Named Entity Recognition for Terahertz Domain Knowledge Graph based on Albert-BiLSTM-CRF
Xiao Zhang, Chuanzhen Li, Huaichang Du · 2020 IEEE 4th Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · 2020
Named entity recognition is a vital part of the question answering system. The current methods for identifying named entities are mostly for short entities. In our terahertz domain question answering system, there are both long and short entities in questions. In this paper, an entity recognition method based on Albert-BiLSTM-CRF is applied to recognize long entity completely. Specifically, a pre-trained model of deep bidirectional representation is applied to fully understand the semantics of the entire sentence. The experimental results on the terahertz domain dataset show that the proposed method improves the accuracy of long domain entity recognition and improves the performance of the terahertz domain question answering system.