Tax knowledge middle platform based on natural language understanding and construction method of platform

Abstract

The invention discloses a tax knowledge middle platform based on natural language understanding and a construction method of the platform. The tax knowledge middle platform comprises a knowledge alignment module, a knowledge understanding module and a knowledge interaction module. When the middle platform is constructed, the three modules are respectively constructed by comprehensively utilizing natural language processing technologies such as a knowledge graph, search and recommendation. The platform can be used as an entrance for knowledge query in the tax policy field, efficient association, updating, retrieval and learning of multi-source tax knowledge are achieved, the platform is suitable for tax-related workers and can also be used as a platform for tax payment individuals or enterprise related personnel to learn tax policy knowledge, and a question-answering system in the tax policy field can also be constructed based on the platform.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to a natural language understanding-based taxation knowledge center and its construction method. Background technique [0002] Compared with knowledge in other fields, tax policies and regulations have the complex characteristics of multiple sources of knowledge, strong policy, strong professionalism, irregular updates, and strong regional characteristics, which brings a lot of inconvenience to tax-related staff. Specifically, on the one hand, tax policies will be temporarily adjusted with the emergence of emergencies such as the new crown epidemic. Tax service workers need to organize and learn relevant policies and regulations in a timely manner to better serve taxpayers. There are three difficulties in the study and search of tax policies and regulations: the contextual relevance of policy documents and changes in local knowledge need to be sorted out manually, and even the...

AI Technical Summary

Problems solved by technology

There are three difficulties in the study and search of tax policies and regulations: the contextual relevance of policy documents and changes in local knowledge need to be sorted out manually, and even the staff need rich practical experience; the diversity of knowledge sources makes knowledge sorting difficult, It is often necessary to go to multiple websites to view relevant files, which is time-consuming and laborious; there are inefficiencies in the retrieval and search of specific knowledge, and often the target file cannot be found after multiple searches

Images