Focusing on differences! Sample framework enhances semantic textual similarity with external knowledge

doi:10.1016/j.eswa.2024.124462

	Focusing on differences! Sample framework enhances semantic textual similarity with external knowledge
	Feng, Jianzhou 1; Liu, Junxin 1; Gu, Chenghan 1; Qi, Haotian 1; Ren, Zhongcan 1; Xu, Kehan 1; Wang, Yuanzhuo 2
	2024-12-01
发表期刊	EXPERT SYSTEMS WITH APPLICATIONS
ISSN	0957-4174
卷号	255 页码:10
摘要	Recently, the widespread application of pre-trained language models (PLMs) such as BERT and RoBERTa has significantly enhanced the performance of tasks related to text semantic similarity. However, methods solely based on PLMs inadequately account for the differential information between sentence pairs, thus underestimating the importance of this information in sentence matching. In this paper, we propose the enriching Differential information with External Knowledge framework (DEK), an approach that explicitly extracts differential information and enriches semantics using external knowledge. Specifically, we devise a module for extracting differential words from sentence pairs, obtain synonyms of differential words from WordNet, and construct a differential information graph. We employ Graph Convolutional Networks (GCNs) to extract features from this graph and subsequently integrate this information into sentence embeddings. In this work, we demonstrate that incorporating differential information enables PLMs-based methods to better focus on the differing aspects of sentences. Moreover, DEK seamlessly adapts to contrastive learning of sentence embeddings models, including SimCSE and PromptBert, among others. Comparing to baseline, our method has improved spearman correlation between 0.22 and 0.64, yielding competitive results in the experiments.
关键词	Semantic similarity task Differential information Sentence embeddings
DOI	10.1016/j.eswa.2024.124462
收录类别	SCI
语种	英语
WOS研究方向	Computer Science ; Engineering ; Operations Research & Management Science
WOS类目	Computer Science, Artificial Intelligence ; Engineering, Electrical & Electronic ; Operations Research & Management Science
WOS记录号	WOS:001345526500001
出版者	PERGAMON-ELSEVIER SCIENCE LTD
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/39510
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Liu, Junxin
作者单位	1.Yanshan Univ, Sch Informat Sci & Engn, Qinhuangdao 066000, Hebei, Peoples R China 2.Chinese Acad Sci, Inst Comp technol, Beijing 100086, Peoples R China
推荐引用方式 GB/T 7714	Feng, Jianzhou,Liu, Junxin,Gu, Chenghan,et al. Focusing on differences! Sample framework enhances semantic textual similarity with external knowledge[J]. EXPERT SYSTEMS WITH APPLICATIONS,2024,255:10.
APA	Feng, Jianzhou.,Liu, Junxin.,Gu, Chenghan.,Qi, Haotian.,Ren, Zhongcan.,...&Wang, Yuanzhuo.(2024).Focusing on differences! Sample framework enhances semantic textual similarity with external knowledge.EXPERT SYSTEMS WITH APPLICATIONS,255,10.
MLA	Feng, Jianzhou,et al."Focusing on differences! Sample framework enhances semantic textual similarity with external knowledge".EXPERT SYSTEMS WITH APPLICATIONS 255(2024):10.