Low Resource Summarization Model Based on Latent Structural Semantic En-hancement
At present, low-resource summary generation tasks are usually processed by data enhancement or pre-training combined with fine-tuning, which cannot make full use of the latent structural semantic information between the source text and the target summary. For this reason, this paper proposes a low r...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Journal of Computer Engineering and Applications Beijing Co., Ltd., Science Press
2023-08-01
|
| Series: | Jisuanji kexue yu tansuo |
| Subjects: | |
| Online Access: | http://fcst.ceaj.org/fileup/1673-9418/PDF/2205064.pdf |
| _version_ | 1827874811923136512 |
|---|---|
| author | LIU Yu, LIU Xiaoming, LIU Weiguang, YANG Guan, LIU Jie |
| author_facet | LIU Yu, LIU Xiaoming, LIU Weiguang, YANG Guan, LIU Jie |
| author_sort | LIU Yu, LIU Xiaoming, LIU Weiguang, YANG Guan, LIU Jie |
| collection | DOAJ |
| description | At present, low-resource summary generation tasks are usually processed by data enhancement or pre-training combined with fine-tuning, which cannot make full use of the latent structural semantic information between the source text and the target summary. For this reason, this paper proposes a low resource summary model based on latent structural semantic enhancement, which enhances the utilization of structured information in the way of graph structure alignment. First of all, the model obtains the latent semantic features of the source text and prediction summary through the structural feature representation layer. Then, the obtained semantic features are aligned with the latent structured alignment module for node alignment and edge alignment, which helps the model to capture the structured information in the semantic features, thus enhancing the model??s use of structured knowledge. Finally, the model uses the structured feature alignment distance between the source text and the prediction summary as the regular term of target loss to assist the model in optimization. Experiments are performed on a low-resource dataset across six domains. The model achieves an average improvement of 0.58 in ROUGE-1 scores relative to the baseline model. The results show that the model can effectively improve the ability of generating low-resource summaries by using latent structured semantic knowledge. |
| first_indexed | 2024-03-12T16:58:14Z |
| format | Article |
| id | doaj.art-df3f5e86cbc243efbd6e586f580f8d67 |
| institution | Directory Open Access Journal |
| issn | 1673-9418 |
| language | zho |
| last_indexed | 2024-03-12T16:58:14Z |
| publishDate | 2023-08-01 |
| publisher | Journal of Computer Engineering and Applications Beijing Co., Ltd., Science Press |
| record_format | Article |
| series | Jisuanji kexue yu tansuo |
| spelling | doaj.art-df3f5e86cbc243efbd6e586f580f8d672023-08-08T00:55:38ZzhoJournal of Computer Engineering and Applications Beijing Co., Ltd., Science PressJisuanji kexue yu tansuo1673-94182023-08-011781961197310.3778/j.issn.1673-9418.2205064Low Resource Summarization Model Based on Latent Structural Semantic En-hancementLIU Yu, LIU Xiaoming, LIU Weiguang, YANG Guan, LIU Jie01. School of Computer Science, Zhongyuan University of Technology, Zhengzhou 450007, China 2. Henan Key Laboratory on Public Opinion Intelligent Analysis, Zhengzhou 450007, China 3. Software College, Zhongyuan University of Technology, Zhengzhou 450007, China 4. School of Information Science, North China University of Technology, Beijing 100144, China 5. China Language Intelligence Research Center, Beijing 102206, ChinaAt present, low-resource summary generation tasks are usually processed by data enhancement or pre-training combined with fine-tuning, which cannot make full use of the latent structural semantic information between the source text and the target summary. For this reason, this paper proposes a low resource summary model based on latent structural semantic enhancement, which enhances the utilization of structured information in the way of graph structure alignment. First of all, the model obtains the latent semantic features of the source text and prediction summary through the structural feature representation layer. Then, the obtained semantic features are aligned with the latent structured alignment module for node alignment and edge alignment, which helps the model to capture the structured information in the semantic features, thus enhancing the model??s use of structured knowledge. Finally, the model uses the structured feature alignment distance between the source text and the prediction summary as the regular term of target loss to assist the model in optimization. Experiments are performed on a low-resource dataset across six domains. The model achieves an average improvement of 0.58 in ROUGE-1 scores relative to the baseline model. The results show that the model can effectively improve the ability of generating low-resource summaries by using latent structured semantic knowledge.http://fcst.ceaj.org/fileup/1673-9418/PDF/2205064.pdflow resources; structured; semantic features; graph structure |
| spellingShingle | LIU Yu, LIU Xiaoming, LIU Weiguang, YANG Guan, LIU Jie Low Resource Summarization Model Based on Latent Structural Semantic En-hancement Jisuanji kexue yu tansuo low resources; structured; semantic features; graph structure |
| title | Low Resource Summarization Model Based on Latent Structural Semantic En-hancement |
| title_full | Low Resource Summarization Model Based on Latent Structural Semantic En-hancement |
| title_fullStr | Low Resource Summarization Model Based on Latent Structural Semantic En-hancement |
| title_full_unstemmed | Low Resource Summarization Model Based on Latent Structural Semantic En-hancement |
| title_short | Low Resource Summarization Model Based on Latent Structural Semantic En-hancement |
| title_sort | low resource summarization model based on latent structural semantic en hancement |
| topic | low resources; structured; semantic features; graph structure |
| url | http://fcst.ceaj.org/fileup/1673-9418/PDF/2205064.pdf |
| work_keys_str_mv | AT liuyuliuxiaomingliuweiguangyangguanliujie lowresourcesummarizationmodelbasedonlatentstructuralsemanticenhancement |