Pith. sign in

REVIEW 2 major objections 6 minor 105 references

MobenFL is the broadest federated-learning benchmark for medical images, pairing 20 algorithms with 22 multi-organ datasets and scoring accuracy, speed, privacy and clinical heterogeneity together.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-10 11:07 UTC pith:NRXWLDD3

load-bearing objection Solid multi-organ FL benchmark that actually ships code and multi-axis tables; the fixed five-client skeleton is a real but not fatal limit on the clinical-realism claim. the 2 major comments →

arxiv 2607.08219 v1 pith:NRXWLDD3 submitted 2026-07-09 cs.CV cs.LG

Benchmark Evaluation of Feredated Learning on Multi-organ Images

Junbin Mao , Xu Tian , Jianchun Zhu , Ludi Li , Jin Liu This is my paper
classification cs.CV cs.LG
keywords Federated LearningMedical Image ClassificationMulti-organ dataPrivacy robustnessAlgorithmic efficiencyCross-device evaluationNon-IID medical data
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Medical AI is blocked by privacy rules and by huge differences in organs, scanners and diseases across hospitals. Federated learning lets centers train together without sharing raw images, yet no shared yardstick exists to compare the flood of new algorithms under realistic clinical messiness. MobenFL fills that gap by running twenty recent federated methods on twenty-two public imaging collections that span twelve body organs and four modalities. Beyond plain accuracy it also times convergence, measures privacy-noise tolerance and tests three real-world mismatches: different diseases of the same organ, different devices imaging the same disease, and different modalities. The resulting scorecards let clinicians pick an algorithm that matches their data diversity, compute budget and privacy needs instead of relying on single-number leaderboards.

Core claim

The paper claims that a single open benchmark covering twenty federated algorithms, twenty-two multi-organ datasets and four evaluation axes (accuracy, efficiency, privacy robustness and clinical-scenario transfer) is both feasible and necessary; without it, claims about which algorithm works in the clinic remain incomparable and untrustworthy.

What carries the argument

MobenFL itself: a unified code framework that standardizes twenty algorithms (grouped as extra-network, regularization, aggregation and split-learning) and twenty-two datasets under fixed ResNet-18 training, Dirichlet non-IID splits, early stopping, and three specialized clinical-heterogeneity test suites.

Load-bearing premise

The claim that five clients, ResNet-18, a fixed learning rate and Dirichlet alpha 0.5 already capture the resource limits and data skews of real multi-hospital deployments.

What would settle it

Re-run the exact same twenty algorithms on a new multi-center collection whose client count, scanner mix and class imbalance differ sharply from the five-client Dirichlet-0.5 setup and check whether the recommended algorithm rankings reverse.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Clinicians can now select algorithms by primary need: SplitFed or HarmoFL when accuracy under heterogeneity matters most, FedGH when wall-clock time is scarce.
  • Privacy-preserving deployments gain a concrete ranking: SplitFed keeps almost full accuracy even under strong differential-privacy noise while most others collapse.
  • Future algorithm papers must report efficiency and privacy numbers on the same multi-organ suite or risk being incomparable.
  • Benchmark users receive ready hyper-parameter defaults (batch 16, three local epochs) that generalize across the twelve organs.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same multi-axis design could be ported to federated segmentation or detection, exposing whether the current accuracy–efficiency trade-offs reverse when pixel-level labels are required.
  • Because the privacy test injects noise only on the uplink, a follow-up that also privatizes the downlink or uses secure aggregation would test whether SplitFed’s advantage survives stronger threat models.
  • The observed organ-specific sensitivity (brain and knee drop hardest under non-IID) suggests that organ-aware client clustering or modality-specific personalization layers may be more fruitful than further generic regularizers.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 6 minor

Summary. The manuscript introduces MobenFL, a federated medical-image classification benchmark that unifies 20 recent FL algorithms (grouped into additional-network, regularization, aggregation-strategy and split-learning families) with 22 public 2-D imaging datasets spanning 12 organs. Under a fixed experimental skeleton (ResNet-18 + Adam, lr = 0.0001, batch 128, five clients, Dirichlet α = 0.5, early-stopping patience 10) the authors report accuracy under IID and non-IID partitions, theoretical and wall-clock efficiency, robustness to simple Gaussian differential-privacy noise, and three clinical-heterogeneity scenarios (cross-organ, cross-disease, cross-device). From these results they derive a recommendation map (Fig. 16) that ranks algorithms for different clinical priorities. Code is released.

