2026-04-17T17:42:24.971Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:10100 2024-12-03T20:10:47Z 3

Generative Adversarial Networks for Synthetic Training Data Replacement in Phishing Email Detection Using Natural Language Processing, Chapter in AIS Algorithms for Intelligent Systems: ICSMDI 2024: International Conference on Smart Data Intelligence, Springer 2024 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/10100 https://link.springer.com/chapter/10.1007/978-981-97-3191-6_46 L. Jovanovic N. Bacanin R. Ravikumar M. Antonijevic G. Radic M. Zivkovic This study explores the convergence of cybersecurity, Machine learning (ML), Generative adversarial networks (GANs), and Natural language processing (NLP) to overcome the threat posed by phishing emails in the digital landscape. The surge in online business models and email communication has fueled the proliferation of malicious content, accentuating phishing emails as a significant cybersecurity challenge. ML and Artificial Intelligence (AI) algorithms present a dynamic solution, adapting to the evolving threat landscape, contingent upon the availability of pertinent data-privacy concerns. To address this, the study investigates the potential of GANs for synthetic data generation in cybersecurity, specifically focusing on phishing emails. A major advantage of utilizing AI to handle phishing email detection is that such a system has an ability to adapt to the dynamic landscape of cybersecurity without the need for explicit directions given by human operators. Text mining is simply used to reformat the data in to a representation suitable to support the application of AI algorithms, that are designed with numerical values in mind. By conducting experiments with real-world datasets, the research evaluates the performance of contemporary ML classifiers, incorporating NLP techniques, and introduces a GAN-based approach to generate synthetic training data. The outcomes aim to contribute to the development of robust intrusion detection techniques, providing insights into mitigating cybersecurity risks in the face of advanced digital threats. bookPart Springer, Singapore 607 618 10.1007/978-981-97-3191-6_46 AIS Algorithms for Intelligent Systems: ICSMDI 2024: Proceedings of International Conference on Smart Data Intelligence