Scholarlink

<p>The protection of IT systems against rapidly evolving threats associated with IoT apps requires the implementation of the best solutions by organizations. This is because the global deployment of IoT apps and their integration with corporate IT systems has raised new security concerns. In order for traditional machine learning (ML) systems to meet the security requirements of sustainable IoT environments in the near future, they must be continually improved as the number of IoT networks and more complex, multi-layered threat/attack vectors increases. Since deep learning (DL) has proven beneficial in many other domains, it can be used to create tools that enhance network intrusion detection systems. However, some new and emerging issues with the precision, efficacy, scalability and reliability of conventional NIDS have surfaced in heterogeneous IoT settings. This study’s deep learning model provides a highly dependable NIDS solution that outperforms many current approaches in terms of performance and reliability. The primary components of this work are: 1) A method that uses the fewest labelled datasets to identify the qualities that best distinguish attack instances from normal examples. (2) Testing the proposed approach using real datasets. The creation of a trustworthy DL-based Enhanced Memory-triggered Model Encoding (IMTE) model for identifying threats and predicting impending hazards. Researchers recently investigated the effectiveness of the proposed NIDS solution incorporating all three components. This was done to confirm the accuracy of the NIDS model. The research and evaluation demonstrated the effectiveness and reliability of the suggested NIDS solution.</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

Designing an improved memory-triggered encoding model for threat prediction using learning approaches

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • Ashwini Gulhane,
  • A. K. Velmurugan

摘要

The protection of IT systems against rapidly evolving threats associated with IoT apps requires the implementation of the best solutions by organizations. This is because the global deployment of IoT apps and their integration with corporate IT systems has raised new security concerns. In order for traditional machine learning (ML) systems to meet the security requirements of sustainable IoT environments in the near future, they must be continually improved as the number of IoT networks and more complex, multi-layered threat/attack vectors increases. Since deep learning (DL) has proven beneficial in many other domains, it can be used to create tools that enhance network intrusion detection systems. However, some new and emerging issues with the precision, efficacy, scalability and reliability of conventional NIDS have surfaced in heterogeneous IoT settings. This study’s deep learning model provides a highly dependable NIDS solution that outperforms many current approaches in terms of performance and reliability. The primary components of this work are: 1) A method that uses the fewest labelled datasets to identify the qualities that best distinguish attack instances from normal examples. (2) Testing the proposed approach using real datasets. The creation of a trustworthy DL-based Enhanced Memory-triggered Model Encoding (IMTE) model for identifying threats and predicting impending hazards. Researchers recently investigated the effectiveness of the proposed NIDS solution incorporating all three components. This was done to confirm the accuracy of the NIDS model. The research and evaluation demonstrated the effectiveness and reliability of the suggested NIDS solution.