Significance. If the rankings prove stable, MobenFL would become the most extensive public reference point for FL medical-image classification, filling a clear gap left by FedCBD and Flamby. The multi-organ coverage, inclusion of split-learning and personalization methods, joint reporting of accuracy/efficiency/privacy, and open-source unified framework are genuine strengths that lower the barrier for subsequent comparative work. The contribution is therefore of clear practical value to the medical-AI community, provided the fixed experimental skeleton is shown not to reverse the reported orderings.

major comments (2)
  1. Section 4.1 freezes every experiment to exactly five clients, Dirichlet α = 0.5, ResNet-18 + Adam (lr = 0.0001), batch size 128 and early-stopping patience 10. All accuracy, efficiency and privacy numbers in Sections 4.2–4.5, and therefore the recommendation map of Fig. 16, are generated under this single skeleton. Real multi-center deployments routinely involve 10–50+ sites with highly unbalanced sample sizes and participation rates far from a uniform five-client Dirichlet draw. Because the central claim of a “comprehensive … evaluation framework for the clinical application of FL” rests on the premise that the skeleton fairly represents clinical conditions, at least a limited sensitivity study (varying client count and α on a representative subset of datasets) is required to establish that the reported rankings do not reverse.
  2. Section 4.5.1 implements differential privacy by adding isotropic Gaussian noise scaled by a free factor ε to the transmitted parameters (Eq. 5). This is not a standard (ε,δ)-DP mechanism with clipping and calibrated noise; consequently the privacy-robustness claims and the special status accorded to SplitFed under “strict privacy” cannot be interpreted as formal DP guarantees. Either a proper DP-SGD implementation with reported privacy budgets should be used, or the language should be revised to “noise robustness” rather than “privacy protection capabilities.”
minor comments (6)
  1. Title and abstract contain the typographical error “Feredated”; correct throughout.
  2. Table 2 lists TurboSVM under both Additional Network Structures and Regularization-Driven; clarify the intended category.
  3. Section 3.2.2 states “FedProx, MOON, PGFed and TurboSVM all fall into this category” while TurboSVM is already placed in the additional-network group; remove the inconsistency.
  4. Figure 4 caption and surrounding text refer to “p-value” without stating the statistical test used; specify the paired test and whether multiple-comparison correction was applied.
  5. Several dataset URLs and accession numbers appear only as footnotes; a consolidated data-availability table would improve reproducibility.
  6. The hyper-parameter study in Section 4.6 is performed only for FedAvg; a short note on whether the recommended (batch=16, local_iter=3) setting transfers to the other 19 algorithms would be useful.

Circularity Check

0 steps flagged

Empirical benchmark paper with no derivation chain; rankings arise from direct runs on held-out public data, not from self-referential definitions or fitted-then-predicted quantities.

full rationale

MobenFL is a pure empirical evaluation framework: 20 published FL algorithms are re-implemented under a fixed skeleton (ResNet-18, Adam, five clients, Dirichlet α=0.5) and scored on 22 public multi-organ datasets for accuracy, wall-clock time, and differential-privacy robustness. No equation defines a quantity in terms of a parameter later presented as a prediction; no uniqueness theorem is imported from the authors’ prior work; self-citations are limited to standard baselines (FedAvg, FedProx, etc.) that are independently reproducible. The experimental skeleton itself is an assumption about clinical realism, not a circular step. Consequently the paper contains zero instances of the six circularity patterns.

