A Deep Learning Approach for Classifying Vulnerability Descriptions Using Self Attention Based Neural Network

Cyber threat intelligence (CTI) refers to essential knowledge used by organizations prevent or mitigate against cyber attacks. Vulnerability databases such as CVE and NVD are crucial intelligence, but also provide information leveraged in hundreds of security products worldwide. However, previous studies have shown that these vulnerability sometimes contain errors inconsistencies which be manually checked professionals. Such could threaten the integrity hamper attack mitigation efforts. Hence, assist community with more accurate time-saving validation data, we propose an automated classification system based on deep learning. Our proposed utilizes a self-attention neural network (SA-DNN) model text mining approach identify category from description contained within report. The performance SA-DNN-based is evaluated using 134,091 reports details website.The experiments performed demonstrates effectiveness our approach, shows SA-DNN outperforms SVM other learning methods i.e. CNN-LSTM graph convolutional networks.