Pith. sign in

REVIEW 6 major objections 4 minor 28 references

A New Perspective on Privacy Protection in Federated Learning with Granular-Ball Computing

T0 review · 6 major / 4 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 Replacing raw images with coarse granular-rectangle graphs before federated training can protect privacy and cut communication costs while preserving accuracy, this paper argues.

desk verdict Interesting input-level privacy angle for federated learning, but the utility claim is unverifiable without accuracy tables and the bias-variance swap in Eq. 5 undercuts the theory. read the letter →

arxiv 2501.04940 v1 pith:7P2GTDI7 submitted 2025-01-09 cs.LG cs.CV

Guannan Lai , Yihui Feng , Xin Yang , Xiaoyu Deng , Hao Yu , Shuyin Xia , Guoyin Wang , Tianrui Li This is my paper · ORCID
classification cs.LGcs.CV
keywords federatedlearninggranular-ballcomputingprivacyprotectiongraphneuralnetworksreconstructionattackimageclassificationcommunicationefficiencyvariance-constrainedsegmentation
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 tries to establish that federated learning can be made more private and more efficient by changing the input data itself rather than by adding defenses to the model or its parameters. It proposes Granular-Ball Federated Learning (GrBFL), which converts each image into a graph of coarse rectangular blocks found by a variance-constrained two-dimensional binary search. A graph neural network trained on these graphs is claimed to withstand reconstruction attacks, reduce communication cost, and keep classification accuracy competitive with competing federated methods. The paper reports experiments on MNIST, CIFAR-10, and CIFAR-100 supporting that balance.

What carries the argument

The load-bearing object is the granular-rectangle graph, in which each node is a rectangular image region described by its center, dimensions, and mean, variance, maximum, and minimum gray values, and edges connect overlapping rectangles. To build it, the paper uses a two-dimensional binary search segmentation driven by a purity constraint: starting from the lowest-gradient pixel, the algorithm grows a rectangle while purity (the fraction of pixels within a gray-level threshold of the center) stays above a bound. The purity function is asserted to be globally monotonic under expansion, which turns the search into a logarithmic procedure. The resulting graph compresses the image, hides exact pixel values from gradient-based attackers, and serves as compact input to a graph neural network.

What would settle it

Take a synthetic grayscale image with a small noisy patch surrounded by a uniform region and run the purity-based binary search: if expanding the rectangle first increases purity (for example by absorbing many pixels within the threshold), the search will stop at a suboptimal rectangle. A measurable failure is a purity-versus-size curve that rises over some interval, which directly contradicts the monotonicity assumption and would invalidate the claimed logarithmic segmentation.

Watch

Extended reading notes

Core claim

GrBFL's central claim is that coarse granularity is sufficient representation for federated image classification while being substantially harder to reconstruct from gradients. The paper argues that removing redundant pixels lowers variance and data noise in the generalization-error decomposition without changing model bias, so utility can be preserved; and that because reconstruction attacks recover information only from what is sent, less input information caps what an attacker can recover. Under a proximal-term aggregation scheme, the graph inputs are said to stabilize federated training and improve communication efficiency. The implemented system is reported to outperform feature-level and model-level privacy defenses on the tested datasets.

Load-bearing premise

The segmentation algorithm assumes that growing a rectangular block in any direction can only make the block less pure, and this monotonicity is used without proof to justify the binary search.

Editorial extensions

If this is right

  • Reconstruction attacks that work on raw-pixel federated learning recover less information when gradients come from granular-rectangle graphs.
  • Communication cost per federated round drops because each client transmits a smaller model and less input-dependent state.
  • Classification accuracy on MNIST, CIFAR-10, and CIFAR-100 is maintained or improved relative to CNN-based federated baselines.
  • The proposed privacy score and PEUM metric offer a single-number way to compare FL methods across privacy, efficiency, and utility.