Axiom & Free-Parameter Ledger

5 free parameters · 3 axioms · 1 invented entities

A benchmark paper rests on a handful of experimental design choices that are free parameters and on standard domain assumptions about federated averaging and medical data heterogeneity; no new physical entities are postulated.

free parameters (5)
  • learning_rate = 0.0001
    Fixed at 0.0001 for every algorithm and dataset (Section 4.1); never tuned per method.
  • batch_size = 128
    Fixed at 128 (Section 4.1); later sensitivity study shows it strongly affects accuracy.
  • num_clients = 5
    Fixed at 5 for all experiments (Section 4.1); real hospital networks vary widely.
  • dirichlet_alpha = 0.5
    Non-IID partitions generated with α = 0.5 (Section 4.1.1); controls degree of label skew.
  • dp_noise_scale_epsilon = 0.01 / 0.1
    Privacy experiments inject Gaussian noise scaled by ε ∈ {0.01, 0.1} (Section 4.5); not a full DP-SGD accountant.
axioms (3)
  • standard math The global objective is the sample-weighted average of local empirical risks (Eq. 1).
    Standard FedAvg derivation; used throughout.
  • domain assumption Medical imaging data across hospitals are non-IID with respect to organ, disease, device and population.
    Stated in Introduction and used to justify Dirichlet partitions and clinical-scenario tests.
  • ad hoc to paper ResNet-18 + Adam is a sufficiently representative backbone for ranking FL algorithms on 2-D medical images.
    Chosen once in Section 4.1 and never varied; may favor methods that interact well with residual networks.
invented entities (1)
  • MobenFL benchmark suite independent evidence
    purpose: Unified code base and evaluation protocol that packages the 20 algorithms and 22 datasets.
    The paper’s central contribution; independent evidence is the public GitHub repository.

pith-pipeline@v1.1.0-grok45 · 48913 in / 2451 out tokens · 35704 ms · 2026-07-10T11:07:00.015065+00:00 · methodology

0 comments
read the original abstract

The privacy requirements of medical data and its substantial variations across organs and modalities hinder the clinical implementation of medical AI. Federated learning (FL) is a feasible approach to overcome these challenges. Due to the continuous emergence of FL algorithms and the highly heterogeneous nature of medical data, objectively evaluating their performance in real-world clinical settings remains difficult. Therefore, a comprehensive federated medical imaging benchmark, serving as a unified evaluation standard, is crucial for advancing the technology toward reliable clinical application. Existing federated medical imaging benchmarks have not yet adequately incorporated state-of-the-art algorithms, are limited to data from single organs or modalities, and overly emphasize model accuracy, making it difficult to comprehensively assess the overall efficacy of FL in real-world medical environments. To address these challenges, we developed the MobenFL benchmark. This benchmark integrates 20 cutting-edge FL algorithms and 22 medical imaging datasets, covering 12 critical organs across the human body, surpassing existing benchmark in breadth. In terms of evaluation dimensions, MobenFL not only assesses performance but also systematically incorporates key metrics such as algorithmic efficiency and privacy protection capabilities. Additionally, it conducts specialized evaluations for complex real-world clinical scenarios involving different diseases, devices, and imaging modalities, thereby providing a comprehensive and in-depth evaluation framework for the clinical application of FL in the medical field.

Figures

Figures reproduced from arXiv: 2607.08219 by Jianchun Zhu, Jin Liu, Junbin Mao, Ludi Li, Xu Tian.

