-assisted medical decision making,” Nature Machine Intelligence, vol. 1, no. 1, pp. 20–23, 2019

  5. [13]

    Transformer-empowered 6G intelligent networks: From massive mimo processing to semantic communication,

    Y . Wang, Z. Gao, D. Zheng, S. Chen, D. Gunduz, and H. V . Poor, “Transformer-empowered 6G intelligent networks: From massive mimo processing to semantic communication,” IEEE Wireless Communica- tions, 2022

  6. [14]

    End-to-end fast training of com- munication links without a channel model via online meta-learning,

    S. Park, O. Simeone, and J. Kang, “End-to-end fast training of com- munication links without a channel model via online meta-learning,” in 2020 IEEE 21st International Workshop on Signal Processing Advances in Wireless Communications (SPAWC). IEEE, 2020, pp. 1–5

  7. [15]

    Deep progressive reinforcement learning-based flexible resource scheduling framework for irs and uav-assisted mec system,

    L. Dong, F. Jiang, M. Wang, Y . Peng, and X. Li, “Deep progressive reinforcement learning-based flexible resource scheduling framework for irs and uav-assisted mec system,” IEEE Transactions on Neural Networks and Learning Systems , pp. 1–13, 2024

  8. [16]

    Beyond transmitting bits: Context, semantics, and task-oriented communications,

    D. G ¨und¨uz, Z. Qin, I. E. Aguerri, H. S. Dhillon, Z. Yang, A. Yener, K. K. Wong, and C.-B. Chae, “Beyond transmitting bits: Context, semantics, and task-oriented communications,” IEEE Journal on Selected Areas in Communications, vol. 41, no. 1, pp. 5–41, 2022

  9. [17]

    Low-latency federated learning over wireless channels with differential privacy,

    K. Wei, J. Li, C. Ma, M. Ding, C. Chen, S. Jin, Z. Han, and H. V . Poor, “Low-latency federated learning over wireless channels with differential privacy,” IEEE Journal on Selected Areas in Communications , vol. 40, no. 1, pp. 290–307, 2021

  10. [18]

    Hierarchical federated learning with social context clustering-based participant selection for internet of medical things applications,

    X. Zhou, X. Ye, K. I.-K. Wang, W. Liang, N. K. C. Nair, S. Shimizu, Z. Yan, and Q. Jin, “Hierarchical federated learning with social context clustering-based participant selection for internet of medical things applications,” IEEE Transactions on Computational Social Systems ,...

  11. [19]

    Analysis of k-means and k-medoids algorithm for big data,

    P. Arora, S. Varshney et al. , “Analysis of k-means and k-medoids algorithm for big data,” Procedia Computer Science , vol. 78, pp. 507– 512, 2016

  12. [20]

    Overcoming forgetting in federated learning on non-iid data,

    N. Shoham, T. Avidor, A. Keren, N. Israel, D. Benditkis, L. Mor-Yosef, and I. Zeitak, “Overcoming forgetting in federated learning on non-iid data,” arXiv preprint arXiv:1910.07796 , 2019

  13. [21]

    Semifreddonets: Partially frozen neural networks for effi- cient computer vision systems,

    L. F. Isikdogan, B. V . Nayak, C.-T. Wu, J. P. Moreira, S. Rao, and G. Michael, “Semifreddonets: Partially frozen neural networks for effi- cient computer vision systems,” in European Conference on Computer Vision. Springer, 2020, pp. 193–208

  14. [22]

    Digital twin enhanced federated reinforcement learning with lightweight knowledge distillation in mobile networks,

    X. Zhou, X. Zheng, X. Cui, J. Shi, W. Liang, Z. Yan, L. T. Yang, S. Shimizu, and K. I.-K. Wang, “Digital twin enhanced federated reinforcement learning with lightweight knowledge distillation in mobile networks,” IEEE Journal on Selected Areas in Communications, vol. 41, no. 1...

  15. [23]

    Trustworthy ai: A computational perspective,

    H. Liu, Y . Wang, W. Fan, X. Liu, Y . Li, S. Jain, Y . Liu, A. K. Jain, and J. Tang, “Trustworthy ai: A computational perspective,” arXiv preprint arXiv:2107.06641, 2021

  16. [24]

    Grad-cam: Why did you say that?

    R. R. Selvaraju, A. Das, R. Vedantam, M. Cogswell, D. Parikh, and D. Batra, “Grad-cam: Why did you say that?” arXiv preprint arXiv:1611.07450, 2016

  17. [25]

    Reluplex made more practical: Leaky relu,

    J. Xu, Z. Li, B. Du, M. Zhang, and J. Liu, “Reluplex made more practical: Leaky relu,” in 2020 IEEE Symposium on Computers and Communications (ISCC), 2020, pp. 1–7

  18. [26]

    Communication-efficient federated learning via knowledge distillation,

    C. Wu, F. Wu, L. Lyu, Y . Huang, and X. Xie, “Communication-efficient federated learning via knowledge distillation,” Nature communications, vol. 13, no. 1, pp. 1–8, 2022. SUBMITTED FOR REVIEW 9

  19. [27]

    A comparative study on disaster detection from social media images using deep learning,

    A. Omar, S. Ashraf, A. Rahman, M. A. Amin, A. A. Ali et al. , “A comparative study on disaster detection from social media images using deep learning,” in Proceedings of the Global AI Congress 2019 . Springer, 2020, pp. 485–499

  20. [28]

    Robust and communication-efficient federated learning from non-iid data,

    F. Sattler, S. Wiedemann, K.-R. M ¨uller, and W. Samek, “Robust and communication-efficient federated learning from non-iid data,” IEEE transactions on neural networks and learning systems , vol. 31, no. 9, pp. 3400–3413, 2019

  21. [29]

    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 , 2020

  22. [30]

    Fedpaq: A communication-efficient federated learning method with periodic averaging and quantization,

    A. Reisizadeh, A. Mokhtari, H. Hassani, A. Jadbabaie, and R. Pedarsani, “Fedpaq: A communication-efficient federated learning method with periodic averaging and quantization,” in International Conference on Artificial Intelligence and Statistics . PMLR, 2020, pp. 2021–2031

  23. [31]

    A joint learning and communications framework for federated learning over wireless networks,

    M. Chen, Z. Yang, W. Saad, C. Yin, H. V . Poor, and S. Cui, “A joint learning and communications framework for federated learning over wireless networks,” IEEE Transactions on Wireless Communications , vol. 20, no. 1, pp. 269–283, 2020

  24. [32]

    Large-group failure mode and effects analysis for risk management of angle grinders in the construction industry,

    Z.-S. Chen, J.-Y . Chen, Y .-H. Chen, Y . Yang, L. Jin, E. Herrera-Viedma, and W. Pedrycz, “Large-group failure mode and effects analysis for risk management of angle grinders in the construction industry,” Information Fusion, vol. 97, p. 101803, 2023. BIOGRAPHIES Li Dong rece...

Pith tools

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