Automatic plane geometry proof question answering method and system

Abstract

The invention discloses an automatic plane geometry proof question answering method and system. The method comprises the following steps of: question input; question understanding: carrying out word segmentation and part-of-speech labelling on a recognized plane geometry proof question text, segmenting a question into n simple clauses and standardizing the clauses; labelling all the geometric entity words in the clauses; expressing each geometric element word as a binary group and expressing each geometric relationship word as a binary group; extracting geometric relationship by adoptions of a syntax-semantic model and forming a binary geometric proposition set by all the geometric relationships; machine solution: reasoning the binary geometric question set to obtain a geometric evidence chain of the question; and quasi-man answer generation: carrying out continuous search through a forward chain according to the evidence chain and a known condition until a conclusion is deduced, so as to generate quasi-man expression. Compared with existing plane geometry proof systems, the method and system are capable of greatly improving the automation degree of answer systems.

Description

technical field [0001] The invention relates to the field of natural language text understanding and machine reasoning in artificial intelligence, in particular to a method and system for automatically answering plane geometry proof questions. Background technique [0002] The two main parts in the automatic solution of elementary mathematics are the automatic solution of the plane geometry theorem proof and the algebra type problem. Since the establishment of the field of artificial intelligence in the 1950s, automatic answering has become an active research problem many times. Driven by the technological progress in related fields and the demand for intelligent education, it has become a research hotspot again in recent years. The geometric expert system, answering system, and online answering system provided by the service have been put into practical application one after another. This patent mainly invents a method, algorithm and system for automatic answering of eleme...

AI Technical Summary

Problems solved by technology

First of all, the existing automatic answering methods come from many experts and scholars in different fields. The established algorithms are often independent of each other, and there are great differences in answering efficiency and adaptability. system

Secondly, the existing topic understanding often uses shallow analysis techniques such as rules or symbolic logic, and lacks the effective use of deep semantic knowledge such as syntax and semantics, resulting in insufficient ability to understand topics

Finally, the type and range of questions that the current automatic answering system can answer are very limited, and the degree of automation of question input and question understanding is not high, which cannot meet the requirements of large-scale promotion and application

Images