Figure 1
Figure 1. Figure 1: Overview of the basic principles of federated learning. Clients send model parameters or model gradients to the server, and the server distributes the aggregated global model to the clients. significant variations in imaging devices, population dis￾tributions, and organ imaging across different institutions. Against this backdrop, the lack of unified evaluation stan￾dards makes it difficult to objectively … view at source ↗
Figure 2
Figure 2. Figure 2: The figure illustrates the multimodal medical data involved in this study. Among these, standard imaging data refers to dermatological and autism facial images captured via conventional cameras. In the field of general computing, widely recognized benchmark frameworks such as FedML1 and FATE2 have emerged. However, the professional evaluation system tai￾lored for the specific domain of medical imaging rema… view at source ↗
Figure 3
Figure 3. Figure 3: A multi-organ medical imaging dataset, covering various body parts such as the brain, abdomen, kidneys, colon, skin, eyes, oral cavity, breast, blood, and knee, with specific annotations showcasing publicly available sample data for diseases including Alzheimer’s disease, diabetic retinopathy, oral squamous cell carcinoma, breast cancer, ulcerative colitis, and knee. and leveraging SVM’s decision boundary … view at source ↗
Figure 4
Figure 4. Figure 4: Performance comparison of single-organ classification: (a) Results under the independently and identically distributed (IID) setting; (b) Results under the non-IID setting. The middle panel of each organ comparison illustrates the statistical significance of performance differences between the two settings. All subfigures are arranged vertically. Mao et al.: Preprint submitted to Elsevier Page 14 of 29 [P… view at source ↗
Figure 5
Figure 5. Figure 5: Detection results from different pathological Datasets. Each box plot summarizes the performance distribution on the test sets of 13 federated learning algorithms. The yellow line in the middle of the box represents the mean performance of the algorithm, the length of the box represents the standard deviation of the performance, and the upper and lower extensions of the whiskers represent the best and wors… view at source ↗
Figure 6
Figure 6. Figure 6: Detection results from different X-ray datasets. Each box plot summarizes the performance distribution on the test sets of 13 federated learning algorithms. The yellow line in the middle of the box represents the mean performance of the algorithm, the length of the box represents the standard deviation of the performance, and the upper and lower extensions of the whiskers represent the best and worst perfo… view at source ↗
Figure 7
Figure 7. Figure 7: Detection results for different brain diseases. Each box plot summarizes the performance distribution on the test sets of 13 federated learning algorithms. The yellow line in the middle of the box represents the mean performance of the algorithm, the length of the box represents the standard deviation of the performance, and the upper and lower extensions of the whiskers represent the best and worst perfor… view at source ↗
Figure 8
Figure 8. Figure 8: Detection results for different eye diseases. Each box plot summarizes the performance distribution on the test sets of 13 federated learning algorithms. The yellow line in the middle of the box represents the mean performance of the algorithm, the length of the box represents the standard deviation of the performance, and the upper and lower extensions of the whiskers represent the best and worst performa… view at source ↗
Figure 9
Figure 9. Figure 9: Detection Results from Different ASD Datasets. Each box plot summarizes the performance distribution on the test sets of 13 federated learning algorithms. The yellow line in the middle of the box represents the mean performance of the algorithm, the length of the box represents the standard deviation of the performance, and the upper and lower extensions of the whiskers represent the best and worst perform… view at source ↗
Figure 10
Figure 10. Figure 10: Detection results for different skin diseases. Each box plot summarizes the performance distribution on the test sets of 13 federated learning algorithms. The yellow line in the middle of the box represents the mean performance of the algorithm, the length of the box represents the standard deviation of the performance, and the upper and lower extensions of the whiskers represent the best and worst perfor… view at source ↗
Figure 11
Figure 11. Figure 11: The theoretical time complexity of federated learn￾ing; a lower position on the time complexity axis indicates a longer theoretical time consumption. the theoretical complexity and actual convergence time of the FL algorithms in MobenFL, with the results shown in [PITH_FULL_IMAGE:figures/full_fig_p022_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Ternary analysis of the efficiency of federated learn￾ing algorithms on the HAM10000 and OrganMNIST datasets. The three vertices of the ternary diagram represent three evaluation dimensions: the number of convergence rounds, the convergence time, and the average per-round runtime of the algorithm. The position of each point in the diagram is determined by the normalized values of these three metrics: the … view at source ↗
Figure 13
Figure 13. Figure 13: a: Performance of accuracy and AUC on the HAM10000 dataset under different noise scaling factors. b: Performance evaluation of accuracy and AUC on the OrganAMNIST dataset under different noise scaling factors. without DP, they demonstrate "mediocre robustness" in DP environments. Their performance degradation is relatively gradual, and under the strictest privacy constraint (𝜖=0.1), their results surpass … view at source ↗
Figure 14
Figure 14. Figure 14: Performance evaluation of accuracy and AUC on the HAM10000 dataset under different noise scaling factors. 16 32 64 128 256 batch_size 16 32 64 128 256 local_iter 78 79 80 81 82 83 84 Accuracy (%) Accuracy 16 32 64 128 256 batch_size 16 32 64 128 256 local_iter 96.0 96.5 97.0 97.5 98.0 98.5 99.0 99.5 100.0 AUC (%) AUC [PITH_FULL_IMAGE:figures/full_fig_p026_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Performance evaluation of accuracy and AUC on the OrganAMNIST dataset under different noise scaling factors. baseline levels and greater robustness, yet the optimal con￾figuration trends still point in the same direction. The impact of batch size exhibits a marginal diminishing characteristic: a batch size of 16 achieved optimal performance in 7 out of 12 experimental groups, but when the batch size incre… view at source ↗
Figure 16
Figure 16. Figure 16: Under the independent and identically distributed (IID) setting, algorithms such as FedNP, SplitFed, PGFed, and HarmoFL, which demonstrate excellent performance, also perform very well under the non-IID setting (as shown in Sec. 4.2), with the exception of PGFed. On the other hand, FedAu and TurboSVM show strong performance under non￾IID conditions while exhibiting suboptimal performance under IID setting… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