Reading between the lines

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

  • If the granular-decomposition idea transfers, the same graph reconstruction could apply to non-image grid data such as medical volumes or sensor arrays, though the paper only tests natural images.
  • The privacy benefit is measured empirically through reconstruction similarity; a formal guarantee would require combining the information-loss argument with a rigorous bound, which the paper does not derive.
  • A stronger attacker who knows the segmentation algorithm's threshold and node-feature semantics might invert coarse blocks back to approximate pixel distributions, so the practical privacy gain depends on keeping those details confidential.
  • The monotonic purity assumption can be tested independently on real image statistics; if it fails, segmentation could be replaced by exhaustive or dynamic-programming search without changing the graph representation.
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

6 major / 4 minor

Summary. The manuscript proposes GrBFL, a federated image-classification framework that replaces raw client images with coarse granular rectangles, maps these rectangles to a graph, and trains GNNs under FedProx aggregation. It introduces a two-dimensional binary search segmentation algorithm based on a purity constraint, and defines privacy, communication-efficiency, and composite metrics (Sp, CE, PEUM). Experiments on MNIST, CIFAR-10, and CIFAR-100 are used to claim that GrBFL simultaneously improves privacy and efficiency while maintaining or outperforming baseline utility relative to CNNFL methods with DP and LotteryFL.

Significance. The idea of protecting federated learning at the input level through granular-ball computing is potentially interesting, because it couples privacy with communication savings without relying on trusted aggregation or added noise. The manuscript also makes its code public and evaluates several GNN backbones, which is useful for reproducibility. However, as submitted, the central utility claim is unsupported: no accuracy numbers are reported, the theoretical utility argument in Eq. (5) rests on a mislabeled bias/variance decomposition, and the privacy and efficiency metrics are defined in ways that make the headline results partly follow from construction. The direction may be worth exploring, but the current paper does not demonstrate the claimed trade-off.

major comments (6)
  1. [§5.3, Table 2] The central utility claim is not supported because the paper never reports classification accuracy for GrBFL or any baseline. PEUM in Eq. (10) is the harmonic mean of Acc, CE, and Sp, yet Acc appears nowhere in the main text; a high PEUM can be produced by high Sp and CE even if accuracy is poor. The abstract's assertion of "robust utility" and "consistently outperforming" other FL methods therefore cannot be checked, and no accuracy table or supplementary pointer with these values is provided.
  2. [§3.3, Eq. (5)] The decomposition of expected generalization error swaps bias and variance. The first term, EDk[(Fk - ED[Fk])^2], is the variance of the predictions, and the second term, (EDk[Fk] - y)^2, is the squared bias; the text labels them in reverse. The subsequent argument that removing redundant information leaves "model bias" unchanged while reducing variance relies on this mislabeling and is not a valid theoretical basis for the utility-preservation claim. The data-noise term is also not clearly defined, since yDk is not specified as a noisy label.
  3. [§4.2] The two-dimensional binary search segmentation algorithm relies on the assertion that purity is globally monotonic under rectangle expansion: "as the granular rectangle expands in any direction, its purity consistently decreases." This property is not proved and is not generally true for natural images, because expanding a rectangle can absorb neighboring pixels that are more similar to the center than the current boundary, which would increase purity. Since the algorithm's logarithmic complexity and the claimed privacy-utility trade-off depend on this monotonicity, the authors need to prove it under the definition in Eq. (6) or provide direct validation that the search returns the intended rectangles.
  4. [§5.2, Eq. (8)] The privacy score Sp is defined directly from the reconstruction MSE. Because GrBFL deliberately discards pixel-level information, its reconstructed images will have large MSE by construction, so Sp is high even if an attacker recovers all available semantic content. This makes the privacy evaluation partly tautological. A meaningful privacy measure should quantify what an adversary can infer under the stated threat model, including attack iterations and knowledge of node-feature semantics, and should be reported together with reconstruction examples and MSE trajectories.
  5. [§5.2, Eq. (9); §5.3] The communication-efficiency metric CE depends on a free scaling factor phi, chosen as 3e6, and on raw time and traffic measurements that are never reported. Without a sensitivity analysis for phi or the raw values, the claim that GrBFL has "significantly higher" CE than CNNFL is not reproducible and could be an artifact of the chosen scaling. This is load-bearing for the efficiency contribution.
  6. [§4.3, Table 1] The table that motivates the choice of FedProx lists five aggregation methods (FedAvg, SCAFFOLD, FedNova, FedProx) but displays only four accuracy values, and the accuracy for FedProx is not among them. The claim that FedProx "performed better" is therefore not supported by the displayed data.
minor comments (4)
  1. [§3.2, Figure 2] Figure 2 is described as validating the privacy theory, but no experimental procedure or data source is given; the figure appears illustrative rather than evidential.
  2. [§4.4] The statement that "common 224x224 image data" is reconstructed in 1ms lacks hardware, software, and measurement details, so the timing is not reproducible from the manuscript.
  3. [§5.3, Figure 4] The figure caption refers to a comparison of "differential privacy performance" while the surrounding text describes reconstruction attacks; the caption should be aligned with the actual experiment.
  4. [§5.3, Parameter Sensitivity] The text refers to the "purity of the constructed granular-receptacle"; this appears to be a typo for "granular rectangle," and the sensitivity analysis does not report numeric values.

