The role of machine learning in predicting zero-day vulnerabilities

AL Rafy *, Md Mashfiquer Rahman, Sharmin Nahar, Md. Najmul Gony and MD IMRANUL HOQUE Bhuiyan

Independent Researcher, Bangladesh.
 
Research Article
International Journal of Science and Research Archive, 2023, 10(01), 1197-1208.
Article DOI: 10.30574/ijsra.2023.10.1.0838
Publication history: 
Received on 05 September 2023; revised on 14 October 2023; accepted on 17 October 2023
 
Abstract: 
Zero-day vulnerabilities keep growing as an important threat in cybersecurity because attackers discover them before security teams can detect them. Signature-based detection methods fail to discover unknown vulnerabilities since they need prior knowledge of known attack techniques. ML technology emerges as the promising tool that predicts zero-day threats before attackers exploit them. This research aims to study the training approach of ML models that detect vulnerabilities by analyzing code structures, behavioral irregularities, and network traffic characteristics. The research examines zero-day exploit prediction effectiveness by implementing anomaly detection systems, classification algorithms, and deep learning frameworks. Research results demonstrate that ML technology implements early warning capabilities, delivering superior identification and response performance over conventional techniques. The proactive stance in cybersecurity through zero-day attack detection could lower the extent of damage these attacks create and establish an enhanced defensive system against advancing cyber threats.
 
Keywords: 
Zero-Day Vulnerabilities; Machine Learning; Anomaly Detection; Deep Learning; Ransomware Detection; Predictive Models
 
Full text article in PDF: 
 
Copyright information: 

Copyright © 2023 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0