PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 13, 2026Internet of Things2 citationsOpen Access

Adversarially Resilient Federated Learning for Heterogeneous Edge Nodes in 5G Networks with Non-IID Data

View Full Paper
SZSaniya ZafarLahore College for Women UniversityPLPhil LeggJWJonathan WhiteUniversity of the West of England

Key Points

  • The aim is to enhance the robustness and efficiency of federated learning in 5G networks dealing with heterogeneous non-IID data.
  • Developed an adversarially robust federated learning (ARFL) mechanism.
  • Utilized hybrid feature selection and adversarial optimization.
  • Implemented a min-max formulation to optimize classifier and adversarial elements.
  • Conducted experiments on a real-world intrusion detection dataset.
  • Standard federated learning performance dropped to 20%-30% under adversarial conditions.
  • ARFL maintained performance above 92% in accuracy, precision, recall, and F1-scores across non-IID distributions.
  • Achieved improvements of 20%-70% points in adversarial accuracy over standard federated learning.
  • Showed marginal performance reduction in clean data scenarios.

Abstract

The rapid deployment of 5 G and Beyond-5 G ( B 5 G ) edge networks introduces unique challenges for federated learning (FL) frameworks deployed at the edge, primarily due to heterogeneous non-IID data distributions and adversarial vulnerabilities. This paper proposes an adversarially robust federated learning (ARFL) mechanism that integrates hybrid feature selection and adversarial optimization to jointly enhance robustness against adversarial perturbations and improve computational efficiency under heterogeneous data distributions. The proposed methodology jointly optimizes classifier and adversary in a min–max formulation to enable robustness against perturbations of varying strengths. Experimental results on a real-world intrusion detection 5G-NIDD dataset demonstrates that standard FL suffers drastic deterioration under adversarial conditions, with accuracy, precision, recall, and F1-scores dropping to 20%–30% at ϵ = 0.3 . In contrast, the proposed ARFL framework consistently sustains performance above 92% across these metrics under all non-IID distributions, highlighting its robustness and reliability. Overall, ARFL achieves absolute adversarial accuracy improvements of 20%–70% points over standard FL while incurring only a marginal reduction in clean performance. Scalability experiments demonstrate the stability and efficiency of the ARFL framework, underscoring its suitability for real-world 5 G edge deployments where robustness and efficiency are paramount.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Zafar et al. (2026) studied this question.

synapsesocial.com/papers/69b3ac3f02a1e69014ccdcf0https://doi.org/10.1016/j.iot.2026.101919
Ask AI
Helpful
Bookmark
Share
View Full Paper