Enhanced Intrusion Detection in Cybersecurity through Explainable Artificial Intelligence

Strategic Insights -

Researchers at Princess Nourah bint Abdulrahman University have made significant advancements in the field of artificial intelligence (AI) for cybersecurity. Their study, published in Scientific Reports, proposes an Enhanced Intrusion Detection in Cybersecurity through Dimensionality Reduction and Explainable Artificial Intelligence with Attention Mechanism in Deep Learning (EIDCDR-XAIADL) model. This innovative approach aims to tackle the growing threats in the cyber world by providing a robust and transparent cybersecurity system. The study's findings have far-reaching implications for the development of AI-powered cybersecurity solutions.

Key Takeaways:

  • The EIDCDR-XAIADL model combines explainable AI (XAI) with machine learning (ML) and deep learning (DL) to enhance intrusion detection in cybersecurity.
  • The proposed model utilizes dimensionality reduction techniques to improve the efficiency of the system and reduces the complexity of the data.
  • Shapley Additive Explanations (SHAP) is used as an XAI technique to provide trustworthy insights into AI-driven security systems.
  • The experimental evaluation of the EIDCDR-XAIADL approach demonstrated superior accuracy values of 99.19% and 99.12% under NSLKDD and CICIDS 2017 datasets.
  • The study emphasizes the importance of interpretability and transparency in AI methods, particularly in cybersecurity applications.
  • The researchers propose the use of antlion optimization (ALO) to adjust the hyperparameter values of the CNN-BiGRU-AM method optimally, resulting in more excellent classification performance.
  • The study's findings have significant implications for the development of AI-powered cybersecurity solutions.
  • The proposed EIDCDR-XAIADL model can be used to detect and prevent various types of cyber attacks, including malware and unauthorized access.
  • The use of SHAP as an XAI technique can enhance threat detection and decision-making in AI-driven security systems.
  • The study highlights the need for more research in the field of AI for cybersecurity, particularly in the development of explainable AI models.

Statistics:

  • Accuracy values of 99.19% and 99.12% were achieved under NSLKDD and CICIDS 2017 datasets, respectively.
  • The EIDCDR-XAIADL approach demonstrated superior performance compared to other existing intrusion detection systems.
  • The use of dimensionality reduction techniques in the proposed model resulted in improved efficiency and reduced complexity of the data.
  • The experimental evaluation of the EIDCDR-XAIADL approach was conducted under dual datasets (NSLKDD and CICIDS 2017).
  • The study emphasized the importance of interpretability and transparency in AI methods, particularly in cybersecurity applications.

Sources:

  • Enhanced intrusion detection in cybersecurity through dimensionality reduction and explainable artificial intelligence. Scientific Reports, 2025,15(1):1-25.
  • Princess Nourah bint Abdulrahman University Researchers Publish Findings in Artificial Intelligence (Enhanced intrusion detection in cybersecurity through dimensionality reduction and explainable artificial intelligence). Information Technology Newsweekly. October 21, 2025; p 581.