2026-05-04T22:06:28.258Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:1:11582 2025-09-09T16:47:53Z 1

Combining Generative Adversarial Networks with Modified Metaheuristic-Optimized Classifiers for Internet of Things Security 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/1/11582 https://ieeexplore.ieee.org/abstract/document/11134924 A. Petrovic L. Jovanovic T. Zivkovic V. Zeljkovic M. Zivkovic N. Bacanin Internet of Things (IoT) devices security issues remain a frequently disregarded concern due to their restricted processing capabilities and the necessity for minimal energy usage. Consequently, safeguarding these systems is often assigned a lower priority. Nonetheless, as IoT technology becomes increasingly embedded in everyday activities, the urgency for resilient protection strategies has intensified. This study tackles the complexities of IoT security by utilizing machine learning methodolo-gies refined through an enhanced metaheuristic algorithm. These approaches are specifically adapted to circumvent the inherent challenges of IoT devices, such as processing limitations and constrained data resources. Furthermore, this research explores the utilisation of Generative Adversarial Networks (GANs) to create and augment datasets, thereby facilitating superior model training while addressing confidentiality issues. By integrating these advancements, the investigation aims to fortify IoT network defenses and alleviate risks associated with data insufficiency and system weaknesses. conferenceObject 1 6 10.1109/ICCSC66714.2025.11134924 2025 International Conference on Circuit, Systems and Communication (ICCSC), IEEE, Fez, Morocco