Search academic texts

Information × Registration Number 0217U000299, 0116U002499 , R & D reports Title Methods for protection against computer attacks based on neural networks, artificial immune systems popup.stage_title Head Sachenko Anatoliy Oleksijovych, Registration Date 26-01-2017 Organization Ternopol National Economic University popup.description2 Object of study - the processing of information in the identification and classification of computer attacks in the information telecommunication networks. Subject of research - the methods of protection against computer attacks in the information telecommunication networks based on artificial neural and immune networks, fuzzy logic and modular arithmetic. Purpose - is to develop a system of protection against computer attacks based on neural and immune networks, fuzzy logic and modular correcting codes to improve the reliability of the detection and classification of computer attacks and improve security. Research methods - methods of system analysis, mathematical statistics, pattern recognition theory, theory of artificial neural networks neyroobchyslen and artificial immune systems, fuzzy logic, modular arithmetic, methods of computer modeling. Results: the analysis of the known methods of protection against cyber attacks; The modified method for constructing the detector detecting computer attacks based on neural networks and artificial immune systems; the method of reducing the dimension of information based on neural networks and deep trust using multi-detector neural network classifier to build hierarchical cyber attack; The generalized architecture intelligent system protection against cyber attacks; Experimental research developed methods and algorithms, which confirmed the reliability of the detection and classification of computer attacks and improve the level of safety; An approach to improving system security protection against computer attacks by realizatsiyiyi neural detectors and FPGA input subsystem decisions based on fuzzy logic Mamdani rules; developed corrective codes based on modular arithmetic, which can effectively correct multiple errors in the data package and increase the reliability and security of data in the information telecommunication networks. Scientific innovation is: A modified method for constructing the detector attacks based on the integration of neural networks and vector quantization artificial immune systems, which allowed using the mechanism of evolution to increase the reliability of detection of computer attacks; improved method of reducing the dimension of information based on neural networks and deep trust using multi-detector neural network classifier to build hierarchical computer attacks that, unlike known, combines the use of deep neural networks to reduce the dimensionality of the input data; union and eliminate conflicts between trained for a certain type of attack neural detectors, thus improving the processing efficiency and classify cyber attacks. Results of research have been tested at international and national conferences. Product Description popup.authors Івасєв Степан Володимирович Биковий Павло Євгенови Головко Володимир Адамович Дорош Віталій Іванович Дубчак Леся Орестівна Загородня Діана Іванівна Карачка Андрій Федорович Комар Мирослав Петрович Кочан Володимир Володимирович Сапожник Григорій Вікторович Скумін Тарас Федорович Цаволик Тарас Григорович Шилінська Інна Федорівна Яцків Василь Васильович popup.nrat_date 2020-04-02 Close