Research on Intelligent Vulnerability Identification Method for Power Information System

With the rapid development of informatization in power systems, the network security threats are increasing day by day. Manual penetration testing is inefficient, while vulnerability scanning tools lack targeting. Therefore, an intelligent, automatic method is urgently needed to identify vulnerabilities. In this paper, an intelligent vulnerability identification method for power information system is proposed. Aiming at the characteristics of power systems, this paper constructs a dedicated vulnerability knowledge database. An end-to-end solution is proposed, extracting features from heterogeneous vulnerability data and modeling them via convolutional and recurrent neural networks to classify Web, system, protocol and firmware vulnerabilities. Experiments show this intelligent solution greatly improves efficiency over manual penetration by promptly discovering vulnerabilities, boosting security of power information system.