Speech semantic recognition method, system, device and medium

Abstract

The present invention provides a voice semantic recognition method, system, equipment and medium, including: acquiring the user's voice signal; using automatic voice recognition technology to convert the voice signal into text data; judging whether the text data conforms to the preset sentence pattern, if so , then respond according to the voice signal in the text data; if not, use natural language processing to analyze the text data to obtain named entities, and perform grammatical analysis to the text data to obtain intention keywords, according to the intention keywords and Name the relationship between entities and get the result of intent recognition. By combining the named entities and intent keywords in the voice information in the text data, the present invention can accurately understand the user's intent when the user does not interact with the speech system, thereby improving the accuracy of semantic recognition and the user's experience .

Description

technical field [0001] The invention belongs to the technical field of voice recognition, and in particular relates to a voice semantic recognition method, system, equipment and medium. Background technique [0002] At present, speech semantic recognition technology is more and more widely used in electronic equipment. Among them, the speech semantic recognition technology uses acoustic and linguistic methods to train the acoustic model in advance, and the acoustic model performs speech recognition on the speech data. In actual products, speech semantic recognition technology will be divided into two stages, namely speech recognition (ASR) and natural language processing (NLP); where speech recognition recognizes audio data into text data, and text data is used as the input of natural speech processing Data; in the NLP process, word slots are extracted according to the set thesaurus, and finally the effect of recognizing audio data is achieved. [0003] However, the tradit...

AI Technical Summary

Problems solved by technology

[0003] However, the traditional voice function is not as intelligent as a real person. The specific speaking skills and usage methods need to be learned by the user to use the voice function better, and the user is not willing to spend time and energy reading the manual, even if they are willing to read the manual. , many words are difficult to remember

Images