Pith. sign in

REVIEW 3 major objections 4 minor 57 references

Privacy-Preserving Chest X-ray Report Generation via Multimodal Federated Learning with ViT and GPT-2

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Krum aggregation tops a centralized ViT–GPT-2 for X-ray reports

desk verdict A modest FL feasibility study with overstated privacy and performance claims; the core experiment is coherent but needs a rewrite of claims and error bars. read the letter →

arxiv 2505.21715 v1 pith:5KXGUQ43 submitted 2025-05-27 eess.IV cs.AIcs.CV

Md. Zahid Hossain , Mustofa Ahmed , Most. Sharmin Sultana Samu , Md. Rakibul Islam This is my paper
classification eess.IVcs.AIcs.CV
keywords MultimodalFederatedLearningPrivacy-PreservingMedicalReportGenerationVision-LanguageModelsChestX-rayViTGPT-2KrumAggregation
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

The paper tries to establish that radiology report generation from chest X-rays can be done with federated learning, so no raw images leave the hospital, without sacrificing report quality. It proposes a ViT encoder and GPT-2 decoder trained across four simulated clients on the IU-Xray dataset, exchanging only model checkpoints through Google Drive and coordinating via Firebase. Comparing FedAvg, Krum aggregation, and a novel loss-weighted averaging (L-FedAvg), the authors find that Krum gives the best lexical and semantic scores. Their headline claim is that federated learning can match or surpass a centralized ViT+GPT-2 baseline, which would make collaborative, privacy-preserving medical AI practical.

What carries the argument

The central object is the federated aggregation rule. FedAvg averages client weights weighted by dataset size; Krum selects the single client update whose squared distance to the other updates is smallest, discarding suspected outliers; and L-FedAvg weights clients by a blend of dataset size and inverse validation loss, controlled by an alpha of 0.5. The model itself is a Vision Transformer (ViT-B16) that encodes image patches, feeding visual features through cross-attention into GPT-2, which autoregressively writes the report. The infrastructure is deliberately lightweight: clients upload 700 to 800 MB checkpoints to a shared Google Drive folder and signal status through Firebase, with no dedicated server hardware.

What would settle it

Run a gradient- or weight-inversion attack on the 700 to 800 MB checkpoints exchanged between clients and the server and try to reconstruct recognizable chest X-ray images or patient-identifying features; if reconstruction succeeds, the paper's privacy claim is falsified. Separately, poisoning one client with corrupted labels and showing that FedAvg degrades while Krum stays stable would support the robustness claim, while observing Krum's advantage disappear under random client splits would weaken it.

Watch

Extended reading notes

Core claim

The paper claims that a four-client federated ViT+GPT-2 system using Krum aggregation generates chest X-ray reports that outperform the centralized ViT+GPT-2 baseline on most reported metrics: ROUGE-1 F1 of 0.306 versus 0.2877, BERTScore F1 of 0.8731 versus 0.8691, and RaTEScore of 62.24 versus 53.47, while the centralized model retains the best BLEU and ROUGE-4 scores. Against published systems, the authors claim that their federated model outperforms all baselines in semantic fidelity, with the highest BERTScore F1 and RaTEScore. The authors interpret these results as showing that decentralized training need not compromise report quality and that robust aggregation can even improve it.

Load-bearing premise

The load-bearing premise is that keeping image data local while exchanging model weights preserves patient privacy, yet the paper adds no differential privacy, secure aggregation, or attack evaluation, so if raw weights can leak patient information the core motivation fails.

Editorial extensions

If this is right

  • Krum aggregation, not FedAvg, should be the default aggregation method for this report-generation task when clients may be heterogeneous or faulty.
  • Hospitals could in principle collaborate on report-generation models without transferring images, using only shared storage and a cloud database.
  • The reported RaTEScore gap of 62.24 versus 53.47 suggests the federated model is more clinically faithful than the centralized baseline, not just lexically similar.
  • The approach is extensible in principle to larger datasets and more clients, as the authors state, since the communication pattern does not depend on the number of sites.

