HRTC: A Triplet Joint Extraction Model Based on Cyber Threat Intelligence

HRTC: A Triplet Joint Extraction Model Based on Cyber Threat Intelligence

연구 분야: Safety

학회: International Conference on Knowledge Science, Engineering and Management

The construction of the network threat intelligence Knowl-edge graph requires a large number of entity relationship triplets, and the open network threat intelligence data set is relatively scarce; Meanwhile, the existing triplet extraction models use traditional pipeline models and cannot share parameters. In response to the above issues, this article constructs a publicly available dataset and proposes the triplet joint extraction model HRTC. HRTC uses XLnet for embedding expression and BiGRU for decoding. The experimental results show that compared with existing joint extraction models, the performance of the baseline model is superior in accuracy, recall, and F1 value. HRTC experiments on universal datasets have shown that it still performs best when dealing with large-scale datasets.