Proskurin D. Information technology of quality assessment generators of sequences of pseudo-random numbers based on machine learning

The information technology proposed in the dissertation is based on the use of modern methods of machine learning, in particular hybrid and convolutional neural networks, which allows to significantly increase the accuracy and speed of assessing the quality of PVC generators even under the conditions of a limited amount of input data. The use of hybrid models allows combining the advantages of different types of neural networks, which provides more effective detection of patterns in data and improves the quality of predictions. In particular, the use of convolutional neural networks allows the analysis of local patterns in sequences, while recurrent neural networks are effective for analyzing sequences with long-term dependencies. The model for predicting the next PVC sequence has been improved, which, due to the use of a hybrid neural network and a limited amount of input data for training, allows predicting next sequences for low-quality PVC generators. The method for evaluating the quality of PVC sequences has received further development, which, using a one-dimensional recurrent neural network and datasets formed by various generators of PVC, allows to quickly evaluate the quality of generators for cryptographic and other applications in the field of computer science. The practical significance of the dissertation lies in the possibility of applying the obtained results in real conditions, where access to a large amount of data is limited, and the requirements for reliability and security are extremely high. For example, the developed models and technologies can be used in the field of mobile communications, in particular to ensure the security of LTE/5G/6G networks, as well as in the field of critical infrastructure protection, where information security is a matter of national importance. In addition, the research findings may find applications in many other industries, including the financial sector, public administration, and the military, where the quality and randomness of PVC generators are critical. The results of the study were implemented in the educational process at the Department of Computer Information Technologies of the National Aviation University, as well as in the research work carried out within the Scientific Research Laboratory for Combating Cyber Threats in the Aviation Industry. In addition, the practical results were implemented in the activities of the Main Directorate of Intelligence of the Ministry of Defense of Ukraine, which emphasizes the significance of the research for national security.