Aug 2026· Frontiers in Digital Health· Vol 8· 0 citations· 25 references
Medicine
TL;DR
TeleZK-FL establishes the feasibility of verifiable, trustless federated learning on commodity telehealth hardware by eliminating the computational bottlenecks of server-side proof generation while incurring only 0.1%–0.3% AUC degradation.
Abstract
Background The rapid expansion of Internet of Medical Things (IoMT) and telehealth platforms has generated vast amounts of patient data suitable for training diagnostic Artificial Intelligence (AI) models. However, strict privacy regulations (HIPAA, GDPR) and the risk of data breaches prevent the centralization of this sensitive information. While Federated Learning (FL) allows for collaborative training without sharing raw patient data, it introduces a critical “trust deficit”: central aggregators cannot verify the integrity of local model updates without inspecting the private data, leaving the system vulnerable to model poisoning and malicious actors. Methods We introduce TeleZK-FL, a privacy-preserving framework designed specifically for resource-constrained telehealth environments. Unlike existing Zero-Knowledge Proof (ZKP) systems that require high-performance computing or offload proof generation to trusted edge servers, TeleZK-FL integrates (1) Post-Training Quantization (PTQ) to compress model gradients from 32-bit floating-point to 8-bit integers, and (2) optimized Look-Up Table (LUT) arguments to generate cryptographic proofs of client-side training integrity directly on edge devices. We evaluated the framework on two clinical modalities—CheXpert (chest X-rays) and PTB-XL (12-lead ECGs)—using simulated medical edge gateways (Raspberry Pi 4) under both homogeneous and heterogeneous data distributions. Results TeleZK-FL accelerates cryptographic proof generation by approximately 25× compared to standard ZK-SNARK implementations, generating full-model proofs in approximately 84 milliseconds per client on edge hardware. It cuts the communication payload by 75% (a 4.0× reduction) while maintaining a diagnostic Area Under the Curve (AUC) of 0.877 on CheXpert and 0.891 on PTB-XL, representing a degradation of only 0.1% and 0.3% respectively compared to the unquantized baselines of 0.878 and 0.894. Conclusion TeleZK-FL establishes the feasibility of verifiable, trustless federated learning on commodity telehealth hardware. By eliminating the computational bottlenecks of server-side proof generation while incurring only 0.1%–0.3% AUC degradation, it provides an efficient, regulation-aligned method for building scalable and secure decentralized medical AI networks. We further discuss an inherent trade-off: the edge-efficient KZG-based construction is classically—not post-quantum—secure, which we position explicitly against recent lattice-based alternatives.
The widespread use of Large Language Models (LLMs) in AI-driven Internet of Medical Things (IoMT) ecosystems has introduced new challenges in areas such as data privacy, communication burden, computational efficiency and collaborative learning in a secure manner. To overcome these challenges, we introduce a novel Trustworthy Large Language Model-Assisted Zero-Trust Federated Learning Framework (TrustFedLLM-ZT) for secure and privacy-preserving healthcare intelligence. The proposed framework embeds Zero-Trust principles within federated learning (FL) and LLM-assisted intelligent model optimization, enhancing the continuous authentication, decentralized privacy protection, secure model aggregation, and trustworthy clinical decision support without compromising sensitive patient data. The framework is tested with the MIMIC-IV healthcare dataset and simulated in the PyTorch Federated Learning (PySyft) environment. Five critical performance metrics, namely, Privacy Preservation Rate, Training Time, Response Time, Communication Cost and Computation Cost, are used for comparison between experimental results and the conventional CNN-LSTM model, Blockchain-Assisted Federated Learning (BAFL), and Zero-Trust Federated Learning (ZT-FL) models. Proposed TrustFedLLM-ZT achieves PPR of 99.1% that is 14.5%, 8.2%, and 4.1% higher than CNN-LSTM, BAFL and ZT-FL, respectively. Moreover, it decreases the Training Time by 27.4%, Response Time by 31.6%, Communication Cost by 35.8% and Computation Cost by 29.7% in comparison with the current approaches. The outcomes illustrate that the TrustFedLLM-ZT approach effectively balances security, computational and communication efficiency, while offering clear benefits in the secure collaborative learning context and highlighting its potential as a robust and scalable solution for future AI-powered medical applications and privacy-aware intelligent healthcare systems.