Circularity Check

2 steps flagged · score 6.0 of 10

Privacy and composite-metric scores are partly generated by the paper's own definitions; the improvement claim rests on a constructed Sp and a free scaling factor, with accuracy unreported.

  1. self definitional [Section 5.2, Eq. (8); Section 5.3, paragraph 'Privacy Protection Experiment']
    "Sp(imgtrue, imgpred) = 1 − 1/(1 + M SE(imgtrue, imgpred)) ... However, since node features only contain coarse pixel information, the attacker can only determine the position and size of each granule rectangle but cannot precisely obtain the specific pixel values."

    Privacy is defined as a monotone function of reconstruction MSE, and GrBFL's granular-rectangle node features deliberately omit pixel-level values. Therefore a large MSE, and hence a high Sp, is guaranteed by the input representation itself. The 'strong privacy protection' result restates the intentional information loss rather than being an independent empirical finding about the method.

  2. other [Section 5.2, Eqs. (9)-(10); Section 5.3, paragraph 'GrBFL performs well in the comprehensive evaluation']
    "PEUM = 1/(1/Acc + 1/CE + 1/Sp) ... Specifically, in the three experimental datasets, when the parameter is set to ϕ = 3 × 10^6, the model efficiency score for GrBFL is significantly higher compared to FL methods based on the CNNFL framework."

    The only reported overall comparison, Table 2, is PEUM, whose components are an unreported accuracy, an efficiency term CE with a freely chosen scaling factor phi, and the by-construction privacy score Sp from the previous step. Since no accuracy numbers for GrBFL or CNNFL baselines appear in the main text, the abstract's 'consistently outperforming' claim is not supported by measured utility; it is generated by the authors' metric definitions and the phi selection.

full rationale

Most of the derivation chain is not circular. The granular-ball graph representation is built on prior published work by the same group, but that prior work is independent, and the FL aggregation experiments (FedProx versus FedAvg, SCAFFOLD, etc.) are external empirical comparisons rather than consequences of the target claim. The unproved monotonicity of purity in Section 4.2 and the mislabeled bias/variance decomposition in Eq. (5) are correctness or assumption problems, not circularities. However, the privacy evaluation is circular in a narrow but load-bearing sense: Sp is defined as a monotone function of reconstruction MSE, and the granular-rectangle representation is defined to omit pixel-level information, so high privacy scores follow by construction. The composite PEUM then combines this by-construction Sp with an efficiency term CE whose scaling factor phi is selected to make GrBFL look better, while accuracy is not reported in the main text. The headline claim that GrBFL 'consistently outperforms' other FL methods therefore reduces in part to the authors' metric choices rather than to an independently verified utility advantage. This is partial circularity; the efficiency and utility aspects retain independent content, so the paper is not wholly tautological.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The framework depends on several hand-chosen hyperparameters (purity threshold, proximal coefficient, communication scaling factor), on the unproved monotonicity of purity, on a specific attack model, and on the assumption that coarse node statistics cannot be inverted. No new physical entities are introduced.

free parameters (3)
  • purity threshold thr
    Eq. 6 defines purity with a threshold thr; no principled selection is given, and sensitivity is only tested as robustness (Figure 6).
  • proximal coefficient mu = not reported in main text
    Eq. 7 introduces mu to stabilize graph-based FL; exact values are deferred to supplementary materials.
  • communication scaling factor phi = 3 x 10^6
    Eq. 9 uses phi to map CE into 0-1; set to 3e6 in the experiments, and since CE enters PEUM it partly determines the ranking.
assumptions (4)
  • domain assumption L-BFGS reconstruction attack is a faithful threat model for privacy leakage.
    All privacy claims are evaluated against this attack from Zhu et al. (2019); other attack families are not considered.
  • ad hoc to paper Purity is globally monotonic under rectangle expansion.
    Asserted in Section 4.2 to justify binary search; no proof is given and the claim is not generally true for arbitrary image neighborhoods.
  • standard math Expected generalization error decomposes into bias, variance, and data noise as in Eq. 5.
    The decomposition is standard, but the paper labels the variance term as 'Model Bias' and the bias term as 'Variance', making the subsequent utility argument unreliable.
  • domain assumption Granular node statistics cannot be inverted to recover exact pixel values.
    Privacy experiments assume attackers with weights, gradients, and node-feature semantics still cannot recover pixels; no inversion attack on the granular statistics is tested.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A New Perspective on Privacy Protection in Federated Learning with Granular-Ball Computing." pith.science (2026). https://pith.science/paper/7P2GTDI7

