Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet
The third-generation semiconductor materials (TGSMs) is a frontier scientific domain, where researchers need to consult extensive literature for the entity information on materials, devices, preparation methods, and experimental performances, and sort out the complex relations between them. However,...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2022-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9733892/ |
| _version_ | 1811271214811840512 |
|---|---|
| author | Xiaobo Jiang Kun He Borui Yang |
| author_facet | Xiaobo Jiang Kun He Borui Yang |
| author_sort | Xiaobo Jiang |
| collection | DOAJ |
| description | The third-generation semiconductor materials (TGSMs) is a frontier scientific domain, where researchers need to consult extensive literature for the entity information on materials, devices, preparation methods, and experimental performances, and sort out the complex relations between them. However, the explosion of relevant papers has far exceeded researchers’ reading ability. In this article, the TGSM-field automatic information extraction is conducted based on entity recognition (ER) and relation extraction (RE) techniques. First, the corpora used for ER and RE in this field are created. Second, aiming at the complexity of the entities, a neural network using domain knowledge (DKNet) is proposed to improve ER performance. It uses the keyword sequence of each entity type as prior knowledge, adds a dedicated embedding to encode entity categories, then combines prior knowledge and encoded vectors with the context through a gated information fusion module to assist recognition. As for the indicative word dependence problem of entity relations, a multi-aspect attention-based network model (MANet) is proposed to enhance the attention to relation-indicative words, thereby improving the RE performance. Finally, F1 scores of 74.5 and 85.9 were achieved on the created ER and RE test sets, outperforming other advanced models by <inline-formula> <tex-math notation="LaTeX">$3.4~\sim ~10.1$ </tex-math></inline-formula>, which is the best performance of the TGSM-field automatic information extraction. |
| first_indexed | 2024-04-12T22:17:10Z |
| format | Article |
| id | doaj.art-3330347bbeba413f9d01762d3d3b4d4f |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-04-12T22:17:10Z |
| publishDate | 2022-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-3330347bbeba413f9d01762d3d3b4d4f2022-12-22T03:14:30ZengIEEEIEEE Access2169-35362022-01-0110293672937610.1109/ACCESS.2022.31593389733892Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANetXiaobo Jiang0https://orcid.org/0000-0003-4865-8613Kun He1https://orcid.org/0000-0002-2951-6102Borui Yang2School of Electronic and Information Engineering, South China University of Technology, Guangzhou, ChinaSchool of Electronic and Information Engineering, South China University of Technology, Guangzhou, ChinaSchool of Electronic and Information Engineering, South China University of Technology, Guangzhou, ChinaThe third-generation semiconductor materials (TGSMs) is a frontier scientific domain, where researchers need to consult extensive literature for the entity information on materials, devices, preparation methods, and experimental performances, and sort out the complex relations between them. However, the explosion of relevant papers has far exceeded researchers’ reading ability. In this article, the TGSM-field automatic information extraction is conducted based on entity recognition (ER) and relation extraction (RE) techniques. First, the corpora used for ER and RE in this field are created. Second, aiming at the complexity of the entities, a neural network using domain knowledge (DKNet) is proposed to improve ER performance. It uses the keyword sequence of each entity type as prior knowledge, adds a dedicated embedding to encode entity categories, then combines prior knowledge and encoded vectors with the context through a gated information fusion module to assist recognition. As for the indicative word dependence problem of entity relations, a multi-aspect attention-based network model (MANet) is proposed to enhance the attention to relation-indicative words, thereby improving the RE performance. Finally, F1 scores of 74.5 and 85.9 were achieved on the created ER and RE test sets, outperforming other advanced models by <inline-formula> <tex-math notation="LaTeX">$3.4~\sim ~10.1$ </tex-math></inline-formula>, which is the best performance of the TGSM-field automatic information extraction.https://ieeexplore.ieee.org/document/9733892/Automatic information extractionentity recognitionrelation extractionthird-generation semiconductor materialsgated information fusionmulti-aspect attention |
| spellingShingle | Xiaobo Jiang Kun He Borui Yang Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet IEEE Access Automatic information extraction entity recognition relation extraction third-generation semiconductor materials gated information fusion multi-aspect attention |
| title | Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet |
| title_full | Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet |
| title_fullStr | Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet |
| title_full_unstemmed | Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet |
| title_short | Automatic Information Extraction in the Third-Generation Semiconductor Materials Domain Based on DKNet and MANet |
| title_sort | automatic information extraction in the third generation semiconductor materials domain based on dknet and manet |
| topic | Automatic information extraction entity recognition relation extraction third-generation semiconductor materials gated information fusion multi-aspect attention |
| url | https://ieeexplore.ieee.org/document/9733892/ |
| work_keys_str_mv | AT xiaobojiang automaticinformationextractioninthethirdgenerationsemiconductormaterialsdomainbasedondknetandmanet AT kunhe automaticinformationextractioninthethirdgenerationsemiconductormaterialsdomainbasedondknetandmanet AT boruiyang automaticinformationextractioninthethirdgenerationsemiconductormaterialsdomainbasedondknetandmanet |