105 extracted references · 105 canonical work pages · 6 internal anchors

  1. [1]

    Nature Medicine , volume=

    Federated learning for predicting clinical outcomes in patients with COVID-19 , author=. Nature Medicine , volume=. 2021 , publisher=

    work page 2021

  2. [2]

    Nature Machine Intelligence , volume=

    Federated benchmarking of medical artificial intelligence with MedPerf , author=. Nature Machine Intelligence , volume=. 2023 , publisher=

    work page 2023

  3. [3]

    Nature Communications , volume=

    Federated learning enables big data for rare cancer boundary detection , author=. Nature Communications , volume=. 2022 , publisher=

    work page 2022

  4. [4]

    Knowledge-Based Systems , volume=

    Development of an AI framework using neural process continuous reinforcement learning to optimize highly volatile financial portfolios , author=. Knowledge-Based Systems , volume=. 2024 , publisher=

    work page 2024

  5. [5]

    IEEE Internet of Things Journal , volume=

    Efficient and lightweight convolutional networks for IoT malware detection: A federated learning approach , author=. IEEE Internet of Things Journal , volume=. 2022 , publisher=

    work page 2022

  6. [6]

    Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages=

    Federatedscope-llm: A comprehensive package for fine-tuning large language models in federated learning , author=. Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages=

    work page

  7. [7]

    IEEE Transactions on Medical Imaging , volume=

    Do gradient inversion attacks make federated learning unsafe? , author=. IEEE Transactions on Medical Imaging , volume=. 2023 , publisher=

    work page 2023

  8. [8]

    Nature Communications , volume=

    Mining multi-center heterogeneous medical data with distributed synthetic learning , author=. Nature Communications , volume=. 2023 , publisher=

    work page 2023

  9. [9]

    International Conference on Medical Image Computing and Computer-Assisted Intervention , pages=

    Fine-tuning network in federated learning for personalized skin diagnosis , author=. International Conference on Medical Image Computing and Computer-Assisted Intervention , pages=. 2023 , organization=

    work page 2023

  10. [10]

    Applied Soft Computing , volume=

    Federated learning for COVID-19 screening from Chest X-ray images , author=. Applied Soft Computing , volume=. 2021 , publisher=

    work page 2021

  11. [11]

    Journal of Medical Internet Research , volume=

    Federated learning on clinical benchmark data: performance assessment , author=. Journal of Medical Internet Research , volume=. 2020 , publisher=

    work page 2020

  12. [12]

    Advances in Neural Information Processing Systems , volume=

    Flamby: Datasets and benchmarks for cross-silo federated learning in realistic healthcare settings , author=. Advances in Neural Information Processing Systems , volume=

    work page

  13. [13]

    2018 , version =

    Tschandl, Philipp , publisher =. 2018 , version =. doi:10.7910/DVN/DBW86T , url =

    work page doi:10.7910/dvn/dbw86t 2018

  14. [14]

    Scientific Data , volume=

    Medmnist v2-a large-scale lightweight benchmark for 2d and 3d biomedical image classification , author=. Scientific Data , volume=. 2023 , publisher=

    work page 2023

  15. [15]

    Medical Image Analysis , volume=

    The liver tumor segmentation benchmark (lits) , author=. Medical Image Analysis , volume=. 2023 , publisher=

    work page 2023

  16. [16]

    IEEE Transactions on Medical Imaging , volume=

    Efficient multiple organ localization in CT image using 3D region proposal network , author=. IEEE Transactions on Medical Imaging , volume=. 2019 , publisher=

    work page 2019

  17. [17]

