Have a personal or library account? Click to login
Research on Construction Site Safety Q&A System Based on BERT Cover
By: Ang Li and  Jianguo Wang  
Open Access
|Dec 2024

Figures & Tables

Figure 1.

Unidirectional context build representation incrementally.
Unidirectional context build representation incrementally.

Figure 2.

Bidirectional context words can “see themselves”.
Bidirectional context words can “see themselves”.

Figure 3.

BERT model input diagram.
BERT model input diagram.

Figure 4.

Model architecture of BERT-based Q&A system.
Model architecture of BERT-based Q&A system.

Figure 5.

Effect of BERT with increasing model size.
Effect of BERT with increasing model size.

Figure 6.

Graphical interface of the construction site Q&A system.
Graphical interface of the construction site Q&A system.

Number of samples in the data set

Data Set NumberDataset Name
training setvalidation settest set
677004775099759975

Experimental environment configuration parameters

Experimental ParametersRetrieve A Value
Learning rate2e-5
Batch Size16
Num of epoch20
Length of Maxseq128

Comparative effects of different baseline models

ModellingEvaluation Metrics
P/%R/%F1/%
LSTM72.7268.6370.61
Text-CNN73.4070.2371.78
BERT80.580.9681.65
Language: English
Page range: 75 - 83
Published on: Dec 31, 2024
Published by: Xi’an Technological University
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2024 Ang Li, Jianguo Wang, published by Xi’an Technological University
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.