Scholarlink

<p>Ransomware detection in cloud Virtual Machines (VMs) is essential because of the frequent attacks on cloud environments that result in major information loss and operational delays. Static Analysis (SA) and Dynamic Analysis (DA) are the traditional methods that suffer from major false positive and negative rates with complex ransomware deviations. The Proposed method introduces a unique method for detecting ransomware in cloud VMs that combines differential privacy, SMPC and Bi LSTM with Deep Neural Networks. This improves privacy and increases detection accuracy by combining Bi LSTMs ability to capture repetitive patterns and DNNs ability to handle complex information formats. This model proved an outstanding capacity in detecting ransomware with small errors and obtained 99.60 accuracy, 99.69 precision, 99.81 recall and 99.75% F1 score. SMPC guarantees that sensitive data is secured during detection and allows collaboration between many entities. This increases ransomware detection but also protects data privacy in working environments. It provides a reliable solution for cloud-based ransomware detection that is both efficient and safe.</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

Differential privacy enhanced ransomware detection in cloud VMs using SMPC and bidirectional LSTM with DNN

  • 期刊论文

  • 发表日期：

    2025年11月11日

查看全文
  • Waeal J. Obidallah,
  • Mubarak Albathan,
  • Zeyad Alshaikh,
  • Riyad Almakki,
  • Qaisar Abbas

摘要

Ransomware detection in cloud Virtual Machines (VMs) is essential because of the frequent attacks on cloud environments that result in major information loss and operational delays. Static Analysis (SA) and Dynamic Analysis (DA) are the traditional methods that suffer from major false positive and negative rates with complex ransomware deviations. The Proposed method introduces a unique method for detecting ransomware in cloud VMs that combines differential privacy, SMPC and Bi LSTM with Deep Neural Networks. This improves privacy and increases detection accuracy by combining Bi LSTMs ability to capture repetitive patterns and DNNs ability to handle complex information formats. This model proved an outstanding capacity in detecting ransomware with small errors and obtained 99.60 accuracy, 99.69 precision, 99.81 recall and 99.75% F1 score. SMPC guarantees that sensitive data is secured during detection and allows collaboration between many entities. This increases ransomware detection but also protects data privacy in working environments. It provides a reliable solution for cloud-based ransomware detection that is both efficient and safe.