Home Mathematics Chapter 8 Hybrid Malware Detection and Classification Using Explainable Deep Neural Network
Chapter
Licensed
Unlicensed Requires Authentication

Chapter 8 Hybrid Malware Detection and Classification Using Explainable Deep Neural Network

  • Deepika Sharma and Manoj Devare
Become an author with De Gruyter Brill
Explore this Subject How to publish with us
Cybersecurity Unlocked
This chapter is in the book Cybersecurity Unlocked

Abstract

Combining the finest aspects of numerous neural network architectures, hybrid deep neural networks (DNNs) provide a fresh approach to identify and categorize malware. This article proposes a hybrid DNN model using generative adversarial networks (GANs) and convolutional neural networks (CNNs) to provide more accurate, robust, and understandable model. Metaheuristic approaches assist to improve it. While the CNN section is excellent in obtaining location information, the GAN component creates false samples to provide a more equitable collection that benefits learning. By selecting the most critical features and altering hyperparameters, metaheuristic planning guarantees that model training operates. By use of explainable artificial intelligence (XAI) techniques such as local interpretable model-agnostic explanations, Shapley additive explanations, and layer-wise relevance propagation, the proposed method enhances the capacity to detect malware and protect against threats from third parties, thereby fostering confidence and simplicity for stakeholders. Among useful applications include deployment in defense systems including IoT security, intrusion detection systems, and API- and CLI tool-based simple integration and utilization. The model is tested using the IoT-23 dataset; the findings reveal that it may be used in real life with great accuracy and low false detection rates. This chapter displays how mixed DNNs might be able to fix issues with scale, explainability, and data imbalance, which would improve security. When new XAI methods are combined with focus on real-world distribution strategies, the gap between academic innovation and practical application is narrowed. This makes it possible for future gains in clever and understandable malware detection systems.

Abstract

Combining the finest aspects of numerous neural network architectures, hybrid deep neural networks (DNNs) provide a fresh approach to identify and categorize malware. This article proposes a hybrid DNN model using generative adversarial networks (GANs) and convolutional neural networks (CNNs) to provide more accurate, robust, and understandable model. Metaheuristic approaches assist to improve it. While the CNN section is excellent in obtaining location information, the GAN component creates false samples to provide a more equitable collection that benefits learning. By selecting the most critical features and altering hyperparameters, metaheuristic planning guarantees that model training operates. By use of explainable artificial intelligence (XAI) techniques such as local interpretable model-agnostic explanations, Shapley additive explanations, and layer-wise relevance propagation, the proposed method enhances the capacity to detect malware and protect against threats from third parties, thereby fostering confidence and simplicity for stakeholders. Among useful applications include deployment in defense systems including IoT security, intrusion detection systems, and API- and CLI tool-based simple integration and utilization. The model is tested using the IoT-23 dataset; the findings reveal that it may be used in real life with great accuracy and low false detection rates. This chapter displays how mixed DNNs might be able to fix issues with scale, explainability, and data imbalance, which would improve security. When new XAI methods are combined with focus on real-world distribution strategies, the gap between academic innovation and practical application is narrowed. This makes it possible for future gains in clever and understandable malware detection systems.