@misc{pith2026250104940,
  author       = {Pith},
  title        = {Pith review of: A New Perspective on Privacy Protection in Federated Learning with Granular-Ball Computing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7P2GTDI7}},
  note         = {Machine review of arXiv:2501.04940}
}
read the original abstract

Federated Learning (FL) facilitates collaborative model training while prioritizing privacy by avoiding direct data sharing. However, most existing articles attempt to address challenges within the model's internal parameters and corresponding outputs, while neglecting to solve them at the input level. To address this gap, we propose a novel framework called Granular-Ball Federated Learning (GrBFL) for image classification. GrBFL diverges from traditional methods that rely on the finest-grained input data. Instead, it segments images into multiple regions with optimal coarse granularity, which are then reconstructed into a graph structure. We designed a two-dimensional binary search segmentation algorithm based on variance constraints for GrBFL, which effectively removes redundant information while preserving key representative features. Extensive theoretical analysis and experiments demonstrate that GrBFL not only safeguards privacy and enhances efficiency but also maintains robust utility, consistently outperforming other state-of-the-art FL methods. The code is available at https://github.com/AIGNLAI/GrBFL.

Figures

Figures reproduced from arXiv: 2501.04940 by the authors.

Figure 1
Figure 1. Comparison between the CNNFL and GrBFL processes. The upper section illustrates [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The Relation Between Input and Privacy During the iteration process, x ′ is updated as follows: x ′ new = x ′ old − α · H−1 · g (3) where H is an approximation of the Hessian matrix, g is the gradient of the current inferred data x ′ , and H−1 is estimated using the changes in parameters and gradients over the past few iterations. This approximation better reflects the local curvature information of the current para… view at source ↗
Figure 3
Figure 3. The flowchart presents the granular-ball-based knowledge reconstruction algorithm. The [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Privacy Protection Experiment. To rigorously evaluate the privacy pro￾tection capabilities of the GrBFL and CNNFL frameworks, this study con￾ducted reconstruction attack experi￾ments under the GCN and ResNet models. The experiments were per￾formed across three datasets…
Figure 5
Figure 5. Figure 5: Comparison of communication efficiency GrBFL performs well in the com￾prehensive evaluation. We con￾ducted a comprehensive assessment of each model using the metrics defined in Equation 10, with the results shown in [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Parameter Sensitivity Study. This experiment was performed on the GAT model using the [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 26 canonical work pages

  1. [1]

    No free lunch theorem for security and utility in federated learning

    [Zhang et al.(2022)] Xiaojin Zhang, Hanlin Gu, Lixin Fan, Kai Chen, and Qiang Yang. No free lunch theorem for security and utility in federated learning. ACM Transactions on Intelligent Systems and Technology, 14(1):1–35,

    work page 2022

  2. [5]

    Untargeted poisoning attack detection in federated learning via behavior attestation

    [Al Mallah et al.(2023)] Ranwa Al Mallah, David Lopez, Godwin Badu-Marfo, and Bilal Farooq. Untargeted poisoning attack detection in federated learning via behavior attestation. IEEE Access,

    work page 2023

  3. [6]

    Communication efficient and differentially private logistic regression under the distributed setting

    [Bao et al.(2023)] Ergute Bao, Dawei Gao, Xiaokui Xiao, and Yaliang Li. Communication efficient and differentially private logistic regression under the distributed setting. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 69–79,

    work page 2023

  4. [7]

    Trading off privacy, utility, and efficiency in federated learning.ACM Transactions on Intelligent Systems and Technology, 14(6):1–32,

    [Zhang et al.(2023)] Xiaojin Zhang, Yan Kang, Kai Chen, Lixin Fan, and Qiang Yang. Trading off privacy, utility, and efficiency in federated learning.ACM Transactions on Intelligent Systems and Technology, 14(6):1–32,

    work page 2023

  5. [8]

