An adaptive language training method and platform

Abstract

The invention discloses an adaptive language training method and platform to realize the evaluation of any voice information input by a user in a selected scene. The method includes: the scheduling module calls the voice recognition module to perform voice recognition on any voice information input by the user in the selected scene; the voice recognition module recognizes any voice information, and inputs the recognition result to the evaluation module; the evaluation module conducts semantic standardization , the evaluation of content completeness and content accuracy, weighting the evaluation results of semantic standardization, content completeness and content accuracy to obtain a comprehensive evaluation result; the scheduling module feeds back the recognition results and comprehensive evaluation results to the user. The invention enables the system to obtain reference answer texts and expected keyword texts when evaluating user expressions, can objectively and accurately reflect the user's language ability, and is beneficial to improving the user's language expression ability.

Description

Technical field [0001] The invention relates to the field of speech recognition, in particular to an adaptive language training method and platform. Background technique [0002] With the development of society, language communication between people is playing an increasingly important role in all aspects of social life, and many issues in the process of language communication have also received increasing attention. Take children with speech and speech impairment as an example. In recent years, relevant institutions have conducted surveys on tens of thousands of students in Beijing and other places and found that the incidence of speech and speech impairment children is 6.42% in ordinary primary schools, 9.24% in schools for the blind, and 30.10% in schools for the mentally handicapped. The incidence of obstacles is relatively high, and these children need professional speech rehabilitation training; for example, police stations in Hong Kong and other places have carried out lan...

AI Technical Summary

Problems solved by technology

[0005] 1) When evaluating a user statement, the text of the user statement must be known in advance, otherwise it cannot be evaluated

The user must express the language according to the text predefined by the system. If the user modifies or repeats the expression inconsistent with the predefined text, the accuracy of the system's evaluation of the user's expression will drop sharply;

[0006] 2) It is difficult to achieve the purpose of comprehensive training of users' ability to understand, express and use language

However, the above-mentioned existing reading-aloud language training limits the user's expression content, so it is impossible to carry out targeted training on the user's understanding, expression and language use ability;

[0007] 3) There is no evaluation of grammar standardization

The above-mentioned existing reading-aloud language expression training only stays at the phoneme and acoustic level, and users cannot express spontaneously, so it is impossible to evaluate the grammatical standardization of user expressions

Images