    Nature Methods , volume=

    Annotated high-throughput microscopy image sets for validation , author=. Nature Methods , volume=

    work page

  18. [18]

    doi:10.5281/zenodo.1214456 , url =

    Kather, Jakob Nikolas and Halama, Niels and Marx, Alexander , title =. doi:10.5281/zenodo.1214456 , url =

    work page doi:10.5281/zenodo.1214456

  19. [19]

    Database , volume=

    Bracs: A dataset for breast carcinoma subtyping in h&e histology images , author=. Database , volume=. 2022 , publisher=

    work page 2022

  20. [20]

    Inflammatory Bowel Diseases , pages=

    Improving the computer-aided estimation of ulcerative colitis severity according to mayo endoscopic score by using regression-based deep learning , author=. Inflammatory Bowel Diseases , pages=. 2022 , publisher=

    work page 2022

  21. [21]

    Proceedings of the IEEE Conference on Computer Cision and Pattern Recognition , pages=

    Chestx-ray8: Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases , author=. Proceedings of the IEEE Conference on Computer Cision and Pattern Recognition , pages=

    work page

  22. [22]

    2020 , publisher =

    Gornale, Shivanand and Patravali, Pooja , title =. 2020 , publisher =. doi:10.17632/t9ndx37v5h.1 , version =

    work page doi:10.17632/t9ndx37v5h.1 2020

  23. [23]

    Neuroimaging Clinics , volume=

    The Alzheimer's disease neuroimaging initiative , author=. Neuroimaging Clinics , volume=. 2005 , publisher=

    work page 2005

  24. [24]

    Artificial Intelligence and Statistics , pages=

    Communication-efficient learning of deep networks from decentralized data , author=. Artificial Intelligence and Statistics , pages=. 2017 , organization=

    work page 2017

  25. [25]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Splitfed: When federated learning meets split learning , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  26. [26]

    FedBN: Federated Learning on Non-IID Features via Local Batch Normalization

    Fedbn: Federated learning on non-iid features via local batch normalization , author=. arXiv preprint arXiv:2102.07623 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  27. [27]

    Proceedings of Machine Learning and Systems , volume=

    Federated optimization in heterogeneous networks , author=. Proceedings of Machine Learning and Systems , volume=

    work page

  28. [28]

    FedCD: Improving Performance in non-IID Federated Learning

    Fedcd: Improving performance in non-iid federated learning , author=. arXiv preprint arXiv:2006.09637 , year=

    work page internal anchor Pith review Pith/arXiv arXiv 2006

  29. [29]

    Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , pages=

    Flop: Federated learning on medical datasets using partial networks , author=. Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining , pages=

    work page

  30. [30]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Model-contrastive federated learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    work page

  31. [31]

    IEEE Journal of Biomedical and Health Informatics , volume=

    Splitavg: A heterogeneity-aware federated deep learning method for medical imaging , author=. IEEE Journal of Biomedical and Health Informatics , volume=. 2022 , publisher=

    work page 2022

  32. [32]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Harmofl: Harmonizing local and global drifts in federated learning on heterogeneous medical images , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  33. [33]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Fedala: Adaptive local aggregation for personalized federated learning , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  34. [34]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    FedNP: Towards non-IID federated learning via federated neural propagation , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  35. [35]

    Nature Communications , volume=

    Decentralized federated learning through proxy model sharing , author=. Nature Communications , volume=. 2023 , publisher=

    work page 2023

  36. [36]

    Proceedings of the 31st ACM International Conference on Multimedia , pages=

    Fedgh: Heterogeneous federated learning with generalized global header , author=. Proceedings of the 31st ACM International Conference on Multimedia , pages=

    work page

  37. [37]

    Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    Gpfl: Simultaneously learning global and personalized feature information for personalized federated learning , author=. Proceedings of the IEEE/CVF International Conference on Computer Vision , pages=

    work page

  38. [38]

    Proceedings of the IEEE/CVF international conference on computer vision , pages=

    Pgfed: Personalize each client's global objective for federated learning , author=. Proceedings of the IEEE/CVF international conference on computer vision , pages=

    work page

  39. [39]

    Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning , year=

    Huang, Wenke and Ye, Mang and Shi, Zekun and Du, Bo , journal=. Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning , year=

    work page

  40. [40]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    TurboSVM-FL: Boosting federated learning through SVM aggregation for lazy clients , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  41. [41]

