id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
fcis-5209	Li, Qingyu; Zhang, Yujun	Improved Text Matching Model Based on BERT	2023	4	.pdf	application/pdf	3088	163	57	With the rapid development of deep learning, a large number of text matching models based on deep learning have emerged, which can be roughly divided into three types: representation based method, interaction based method and pretraining language model based method. However, the text semantic information obtained by fine-tuning specific text matching tasks using BERT model alone is insufficient, Therefore, based on this problem, this paper proposes a method based on pretraining BERT model fusion of BiLSTM and multi-scale CNN to remodel the text matching task, further extract global semantic information and local semantic information to improve the accuracy of text matching.	cache/fcis-5209.pdf	txt/fcis-5209.txt