Dissertation on malware detection discussed

The Dissertation Council held a scientific seminar at the Institute of Information Technology of the Ministry of Science and Education. Opening the event with introductory remarks, the chairman of the meeting, Doctor of Technical Sciences Mutallim Mutallimov, stated that the seminar was dedicated to discussing the dissertation submitted by the Institute's doctoral candidate, Elshan Baghirov, specializing in 3339.01 – “Methods and systems of information protection, information security.”

The doctoral candidate, Elshan Baghirov, presented his dissertation entitled “Development of Intelligent Analysis Methods and Algorithms for Malware Detection,” outlining its relevance, objectives, scientific novelty, practical significance, and key findings. He noted that according to statistical data, the volume of malware samples, including novel variants, is growing exponentially, leading to highly destructive consequences. He emphasized that as digital systems become increasingly complex, security-integrated software engineering plays a strategic role in preventing malware and building resilient architectures. Furthermore, cybercriminals are increasingly employing detection-evasion techniques to expand both the volume and diversity of new malware strains, emphasizing the necessity of research in this domain.

The candidate also highlighted that the relevance of the dissertation is driven by the potential of applying intelligent analysis techniques, such as machine learning, deep learning, and generative adversarial networks, to malware detection, as well as national objectives aimed at strengthening technical, cryptographic, software, and other information protection mechanisms.

Elshan Baghirov noted that the primary objective of his research is to develop high-accuracy intelligent analysis methods and algorithms for detecting malicious executable software within Windows and Android operating system environments.

Presenting the main scientific results obtained in his study, he explained that a novel method was developed for selecting optimal features from static and dynamic feature spaces to build machine learning-based malware detection models. Additionally, machine learning approaches based on classification and clustering were designed to detect malware and malicious network traffic across Windows and Android platforms using diverse static and dynamic features. A high-accuracy detection algorithm based on transfer learning was also introduced, capable of identifying malware by transforming it into image representations. To address data scarcity during training, a conditional generative adversarial network approach was developed to augment malware samples; evaluating machine learning models on this augmented dataset demonstrated improved performance metrics. Furthermore, to assess the explainability of the proposed detection models, experiments were conducted using three distinct datasets, and feature drift was experimentally evaluated in real-world application scenarios.

Following the candidate's presentation, the reports of the official reviewers, Doctor of Technical Sciences, Professor Shahla Huseynzada, and Doctor of Philosophy (PhD) in Technical Sciences, Associate Professor Nigar Ismayilova, were heard. The reviewers emphasized that comprehensive research had been conducted, the formulated research problems were successfully resolved, and the study demonstrated significant scientific novelty and practical value. They concluded that the dissertation fully complies with the requirements set by the Supreme Attestation Commission under the President of the Republic of Azerbaijan for specialty 3339.01 – “Methods and systems of information protection, information security,” and recommended the work for defence to qualify for the degree of Doctor of Philosophy in Technical Sciences.

After the applicant addressed the suggestions and remarks raised in the reviews, discussions were held regarding the dissertation. Doctor of Technical Sciences Mutallim Mutallimov, Doctor of Technical Sciences Associate Professor Fahrad Pashayev, and Doctors of Philosophy in Technical Sciences Irada Alakbarova, Babak Nabiyev, and Afruz Gurbanova participated in the discussions.

   © All rights reserved. When using these materials, reference to www.iсt.az is required.