Reading between the lines

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

  • The privacy guarantee as stated is weaker than the title suggests: exchanging full model weights without differential privacy or secure aggregation leaves the system open to model-inversion and membership-inference attacks, which the paper's own cited related work documents.
  • The comparison to published baselines is apples-to-oranges because those systems are trained and evaluated under different data splits and metrics; the semantic-fidelity lead may reflect evaluation differences as much as model quality.
  • A direct test of Krum's robustness hypothesis would be to poison one client's updates and show that Krum preserves report quality while FedAvg degrades; the paper does not perform this attack.
  • Because all four clients are simulated by partitioning the same IU-Xray dataset, real-world non-IID drift across hospitals is not tested; performance under genuine distribution shift remains an open question.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a federated learning framework for chest X-ray report generation using a Vision Transformer (ViT) encoder and a GPT-2 decoder. The IU-Xray dataset is partitioned across four simulated clients, and three aggregation strategies are compared: FedAvg, Krum aggregation, and a proposed Loss-aware Federated Weighted Averaging (L-FedAvg). Tables 5 and 6 report n-gram and semantic metrics on the IU-Xray test split, and Table 7 compares the authors' Krum-based federated model with prior report-generation systems. The paper concludes that Krum aggregation performs best and that the framework is privacy-preserving because raw image data is not shared.

Significance. If the empirical claims were supported, this would be a useful, low-cost demonstration that generative vision-language models can be trained in a decentralized manner for radiology report generation. The manuscript has some strengths: the client partitions and local training hyperparameters are stated explicitly, the L-FedAvg algorithm is written out precisely, and the evaluation covers lexical, semantic, and clinical metrics. However, the central privacy-preservation claim is not supported by any privacy mechanism, formal accounting, or adversarial evaluation, and the performance claims rest on single-run comparisons with very small metric margins. The contribution is therefore better described as a decentralized-training feasibility study than as a privacy-preserving system. No code, data splits, or checkpoints are provided, despite the claimed lightweight replicability.

major comments (3)
  1. [Section 1 and Section 3 (parameter exchange)] The title, abstract, and contribution list describe the framework as 'privacy-preserving' and claim it works 'without compromising data confidentiality', but the implemented system exchanges raw model weights of 700-800 MB via a shared Google Drive folder and adds no differential privacy, secure aggregation, secure multiparty computation, or privacy accounting. The paper's own related work ([12], [17]) documents reconstruction attacks against shared-update federated systems. As it stands, the privacy premise is unsupported; either add a concrete privacy mechanism with an adversarial evaluation or remove the privacy claim from the title and abstract and list the absence of privacy guarantees as an explicit limitation.
  2. [Tables 5 and 6; Section 4] All metric results appear to be from single runs: no standard deviations, numbers of seeds, or significance tests are reported. The superiority claim for Krum rests on differences such as BERTScore F1 0.8731 versus 0.8720 and RaTEScore 62.24 versus 61.99, which are far smaller than typical run-to-run variation in generative models. In addition, Section 4 omits the total number of federated rounds T and the actual client-selection scheme used in Algorithm 1, and no random seeds are reported. These omissions make the central performance claims non-reproducible and statistically unsupported; please add repeated runs, variance estimates, and a significance test, or weaken the 'superior performance' wording accordingly.
  3. [Table 7 and Section 5] The comparison with prior work in Table 7 is not valid as presented. The cited systems (MAIRA-2, CXRMate, EAST, etc.) were trained and evaluated on different corpora and with different metric configurations, so the statement that the federated model 'significantly outperforms all baselines in semantic fidelity' is unsupported. The table is also internally inconsistent: the same Krum condition is reported with BLEU 0.0426 in Table 7 but BLEU 0.0395 in Table 5. Please restrict comparisons to the centralized baseline reported in Table 5, provide matched evaluation protocols for any external systems, and reconcile the duplicated BLEU values.
minor comments (4)
  1. [Section 3 (metric introduction) and Section 5] The metric name is spelled 'RateScore' in one passage and 'RaTEScore' elsewhere; please use the consistent name 'RaTEScore' throughout.
  2. [Section 3.1.2] There is a typo in the text: 'vainlla' should be 'vanilla'.
  3. [Section 3.1.3 and Table 3] L-FedAvg is evaluated only with alpha fixed at 0.5 and Krum only with f fixed at 1, with no sensitivity analysis. A brief ablation or discussion would help justify these choices and show that the results are not artifacts of a single hand-set configuration.
  4. [Section 4 and Algorithm 1] Algorithm 1 includes a client fraction C, but Section 4 does not state whether all four clients participate in every round or whether a random subset is selected; please clarify the actual experimental protocol.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: evaluation uses external IU-Xray ground truth and no test metric is fed into aggregation, but the centralized baseline supporting the headline 'FL can match or surpass centralized' claim is the authors' own prior work ([51]).

