Research on Chinese short text semantic matching based on lightweight ERNIE
Citations Over Time
Abstract
Semantic matching of short texts is a hot research problem in the field of natural language processing, which has a wide range of application needs in information retrieval, dialogue system, text paraphrase questions, and question answering system etc. Short texts have the characteristics of less information and lack of contextual background. The existing semantic matching methods for short texts also generally have the problem of low matching accuracy. With the speedy development of deep learning technology, various models of deep learning have been extensively used in natural language processing and achieved good results. In this paper, we design a FERNIE model based on a lightweight pre-training model ERNIE3.0-medium of BERT architecture, which integrates low-level features and high-level features. Experiments on several datasets show that the FERNIE model has good results in short text semantic matching, and the accuracy is further improved compared to ERNIE3.0-medium.
Related Papers
- IITD-IBMIRL System for Question Answering Using Pattern Matching, Semantic Type and Semantic Category Recognition.(2007)
- → QuesBELM: A BERT based Ensemble Language Model for Natural Questions(2020)3 cited
- → Testing Paraphrase Models on Recognising Sentence Pairs at Different Degrees of Semantic Overlap(2023)1 cited
- → Find the Best to Share(2006)
- → A Language Tutoring Tool based on AI and Paraphrase Detection(2021)