Chapters in this book

  1. Frontmatter I
  2. Contents V
  3. Chapter 1 Emerging Cyber Threats: Challenges, Impacts, and Proactive Defenses in the Digital Age 1
  4. Chapter 2 Silent Guardians: Proactive Approaches to Modern Cyber Threats 31
  5. Chapter 3 Data Science for Threat Detection and Analysis 59
  6. Chapter 4 An Integrated Approach: Merging Cybersecurity, AI, and Threat Detection 87
  7. Chapter 5 Cybersecurity Analytics: A Review of Challenges and the Role of Machine Learning and Deep Learning in Threat Detection 103
  8. Chapter 6 Hardware-Based Authentication Techniques for Secure Data Transmission in IoT Edge Computing 141
  9. Chapter 7 Securing the IoT Networks Using a Deep Learning Paradigm for Intrusion Detection 161
  10. Chapter 8 Hybrid Malware Detection and Classification Using Explainable Deep Neural Network 177
  11. Chapter 9 Light POW for Smart Grid Communication 201
  12. Chapter 10 Zero Trust Architecture – A Beginner’s Guide 227
  13. Chapter 11 Post-quantum Cryptography for Enhanced Authentication in Mobile Data Communication: Resilience Against Quantum Attacks 265
  14. Chapter 12 Two-Factor Authentication (2FA) and Multi-factor Authentication (MFA) Solutions for Secure Mobile Data Communication 287
  15. Chapter 13 Artificial Intelligence and Machine Learning in Cybersecurity 313
  16. Chapter 14 Enhancing IoT Security with Zero Trust Networking: Protecting Wireless Sensors, Edge Devices, and Cloud Environments 343
  17. Chapter 15 Biometric Authentication Methods for Mobile Devices: Exploring Fingerprint, Face Recognition, and Iris Scanning 365
  18. Chapter 16 Robust Dynamic Voice-Based Key Generation Using Novel Fuzzy Extraction, Averaged Thresholding, and Hamming Enhancement Techniques 385
  19. Chapter 17 Enhancing Cybersecurity with Artificial Intelligence and Machine Learning Techniques 413
  20. Chapter 18 Firewall and IDS in Cybersecurity 439
  21. Index
https://doi.org/10.1515/9783111712895-008

Chapters in this book

  1. Frontmatter I
  2. Contents V
  3. Chapter 1 Emerging Cyber Threats: Challenges, Impacts, and Proactive Defenses in the Digital Age 1
  4. Chapter 2 Silent Guardians: Proactive Approaches to Modern Cyber Threats 31
  5. Chapter 3 Data Science for Threat Detection and Analysis 59
  6. Chapter 4 An Integrated Approach: Merging Cybersecurity, AI, and Threat Detection 87
  7. Chapter 5 Cybersecurity Analytics: A Review of Challenges and the Role of Machine Learning and Deep Learning in Threat Detection 103
  8. Chapter 6 Hardware-Based Authentication Techniques for Secure Data Transmission in IoT Edge Computing 141
  9. Chapter 7 Securing the IoT Networks Using a Deep Learning Paradigm for Intrusion Detection 161
  10. Chapter 8 Hybrid Malware Detection and Classification Using Explainable Deep Neural Network 177
  11. Chapter 9 Light POW for Smart Grid Communication 201
  12. Chapter 10 Zero Trust Architecture – A Beginner’s Guide 227
  13. Chapter 11 Post-quantum Cryptography for Enhanced Authentication in Mobile Data Communication: Resilience Against Quantum Attacks 265
  14. Chapter 12 Two-Factor Authentication (2FA) and Multi-factor Authentication (MFA) Solutions for Secure Mobile Data Communication 287
  15. Chapter 13 Artificial Intelligence and Machine Learning in Cybersecurity 313
  16. Chapter 14 Enhancing IoT Security with Zero Trust Networking: Protecting Wireless Sensors, Edge Devices, and Cloud Environments 343
  17. Chapter 15 Biometric Authentication Methods for Mobile Devices: Exploring Fingerprint, Face Recognition, and Iris Scanning 365
  18. Chapter 16 Robust Dynamic Voice-Based Key Generation Using Novel Fuzzy Extraction, Averaged Thresholding, and Hamming Enhancement Techniques 385
  19. Chapter 17 Enhancing Cybersecurity with Artificial Intelligence and Machine Learning Techniques 413
  20. Chapter 18 Firewall and IDS in Cybersecurity 439
  21. Index
Sign up now to receive a 20% welcome discount
Institutional Access
How does access work?
Have an idea on how to improve our website?
Please write us.
© 2026 De Gruyter Brill
Downloaded on 24.1.2026 from https://www.degruyterbrill.com/document/doi/10.1515/9783111712895-008/html
Scroll to top button