full rationale

The paper's derivation chain is not circular. All reported metrics (ROUGE, BLEU, BERTScore, RaTEScore) are computed by comparing generated reports against the external IU-Xray ground-truth reports; no test-set metric is fed back into any aggregation rule. L-FedAvg's weighting uses local validation loss and client data sizes (Algorithm 3), with alpha fixed at 0.5, and Krum's fault tolerance is fixed at 1 (Table 3), so the aggregation outcomes are not statistically forced by the evaluation targets. The only noticeable self-citation is the centralized baseline [51], which is the same authors' prior ViT+GPT-2 model and is used to support the headline claim that federated learning can match or surpass centralized models; this is a benchmark comparison rather than a derivation input, so it is a minor self-citation issue, not a circular step. The paper's 'privacy-preserving' framing is asserted rather than demonstrated (no differential privacy, secure aggregation, or attack evaluation, despite raw 700-800 MB checkpoints being shared via Google Drive), but that is an unsupported premise and a validity concern, not a circular derivation. Table 7 also compares against prior methods on different setups and uses missing metrics to claim semantic superiority, which is a comparability flaw, not circularity. Overall, the central results are self-contained against an external benchmark and no prediction reduces by construction to a fitted input or to the authors' prior work.

Assumptions & free parameters 5 free parameters · 5 assumptions · 1 invented entities

The central results rest on standard fine-tuning choices and on an unexamined privacy premise: raw checkpoints are shared with no differential privacy or secure aggregation. One new construct (L-FedAvg) is introduced with no independent evidence and underperforms the FedAvg baseline in the paper's own tables. Several load-bearing choices are unreported, most notably the number of federated rounds and the BERTScore configuration, which prevents external reproduction of the exact numbers.

free parameters (5)
  • L-FedAvg alpha = 0.5
    Hand-chosen weighting between validation-loss weight and training-data-size weight (Table 3). No sensitivity analysis is reported, and the paper's own Tables 5-6 show L-FedAvg trailing FedAvg on most metrics, so the choice is not validated.
  • Krum fault tolerance f = 1
    Set to tolerate up to 1 faulty client (Table 3). No Byzantine attack simulation or sensitivity analysis justifies this value.
  • Local training hyperparameters = 3 epochs, batch 8, LR 5e-5, weight decay 0.01
    Table 2 sets these standard fine-tuning values by hand without a reported sweep. They affect convergence and every downstream metric.
  • Number of federated rounds T = not reported
    Algorithm 1 requires T, but the paper never states how many rounds were run. This unreported choice directly shapes the reported losses and metric values.
  • BERTScore and RaTEScore configuration = not reported
    BERTScore values depend on the underlying BERT checkpoint and rescaling; RaTEScore depends on entity-extraction settings. The paper does not report these, which makes the Table 7 cross-paper comparison non-comparable.
assumptions (5)
  • ad hoc to paper Raw model-parameter exchange among trusted clients preserves patient privacy and satisfies HIPAA and GDPR.
    Asserted in Section 1 and operationalized in Section 3 (700-800 MB checkpoints via Google Drive). No differential privacy or secure aggregation is used, and the paper's own citations [12] and [17] document reconstruction attacks on such shared-update systems.
  • domain assumption IU-Xray reference reports are reliable ground truth for training and evaluation.
    Standard assumption in the field; Figure 2 treats the paired findings as ground truth for all metrics.
  • domain assumption A single public dataset split into 4 disjoint shards simulates a realistic multi-institution federated setting.
    Section 3 partitions IU-Xray among 4 simulated Colab clients; this is a same-dataset simulation, so non-IID realism, communication costs, and privacy realism are assumed rather than demonstrated.
  • domain assumption GPT-2 with cross-attention over ViT features is a valid image-to-text decoder for this task.
    Section 3 says 'a cross-attention mechanism is incorporated' without architectural detail; the implementation presumably follows the standard VisionEncoderDecoder pattern, but that is assumed, not validated.
  • domain assumption Metric values computed by the authors are directly comparable to the same metric names in prior publications.
    Table 7 compares BERTScore and RaTEScore across papers with different evaluation pipelines; BERTScore values of 0.32-0.56 in cited works versus 0.87 here signal incompatible configurations, so comparability is assumed without evidence.
invented entities (1)
  • L-FedAvg (Loss-aware Federated Weighted Averaging)
    purpose: Aggregates client updates weighted by inverse validation loss and training data size (alpha=0.5) to prioritize clients aligned with the global objective.
    The only new construct in the paper. It has no falsifiable handle outside the paper, and the paper's own Table 5 shows it underperforming FedAvg on ROUGE-1/2/3/L and BERTScore F1, so no independent evidence supports its claimed benefit.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Privacy-Preserving Chest X-ray Report Generation via Multimodal Federated Learning with ViT and GPT-2." pith.science (2026). https://pith.science/paper/5KXGUQ43

