| Journal of Cybersecurity and Intelligent Systems
Received: 25 July 2026; Revised: 04 September 2026; Accepted: 17 September 2026; Published Online: 22 September 2026
J. Cybersecur. Intell. Syst., 2026, 1(1), 26504 | Volume 1 Issue 1 (September 2026) | DOI: https://doi.org/10.64189/cis.26504
© The Author(s) 2026
This article is licensed under Creative Commons Attribution NonCommercial 4.0 International (CC-BY-NC 4.0)
Decentralized Data Trust System for Privacy-
Preserving Federated Learning
Mohd Salique Khan,
1, *
Samiya Khan,
2
Sania Shaikh,
2
and Luminita Moraru
3
1
Department of Computer Engineering, M.H. Saboo Siddik College of Engineering, Mumbai, Maharashtra, 400008, India
2
Department of Computer Science Engineering (Data Science), Kalsekar Technical Campus, Navi Mumbai, Maharashtra, 410206, India
3
Faculty of Sciences and Environment, the Modelling & Simulation Laboratory, Dunarea de Jos University of Galati, 47 Domneasca
Str., Galati, 800008 Romania
*Email : msaliquekhan@gmail.com (Mohd Salique Khan)
Abstract
The increasing use of Artificial Intelligence (AI) in sensitive and regulated domains has raised concerns
regarding data privacy, trust, governance, accountability, and fair value distribution. Although Federated
Learning (FL) enables collaborative model training without transferring raw data, conventional FL does not
inherently provide decentralized auditability, trust management, or contribution-aware incentives. In this
paper, a decentralized data trust system for privacy-preserving federated learning based on a four-layer
architecture integrating federated training, privacy protection, blockchain-based governance, and contribution-
aware incentives is proposed. The framework uses secure aggregation and differential privacy for privacy
protection, blockchain and smart contracts for verifiable governance and auditability, and an approximate
Shapley value mechanism for contribution-based reward allocation. The proposed system connects participant
activity, trust evaluation, contribution assessment, and reward distribution within a unified workflow. A
comparative analysis with conventional FL, blockchain-enabled FL, and data marketplace approaches
demonstrates the broader functional coverage of the proposed architecture across privacy, auditability, trust,
incentives, and regulatory traceability. As this work is conceptual and architectural, quantitative prototype
validation remains part of future work to evaluate model utility, communication overhead, scalability, and
incentive-computation cost.
Keywords: Federated learning; Blockchain; Data privacy; Trust management; Incentive mechanism; Shapley value;
Healthcare.
1. Introduction
-
model collaboration, which are techniques that collectively allow multiple parties to jointly train models without
exposing
ecosystems:
• breaches.
• usage.
•
•
ecosystems.
2. Literature survey
2.1 Literature review
research
observed:
•
•
•
• conditions.
• contribution.
• updates.
1.
Problem
Cause
/ (Reported)
Data Breaches
Centralized AI Databases
Privacy Misuse
tracking
use
Financial Loss
Large-scale breaches
(avg)
Loss
Black-box AI decisions
systems
Systemic Failure
Single control authority
users
challenges
ecosystems
•
•
• systems
•
•
privacy.
•
• a
2.
coordination.
limitations
•
• (FL)
• firms
• initiatives
• mechanisms
• learning the
updates.
complexity.
• models
solutions:
• landscape
•
•
• risks
• compliance
the system
Evaluation
Parameter
Federated
Learning
FL
Data Marketplaces
System
Exposure
None
None
Depends on policy
None
Auditability
Low
High
Medium
High
Contribution-Based
Reward
Supported
Limited
Platform-
Dependent
Automated
Based)
PrivacyLeakage
Resistance
Medium
Medium
Medium
Privacy
Aggregation)
Regulatory
Traceability
Low
Medium
Low
High
Scoring
Available
Limited
Reputation-Based
Model
Coordination
Overhead
Low
High
Medium
Medium
Long-Term
Participation
Incentive
Low
Low
Medium
High
(FL)
• system
• (FL)
• enforcement:
•
3.
Fig. 3: System architecture of the proposed decentralized data trust system.
input:
Output:
M0.
do
3: 4:
5: 6:
7: 8:
9:
10: for
11:
12:
13: do
14:
15:
16:
for
blockchain.
while
processing.
framework.
modeling
aggregation
where,
participants
datasets.
score
the basis of
accuracy.
to
= Noise parameter
Increasing the value of improves privacy protection but may slightly reduce model accuracy.
