MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction.
Considering that the traditional deep learning event extraction method ignores the correlation between word features and sequence information, it cannot fully explore the hidden associations between events and events and between events and primary attributes. To solve these problems, we developed a...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2024-01-01
|
| Series: | PLoS ONE |
| Online Access: | https://doi.org/10.1371/journal.pone.0306673 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850049918856593408 |
|---|---|
| author | Guangwei Zhang Fei Xie Lei Yu |
| author_facet | Guangwei Zhang Fei Xie Lei Yu |
| author_sort | Guangwei Zhang |
| collection | DOAJ |
| description | Considering that the traditional deep learning event extraction method ignores the correlation between word features and sequence information, it cannot fully explore the hidden associations between events and events and between events and primary attributes. To solve these problems, we developed a new framework for event extraction called the masked attention-guided dynamic graph aggregation network. On the one hand, to obtain effective word representation and sequence representation, an interaction and complementary relationship are established between word vectors and character vectors. At the same time, a squeeze layer is introduced in the bidirectional independent recurrent unit to model the sentence sequence from both positive and negative directions while retaining the local spatial details to the maximum extent and establishing practical long-term dependencies and rich global context representations. On the other hand, the designed masked attention mechanism can effectively balance the word vector features and sequence semantics and refine these features. The designed dynamic graph aggregation module establishes effective connections between events and events, and between events and essential attributes, strengthens the interactivity and association between them, and realizes feature transfer and aggregation on graph nodes in the neighborhood through dynamic strategies to improve the performance of event extraction. We designed a reconstructed weighted loss function to supervise and adjust each module individually to ensure the optimal feature representation. Finally, the proposed MaskDGNets framework is evaluated on two baseline datasets, DuEE and CCKS2020. It demonstrates its robustness and event extraction performance, with F1 of 81.443% and 87.382%, respectively. |
| format | Article |
| id | doaj-art-8c339e66f7694d32ae09f07bd233dbb9 |
| institution | DOAJ |
| issn | 1932-6203 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS ONE |
| spelling | doaj-art-8c339e66f7694d32ae09f07bd233dbb92025-08-20T02:53:37ZengPublic Library of Science (PLoS)PLoS ONE1932-62032024-01-011911e030667310.1371/journal.pone.0306673MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction.Guangwei ZhangFei XieLei YuConsidering that the traditional deep learning event extraction method ignores the correlation between word features and sequence information, it cannot fully explore the hidden associations between events and events and between events and primary attributes. To solve these problems, we developed a new framework for event extraction called the masked attention-guided dynamic graph aggregation network. On the one hand, to obtain effective word representation and sequence representation, an interaction and complementary relationship are established between word vectors and character vectors. At the same time, a squeeze layer is introduced in the bidirectional independent recurrent unit to model the sentence sequence from both positive and negative directions while retaining the local spatial details to the maximum extent and establishing practical long-term dependencies and rich global context representations. On the other hand, the designed masked attention mechanism can effectively balance the word vector features and sequence semantics and refine these features. The designed dynamic graph aggregation module establishes effective connections between events and events, and between events and essential attributes, strengthens the interactivity and association between them, and realizes feature transfer and aggregation on graph nodes in the neighborhood through dynamic strategies to improve the performance of event extraction. We designed a reconstructed weighted loss function to supervise and adjust each module individually to ensure the optimal feature representation. Finally, the proposed MaskDGNets framework is evaluated on two baseline datasets, DuEE and CCKS2020. It demonstrates its robustness and event extraction performance, with F1 of 81.443% and 87.382%, respectively.https://doi.org/10.1371/journal.pone.0306673 |
| spellingShingle | Guangwei Zhang Fei Xie Lei Yu MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction. PLoS ONE |
| title | MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction. |
| title_full | MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction. |
| title_fullStr | MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction. |
| title_full_unstemmed | MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction. |
| title_short | MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction. |
| title_sort | maskdgnets masked attention guided dynamic graph aggregation network for event extraction |
| url | https://doi.org/10.1371/journal.pone.0306673 |
| work_keys_str_mv | AT guangweizhang maskdgnetsmaskedattentionguideddynamicgraphaggregationnetworkforeventextraction AT feixie maskdgnetsmaskedattentionguideddynamicgraphaggregationnetworkforeventextraction AT leiyu maskdgnetsmaskedattentionguideddynamicgraphaggregationnetworkforeventextraction |