Zobraziť v EDS

Malware Detection Using Hebbian Spike Deep Belief Network with Cosine Similarity Loss Function.

Uložené v:
Podrobná bibliografia
Názov: Malware Detection Using Hebbian Spike Deep Belief Network with Cosine Similarity Loss Function.
Autori: Jayaprakash, Parvathi Sathenahalli, Chandrashekaraiah, Yogeesh Amabalagere
Zdroj: International Journal of Intelligent Engineering & Systems; 2025, Vol. 18 Issue 10, p379-398, 20p
Predmety: MALWARE, DEEP learning, LOSS functions (Statistics), MACHINE learning, INTERNET security, ARTIFICIAL neural networks, SIGNAL detection
Abstrakt: Malware detection involves identifying harmful software that steals sensitive information or damages systems. It analyzes the behavior, characteristics, and patterns of software to distinguish malicious activity from legitimate operations. Despite ongoing advancements, malware attacks continue to pose significant threats to cybersecurity due to the emergence of new variants and increasing malware sophistication. To address this challenge, a Hebbian Spike Deep Belief Network with Cosine Similarity Loss Function (HSDBN-CSLF) is proposed for effective malware detection and classification, improving both model performance and generalization capability. The method incorporates spiking neurons modelled using Siegert dynamics to capture spatial and event-driven behavior. Hebbian learning strengthens synaptic connections based on correlated activation, enabling adaptive learning of evolving malware patterns. The CSLF enhances intra-class feature cohesion and inter-class separation by leveraging angular similarity between feature vectors, thereby enhancing overall performance. The HSDBN-CSLF model achieves high accuracies of 99.96%, 99.83%, and 99.81% on Malimg, Microsoft BIG2015, and Malevis datasets outperforming existing methods such as the Convolutional Neural Network. [ABSTRACT FROM AUTHOR]
Copyright of International Journal of Intelligent Engineering & Systems is the property of Intelligent Networks & Systems Society and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Databáza: Complementary Index
Popis
Abstrakt:Malware detection involves identifying harmful software that steals sensitive information or damages systems. It analyzes the behavior, characteristics, and patterns of software to distinguish malicious activity from legitimate operations. Despite ongoing advancements, malware attacks continue to pose significant threats to cybersecurity due to the emergence of new variants and increasing malware sophistication. To address this challenge, a Hebbian Spike Deep Belief Network with Cosine Similarity Loss Function (HSDBN-CSLF) is proposed for effective malware detection and classification, improving both model performance and generalization capability. The method incorporates spiking neurons modelled using Siegert dynamics to capture spatial and event-driven behavior. Hebbian learning strengthens synaptic connections based on correlated activation, enabling adaptive learning of evolving malware patterns. The CSLF enhances intra-class feature cohesion and inter-class separation by leveraging angular similarity between feature vectors, thereby enhancing overall performance. The HSDBN-CSLF model achieves high accuracies of 99.96%, 99.83%, and 99.81% on Malimg, Microsoft BIG2015, and Malevis datasets outperforming existing methods such as the Convolutional Neural Network. [ABSTRACT FROM AUTHOR]
ISSN:2185310X
DOI:10.22266/ijies2025.1130.25