@misc{pith2026250521715,
  author       = {Pith},
  title        = {Pith review of: Privacy-Preserving Chest X-ray Report Generation via Multimodal Federated Learning with ViT and GPT-2},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5KXGUQ43}},
  note         = {Machine review of arXiv:2505.21715}
}
read the original abstract

The automated generation of radiology reports from chest X-ray images holds significant promise in enhancing diagnostic workflows while preserving patient privacy. Traditional centralized approaches often require sensitive data transfer, posing privacy concerns. To address this, the study proposes a Multimodal Federated Learning framework for chest X-ray report generation using the IU-Xray dataset. The system utilizes a Vision Transformer (ViT) as the encoder and GPT-2 as the report generator, enabling decentralized training without sharing raw data. Three Federated Learning (FL) aggregation strategies: FedAvg, Krum Aggregation and a novel Loss-aware Federated Averaging (L-FedAvg) were evaluated. Among these, Krum Aggregation demonstrated superior performance across lexical and semantic evaluation metrics such as ROUGE, BLEU, BERTScore and RaTEScore. The results show that FL can match or surpass centralized models in generating clinically relevant and semantically rich radiology reports. This lightweight and privacy-preserving framework paves the way for collaborative medical AI development without compromising data confidentiality.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

57 extracted references · 46 canonical work pages

  1. [30]

    Computerized Medical Imaging and Graphics 113, 102342 (2024)

    Chen, J., Pan, R.: Medical report generation based on multimodal federated learning. Computerized Medical Imaging and Graphics 113, 102342 (2024)

    work page 2024

  2. [12]

    Sensors 22(14), 5195 (2022)

    Ziegler, J., Pfitzner, B., Schulz, H., Saalbach, A., Arnrich, B.: Defending against reconstruction attacks through differentially private federated learning for classification of heterogeneous chest x-ray data. Sensors 22(14), 5195 (2022)

    work page 2022

  3. [17]

    Sensors 22(10), 3728 (2022)

    Ho, T.-T., Tran, K.-D., Huang, Y.: Fedsgdcovid: Federated sgd covid-19 detec- tion under local differential privacy using chest x-ray images and symptom information. Sensors 22(10), 3728 (2022)

    work page 2022

  4. [1]

    Medical Image Analysis 42, 60–88 (2017) https://doi.org/10.1016/j.media.2017.07.005

    Litjens, G., Kooi, T., Bejnordi, B.E., Setio, A.A.A., Ciompi, F., Ghafoorian, M., van der Laak, J.A.W.M., van Ginneken, B., S´ anchez, C.I.: A survey on deep learning in medical image analysis. Medical Image Analysis 42, 60–88 (2017) https://doi.org/10.1016/j.media.2017.07.005

    doi:10.1016/j.media.2017.07.005 2017

  5. [2]

    arXiv preprint arXiv:2204.13258 (2022)

    Chen, Z., Shen, Y., Song, Y., Wan, X.: Cross-modal memory networks for radiology report generation. arXiv preprint arXiv:2204.13258 (2022)

    arXiv 2022

  6. [3]

    arXiv preprint arXiv:1711.08195 (2017)

    Jing, B., Xie, P., Xing, E.: On the automatic generation of medical imaging reports. arXiv preprint arXiv:1711.08195 (2017)

    arXiv 2017

  7. [4]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Li, C.Y., Liang, X., Hu, Z., Xing, E.P.: Knowledge-driven encode, retrieve, paraphrase for medical image report generation. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 33, pp. 6666–6673 (2019)

    work page 2019

  8. [5]