4.2.5 Trust score computation
Each participant is assigned a trust score on the basis of its contribution, security behavior, and historical
reputation.
subject to
where,
= Trust score of participant
= Contribution score
= Security score
= Historical reputation score
, , and = Weighting coefficients
Participants with higher trust scores are given greater preference during future rounds of collaboration.
4.2.6 Overall objective function
The objective of the proposed framework is to maximize model performance and participant trust while
minimizing privacy risk.
subject to
where,
= Global model accuracy
= Overall trust score
= Privacy risk
1
,
2
,
3
= Weight coefficients balancing accuracy, trust, and privacy
The optimization objective is to achieve an effective balance between learning performance, trustworthiness,
and privacy preservation in decentralized federated learning environments.
4.3 Computational complexity and scalability analysis
Compared with conventional federated learning, the proposed framework introduces additional computational
and communication operations. The overall computational cost is associated mainly with local model training,
privacy-preserving processing, secure aggregation, blockchain transaction handling, trust score computation,
and contribution valuation. Local model training remains the dominant computation at participating
institutions and depends primarily on the model architecture, dataset size, and number of local training epochs.
The blockchain layer introduces additional overhead for recording participant-related metadata, update hashes,
timestamps, and reward transactions. However, sensitive model updates and raw training data are not stored
directly on the blockchain. This design limits blockchain storage requirements to verification and governance-
related information rather than complete model or dataset contents.
Contribution valuation represents another important scalability consideration. Exact Shapley-value
computation can require evaluating a large number of participant coalitions and therefore becomes
computationally expensive as the number of participants increases. The proposed framework consequently
uses an approximate Shapley value mechanism to reduce this computational burden while retaining
contribution-aware reward allocation.
The communication overhead is also greater than that of conventional federated learning because the proposed
workflow requires additional interactions associated with blockchain verification, privacy-preserving
operations, and incentive management. Nevertheless, the decentralized architecture distributes computation
across participating institutions rather than relying on a single centralized coordinator. The framework is
therefore designed to support scalability through distributed processing, lightweight on-chain metadata, and
approximate contribution evaluation.
In practical deployment, scalability should be evaluated by varying the number of participating clients,
federation rounds, model update sizes, and levels of data heterogeneity. Important measurements include
training time per round, communication volume, blockchain transaction latency, blockchain storage growth,
contribution evaluation time, and total system execution time. These measurements provide quantitative
evidence for determining the scalability of the proposed framework under realistic federated learning
workloads.
5. Case study: Multi-institutional brain tumor segmentation using federated learning
The corresponding workflow is depicted in Fig. 4.
Fig. 4: Case study flow-multi-institution brain tumor segmentation.
Case study: Multi-institutional brain tumor segmentation (medical imaging)
Sheller et al. reported a feasibility study in which federated learning was applied for multi-institutional brain
tumor segmentation, enabling deep learning collaboration without the need to share patient data.
[28]
Their setup
-preserving training can work while
coordination often remains centralized in practice. Follow-up analysis in related federated tumor segmentation
research indicates that differences in data distributions across sites can significantly impact federated model
performance, illustrating a key limitation for cross-silo real deployments.
Sheller et al. reported that federated models trained across four institutions achieved segmentation
performance within approximately 14% of a centrally trained baseline, despite the added coordination
overhead, demonstrating the feasibility of federated learning while preserving data locality.
[19,28]
In the proposed decentralized data trust framework, each hospital can define participation terms, including
training permissions, model purpose, and privacy constraints, while federated learning enables local training
with only model updates and relevant metrics being exchanged. Secure aggregation and differential privacy can
be applied to reduce leakage from shared updates, whereas blockchain-based coordination can provide
verifiable records of contributions and training-round participation. Incentives can be linked to measurable
participant contributions using Shapley value-based valuation. The case study also highlights the practical need
to address non-IID data and mitigate adversarial or low-quality contributions in decentralized healthcare
environments.
6. Results and discussion
6.1 Market relevance and adoption of trustworthy AI
As artificial intelligence adoption accelerates across industries, trustworthy AI has emerged as a critical market
requirement rather than an optional feature. Organizations increasingly demand AI systems that are
transparent, reliable, and accountable, particularly in regulated and safety-critical domains. While global
investment in AI continues to grow rapidly, a lack of trust remains a major barrier to large-scale deployment.
