AI Networking Cookbook: Practical recipes for AI-assisted network automation and development
TL;DR
A integrated focus on privacy-preserving dyslexia detection using QEEG features is integrated, and explainability in the encrypted domain is achieved via a perturbation-based SHapley Additive exPlanations (SHAP)-like approximation.
Internet of Things (IoT) has revolutionized mobile healthcare applications, increases diagnosis speed andaccuracy. Disease prediction systems (DPS) improve healthcare quality, but they raise privacy concerns due to sensitivedata. These concerns include illegal sharing, misuse, and exposure of sensitive information. However, developing newtechniques does not provide improved protection from attackers and fraudsters. This paper suggests an effectiveprivacy-preserving strategy for patient healthcare data from IoT devices in order to anticipate diseases in thecontemporary medical field. Heart failure prediction health data is utilized as an input in this proposed approach.Initially, elastic net (EN) is used to reduce the dimensionality of the input raw dataset. Hyper parameter in EN isoptimally selected using the Zebra Optimization Algorithm (ZOA). After dimensionality reduction, data is encryptedusing the ElGamal technique. During the encryption procedure, the True Random Number Generator-Pseudo RandomNumber Generator (TRNG-PRNG) encryption method is used to generate the secret key. These encrypted data issecurely stored in cloud. Finally, Attention Mechanism based Bi-directional Gated Recurrent Unit (AT-BiGRU)technique is employed to predict heart disease. The ElGamal-TRPRNG method strengths privacy and security byachieving encryption and decryption times of 0.30 sec and 0.12 sec, respectively. The suggested model is evaluated andcontrasted with current methods using encryption data performance measures. Model achieved 93.47% accuracy, 6.53%error, and 93.45% precision in encrypted data performance metrics. Therefore, this suggested method is the mosteffective way to effectively safeguard health care records.
FHEON is presented, an open-source configurable framework for developing privacy-preserving neural network models for inference using the CKKS scheme of HE, and outperform all state-of-the-art HE inference works in both latency and memory utilization.
This work proposes a practical non-interactive encrypted retrieval framework for RAG based on threshold selection, and introduces a precision-stable mask polarization method that ensures accurate recovery of selected documents.
Big data analysis is required to predict future trends using machine learning. A lot of processing and computation is required to make this analysis. Conventional encryption methods may help to achieve privacy during the resting stage of data over the cloud. The computation of data needs to be in its original form or needs to be decrypted before any computation, which leaves the data again at risk and makes it more vulnerable. The latest homomorphic encryption algorithms provide a way to perform the computation over data in its encrypted form and help to analyze big data without compromising the privacy and security of data, even when data is in use. In this paper, we implemented the linear regression model over plain data and encrypted data to measure and analyze the fitness, MSE, and RMSE scores of models and the space complexity of plain data and encrypted data. The results show that the fitness score of the linear regression model for plain data and encrypted data is almost the same and differs by only 0.00003551, whereas the RMSE score is 386.9949011 and 364.03753641, respectively. It is observed that the overall performance of the linear regression model improved for encrypted data, and the error is also reduced while maintaining data privacy
Abstract Background The development of robust medical AI for knowledge discovery and decision support commonly necessitates large-scale datasets from multiple institutions. However, such data aggregation is severely constrained by privacy regulations and the inherent risk of sensitive information leakage, making it difficult to navigate the utility-privacy trade-off. Objective We aimed to design a secure multiparty deep learning system that enables privacy-preserving modeling from distributed medical time-series data without centralizing raw information or exposing model parameters. Our goal was to achieve predictive accuracy comparable to nonsecure models while providing strong security and efficiency. Methods We developed a framework using threshold homomorphic encryption to securely train recurrent neural networks on distributed longitudinal data. To improve the efficiency, we proposed an optimized encrypted matrix multiplication scheme, a secure ciphertext refresh protocol, and used lightweight encryption parameters and low-degree approximated activation polynomials. The system was evaluated on 4 real-world intensive care unit datasets for tasks like mortality and sepsis prediction. Results The system demonstrated practical efficiency, requiring approximately 1 minute per training iteration for processing 125 local batches over 39 variables and 48 time steps, and scaling well with data size and participant number. Securely trained models achieved predictive performance that was comparable to, and in some cases superior to, nonsecure centralized models, highlighting their ability to learn generalizable patterns in different unseen data distributions. For example, on the PhysioNet Challenge 2012 dataset, our secure model achieved an area under the curve (AUC) of 0.8480, outperforming the nonsecure baseline AUC of 0.8404. Conclusions This work provides a viable and efficient solution for cross-institutional, privacy-preserving analysis of longitudinal medical data. The framework successfully bridges the utility-privacy gap, facilitating safer collaborative research and enabling robust knowledge discovery and decision support while adhering to strict data protection standards.
Experimental results indicate that TMI-VFL achieves an effective trade-off between privacy protection and model utility, providing a practical solution for secure VFL.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.