    BMC Medical Ethics 22(1), 122 (2021) https://doi

    Murdoch, B.: Privacy and artificial intelligence: challenges for protecting health information in a new era. BMC Medical Ethics 22(1), 122 (2021) https://doi. 19 org/10.1186/s12910-021-00687-3

    work page doi:10.1186/s12910-021-00687-3 2021

Show all 57 references
  1. [6]

    In: Artificial Intelligence and Statistics, pp

    McMahan, B., Moore, E., Ramage, D., Hampson, S., Arcas, B.A.: Communication-efficient learning of deep networks from decentralized data. In: Artificial Intelligence and Statistics, pp. 1273–1282 (2017). PMLR

    2017

  2. [7]

    https://arxiv.org/abs/1912.04977

    Kairouz, P., McMahan, H.B., Avent, B., Bellet, A., Bennis, M., Bhagoji, A.N., Bonawitz, K., Charles, Z., Cormode, G., Cummings, R., D’Oliveira, R.G.L., Eich- ner, H., Rouayheb, S.E., Evans, D., Gardner, J., Garrett, Z., Gasc´ on, A., Ghazi, B., Gibbons, P.B., Gruteser, M., Har...

    2021 arXiv

  3. [8]

    In: 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), pp

    Chakravarty, A., Kar, A., Sethuraman, R., Sheet, D.: Federated learning for site aware chest radiograph screening. In: 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), pp. 1077–1081 (2021). IEEE

    2021

  4. [9]

    In: AI 2020: Advances in Artificial Intelligence: 33rd Australasian Joint Conference, AI 2020, Canberra, ACT, Australia, November 29–30, 2020, Proceedings 33, pp

    Banerjee, S., Misra, R., Prasad, M., Elmroth, E., Bhuyan, M.H.: Multi-diseases classification from chest-x-ray: A federated deep learning approach. In: AI 2020: Advances in Artificial Intelligence: 33rd Australasian Joint Conference, AI 2020, Canberra, ACT, Australia, November...

    2020

  5. [10]

    Scientific reports 10(1), 12598 (2020)

    Sheller, M.J., Edwards, B., Reina, G.A., Martin, J., Pati, S., Kotrotsou, A., Milchenko, M., Xu, W., Marcus, D., Colen, R.R., et al.: Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data. Scientific reports 10(1), 12598 (2020)

    2020

  6. [11]

    In: Interna- tional Conference on Computational Science, pp

    ´Slazyk, F., Jab lecki, P., Lisowska, A., Malawski, M., P lotka, S.: Cxr-fl: deep learning-based chest x-ray image analysis using federated learning. In: Interna- tional Conference on Computational Science, pp. 433–440 (2022). Springer

    2022

  7. [13]

    In: Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pp

    Papineni, K., Roukos, S., Ward, T., Zhu, W.-J.: Bleu: a method for automatic evaluation of machine translation. In: Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pp. 311–318 (2002)

    2002

  8. [14]

    In: Text Summarization Branches Out, pp

    Lin, C.-Y.: Rouge: A package for automatic evaluation of summaries. In: Text Summarization Branches Out, pp. 74–81 (2004) 20

    2004

  9. [15]

    arXiv preprint arXiv:1904.09675 (2019)

    Zhang, T., Kishore, V., Wu, F., Weinberger, K.Q., Artzi, Y.: Bertscore: Evaluat- ing text generation with bert. arXiv preprint arXiv:1904.09675 (2019)

    2019 arXiv

  10. [16]

    In: Al-Onaizan, Y., Bansal, M., Chen, Y.-N

    Zhao, W., Wu, C., Zhang, X., Zhang, Y., Wang, Y., Xie, W.: RaTEScore: A met- ric for radiology report generation. In: Al-Onaizan, Y., Bansal, M., Chen, Y.-N. (eds.) Proceedings of the 2024 Conference on Empirical Methods in Natural Lan- guage Processing, pp. 15004–15019. Assoc...