Studies indicate that poorly calibrated trust in AI systems can lead to financial losses, legal liabilities, safety
failures, and reduced user acceptance. Consequently, trustworthiness is becoming a key differentiator for
enterprise and public-sector AI solutions.
6.2 Applications of trustworthy AI across domains
• Increased financial and operational losses
• Higher regulatory and legal exposure
• Safety risks in high-stakes environments
• Slower adoption of AI-driven systems
These risks highlight the growing market demand for AI systems that embed trust and accountability by design.
6.3 Applications of trustworthy AI across domains
• Healthcare: AI systems support diagnosis, medical imaging, and treatment planning. Over-trusting AI can
result in incorrect clinical decisions, while under trusting reduces efficiency. Trustworthy AI ensures that
AI functions as a decision-support system, maintaining clinician oversight.
• Autonomous vehicles: Self-driving systems require precise human-AI trust calibration. Misaligned trust
has already contributed to real-world accidents. Trust-aware AI mechanisms improve safety and public
acceptance.
• Banking and finance: AI is widely used for credit scoring, fraud detection, and risk analysis. Trustworthy
AI helps professionals understand when automated decisions require manual review, reducing financial
and compliance risks.
• Public services and governance: AI is being increasingly applied in welfare distribution, policing, and
public decision-making. Trustworthy AI is essential for ensuring fairness, accountability, transparency, and
ethical use.
6.4 Comparative evaluation and discussion
The comparative evaluation presented in Fig. 5 examines the proposed system against plain federated learning,
blockchain-logged federated learning, and data marketplaces across auditability, contribution-based reward,
privacy leakage resistance, regulatory traceability, trust scoring, coordination overhead, and long-term
participation incentives. The comparison is based on the functional capabilities defined in the proposed
architecture and existing solution landscape rather than on experimentally measured benchmark values. Plain-
federated learning provides the advantage of keeping raw data local and maintaining relatively low coordination
overhead; however, it does not inherently provide immutable auditability, contribution-based rewards, or
explicit trust scoring. Blockchain-logged FL improves auditability by recording training-related events but
introduces additional coordination requirements and does not by itself establish a complete mechanism for
contribution-based reward allocation. Data marketplace approaches support data valuation and economic
exchange, but governance, privacy enforcement, and regulatory traceability may remain dependent on the
underlying platform.
Fig. 5: Comparative analysis of the proposed system and existing approaches.
The proposed system provides broader functional coverage by integrating privacy preservation, decentralized
governance, trust management, contribution valuation, and incentive distribution within a common workflow.
Differential privacy and secure aggregation are used to strengthen protection against privacy leakage, while the
blockchain layer provides verifiable records of participation and model-update metadata. The approximate
Shapley value mechanism further connects participant contribution with reward allocation, providing a basis
for transparent and contribution-aware incentives. These characteristics correspond to the higher capability
levels shown for the proposed system in the comparative evaluation.
The comparison also highlights important trade-offs. Compared with conventional federated learning, the
additional blockchain, privacy, trust, and contribution-valuation mechanisms can increase computational and
communication requirements. In particular, Shapley value computation may become costly as the number of
participants increases, which motivates the use of an approximate valuation approach. Similarly, privacy-
enhancing mechanisms may introduce a trade-off between privacy protection and model utility. Therefore, the
proposed framework should be interpreted as a system-level improvement in governance, accountability,
privacy, and fairness rather than as a claim of superior numerical model accuracy.
Overall, the comparative analysis supports the central objective of the proposed framework: integrating
functions that are commonly treated separately in existing approaches. The framework connects participant
activity, privacy protection, trust evaluation, contribution assessment, auditability, and reward allocation within
a unified architecture. However, because the present study is conceptual and architectural, quantitative
validation on standard federated learning datasets is needed to measure model accuracy, communication
overhead, scalability, and incentive-computation cost in future work.
6.5 Future outlook for decentralized data trust systems and federated data
• Learning enterprise-scale adoption: Healthcare, banking, insurance, autonomous systems, and smart
cities will increasingly adopt decentralized data trust systems to enable secure, privacy-preserving AI
collaboration across organizations.
• Blockchain-driven AI governance: Blockchain will evolve from passive logging to active governance,
enforcing data usage policies, auditability, contribution tracking, and regulatory compliance through smart
contracts.
• Privacy-enhancing AI techniques: Advanced federated learning, secure aggregation, and differential
privacy mechanisms will mature to further reduce data leakage while maintaining high model performance.