    Shi, Yujun and Liang, Jian and Zhang, Wenqing and Xue, Chuhui and Tan, Vincent Y. F. and Bai, Song , journal=. Understanding and Mitigating Dimensional Collapse in Federated Learning , year=

    work page

  42. [42]

    The Twelfth International Conference on Learning Representations , year=

    A Lightweight Method for Tackling Unknown Participation Statistics in Federated Averaging , author=. The Twelfth International Conference on Learning Representations , year=

    work page

  43. [43]

    Journal of Machine Learning Research , volume=

    Latent dirichlet allocation , author=. Journal of Machine Learning Research , volume=

    work page

  44. [44]

    International Colloquium on Automata, Languages, and Programming , pages=

    Differential privacy , author=. International Colloquium on Automata, Languages, and Programming , pages=. 2006 , organization=

    work page 2006

  45. [45]

    ACM Computing Surveys (Csur) , volume=

    A survey on homomorphic encryption schemes: Theory and implementation , author=. ACM Computing Surveys (Csur) , volume=. 2018 , publisher=

    work page 2018

  46. [46]

    Nature , year=

    Swarm Learning for decentralized and confidential clinical machine learning , author=. Nature , year=

    work page

  47. [47]

    Nature Medicine , year=

    Swarm learning for decentralized artificial intelligence in cancer histopathology , author=. Nature Medicine , year=

    work page

  48. [48]

    Computer Communications , volume=

    A survey on federated learning for security and privacy in healthcare applications , author=. Computer Communications , volume=. 2023 , publisher=

    work page 2023

  49. [49]

    2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) , pages=

    Differential private federated transfer learning for mental health monitoring in everyday settings: A case study on stress detection , author=. 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) , pages=. 2024 , organization=

    work page 2024

  50. [50]

    2024 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) , pages=

    Differentially Private Multi-Site Treatment Effect Estimation , author=. 2024 IEEE Conference on Secure and Trustworthy Machine Learning (SaTML) , pages=. 2024 , organization=

    work page 2024

  51. [51]

    Information Fusion , volume=

    Increasing trust in AI through privacy preservation and model explainability: Federated Learning of Fuzzy Regression Trees , author=. Information Fusion , volume=. 2025 , publisher=

    work page 2025

  52. [52]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    LR-XFL: logical reasoning-based explainable federated learning , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  53. [53]

    IEEE Transactions on Wireless Communications , year=

    Tackling Class Imbalance and Client Heterogeneity for Split Federated Learning in Wireless Networks , author=. IEEE Transactions on Wireless Communications , year=

    work page

  54. [54]

    IEEE Transactions on Mobile Computing , year=

    Exploring amplified heterogeneity arising from heavy-tailed distributions in federated learning , author=. IEEE Transactions on Mobile Computing , year=

    work page

  55. [55]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Clip-guided federated learning on heterogeneity and long-tailed data , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  56. [56]

    IEEE Transactions on Information Forensics and Security , year=

    NSPFL: A novel secure and privacy-preserving federated learning with data integrity auditing , author=. IEEE Transactions on Information Forensics and Security , year=

    work page

  57. [57]

    IEEE Transactions on Sustainable Computing , year=

    Using third-party auditor to help federated learning: An efficient Byzantine-robust federated learning , author=. IEEE Transactions on Sustainable Computing , year=

    work page

  58. [58]

    Transactions on Machine Learning Research , year=

    Hybrid federated learning for feature & sample heterogeneity: Algorithms and implementation , author=. Transactions on Machine Learning Research , year=

    work page

  59. [59]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Unlocking the potential of prompt-tuning in bridging generalized and personalized federated learning , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    work page

  60. [60]

    Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence , pages=

    FedSampling: a better sampling strategy for federated learning , author=. Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence , pages=

    work page

  61. [61]

    Robust Model Aggregation for Heterogeneous Federated Learning: Analysis and Optimizations

    Robust Model Aggregation for Heterogeneous Federated Learning: Analysis and Optimizations , author=. arXiv preprint arXiv:2405.06993 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  62. [62]

    FedMap: Iterative Magnitude-Based Pruning for Communication-Efficient Federated Learning