Harika B, Vishwesh Nagamalla, Kodipaka Rajeshwar Rao et al.· 2026 International Conferenc...· 0 citations
Federated learning (FL) enables collaborative model training without sharing raw data, making it well suited to privacy-sensitive healthcare applications. However, malicious model updates and anonymous participation abuse can disrupt optimization and undermine the reliability of downstream medical AI systems. Existing defenses often address robustness or privacy separately, with limited support for accountable authentication and client revocation. To address these limitations, we propose TGSFL, a Trustworthy Group-Signature-Based Federated Learning framework for healthcare large language models (LLMs). The proposed framework integrates dynamic group-signature authentication, per-client local differential privacy, and dual-layer malicious-update detection to identify and suppress adversarial updates before aggregation. It further supports anonymous yet traceable participation and dynamic revocation of persistently malicious clients. Theoretical analysis establishes conditional traceability, dynamic revocation, and per-client local differential privacy. Experiments under multiple Byzantine attack scenarios show that TGSFL improves robustness and training stability over existing baselines while introducing only modest computational overhead, demonstrating its practicality for secure and privacy-preserving medical AI training.
Xinying Liu, Chunhua Jin, Yu-Peng Wu et al.· Journal of King Saud Univers...· 0 citations
Results indicate that a federated TinyML architecture with lightweight patient-state tracking, validation-based ensemble filtering, differential privacy, and post-quantum-secure communication can support privacy-aware and attack-resilient ICU monitoring experiments in resource-constrained IoMT settings.
Umar Hayat Khan, Rahim Khan, Tahani Alsaedi et al.· Scientific Reports· 0 citations
Quantum computing is on the horizon and will destroy existing cryptography standards, putting digital healthcare system security and patient privacy at danger. For very private and secure communication in fog healthcare settings, this article presents a new Quantum-Resistant Federated Deep Learning (QR-FDL) Framework. The QR-FDL architecture incorporates a lattice-based technique (like Kyber) into the model aggregation stage of Federated Learning (FL), which is a kind of Post-Quantum Cryptography (PQC). We provide a solution that uses Differential Privacy (DP) to prevent attacks based on gradients in data reconstruction and guarantees quantum-era security for weights of models with gradients sent between the central server and decentralized fog nodes. The practical use of QR-FDL is shown by its rapid convergence rate and excellent classification accuracy of 91.5% in an empirical assessment including a medical imaging job, such as tumor classification. Critically, we measure the cryptographic overhead and demonstrate that, even if PQC causes a regulated latency rise the total FL round time is still tolerable for real-time fog settings. In order to implement quantum-secure along with privacy-preserving Deep Learning in mission-critical healthcare communications, this study presents the first proven, end-to-end solution. In addition to showing that federated medical facilities can be securely encrypted from end to end, this architecture uses differential privacy during model training to protect patients’ personal information. Through the research, we demonstrate its practical implementation in the near future for it is hard to balance the precision of modeled obstructions with cryptographic overhead that still remains acceptable. The results of this research lay the groundwork for medical AI systems of the future that are both secure and respectful of patients’ privacy.
N. Kannan, K. Balasubramanian· International Journal of Com...· 0 citations
A Federated Learning and LLM-Driven Threat Intelligence for Zero Trust IoT Architecture is presented, with FL for anomaly detection integrating privacy-preserving distributed learning, continuous identity verification, and LLM-driven autonomous threat response into a unified pipeline.
Healthcare Internet of Things (HIoT) deployments generate sensitive patient telemetry data on resource-constrained edge devices, which are prime targets for network intrusions. Centralizing raw telemetry for training intrusion detection system (IDS) models violates patient privacy and contravenes data-protection regulations such as HIPAA and GDPR. This paper proposes PPFL-IDS, a Privacy-Preserving Federated Learning framework for intrusion detection in HIoT environments. PPFL-IDS combines federated model aggregation with differential privacy noise injection and secure aggregation protocols to train a lightweight gradient-boosted ensemble IDS without exposing local device data. A heterogeneity-aware client selection mechanism addresses the challenge of non-independent and identically distributed (non-IID) data inherent in multi-site HIoT deployments. Evaluated on the UNSW-NB15 and a synthetic HIoT dataset spanning five attack categories, PPFL-IDS achieves a weighted F1-score of 0.938 and a mean detection latency of 20.3 ms, outperforming FedAvg, FedProx, and SCAFFOLD baselines while satisfying an ε-differential privacy budget of 1.2. Results demonstrate that strong privacy guarantees and high detection accuracy can be achieved simultaneously in federated HIoT security architectures.
Nutan Gusain, J. Alzubi· International Journal on Com...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
