Scholarlink

The trend of digitization in almost every aspect of daily human life has raised serious concerns about security in the digital world. With new technologies, solutions, and tools emerging daily, new vulnerabilities also arise. Botnets are among the most widespread cyberthreats in the modern digital landscape, as they can breach and affect entire organizations or domains by infecting just a single device in a network. This study involves the design and implementation of a generative adversarial network (the so-called BotNetGAN - BNGAN) to synthetically generate botnet attack data samples, which are assessed for both quality and quantity using specific data quality indicators. The quality assessment results show that the produced data are very similar to the original ones. Therefore, the significance of GANs in data generation processes is almost undeniable. Furthermore, increasing the volume of annotated data can lead to the improvement and enhancement of AI-based cybersecurity solutions that heavily rely on data availability.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Generative Adversarial Network (GAN) Solution for Synthetically Generated Botnet Attack Data Samples

  • 章节

  • 出版模式：

    开放获取

  • 发表日期：

    2025年

查看全文
  • Nikolaos Peppes,
  • Theodoros Alexakis,
  • Emmanouil Daskalakis,
  • Evgenia Adamopoulou,
  • Konstantinos Demestichas

摘要

The trend of digitization in almost every aspect of daily human life has raised serious concerns about security in the digital world. With new technologies, solutions, and tools emerging daily, new vulnerabilities also arise. Botnets are among the most widespread cyberthreats in the modern digital landscape, as they can breach and affect entire organizations or domains by infecting just a single device in a network. This study involves the design and implementation of a generative adversarial network (the so-called BotNetGAN - BNGAN) to synthetically generate botnet attack data samples, which are assessed for both quality and quantity using specific data quality indicators. The quality assessment results show that the produced data are very similar to the original ones. Therefore, the significance of GANs in data generation processes is almost undeniable. Furthermore, increasing the volume of annotated data can lead to the improvement and enhancement of AI-based cybersecurity solutions that heavily rely on data availability.