EUVIMEDEUVIMED entdecken
Crossref

TRUST-GUIDED LAYER-ADAPTIVE FEDERATED LEARNING FOR PRIVACYPRESERVING CLOUD-EDGE-END INTELLIGENCE

S. Pavaimalar, Booharan S, Rajkamal N

International Journal of Drug Delivery Technology · 01.07.2026 · 16 (60s)

DOI: 10.25258/ijddt.16.60s.56 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 01.07.2026. Verfasst von S. Pavaimalar und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The distributed computing is rapidly growing over the cloud, edge, and end devices and allows carrying out intelligent realtime analytics but makes sensitive data vulnerable to privacy and security threats. In this paper, we provide the Layer-Adaptive Secure Federated Learning (LASFL) which is a new framework that aims to facilitate secure and privacy-conscious collaborative learning in heterogeneous settings. Materials and Methods LASFL implements layer-adaptive gradient perturbation that injects noise selectively basing on the sensitivity of local data and device reliability, and dynamic contribution weighting, which gives trustworthy devices a greater priority in improving the performance of global models without jeopardizing the privacy. The framework is capable of supporting scalable and efficient distributed training by encrypting the edge-layer model aggregation to ensure that intermediate updates are not leaked. LASFL dynamically strikes a balance of utility and privacy of the model and data privacy among end, edge, and cloud layers to alleviate inference attacks without sacrificing accuracy.

Results

Experimental analysis of benchmark datasets proves that LASFL can decrease by up to 40 percent the privacy leakage with respect to the traditional federated learning methods, with little effect on the model performance.

Conclusion

The proposed system offers a software-intensive framework of privacy-guaranteeing distributed intelligence, combining adaptive privacy, trust-based updates, and encrypted aggregation into a single and scalable framework.

Zitieren

APA

Pavaimalar, S., S, B., & N, R. (2026). TRUST-GUIDED LAYER-ADAPTIVE FEDERATED LEARNING FOR PRIVACYPRESERVING CLOUD-EDGE-END INTELLIGENCE. International Journal of Drug Delivery Technology, 16(60s). https://doi.org/10.25258/ijddt.16.60s.56

Vancouver

Pavaimalar S, S B, N R. TRUST-GUIDED LAYER-ADAPTIVE FEDERATED LEARNING FOR PRIVACYPRESERVING CLOUD-EDGE-END INTELLIGENCE. International Journal of Drug Delivery Technology. 2026;16(60s). doi:10.25258/ijddt.16.60s.56

BibTeX
@article{Pavaimalar2026,
  author = {Pavaimalar, S. and S, Booharan and N, Rajkamal},
  title = {TRUST-GUIDED LAYER-ADAPTIVE FEDERATED LEARNING FOR PRIVACYPRESERVING CLOUD-EDGE-END INTELLIGENCE},
  journal = {International Journal of Drug Delivery Technology},
  year = {2026},
  volume = {16},
  number = {60s},
  doi = {10.25258/ijddt.16.60s.56}
}
Kostenlos starten →