A Systematic Analysis of Multi-Task Learning Frameworks for Advanced Cybersecurity Applications
Volume 20, Issue 1, 2026
Download| Author(s): |
Umair Ayaz Kamangar Sukkur IBA University, Pakistan, umair.ayaz@iba-suk.edu.pk Sanaullah Khaskheli* Sukkur IBA University, Pakistan, sanaullah@iba-suk.edu.pk Zakria Jamali Sukkur IBA University, Pakistan, zakria@iba-suk.edu.pk Zainab Umair Kamangar Sukkur IBA University, Pakistan, zainabumair.phdcss22@iba-suk.edu.pk Abdul Sattar Chan Sukkur IBA University, Pakistan, abdul.sattar@iba-suk.edu.pk Muhammad Ahsan Kareem Sukkur IBA University, Pakistan, ahsanmughal.becsef22@iba-suk.edu.pk Sibgha Mursaleen Sukkur IBA University, Pakistan, sibghamursaleen.becsef22@iba-suk.edu.pk Muhammad Usman Sarwar Sukkur IBA University, Pakistan, muhammadusman.becsef22@iba-suk.edu.pk Soyam Kapoor Sukkur IBA University, Pakistan, soyamkapoor.becsef22@iba-suk.edu.pk |
|---|---|
| Abstract | Increasing reliance on technology has elevated the possibility of cyber-attacks directed towards financial, military and political interests. The significance of cybersecurity to the IT industry can no longer be under-estimated, and the security of data is a paramount concern. The Multi-task Learning (MTL) in cybersecurity offers a promising solution. In this paper, the potential of MTL is being explored as a solution to the sophisticated attacks against financial, defense and governmental system with the main goal being to determine how MTL improves the adaptive and intelligent cyber defense by learning on different, but related, security tasks at once. A SLR methodology was applied using the selected research articles on MTL in cybersecurity, which covered 40 articles. All selected papers were systematically reviewed to identify the application fields as well as common ML techniques. The findings have shown that there are 5 major fields where MTL can be applied: intrusion detection, malware classification, cyber threat detection, critical infrastructure security and online abuse detection. The research conducted concludes that supervised learning models, especially DL models like CNN, RNN and LSTM are the most applied models in this research area. MTL can greatly improve cybersecurity by providing adaptability, learn efficiency and shared representation across various tasks. It contributed to create a consolidated analysis of current MTL use cases in cybersecurity and highlight research challenges for future work in developing robust, scalable and intelligent security defense. |
| Keywords | cybersecurity, multi-task learning, cyber threats, deep learning |
| Year | 2026 |
| Volume | 20 |
| Issue | 1 |
| Type | Research paper, manuscript, article |
| Journal Name | Journal of Information & Communication Technology | Publisher Name | ILMA University | Jel Classification | - | DOI | - | ISSN no (E, Electronic) | 2075-7239 | ISSN no (P, Print) | 2415-0169 | Country | Pakistan | City | Karachi | Institution Type | University | Journal Type | Open Access | Manuscript Processing | Blind Peer Reviewed | Format | Paper Link | https://jict.ilmauniversity.edu.pk/journal/jict/20.1/1.pdf | Page | 1-10 |