Scholarlink

<p>This study presents an advanced deep learning-based Intrusion Detection System (IDS) designed to enhance security in Internet of Things (IoT) networks, where traditional security mechanisms often fall short. The proposed IDS utilizes a bidirectional gated recurrent unit (Bi-GRU) architecture combined with an attention mechanism to effectively capture both short-term and long-term dependencies and emphasize critical features relevant to intrusion detection. The attention mechanism enhances the model’s ability to focus on the most impactful features in real-time data streams, improving detection accuracy. To address the interpretability issue commonly associated with deep learning models, we integrate the SHapley Additive exPlanations (SHAP) method, which provides a comprehensive breakdown of feature importance, allowing security analysts to gain deeper insights into the decision-making process of the IDS. We rigorously evaluate the model on the CICIDS2017 dataset, achieving accuracies of 99.57% for binary classification and 99.44% for multi-class classification, outperforming current state-of-the-art solutions. These results demonstrate that the proposed IDS not only improves detection accuracy but also provides transparency in model decisions, making it more suitable for practical deployment in complex IoT environments prone to evolving cyber threats.</p>

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 cyber-resilient and explainable intrusion detection system for Internet of Things networks

  • 期刊论文

  • 发表日期：

    2025年4月28日

查看全文
  • Muhammad Bilal,
  • Rongfei Zeng,
  • Owais Muhammad,
  • Muhammad Adil,
  • Shifa Shoukat

摘要

This study presents an advanced deep learning-based Intrusion Detection System (IDS) designed to enhance security in Internet of Things (IoT) networks, where traditional security mechanisms often fall short. The proposed IDS utilizes a bidirectional gated recurrent unit (Bi-GRU) architecture combined with an attention mechanism to effectively capture both short-term and long-term dependencies and emphasize critical features relevant to intrusion detection. The attention mechanism enhances the model’s ability to focus on the most impactful features in real-time data streams, improving detection accuracy. To address the interpretability issue commonly associated with deep learning models, we integrate the SHapley Additive exPlanations (SHAP) method, which provides a comprehensive breakdown of feature importance, allowing security analysts to gain deeper insights into the decision-making process of the IDS. We rigorously evaluate the model on the CICIDS2017 dataset, achieving accuracies of 99.57% for binary classification and 99.44% for multi-class classification, outperforming current state-of-the-art solutions. These results demonstrate that the proposed IDS not only improves detection accuracy but also provides transparency in model decisions, making it more suitable for practical deployment in complex IoT environments prone to evolving cyber threats.