    2024 doi

  11. [18]

    Scientific Reports 12(1), 3551 (2022)

    Linardos, A., Kushibar, K., Walsh, S., Gkontra, P., Lekadir, K.: Federated learn- ing for multi-center imaging diagnostics: a simulation study in cardiovascular disease. Scientific Reports 12(1), 3551 (2022)

    2022

  12. [19]

    Scientific reports 12(1), 1953 (2022)

    Adnan, M., Kalra, S., Cresswell, J.C., Taylor, G.W., Tizhoosh, H.R.: Federated learning and differential privacy for medical image analysis. Scientific reports 12(1), 1953 (2022)

    2022

  13. [20]

    Scientific Reports 13(1), 22576 (2023)

    Tayebi Arasteh, S., Kuhl, C., Saehn, M.-J., Isfort, P., Truhn, D., Nebelung, S.: Enhancing domain generalization in the ai-based analysis of chest radiographs with federated learning. Scientific Reports 13(1), 22576 (2023)

    2023

  14. [21]

    In: Medical Imaging 2023: Computer-Aided Diagnosis, vol

    Chowdari, D.K., Radhasyam, N., Pal, A., Paul, A.: Federated learning using multi-institutional data for generalizable chest x-ray diagnosis. In: Medical Imaging 2023: Computer-Aided Diagnosis, vol. 12465, pp. 81–87 (2023). SPIE

    2023

  15. [22]

    International Journal of Machine Learning and Cybernetics 14(8), 2659–2670 (2023)

    Makkar, A., Santosh, K.: Securefed: federated learning empowered medical imag- ing technique to analyze lung abnormalities in chest x-rays. International Journal of Machine Learning and Cybernetics 14(8), 2659–2670 (2023)

    2023

  16. [23]

    IEEE Access 11, 28628–28644 (2023)

    Sohan, M.F., Basalamah, A.: A systematic review on federated learning in medical image analysis. IEEE Access 11, 28628–28644 (2023)

    2023

  17. [24]

    In: 2023 4th IEEE Global Conference for Advancement in Technology (GCAT), pp

    Jindal, V., Kukreja, V., Singh, D.P., Vats, S., Mehta, S.: Pushing diagnostic fron- tiers: Federated learning cnn for diverse lung disease. In: 2023 4th IEEE Global Conference for Advancement in Technology (GCAT), pp. 1–6 (2023). IEEE

    2023

  18. [25]

    In: 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp

    Liu, C., Luo, Y., Xu, Y., Du, B.: Fedarc: Federated learning for multi-center tuber- culosis chest x-ray diagnosis with adaptive regularizing contrastive representation. In: 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), pp. 2125–2128 (2023). IEEE

    2023

  19. [26]

    In: Proceedings of the 2023 9th International Conference on Computer Technology Applications, pp

    Zubair Nafis, K.F., Maisha Tarannum, S., Haque Charu, K., Kabir Mehedi, 21 M.H., Alim Rasel, A.: Comparative analysis of federated learning and central- ized approach for detecting different lung diseases. In: Proceedings of the 2023 9th International Conference on Computer Te...

    2023

  20. [27]

    IEEE Journal of Biomedical and Health Informatics (2023)

    Ullah, F., Srivastava, G., Xiao, H., Ullah, S., Lin, J.C.-W., Zhao, Y.: A scalable federated learning approach for collaborative smart healthcare systems with inter- mittent clients using medical imaging. IEEE Journal of Biomedical and Health Informatics (2023)

    2023

  21. [28]

    Sensors 23(2), 743 (2023)

    Malik, H., Naeem, A., Naqvi, R.A., Loh, W.-K.: Dmfl net: A federated learning- based framework for the classification of covid-19 from multiple chest diseases using x-rays. Sensors 23(2), 743 (2023)

    2023

  22. [29]

    Diagnostics 13(9), 1532 (2023)

    Nazir, S., Kaleem, M.: Federated learning for medical image analysis with deep neural networks. Diagnostics 13(9), 1532 (2023)

    2023

  23. [31]

    In: 2024 International Con- ference on Artificial Intelligence in Information and Communication (ICAIIC), pp

    Khan, S., Palanisamy, L.S., Raghuraman, M.: Federated learning-a novel approach for predicting diseases in unprecented areas. In: 2024 International Con- ference on Artificial Intelligence in Information and Communication (ICAIIC), pp. 058–063 (2024). IEEE

    2024

  24. [32]

    In: 2024 International Wireless Communications and Mobile Computing (IWCMC), pp

    Babar, F.F., Jamil, F., Alsboui, T., Babar, F.F., Ahmad, S., Alkanhel, R.I.: Fed- erated active learning with transfer learning: Empowering edge intelligence for enhanced lung cancer diagnosis. In: 2024 International Wireless Communications and Mobile Computing (IWCMC), pp. 13...

