Vietnam Academy of Science and Technology

Project's information

Project's title	Research and development of hybrid intelligent methods for detecting rare computer attacks in wireless local area networks
Project’s code	QTRU01.14/21-22
Research hosting institution	Institute of Information Technology
Coordinating unit, co-chair	The department “Software Engineering” - Higher School of Economics National Research University, Russian Federation
Project leader’s name	Dr. T.T. Quyen Bui and Assc. Prof. Dr. Avdoshin S. M.
Project duration	01/05/2021 - 31/10/2023
Project’s budget	200 million VND
Classify	Fair
Goal and objectives of the project	- Developing hybrid methods to improve the effectiveness of detecting rare attacks in wireless local area networks. - Training scientific and technological human resources, strengthening cooperation, integration, and international exchange.
Main results	Research contribution 1) Analysis, comparison and evaluation of popular datasets for general attack detection problems and datasets for local wireless network attack detection problems. In addition, the research also points out the issue of data imbalance between rare attacks and common attacks. From there, it shows that proposing solutions for rare attack detection problems is urgent and important. 2) A new activation function, SegRelu, for convolution neural networks (CNNs) is proposed in order to improve the quality of classification of rare types of attacks. Experimental results show that SegRelu is positive compared to other activation functions. 3) A new semi-supervised density peak clustering method that improves clustering results with a small set of constraints expressed as must-link and cannot-link is proposed. This approach conbines constraints and a k-nearest neighbor graph to filter out peaks and find the center for each cluster. Constraints are used to support label assignment during the clustering procedure. The efficacy of this method is demonstrated through experiments on data sets from UCI and benchmarked against comtemporary semi-supervised clustering techniques. Publications Publications of the project are: 01 SCIE article, 01 international conference paper, and 01 national journal. Education Instructing 01 Computer Science student to sucessfully defend a graduation project.
Novelty and actuality and scientific meaningfulness of the results	- Proposed a new activation function for convolutional neural networks to increase the ability to classify rare attacks. - Proposed an anomaly detection method based on clustering (Density peaks clustering with constraints), which allows detecting anomalies for data with different structures and densities as well as any size.
Products of the project	- Publications: 01 SCIE article, 01 international conference paper, and 01 national journal, details are as follows: [1] Vu, V., Bui, T. T., Nguyen, T. L., Tran, D., Do, H., Vu, V., & Avdoshin, S. M. (Aug. 2023). Constrained Density Peak Clustering. International Journal of Data Warehousing and Mining (IJDWM), 19(1), 1-19. (SCIE, ISSN: 1548-3924). [2]Viet-Thang Vu, Thanh Quyen Bui Thi, Hong-Seng Gan, Viet-Vu Vu, Do Manh Quang, Vu Thanh Duc, Dinh-Lam Pham. Activation functions for deep learning: an application for rare attack detection in wireless local area network (WLAN). 25th International Conference on Advanced Communication Technology (ICACT), Pyeongchang, Korea, Republic of, 2023, pp. 59-64. [3] Thang, V. V., Pantiukhin, D. V. ., Quyen, B. T. T. ., & Vu, V. V. (Dec. 2023). A review of neural networks for rare intrusions detection in wireless networks. Journal of Science and Technology on Information Security, 3(20), 23-34. (ISSN: 2615-9570). - Patents: - Products: A full report of the project. - Education: Instructing 01 Computer Science student to successfully defend the graduation project.
Research region	Propose to apply the research results of the project to teach the subject "Cyber Security" at the Faculty of Information Technology, Hanoi University of Industry.
Recommendations	Continue to cooperate with the research team of the partners (Department “Software Engineering” - Higher School of Economics National Research University, Russian Federation) on the application and implementation of proposed solutions in solving specific practical problems.
Images of project