Abstract
In the rapid development of various technologies at the present stage, representative artificial intelligence technology has developed more prominently. Therefore, it has been widely applied in various social service areas. The application of artificial intelligence technology in tax consultation can optimize the application scenarios and update the application mode, thus further improving the efficiency and quality of tax data inquiry. In this paper, we propose a novel model, named RDN-MESIM, for paraphrase identification tasks in the tax consulting area. The main contribution of this work is designing the RNN-Dense network and modifying the original ESIM to adapt to the RDN structure. The results demonstrate that RDN-MESIM obtained a better performance as compared to other existing relevant models and archived the highest accuracy, of up to 97.63%.
Highlights
With the rapid development of computer technology and the overall popularization of the Internet, the global digital information storage capacity has been growing explosively
All parameters in each model have been set based on their original setting, and they have been fine-tuned into better performance on our tax question pairs’ data. e results show that RDN-MESIM outperformed other models
1,933,570 3,257,602 4,977,922 7,094,530 9,607,426 layers. ese three layers are Encoding Layer (EL), Local Inference Layer (LIL), and Inference Composition Layer (ICL). e dense block and transition layer pair size are set into 2, which will be modified in the following table. e experiments’ results show that the model achieved better performance when our RDN structure is only added into EL of ESIM as compared to other positions of ESIM
Summary
With the rapid development of computer technology and the overall popularization of the Internet, the global digital information storage capacity has been growing explosively. Ese text data are sparse, real-time, nonnormative, and other characteristics, resulting in the manual processing of this massive text information which is extremely difficult. E paraphrase identification (PI) for user-generated noisy text is an important task in natural language processing, for example, question answering, semantic disambiguation, text summarization, information extraction, and recommendation systems. Sentences with the same meaning are called paraphrase pairs, and sentences with different meanings are called nonparaphrase pairs [1] It has been considered for determining different linguistic expressions with similar meanings. In other words, they are annotated with the binary classification task: given two sentences (S1 and S2), the result of measuring whether they have the same represented intention will be defined as 0 or 1 [2]. Deep learning has shown its powerful advantages
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
