<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>Because, the existing named entity recognition models lack the Specificity of the field, and most of their combine the prediction of entity Location and entity category, which results in the accumulation of errors. So, named entity recognition model based on multi-task learning and pointer network is proposed, and innovations are made in the task Construction and domain entity information utilization in the model of Named entity recognition. This model is based on Transformer with multi-Head attention mechanism, and decomposes traditional tasks in entity recognition tasks and entity classification tasks, and carries out multi-tasks Task learning to reduce the accumulation of errors between tasks. Model in this paper also uses the similarity calculation based on the Comprehensive description of the entity category in entity classification Task for better pertinence of the domain entity. Experiments are conducted on public datasets and domain datasets to prove the advancement of the Model.</p>
</abstract>

Because, the existing named entity recognition models lack the Specificity of the field, and most of their combine the prediction of entity Location and entity category, which results in the accumulation of errors. So, named entity recognition model based on multi-task learning and pointer network is proposed, and innovations are made in the task Construction and domain entity information utilization in the model of Named entity recognition. This model is based on Transformer with multi-Head attention mechanism, and decomposes traditional tasks in entity recognition tasks and entity classification tasks, and carries out multi-tasks Task learning to reduce the accumulation of errors between tasks. Model in this paper also uses the similarity calculation based on the Comprehensive description of the entity category in entity classification Task for better pertinence of the domain entity. Experiments are conducted on public datasets and domain datasets to prove the advancement of the Model.

A Named Entity Recognition Model Based on Multi-Task Learning and Cascading Pointer Network

School of Computer Science and Engineering

International Journal of Advanced Network, Monitoring and Controls

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Because, the existing named entity recognition models lack the Specificity of the field, and most of their combine the prediction of entity Location and entity...

Parameter	Value
Optimizer	SGD
Learning rate	5e–6
Activate function	ReLU
Entity category length limit	16
Enter length limit	128
Batch size	6
Deep learning framework	Pytorch
Number of GPUs	1

CMeEE
Model	P (%)	R (%)	F1 (%)
BiLSTM-CRF	56.41	49.52	52.74
BERT-BiLSTM-CRF	68.98	66.25	67.59
BERT-MRC	71.26	69.34	70.29
MTL-NER	73.13	70.68	71.89

MSRA
Model	P (%)	R (%)	F1 (%)
BiLSTM-CRF	87.47	85.23	83.34
BERT-BiLSTM-CRF	95.15	94.85	95.00
BERT-MRC	96.28	95.74	96.01
MTL-NER	97.07	96.43	96.75

OntoNotes4.0
Model	P (%)	R (%)	F1 (%)
BiLSTM-CRF	73.45	60.07	61.71
BERT-BiLSTM-CRF	79.23	79.58	79.40
BERT-MRC	82.49	81.23	81.56
MTL-NER	84.87	82.56	83.70

Entity Category	Comprehensive description of entity category
color	Yellow green blue, environment-friendly color, hue, lightness, saturation and various phenomena of light
Environmental-friendly	Characteristic value, protection, positive evaluation, low carbon, energy conservation and emission reduction, life, agriculture, circular economy, wind and solar power generation
O	General text

A Named Entity Recognition Model Based on Multi-Task Learning and Cascading Pointer Network

Published Online: May 24, 2023

Page range: 52 - 62

DOI: https://doi.org/10.2478/ijanmc-2022-0027

Keywords
Named Entity Recognition, Multi-Tasks Task Learning, Transformer

© 2022 Chaoyang Geng et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Experimental parameters

CMEeE dataset model indicators

MSRA dataset model indicators

OntoNotes4.0 dataset model indicators

Example of comprehensive description of domain entity categories

CLUENER2020 dataset model indicators

CLUENER2020
Model	P (%)	R (%)	F1 (%)
BiLSTM-CRF	67.23	65.42	66.31
BERT-BiLSTM-CRF	77.42	78.15	77.78
BERT-MRC	79.04	80.26	79.65
MTL-NER	82.14	80.79	81.46

A Named Entity Recognition Model Based on Multi-Task Learning and Cascading Pointer Network

Published Online: May 24, 2023

Page range: 52 - 62

DOI: https://doi.org/10.2478/ijanmc-2022-0027

KeywordsNamed Entity Recognition, Multi-Tasks Task Learning, Transformer

© 2022 Chaoyang Geng et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Experimental parameters

CMEeE dataset model indicators

MSRA dataset model indicators

OntoNotes4.0 dataset model indicators

Example of comprehensive description of domain entity categories

CLUENER2020 dataset model indicators

Keywords
Named Entity Recognition, Multi-Tasks Task Learning, Transformer