.png){kind=icon}
返回
SEASum: Syntax-Enriched Abstractive Summarization
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2022.116819.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Compared to traditional RNN-based models, abstractive summarization systems based on Pre-trained LanguageModels (PTMs) achieve dramatic improvements in readability. Thus, in the field of abstractive summarization,more attention should be devoted to the faithfulness issue that predicted summaries are not factuallyconsistent with source texts. To alleviate this disadvantage, we propose a novelSyntax-EnrichedAbstractiveSummarization (SEASum) framework, which utilizes graph attention networks (GATs) to introduce syntacticfeatures of source texts to generate faithful summaries. In the SEASum framework, the PTM-based semanticencoder encodes word sequence, while the GAT-based syntactic encoder captures explicit syntax, i.e., part-of-speech tags, parse trees, and dependency-based relative positions of source documents. A feature fusionmodule is introduced to incorporate encoded syntactic features into the summarization framework. Basedon the proposed SEASum framework, we develop two summarization models: 1) parallel SEASum model,in which the semantic encoder and syntactic encoder work in parallel, a multi-head attention modulefused two-stream features for the following decoding process; 2) cascaded SEASum model, which takescontextual word embeddings from semantic encoder as node embeddings for the syntactic encoder and employshighway networks to regulate information flow. Experimental results on CNN/DailyMail and Reddit-TIFU(short) datasets show our parallel SEASum model and cascaded SEASum model outperform state-of-the-artabstractive summarization approaches in the faithfulness measurement. The results also demonstrate thatcascaded SEASum model is more effective than parallel SEASum model in boosting faithfulness.
Keyword:
Abstractive summarization
Syntax-enriched summarization
Pre-trained language model
Graph neural networks
Deep learning
期刊
IF:
7.5
论文数:
3.0W
被引数:
10.2W
机构
引用论文
Effect of colorspace transformation,the illuminance component,and color modeling on skin detection
OAAI
OAAI分享收藏分享收藏The RNA-binding protein SRSF3 has an essential role in megakaryocyte maturation and platelet production.png)
Blood
OAAI
BloodIF0
OAAI分享收藏
查看更多内容