TY - JOUR
T1 - Chinese sentence semantic matching based on multi-level relevance extraction and aggregation for intelligent human–robot interaction
AU - Lu, Wenpeng
AU - Zhao, Pengyu
AU - Li, Yifeng
AU - Wang, Shoujin
AU - Huang, Heyan
AU - Shi, Shumin
AU - Wu, Hao
PY - 2022/12
Y1 - 2022/12
N2 - With the development of Internet of Things and cloud computing, intelligent question-answering (QA) has brought great convenience to human's daily activities. As one of the core technologies, sentence semantic matching (SSM) plays a critical role in a variety of intelligent QA systems. However, existing SSM methods usually first encode sentences on either character or word level, and then model semantic interactions on sentence level. Consequently, they fail to capture the rich interactions on multi-levels (i.e., character, word and sentence levels). In this paper, we propose Chinese sentence semantic matching based on Multi-level Relevance Extraction and Aggregation (MREA) for intelligent QA. MREA can comprehensively capture and aggregate various semantic relevance on character, word and sentence levels respectively based on multiple attention mechanisms. Extensive experiments on two real-world datasets demonstrate that MREA outperforms the best-performing baselines by 0.5% and 0.89% w.r.t. ACC. and F1 respectively, and achieves comparable performance with BERT-based methods.
AB - With the development of Internet of Things and cloud computing, intelligent question-answering (QA) has brought great convenience to human's daily activities. As one of the core technologies, sentence semantic matching (SSM) plays a critical role in a variety of intelligent QA systems. However, existing SSM methods usually first encode sentences on either character or word level, and then model semantic interactions on sentence level. Consequently, they fail to capture the rich interactions on multi-levels (i.e., character, word and sentence levels). In this paper, we propose Chinese sentence semantic matching based on Multi-level Relevance Extraction and Aggregation (MREA) for intelligent QA. MREA can comprehensively capture and aggregate various semantic relevance on character, word and sentence levels respectively based on multiple attention mechanisms. Extensive experiments on two real-world datasets demonstrate that MREA outperforms the best-performing baselines by 0.5% and 0.89% w.r.t. ACC. and F1 respectively, and achieves comparable performance with BERT-based methods.
KW - Sentence semantic matching
KW - Human–robot interface
KW - Attention mechanism
KW - Feature extraction and aggregation
UR - http://www.scopus.com/inward/record.url?scp=85142156051&partnerID=8YFLogxK
U2 - 10.1016/j.asoc.2022.109795
DO - 10.1016/j.asoc.2022.109795
M3 - Article
AN - SCOPUS:85142156051
SN - 1568-4946
VL - 131
SP - 1
EP - 11
JO - Applied Soft Computing
JF - Applied Soft Computing
M1 - 109795
ER -