    Springer Nature,

    [Yang et al.(2020)] Qiang Yang, Lixin Fan, and Han Yu.Federated Learning: Privacy and Incentive, volume 12500. Springer Nature,

    work page 2020

  6. [10]

    Differentially private filtering

    [Le Ny and Pappas(2013)] Jerome Le Ny and George J Pappas. Differentially private filtering. IEEE Transactions on Automatic Control, 59(2):341–354,

    work page 2013

  7. [12]

    On data distribution leakage in cross-silo federated learning.IEEE Transactions on Knowledge and Data Engineering,

    [Jiang et al.(2024)] Yangfan Jiang, Xinjian Luo, Yuncheng Wu, Xiaochen Zhu, Xiaokui Xiao, and Beng Chin Ooi. On data distribution leakage in cross-silo federated learning.IEEE Transactions on Knowledge and Data Engineering,

    work page 2024

  8. [13]

    Practical secure aggregation for privacy-preserving machine learning

    [Bonawitz et al.(2017)] Keith Bonawitz, Vladimir Ivanov, Ben Kreuter, Antonio Marcedone, H Bren- dan McMahan, Sarvar Patel, Daniel Ramage, Aaron Segal, and Karn Seth. Practical secure aggregation for privacy-preserving machine learning. In proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security, pages 1175–1191,

    work page 2017

Show all 28 references
  1. [15]

    Privacy- preserving deep learning via additively homomorphic encryption

    [Aono et al.(2017)] Yoshinori Aono, Takuya Hayashi, Lihua Wang, Shiho Moriai, et al. Privacy- preserving deep learning via additively homomorphic encryption. IEEE Transactions on Information Forensics and Security, 13(5):1333–1345,

    2017

  2. [16]

    Privacy-preserving federated deep learning with irregular users

    [Xu et al.(2020)] Guowen Xu, Hongwei Li, Yun Zhang, Shengmin Xu, Jianting Ning, and Robert H Deng. Privacy-preserving federated deep learning with irregular users. IEEE Transactions on Dependable and Secure Computing, 19(2):1364–1381,

    2020

  3. [17]

    Federated optimization in heterogeneous networks

    [Li et al.(2020a)] Tian Li, Anit Kumar Sahu, Manzil Zaheer, Maziar Sanjabi, Ameet Talwalkar, and Virginia Smith. Federated optimization in heterogeneous networks. Proceedings of Machine Learning and Systems, 2:429–450, 2020a. 11 [Ahmed et al.(2024)] Syed Thouheed Ahmed, V Vino...

    2020

  4. [18]

    Scaffold: Stochastic controlled averaging for federated learning

    [Karimireddy et al.(2020)] Sai Praneeth Karimireddy, Satyen Kale, Mehryar Mohri, Sashank Reddi, Sebastian Stich, and Ananda Theertha Suresh. Scaffold: Stochastic controlled averaging for federated learning. In International Conference on Machine Learning, pages 5132–5143. PMLR,

    2020

  5. [19]

    Tack- ling the objective inconsistency problem in heterogeneous federated optimization

    [Wang et al.(2020)] Jianyu Wang, Qinghua Liu, Hao Liang, Gauri Joshi, and H Vincent Poor. Tack- ling the objective inconsistency problem in heterogeneous federated optimization. Advances in Neural Information Processing Systems, 33:7611–7623,

    2020

  6. [20]

    Topological structure in visual perception

    [Chen(1982)] Lin Chen. Topological structure in visual perception. Science, 218(4573):699–700,

    1982

  7. [22]

    A fast adaptive k-means with no bounds

    [Xia et al.(2020)] Shuyin Xia, Daowan Peng, Deyu Meng, Changqing Zhang, Guoyin Wang, Elis- abeth Giem, Wei Wei, and Zizhong Chen. A fast adaptive k-means with no bounds. IEEE Transactions on Pattern Analysis and Machine Intelligence,

    2020

  8. [23]

    An efficient and adaptive granular-ball generation method in classification problem

    [Xia et al.(2022)] Shuyin Xia, Xiaochuan Dai, Guoyin Wang, Xinbo Gao, and Elisabeth Giem. An efficient and adaptive granular-ball generation method in classification problem. IEEE Transactions on Neural Networks and Learning Systems,

    2022

  9. [24]

