Kamal Acharya

Ph.D. Candidate at UMBC, AI Researcher

Back to Publications

Conference Paper

Decoding Android Malware with a Fraction of Features: An Attention-Enhanced MLP-SVM Approach

2025 NSS 2024 (LNCS 15564, 2025) DOI: 10.1007/978-981-96-3531-3_10

Android Malware Cybersecurity Machine Learning

DOI Cite

Abstract

This paper proposes an attention-enhanced MLP-SVM framework for Android malware detection and classification using a compact feature set. It reduces thousands of available features to a small discriminative subset, applies attention and dimensionality reduction, and uses an RBF-kernel SVM to classify malware families with high reported accuracy and reduced computational complexity.

Plain-Language Summary

The paper shows that Android malware can be detected accurately without using thousands of features, making the model more efficient and easier to deploy.

Why This Paper Matters

The work supports efficient malware detection pipelines that can remain accurate while reducing feature and computation requirements.

Research Summary

This paper addresses Android malware detection under a practical constraint: high-dimensional malware feature sets can be expensive and difficult to deploy. The goal is to classify malware accurately while using only a fraction of the available features.

The method combines an attention-enhanced multilayer perceptron with a support vector machine. Attention helps identify discriminative features, dimensionality reduction compresses the representation, and the SVM performs the final classification.

The paper contributes to efficient cybersecurity modeling by showing that compact feature representations can still support strong malware detection and classification performance.

Key Contributions

  • Combines an attention-enhanced multilayer perceptron with an SVM classifier.
  • Reduces the feature set from thousands of Android malware features to a compact subset.
  • Uses dimensionality reduction before malware family classification.
  • Evaluates the approach with accuracy, precision, recall, and F1-score.

Publication Details

Type
Conference Paper
Venue
NSS 2024 (LNCS 15564, 2025)
Year
2025

Authors

Hakim S. B., Adil M., Acharya K., Song H. H.

Research Topics

Android Malware Cybersecurity Machine Learning

Citation

@inproceedings{hakim2025-decoding-android-malware-with-a-fraction-of-features-an-attention-enhanced-mlp-svm-approach,
  title = {Decoding Android Malware with a Fraction of Features: An Attention-Enhanced MLP-SVM Approach},
  author = {Hakim S. B. and Adil M. and Acharya K. and Song H. H.},
  booktitle = {NSS 2024 (LNCS 15564, 2025)},
  year = {2025},
  doi = {10.1007/978-981-96-3531-3_10}
}