    2024

  25. [33]

    Pattern Recognition, 110424 (2024)

    Guan, H., Yap, P.-T., Bozoki, A., Liu, M.: Federated learning for medical image analysis: A survey. Pattern Recognition, 110424 (2024)

    2024

  26. [34]

    International Journal of Interactive Mobile Technologies 17(1) (2023)

    Tabassum, N., Ahmed, M., Shorna, N.J., Sowad, U.R., Mejbah, M., Haque, H.: Depression detection through smartphone sensing: A federated learning approach. International Journal of Interactive Mobile Technologies 17(1) (2023)

    2023

  27. [35]

    https://arxiv.org/abs/2410.13709

    Ahmed, M., Muntakim, A., Tabassum, N., Rahim, M.A., Shah, F.M.: On-device Federated Learning in Smartphones for Detecting Depression from Reddit Posts (2025). https://arxiv.org/abs/2410.13709

    2025 arXiv

  28. [36]

    Alexandria Engineering Journal 122, 344–354 (2025) 22

    Alruwais, N., Elhessewi, G.M.S., Saeed, M.K., Alshammeri, M., Alrusaini, O., Alkharashi, A., Al Zanin, S., Said, Y.: Federated learning and gwo-enabled consumer-centric healthcare internet of things for pancreatic tumour. Alexandria Engineering Journal 122, 344–354 (2025) 22

    2025

  29. [37]

    Alexandria Engineering Journal 126, 131–142 (2025)

    Kumari, A., Patadia, D., Tanwar, S., Pau, G., Alqahtani, F., Tolba, A.: Cnn- based cancer prediction scheme using 5g-assisted federated learning for healthcare industry 5.0. Alexandria Engineering Journal 126, 131–142 (2025)

    2025

  30. [38]

    Alexandria Engineering Journal 108, 498–508 (2024)

    Raju, V.N., Saravanakumar, R., Yusuf, N., Pradhan, R., Hamdi, H., Saravanan, K.A., Rao, V.S., Askar, M.A.: Enhancing emotion prediction using deep learning and distributed federated systems with smote oversampling technique. Alexandria Engineering Journal 108, 498–508 (2024)

    2024

  31. [39]

    Alexandria Engineering Journal 114, 1–11 (2025)

    Onaizah, A.N., Xia, Y., Hussain, K.: Fl-sicnn: An improved brain tumor diagnosis using siamese convolutional neural network in a peer-to-peer federated learning approach. Alexandria Engineering Journal 114, 1–11 (2025)

    2025

  32. [40]

    Alexandria Engineering Journal 105, 105–119 (2024)

    Alalwan, N., Alwadain, A., Alzahrani, A.I., Al-Bayatti, A.H., Abozeid, A., Abd El-Aziz, R.M.: Advancements in brain tumor identification: Integrating synthetic gans with federated-cnns in medical imaging analysis. Alexandria Engineering Journal 105, 105–119 (2024)

    2024

  33. [41]

    Alexandria Engineering Journal 120, 215–228 (2025)

    Alanazi, S., Alanazi, R.: Enhancing diabetic retinopathy detection through fed- erated convolutional neural networks: Exploring different stages of progression. Alexandria Engineering Journal 120, 215–228 (2025)

    2025

  34. [42]

    Alexandria Engineering Journal 125, 441–448 (2025)

    Rathee, G., Garg, S., Kaddoum, G., Alzanin, S.M., Hassan, M.M.: An improved and decentralized/distributed healthcare framework for disabled people through ai models. Alexandria Engineering Journal 125, 441–448 (2025)

    2025

  35. [43]

    IEEE Transactions on Emerging Topics in Computational Intelligence (2024)

    Naz, S., Phan, K., Chen, Y.-P.P.: Centralized and federated learning for covid- 19 detection with chest x-ray images: Implementations and analysis. IEEE Transactions on Emerging Topics in Computational Intelligence (2024)

    2024

  36. [44]

    In: 2024 5th International Conference on Electronics and Sustainable Communication Systems (ICESC), pp

    Muthalakshmi, M., Jeyapal, K., Vinoth, M., Dinesh, P., Murugan, N.S., Sheela, K.S.: Federated learning for secure and privacy-preserving medical image analysis in decentralized healthcare systems. In: 2024 5th International Conference on Electronics and Sustainable Communicati...