    Graph-based representation for image based on granular-ball

    [Shuyin et al.(2023)] Xia Shuyin, Dai Dawei, Yang Long, Zhany Li, Lan Danf, Wang Guoy, et al. Graph-based representation for image based on granular-ball. arXiv preprint arXiv:2303.02388,

    2023 arXiv

  10. [25]

    Deep leakage from gradients

    [Zhu et al.(2019)] Ligeng Zhu, Zhijian Liu, and Song Han. Deep leakage from gradients. Advances in neural information processing systems, 32,

    2019

  11. [27]

    Communication-efficient learning of deep networks from decentralized data

    [McMahan et al.(2017)] Brendan McMahan, Eider Moore, Daniel Ramage, Seth Hampson, and Blaise Aguera y Arcas. Communication-efficient learning of deep networks from decentralized data. In Artificial Intelligence and Statistics, pages 1273–1282. PMLR,

    2017

  12. [28]

    Lotteryfl: Personalized and communication-efficient federated learning with lottery ticket hypothesis on non-iid datasets

    [Li et al.(2020b)] Ang Li, Jingwei Sun, Binghui Wang, Lin Duan, Sicheng Li, Yiran Chen, and Hai Li. Lotteryfl: Personalized and communication-efficient federated learning with lottery ticket hypothesis on non-iid datasets. arXiv preprint arXiv:2008.03371, 2020b. 12

    2020 arXiv

  13. [1982]

    Dgcc: data-driven granular cognitive computing.Granular Computing, 2(4):343–355,

    [Wang(2017)] Guoyin Wang. Dgcc: data-driven granular cognitive computing.Granular Computing, 2(4):343–355,

    2017

  14. [2013]

    Differentially private federated learning on non-iid data: Convergence analysis and adaptive optimization

    [Chen et al.(2024)] Lin Chen, Xiaofeng Ding, Zhifeng Bao, Pan Zhou, and Hai Jin. Differentially private federated learning on non-iid data: Convergence analysis and adaptive optimization. IEEE Transactions on Knowledge and Data Engineering,

    2024

  15. [2017]

    Secure single-server aggregation with (poly) logarithmic overhead

    [Bell et al.(2020)] James Henry Bell, Kallista A Bonawitz, Adrià Gascón, Tancrède Lepoint, and Mariana Raykova. Secure single-server aggregation with (poly) logarithmic overhead. In Proceedings of the 2020 ACM SIGSAC Conference on Computer and Communications Security, pages 1253–1269,

    2020

  16. [2019]

    Granular ball computing classifiers for efficient, scalable and robust learning

    [Xia et al.(2019)] Shuyin Xia, Yunsheng Liu, Xin Ding, Guoyin Wang, Hong Yu, and Yuoguo Luo. Granular ball computing classifiers for efficient, scalable and robust learning. Information Sciences, 483:136–152,

    2019

  17. [2020]

    A survey on differentially private machine learning

    [Gong et al.(2020)] Maoguo Gong, Yu Xie, Ke Pan, Kaiyuan Feng, and Alex Kai Qin. A survey on differentially private machine learning. IEEE Computational Intelligence Magazine, 15(2): 49–64,

    2020

  18. [2022]

    Feddefender: Client-side attack-tolerant federated learning

    [Park et al.(2023)] Sungwon Park, Sungwon Han, Fangzhao Wu, Sundong Kim, Bin Zhu, Xing Xie, and Meeyoung Cha. Feddefender: Client-side attack-tolerant federated learning. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages 1850–1861,

    2023

  19. [2023]

    Vertical federated learning: Concepts, advances, and challenges

    [Liu et al.(2024)] Yang Liu, Yan Kang, Tianyuan Zou, Yanhong Pu, Yuanqin He, Xiaozhou Ye, Ye Ouyang, Ya-Qin Zhang, and Qiang Yang. Vertical federated learning: Concepts, advances, and challenges. IEEE Transactions on Knowledge and Data Engineering,

    2024

  20. [2024]

    Survey on federated learning threats: Concepts, taxonomy on attacks and defences, experimental study and challenges

    [Rodríguez-Barroso et al.(2023)] Nuria Rodríguez-Barroso, Daniel Jiménez-López, M Victoria Luzón, Francisco Herrera, and Eugenio Martínez-Cámara. Survey on federated learning threats: Concepts, taxonomy on attacks and defences, experimental study and challenges. Information Fu...

    2023

Pith tools

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