Scholarlink

We propose hybrid deep learning architectures for improving the accuracy of malware detection. The proposed model combines the strengths of standard DL architectures to build hybrid models. The first model combines MobileNetV2, and EfficientNetB0, whereas the second proposed model combines DenseNet121 and ResNet-50. By leveraging these hybrid approaches, the proposed models achieve superior performance in malware detection and classification, ensuring robustness against both known and unknown threats. Experimental results demonstrate the effectiveness of the hybrid architectures in accurately identifying malware variants, paving the way for more resilient and efficient malware classification systems.

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

Late-Fusion-Based Hybrid Deep Learning Architectures for Malware Detection

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Alekha Kumar Mishra,
  • Sabyasachi Biswal,
  • Penta Sanjana Naidu

摘要

We propose hybrid deep learning architectures for improving the accuracy of malware detection. The proposed model combines the strengths of standard DL architectures to build hybrid models. The first model combines MobileNetV2, and EfficientNetB0, whereas the second proposed model combines DenseNet121 and ResNet-50. By leveraging these hybrid approaches, the proposed models achieve superior performance in malware detection and classification, ensuring robustness against both known and unknown threats. Experimental results demonstrate the effectiveness of the hybrid architectures in accurately identifying malware variants, paving the way for more resilient and efficient malware classification systems.