    2024

  37. [45]

    arxiv 2024

    Adhikari, R., Settles, C.: Secure federated learning approaches to diagnosing covid-19. arxiv 2024. arXiv preprint arXiv:2401.12438

    2024 arXiv

  38. [46]

    In: 2024 36th Conference of Open Innovations Association (FRUCT), pp

    Ram, S., Kiran, Y.N., Bhute, A., Khare, T.: Federated learning for accurate label- ing of chest x-ray scans. In: 2024 36th Conference of Open Innovations Association (FRUCT), pp. 649–654 (2024). IEEE

    2024

  39. [47]

    arXiv preprint arXiv:2010.16056 (2020)

    Chen, Z., Song, Y., Chang, T.-H., Wan, X.: Generating radiology reports via memory-driven transformer. arXiv preprint arXiv:2010.16056 (2020)

    2020 arXiv

  40. [48]

    arXiv preprint arXiv:2010.11929 (2020)

    Dosovitskiy, A.: An image is worth 16x16 words: Transformers for image 23 recognition at scale. arXiv preprint arXiv:2010.11929 (2020)

    2020 arXiv

  41. [49]

    OpenAI blog 1(8), 9 (2019)

    Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., Sutskever, I., et al.: Language models are unsupervised multitask learners. OpenAI blog 1(8), 9 (2019)

    2019

  42. [50]

    Advances in neural information processing systems 30 (2017)

    Blanchard, P., El Mhamdi, E.M., Guerraoui, R., Stainer, J.: Machine learn- ing with adversaries: Byzantine tolerant gradient descent. Advances in neural information processing systems 30 (2017)

    2017

  43. [51]

    https://arxiv.org/abs/2501.12356

    Islam, M.R., Hossain, M.Z., Ahmed, M., Samu, M.S.S.: Vision-Language Models for Automated Chest X-ray Interpretation: Leveraging ViT and GPT-2 (2025). https://arxiv.org/abs/2501.12356

    2025 arXiv

  44. [52]

    arXiv preprint arXiv:2406.04449 (2024)

    Bannur, S., Bouzid, K., Castro, D.C., Schwaighofer, A., Thieme, A., Bond-Taylor, S., Ilse, M., P´ erez-Garc ´ ıa, F., Salvatelli, V., Sharma, H., et al.: Maira-2: Grounded radiology report generation. arXiv preprint arXiv:2406.04449 (2024)

    2024 arXiv

  45. [53]

    Informatics in Medicine Unlocked 50, 101585 (2024)

    Nicolson, A., Dowling, J., Anderson, D., Koopman, B.: Longitudinal data and a semantic similarity reward for chest x-ray report generation. Informatics in Medicine Unlocked 50, 101585 (2024)

    2024

  46. [54]

    arXiv preprint arXiv:2408.03500 (2024)

    Nicolson, A., Liu, J., Dowling, J., Nguyen, A., Koopman, B.: e-health csiro at rrg24: entropy-augmented self-critical sequence training for radiology report generation. arXiv preprint arXiv:2408.03500 (2024)

    2024 arXiv

  47. [55]

    Heliyon 10(7) (2024)

    Magalh˜ aes, G.V., Santos, R.L.d.S., Vogado, L.H., Paiva, A.C., Santos Neto, P.d.A.: Xrayswingen: Automatic medical reporting for x-ray exams with multi- modal model. Heliyon 10(7) (2024)

    2024

  48. [56]

    In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

    Jin, H., Che, H., Lin, Y., Chen, H.: Promptmrg: Diagnosis-driven prompts for medical report generation. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, pp. 2607–2615 (2024)

    2024

  49. [57]

    In: Machine Learning for Healthcare Conference, pp

    Liu, G., Hsu, T.-M.H., McDermott, M., Boag, W., Weng, W.-H., Szolovits, P., Ghassemi, M.: Clinically accurate chest x-ray report generation. In: Machine Learning for Healthcare Conference, pp. 249–269 (2019). PMLR 24

    2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.