• Fair data economies: Data contribution valuation models will enable transparent and incentive-driven data
sharing, encourage long-term participation and reduce data monopolization.
6.6 Prototype validation and experimental evaluation plan
The proposed framework is currently presented as a conceptual and architectural system; therefore, the present
study does not claim experimental validation or measured performance improvements. To validate the
proposed design in future work, a prototype implementation will be developed using standard federated
learning datasets and multiple simulated participant nodes. The planned prototype will implement the four
proposed layers, including local federated training, privacy protection, blockchain-based governance, and
contribution-aware incentive computation.
The experimental evaluation will measure the model utility, privacyutility trade-offs, communication
overhead, blockchain transaction and storage overhead, computational cost of contribution evaluation, trust
score behavior, and reward distribution. Standard federated learning performance measures such as accuracy,
precision, recall, F1 score, and convergence behavior are used where appropriate for the selected application
dataset.
The proposed framework is compared with conventional federated learning, blockchain-enabled federated
learning, and other relevant decentralized approaches under comparable experimental conditions. Different
numbers of participants and data distributions are considered to examine scalability and the effect of non-IID
data. The resulting measurements provide empirical evidence for determining whether the proposed
integration of privacy, trust, governance, and incentives provides practical benefits while maintaining
acceptable computational and communication costs.
6.7 Quantitative evaluation methodology
The quantitative evaluation will be conducted using standard federated learning datasets and multiple
simulated clients under comparable experimental conditions. The proposed framework is evaluated against
conventional federated learning and blockchain-enabled federated learning baselines. Model performance will
be assessed using accuracy, precision, recall, F1 score, and convergence behavior, while system-level
performance will be evaluated using communication overhead, computational cost, blockchain transaction and
storage overhead, and contribution evaluation time.
To examine scalability, experiments will be conducted with different numbers of participating clients and
varying levels of non-IID data distribution. The effects of privacy-preserving mechanisms will also be evaluated
by analyzing the trade-off between model utility and privacy protection. These measurements provide a
quantitative basis for assessing the practical benefits and overheads introduced by the proposed decentralized
trust, privacy, and incentive mechanisms.
7. Conclusion
This research set out to address a gap that current federated learning systems leave open: privacy alone is not
enough without the trust, accountability, and fair incentives built into the system by design. Federated learning
solves the data-sharing problem, but on its own, it does not provide mechanisms to verify training activities,
maintain accountability, or reward participants according to their contribution. The proposed decentralized
data trust system addresses this gap by combining three key components that existing approaches often
separate. Federated learning keeps raw data with its owner. A blockchain layer provides an immutable record
of training activities, supporting transparency and auditability. A Shapley value-based incentive mechanism
links participant rewards to their estimated contribution to model performance, supporting transparent and
contribution-aware reward allocation. The literature review and comparative analysis in this paper show that
existing solutions generally address privacy, blockchain-based auditability, data valuation, or incentives as
separate components. The proposed system is designed to integrate these functions with trust management
and regulatory traceability within a unified federated learning workflow. This integration provides a system-
level approach to privacy, accountability, fairness, and governance rather than treating these requirements
independently. Notably, this work represents a conceptual and architectural contribution rather than an
empirically benchmarked system. Future work should therefore focus on prototype implementation and
evaluation using standard federated datasets, particularly in domains such as healthcare and finance. Such
evaluation determines whether the proposed integration of privacy, trust, fairness, and regulatory traceability
provides practical benefits while maintaining acceptable computational and communication overhead.
Acknowledgment
The authors would like to thank M.H. Saboo Siddik College of Engineering and Kalsekar Technical Campus for
providing the necessary academic support and facilities to carry out this research.
CRediT Author Contribution Statement
Mohd Salique Khan: Conceptualization, Methodology, Supervision, Project Administration, Writing Original
draft. Samiya Khan: Investigation, Formal Analysis, Methodology, Validation, Writing Review & editing. Sania
Shaikh: Investigation, Formal Analysis, Visualization, Validation, Writing Review & editing. Luminita Moraru:
Methodology, Formal Analysis, Writing Review & editing. All authors have read and agreed to the published
version of the manuscript.
Funding Declaration
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-
profit sectors.
Data Availability Statement
No data were generated or analysed during the current study. Therefore, data sharing is not applicable to this
article.
Conflict of Interest
The authors declare that there are no conflicts of interest.
