2026· Computers, Materials & Continua· 0 citations· 29 references
TL;DR
FedE, a multi-precision, multi-source, heterogeneous privacy-preserving federated learning training method based on functional encryption that enhances numerical adaptation during ciphertext computation and prevents model parameter updates from easily compromising privacy in cross-institutional federated learning.
Abstract
: With the rapid development of artificial intelligence technologies in machine learning-as-a-service (MLaaS), deep learning-based intelligent models have demonstrated high value in applications such as trend prediction and risk assessment. However, MLaaS data are typically highly sensitive and contain private information. In cross-institutional collaborative modeling scenarios, different departments and local centers often hold partial, heterogeneous data resources on MLaaS platforms. Given data security, privacy, and compliance requirements, raw data are difficult to share directly, which makes it challenging to apply centralized model training methods. This paper proposes FedE, a multi-precision, multi-source, heterogeneous privacy-preserving federated learning training method based on functional encryption. By introducing a multi-precision joint-computation mechanism, this method enhances numerical adaptation during ciphertext computation. Besides, combined with differential privacy techniques, it prevents model parameter updates from easily compromising privacy in cross-institutional federated learning. Based on the federated learning training process, a secure training framework for sensitive MLaaS data is constructed. Finally, experiments demonstrate the security and efficiency of the proposed approach.
Deep learning is becoming popular in cloud applications and serves to provide intelligent services; data aggregation in a central location makes sensitive information vulnerable to privacy breaches, regulatory infractions, and adversarial manipulation. All modern privacy mechanisms offer partial protection and frequently lack accuracy, scalability, or practicality in their operations. To overcome these limitations, a federated deep learning model is formulated so that secure joint learning can occur without transferring raw data across the domains of ownership. The framework incorporates training that is decentralized, training that uses differential privacy, training that uses secure aggregation, training that uses encrypted communication, and training that uses trust-based anomaly defense to defend against leakage, poisoning, and inference attacks. It also supports heterogeneous and highly non-IID datasets using adaptive coordination and stability-relevant participation regulation and meets emerging data protection requirements. The methods of resource-conscious orchestration and the optimization of communication eliminate overhead without obstructing the effectiveness of learning. The paradigm has therefore formed a privacy-by-design intelligent cloud ecosystem which ensures confidentiality, maintains performance, enhances robustness, and ensures responsible AI implementation in privacy-related sectors of healthcare, finance, governance, and smart infrastructure.
Sribidhya Mohanty, Pallavi Gupta, Anil Pratap Singh et al.· 2026 International Conferenc...· 0 citations
The use of cloud-based web systems has increased the issue of data privacy, compliance on regulations and secure control of distributed enterprise information. The traditional centralized machine learning models mandate aggregation of data in one server which makes it more dangerous to expose sensitive information. To overcome these limitations, model combines TF-IDF-based features extractions with a deep neural network classifier trained under a federated learning system, and no longer needs raw this paper suggests a Hybrid Federated Privacy-Aware Deep Neural Network (HFPA-DNN) to learn the features of privacy-controllable data on the cloud. The data sharing among distributed cloud nodes is proposed. The model uses the Adam optimization algorithm, binary cross-entropy loss and federated averaging (FedAvg) to aggregate global models. A non-compulsory differentiation privacy system is involved in order to increase resistance to inference and reconstruction attacks. The HFPA-DNN performance is compared to the traditional machine learning baselines, which are Logistic Regression, Support Vector Machines, Random Forest, and centralized deep neural networks. As experimental findings indicate, the suggested strategy can attain a better accuracy, F1-score, and ROC-AUC, and a significant decrease in the risks of privacy exposure and preserving in scalable training performance. Through training curves, confusion matrices, ROC and precision- recall analysis, ablation studies and statistical comparison, the strength and use of the framework in ensuring data governance in the cloud in a secure setting is always verified. The results demonstrate that HFPA-DNN is an efficient trade-off to guarantee predictive performance, low computational cost, and privacy in distributed web system.
Divya sai Jaladi, Ashok Mallempati, Dr. B. Jegajothi· 2026 4th International Confe...· 0 citations
The research findings suggest that improved federated learning can achieve an optimal predictive performance, privacy protection, and secure collaborative learning, which makes it a viable method for next-generation distributed AI systems.
Sheetal Bawane, Leeladhar Chourasiya, S. Jain et al.· International journal of com...· 0 citations
An in-depth analysis of federated learning methods and paying special attention to the issue of privacy is provided, which examines new developments, concerns and tradeoffs connected with privacy, effectiveness of communication, model noise, and scalability of systems.
Aarav Mehta· International Journal of App...· 0 citations
The need to train distributed models on cloud and edge devices is of growing concern in terms of security, privacy, and efficiency. Federated learning (FL) is a potential solution to training machine learning models without causing data decentralization, but its implementation in cloud-edge systems creates issues like data heterogeneity, communication overhead, and susceptibility to security attacks. To counteract this, the paper introduces a stronger federated learning architecture that guarantees the security and privacy-friendly model training at cloud and edge layers. The frameworks are new algorithms, which take into consideration the latest encryption algorithms and differential privacy schemes, and optimization of communication protocols to provide better efficiency. This solution creates secure aggregation and a hybrid edge-cloud interaction model, and it reduces the risks of data leakage and unauthorized access to the information during the training process. The results of the experiments show that the improved federated learning framework attains a notable degree of balance concerning the model accuracy, privacy preservation, and communication cost, which is better in security and computational efficiency than the current federated learning frameworks. The present paper adds a powerful framework of safe, privacy-confidential zed, and distributed cloud-edge machine learning as a basis of future developments in the field.
Sanjay Kumar, Sapna Bawankar· International Conference Com...· 0 citations
It was concluded that strategies such as the careful selection of differential privacy parameters and training settings, along with the use of larger datasets, can improve the efficiency of FL and demonstrate that privacy-preserving and high-performance artificial intelligence systems can be securely applied in sensitive domains such as healthcare and finance.
Cagdas Karatas, Hibanur Karadogan, A. Ertug et al.· 0 citations