Research on Named Entity Recognition of Laboratory Safety Knowledge based on Deep Learning

Laboratory safety is an important subject of education safety. In order to strengthen its application and provide the groundwork for knowledge search, the effect of laboratory safety entity recognition is studied using deep learning. This paper proposes to integrate ALBERT into the conventional BiLS...

Full description

Saved in:
Bibliographic Details
Published in2023 4th International Conference on Big Data & Artificial Intelligence & Software Engineering (ICBASE) pp. 225 - 230
Main Authors Han, Cong, Wen, Bin, Wan, Xin
Format Conference Proceeding
LanguageEnglish
Published IEEE 25.08.2023
Subjects
ALBERT
Annotations
BERT
Data models
Deep learning
Knowledge based systems
laboratory safety knowledge
Manuals
Named Entity Recognition
Safety
Text recognition
Online AccessGet full text

Cover

More Information
Summary:Laboratory safety is an important subject of education safety. In order to strengthen its application and provide the groundwork for knowledge search, the effect of laboratory safety entity recognition is studied using deep learning. This paper proposes to integrate ALBERT into the conventional BiLSTM-CRF model to form the ALBERT-BiLSTM-CRF model. The model uses ALBERT as the embedding layer, obtains the character position information through BiLSTM and the optimal sequence is obtained by CRF. The final model increases the recognition effect of complex entities of laboratory safety. Because this field lacks an annotated corpus, this paper uses manual annotation to construct a corpus. Experiments demonstrate that the ALBERT-BiLSTM-CRF model's laboratory safety entity recognition effect is superior to other conventional models with F1 value of 93.24%.
DOI:10.1109/ICBASE59196.2023.10303220