Artificial Intelligence (AI) Use Disclosure
The authors confirm that no artificial intelligence (AI)-assisted technologies were used in the writing of the
manuscript, and no images were generated or manipulated using AI. AI-based tools were used solely for
language editing to improve grammar, clarity, and readability, in accordance with journal policy. The authors
take full responsibility for the accuracy, originality, and integrity of the work.
Supporting Information
Not applicable.
References:
[1]
B. McMahan, E. Moore, D. Ramage, S. Hampson, B. A. Arcas, Communication-Efficient Learning of Deep
Networks from Decentralized Data, Proceedings of the 20th International Conference on Artificial
Intelligence and Statistics, PMLR, 2017, 54, 1273-1282.
[2]
K. Wei, J. Li, M. Ding, C. Ma, H. H. Yang, F. Farokhi, S. Jin, T. QS Quek, H. V. Poor, Federated learning with
differential privacy: Algorithms and performance analysis, IEEE transactions on information forensics
and security, 2020, 15, 3454-3469, doi: 10.1109/TIFS.2020.2988575.
[3]
L. S. Shapley, A Value for n-Person Games, in Contributions to the Theory of Games, Princeton
University Press, 1953, 307317.
[4]
Europen. Union, Artificial intelligence act. European Parliament: European Parliamentary Research
Service, 2024, 1689, https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng, Accessed 29 June 2026.
[5]
European Parliament and Council of the European Union, Regulation (EU) 2023/2854 (Data Act),
Official Journal of the European Union, 2023, https://eur-lex.europa.eu/eli/reg/2023/2854/oj,
Accessed 29 June 2026.
[6]
J. Kelly, S. A. Zafar, L. Heidemann, J. -V. Zacchi, D. Espinoza, N. Mata, Navigating the EU AI Act: A
Methodological Approach to Compliance for Safety-critical Products, IEEE Conference on Artificial
Intelligence (CAI), 2024, 979-984, doi: 10.1109/CAI59869.2024.00179.
[7]
S. K. M. Sameera, S. Nicolazzo, M. Arazzi, A. Nocera, R. R. K. A., V. P., M. Conti, Privacy-Preserving in
Blockchain-Based Federated Learning Systems, Computer Communications, 2024, 222, 3867, doi:
10.1016/j.comcom.2024.04.024.
[8]
K. Li, A Blockchain-Integrated Federated Learning Approach for Secure Data Sharing and Privacy
Protection in Multi-Device Communication, Applied Artificial Intelligence, 2025, 39, doi:
10.1080/08839514.2024.2442770.
[9]
X. H. Yang, T. C. Li, A Blockchain-Based Federated Learning Framework for Secure Aggregation and
Fair Incentives, Connection Science, 2024, 36, 2316018, doi: 10.1080/09540091.2024.2316018.
[10]
Y. Li, C. Chen, N. Liu, H. Huang, Z. Zheng, Q. Yan, A Blockchain-Based Decentralized Federated Learning
Framework with Committee Consensus, IEEE Network, 2021, 35, 234-241, doi:
10.1109/MNET.011.2000263.
[11]
A. Yazdinejad, A. Dehghantanha, H. Karimipour, G. Srivastava, R. M. Parizi, A Robust Privacy-
Preserving Federated Learning Model Against Model Poisoning Attacks, IEEE Transactions on
Information Forensics and Security, 2024, 19, 6693-6708, doi: 10.1109/TIFS.2024.3420126.
[12]
J. Liu, C. Chen, Y. Li, L. Sun, Y. Song, J. Zhou, B. Jing, D. Dou, Enhancing Trust and Privacy in Distributed
Networks: A Comprehensive Survey on Blockchain-Based Federated Learning, Knowledge and
Information Systems, 2024, 66, 43774403, doi: 10.1007/s10115-024-02117-3.
[13]
Y. Tang, Y. Zhang, T. Niu, Z. Li, Z. Zhang, H. Chen, L. Zhang, A Survey on Blockchain-Based Federated
Learning: Categorization, Application and Analysis, Computer Modeling in Engineering & Sciences
(CMES), 2024, 139, 2451, doi: 10.32604/cmes.2024.030084.
[14]
Y. Lu, X. Huang, K. Zhang, S. Maharjan, Y. Zhang, Blockchain and Federated Learning for 5G Beyond,
IEEE Network, 2021, 35, 1, 219-225, doi: 10.1109/MNET.011.1900598.
[15]
P. Guo, Y. Yang, W. Guo, Y. Shen, A Fair Contribution Measurement Method for Federated Learning,
Sensors, 2024, 24, 4967, doi: 10.3390/s24154967.