    FedMap: Iterative Magnitude-Based Pruning for Communication-Efficient Federated Learning , author=. arXiv preprint arXiv:2406.19050 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  63. [63]

    CoRR , year=

    Comb, Prune, Distill: Towards Unified Pruning for Vision Model Compression , author=. CoRR , year=

    work page

  64. [64]

    Information Sciences , volume=

    Communication-efficient federated learning via personalized filter pruning , author=. Information Sciences , volume=. 2024 , publisher=

    work page 2024

  65. [65]

    FedCod: An Efficient Communication Protocol for Cross-Silo Federated Learning with Coding

    FedCod: An Efficient Communication Protocol for Cross-Silo Federated Learning with Coding , author=. arXiv preprint arXiv:2501.00216 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  66. [66]

    Proceedings of the 53rd International Conference on Parallel Processing , pages=

    Bandwidth-Aware and Overlap-Weighted Compression for Communication-Efficient Federated Learning , author=. Proceedings of the 53rd International Conference on Parallel Processing , pages=

    work page

  67. [67]

    IEEE Internet of Things Journal , year=

    Federated Continual Learning Based on Weakly Supervised Diffusion Models for Disease Diagnosis , author=. IEEE Internet of Things Journal , year=

    work page

  68. [68]

    IEEE Transactions on Parallel and Distributed Systems , year=

    Loci: Federated Continual Learning of Heterogeneous Tasks at Edge , author=. IEEE Transactions on Parallel and Distributed Systems , year=

    work page

  69. [69]

    IEEE Transactions on Information Forensics and Security , year=

    Lightweight and Dynamic Privacy-Preserving Federated Learning via Functional Encryption , author=. IEEE Transactions on Information Forensics and Security , year=

    work page

  70. [70]

    Computers in Industry , volume=

    FedCPG: A class prototype guided personalized lightweight federated learning framework for cross-factory fault detection , author=. Computers in Industry , volume=. 2025 , publisher=

    work page 2025

  71. [71]

    PraVFed: Practical Heterogeneous Vertical Federated Learning via Representation Learning , year=

    Wang, Shuo and Gai, Keke and Yu, Jing and Zhang, Zijian and Zhu, Liehuang , journal=. PraVFed: Practical Heterogeneous Vertical Federated Learning via Representation Learning , year=

    work page

  72. [72]

    Medical image analysis , volume=

    Refuge challenge: A unified framework for evaluating automated methods for glaucoma assessment from fundus photographs , author=. Medical image analysis , volume=. 2020 , publisher=

    work page 2020

  73. [73]

    Journal of microscopy , volume=

    Textural pattern classification for oral squamous cell carcinoma , author=. Journal of microscopy , volume=. 2018 , publisher=

    work page 2018

  74. [74]

    Scientific data , volume=

    A patient-centric dataset of images and metadata for identifying melanomas using clinical context , author=. Scientific data , volume=. 2021 , publisher=

    work page 2021

  75. [75]

    Data in brief , volume=

    PAD-UFES-20: A skin lesion dataset composed of patient data and clinical images collected from smartphones , author=. Data in brief , volume=. 2020 , publisher=

    work page 2020

  76. [76]

    Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    He, Kaiming and Zhang, Xiangyu and Ren, Shaoqing and Sun, Jian , title =. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , month =

    work page

  77. [77]

    Towards Personalized Federated Learning , year=

    Tan, Alysa Ziying and Yu, Han and Cui, Lizhen and Yang, Qiang , journal=. Towards Personalized Federated Learning , year=

    work page

  78. [78]

    Distributed learning of deep neural network over multiple agents , journal =

    Otkrist Gupta and Ramesh Raskar , keywords =. Distributed learning of deep neural network over multiple agents , journal =. 2018 , issn =. doi:https://doi.org/10.1016/j.jnca.2018.05.003 , url =

    work page doi:10.1016/j.jnca.2018.05.003 2018

  79. [79]

    IEEE Communications Surveys & Tutorials , volume=

    Decentralized federated learning: Fundamentals, state of the art, frameworks, trends, and challenges , author=. IEEE Communications Surveys & Tutorials , volume=. 2023 , publisher=

    work page 2023

  80. [80]

    Adaptive Federated Optimization , author=

    work page

Showing first 80 references.