EUVIMEDScopri EUVIMED
Crossref

Adversarial Robust Deep Autoencoder for Intrusion Detection in Blockchain-Secured Nano Virtual Health Record Systems

Daniel Kareem, Sofia Haddad

Journal of Nano Molecular Intelligence and Virtual Health Systems · 2026-01-30 · 2 (1) · 33

DOI: 10.70023/nmivhs.213 ↗

At a glance

Journal article in Journal of Nano Molecular Intelligence and Virtual Health Systems, published online on 2026-01-30. Written by Daniel Kareem and 1 co-authors.

AbstractDati bibliograficiCiteRelated studies

Abstract (original language)

Blockchain-secured nano virtual health record (NVHR) systems safely, transparently, and discreetly exchange medical institution data for decentralized healthcare data management. NVHR systems are subject to sophisticated cyberattacks and malicious manipulations that might compromise data integrity and service dependability when IoMT devices and blockchain infrastructure merge. Intrusion detection systems are vulnerable to adversary disturbances and lose latent traffic characteristics under threat pressure. For blockchain-secured NVHR systems, the Adversarially Robust Deep Autoencoder for Intrusion Detection (ARDA-ID) architecture overcomes these issues in this work. Adversarial consistency regularization and hierarchical feature reconstruction improve embedding stability and perturbation resistance. Latent divergence-guided anomaly separation and blockchain-aware trust verification improve transaction authenticity and network behavior discrimination. While maintaining learnt representations, they protect against invasions. The experiments reveal that ARDA-ID works in numerous attack circumstances. Beyond classification metrics, the framework is evaluated using the Adversarial Representation Stability Index (ARSI), Latent Feature Preservation Rate (LFPR), Reconstruction Divergence Suppression Score (RDSS), and Threat Isolation Efficiency (TIE). The proposed framework has ARSI of 0.968, LFPR of 97.83%, RDSS of 96.74%, and TIE of 97.56%, indicating excellent representation stability, latent information preservation, anomaly separation capability, and adversarial resistance

Cite

APA

Kareem, D., & Haddad, S. (2026). Adversarial Robust Deep Autoencoder for Intrusion Detection in Blockchain-Secured Nano Virtual Health Record Systems. Journal of Nano Molecular Intelligence and Virtual Health Systems, 2(1), 33. https://doi.org/10.70023/nmivhs.213

Vancouver

Kareem D, Haddad S. Adversarial Robust Deep Autoencoder for Intrusion Detection in Blockchain-Secured Nano Virtual Health Record Systems. Journal of Nano Molecular Intelligence and Virtual Health Systems. 2026;2(1):33. doi:10.70023/nmivhs.213

BibTeX
@article{Kareem2026,
  author = {Kareem, Daniel and Haddad, Sofia},
  title = {Adversarial Robust Deep Autoencoder for Intrusion Detection in Blockchain-Secured Nano Virtual Health Record Systems},
  journal = {Journal of Nano Molecular Intelligence and Virtual Health Systems},
  year = {2026},
  volume = {2},
  number = {1},
  pages = {33},
  doi = {10.70023/nmivhs.213}
}
Start for free →