[16]
H. Zhu, Z. Li, D. Zhong, C. Li, Y. Yuan, Shapley-value-based Contribution Evaluation in Federated
Learning: A Survey, 2023 IEEE 3rd International Conference on Digital Twins and Parallel Intelligence
(DTPI), 2023, 1-5, doi: 10.1109/DTPI59677.2023.10365410.
[17]
R. H. L. Sim, X. Xu, B. K. H. Low, Data Valuation in Machine Learning: Ingredients, Strategies, and Open
Challenges, in Proc. IJCAI, 2022, 56075614.
[18]
N. Tastan, S. Fares, T. Aremu, S. Horvath, K. Nandakumar, Redefining contributions: Shapley-driven
federated learning, arXiv preprint, 2024, doi: 10.48550/arXiv.2406.00569.
[19]
S. Zheng, Y. Cao, M. Yoshikawa, Secure Shapley Value for Cross-Silo Federated Learning, Proc. VLDB
Endowment, 2023, 16, 16571670, doi: 10.14778/3587136.3587141.
[20]
J. T. Wang, R. Jia, Data Banzhaf: A Robust Data Valuation Framework for Machine Learning,
Proceedings of the 26th International Conference on Artificial Intelligence and Statistics, 2023, 6388
6421.
[21]
Z. Liu, Y. Chen, H. Yu, Y. Liu, L. Cui, GTG-Shapley: Efficient and Accurate Participant Contribution
Evaluation in Federated Learning, ACM Transactions on intelligent Systems and Technology, 2022, 13,
21, doi: 10.1145/3501811.
[22]
Q. Sun, X. Li, J. Zhang, L. Xiong, W. Liu, J. Liu, Z. Qin, K. Ren, ShapleyFL: Robust Federated Learning
Based on Shapley Value, In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery
and Data Mining (KDD '23), Association for Computing Machinery, 2023, 20962108, doi:
10.1145/3580305.3599500.
[23]
Q. P. Nguyen, B. K. H. Low, P. Jaillet, Trade-off Between Payoff and Model Rewards in Shapley-Fair
Collaborative Machine Learning, in Advances in Neural Information Processing Systems (NeurIPS),
2022, 35, 3054230553, doi: 10.52202/068431-2214.
[24]
K. Chen, J. Zhang, X. Liu, Z. Feng, X. Yang,
learning based on blockchain, Blockchain: Research and Applications, 2025, 6, 100271, doi:
10.1016/j.bcra.2024.100271.
[25]
F. Zhao, B. Yang, Z. Su, C. Li, Y. Ding, A Blockchain-Enabled Privacy-Preserving and Incentive
Mechanism-Driven Federated Learning Scheme for IoV, 2025, Computer Networks, 264, 111262, doi:
10.1016/j.comnet.2025.111262.
[26]
N. Liu, Y. Zhou, Y. Qin, Q. Feng, H. Qi, COZO: A Secure and Efficient Blockchain-Enhanced Federated
Learning Paradigm with Optimized Storage and Equitable Contribution Valuation, Future Generation
Computer Systems, 2026, 179, 108362, doi: 10.1016/j.future.2025.108362.
[27]
C. Dwork, Differential Privacy, in Proc. ICALP, Springer, 2006, doi: 10.1007/11787006_1.
[28]
M. J. Sheller, B. Edwards, G. A. Reina, J. Martin, S. Pati, A. Kotrotsou, M. Milchenko, W. Xu, D. Marcus, R.
R. Colen, S. Bakas, Federated Learning in Medicine: Facilitating Multi-Institutional Collaborations
Without Sharing Patient Data, Scientific Reports, 2020, 10, 12598, doi: 10.1038/s41598-020-69250-
1.
Publisher Note: The views, statements, and data in all publications solely belong to the authors and
contributors. GR Scholastic is not responsible for any injury resulting from the ideas, methods, or products
mentioned. GR Scholastic remains neutral with respect to jurisdictional claims in published maps and
institutional affiliations.
Open Access
This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which
permits noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as
long as appropriate credit to the original author(s) and the source is given by providing a link to the Creative
Commons License, and changes need to be indicated if there are any. The images or other third-party material
in this article are included in the article's Creative Commons License, unless indicated otherwise in a credit line
to the material. If material is not included in the article's Creative Commons License and your intended use is
not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this License, visit: https://creativecommons.org/licenses/by-
nc/4.0/
© The Author(s) 2026
