Back to Search
Start Over
UGDAS: Unsupervised graph-network based denoiser for abstractive summarization in biomedical domain.
- Source :
-
Methods . Jul2022, Vol. 203, p160-166. 7p. - Publication Year :
- 2022
-
Abstract
- • A novel model called UGDAS is proposed for the biomedical-domain abstractive summarization task. • The model is consisted of a graph-network based denoiser and an auto-regressive generator, and it is more effective to capture the sentence relationship. • The denoiser utilizes domain knowledge to represent the rich information of the sentence in the biomedical domain. • The ranking process is optimized by the integration of sentence position information. • Our model achieves the state-of-the-art result on a recently-introduced CORD-19 dataset and outperforms the related abstractive models on the PubMed dataset. Abstractive summarization models can generate summary auto-regressively, but the quality is often impacted by the noise in the text. Learning cross-sentence relations is a crucial step in this task and the graph-based network is more effective to capture the sentence relationship. Moreover, knowledge is very important to distinguish the noise of the text in special domain. A novel model structure called UGDAS is proposed in this paper, which combines a sentence-level denoiser based on an unsupervised graph-network and an auto-regressive generator. It utilizes domain knowledge and sentence position information to denoise the original text and further improve the quality of generated summaries. We use the recently-introduced dataset CORD-19 (COVID-19 Open Research Dataset) on text summarization task, which contains large-scale data on coronaviruses. The experimental results show that our model achieves the SOTA (state-of-the-art) result on CORD-19 dataset and outperforms the related baseline models on the PubMed Abstract dataset. [ABSTRACT FROM AUTHOR]
- Subjects :
- *CORONAVIRUSES
*COVID-19
*NOISE
*TASKS
Subjects
Details
- Language :
- English
- ISSN :
- 10462023
- Volume :
- 203
- Database :
- Academic Search Index
- Journal :
- Methods
- Publication Type :
- Academic Journal
- Accession number :
- 157503619
- Full Text :
- https://doi.org/10.1016/j.ymeth.2022.03.012