Using neuroevolution methods for automating optimization of cyber security algorithms in cognitive information centers

Background. Cognitive data centers such as smart cities and advanced data centers are increasingly becoming targets of cyber threats. Traditional cybersecurity measures often fail to keep pace with the evolution of these threats. Materials and methods. This paper presents a security model for cognitive information centers using neuroevolution, a method combining neural networks and genetic algorithms. The model uses neuroevolution to create adaptable cybersecurity models that can learn in response to a wide range of cyber threats, including malware, phishing, and denial of service (DoS) attacks. Results. Key performance metrics such as detection accuracy, false positive rate, and response time were examined. The analysis showed that the model can effectively learn and adapt to cyber threats, providing a reliable basis for protecting cognitive information centers. In addition, possibilities for optimizing the model are considered, trends in training losses and the distribution of neural network parameters are studied. Conclusions. The findings suggest that neuroevolution is a promising approach to cybersecurity that provides flexibility and adaptability in the face of a rapidly changing threat landscape.