Multi-round dialogue intelligent voice interaction system and device

Abstract

The invention discloses a multi-round dialogue intelligent voice interaction system and device. The system comprises a hybrid semantic understanding module, a semantic understanding adaptive module and an automatic dialogue management module. The voice input is converted into a text and input to a hybrid semantic understanding module after being subjected to voice recognition; wherein the hybrid semantic understanding module is used for understanding user intention and extracting corresponding state information, an automatic dialogue management module is used for guiding a dialogue process, outputting dialogue texts and converting the dialogue texts into voice output based on the user intention to realize dialogue, and the semantic understanding self-adaptive module is used for optimized learning of the hybrid semantic understanding module. According to the invention, a plurality of modules such as speech recognition, natural language understanding, natural language generation, speechsynthesis and dialogue management are integrated to form a whole set of multi-round dialogue intelligent speech interaction system which is easy to expand and configure and can be applied to any scene.

Description

technical field [0001] The invention belongs to the field of computer technology, relates to the fields of natural language processing and artificial intelligence, and is a multi-round dialog intelligent voice interaction system. Background technique [0002] How to make computers understand human language has always been a hot research direction in artificial intelligence and natural language processing, and it is also one of the core problems to be solved in the field of modern artificial intelligence. With the application of speech recognition and image recognition technology becoming more and more mature, although deep learning technology is very popular in the field of semantic understanding, there are still very few artificial intelligence products for real spoken dialogue. The common voice robots on the market are mostly voice assistant robots, which are often based on keyword matching. It is not only very simple to recognize and understand the content, but it is also...

AI Technical Summary

Problems solved by technology

[0013] The problem to be solved by the present invention is: the existing intelligent speech recognition technology has a single recognition method, and the recognition effect is not good for spoken language that is not very rigorous in grammar, and cannot recognize complex speech content, or requires a large number of manual annotations in the recognition study, and automatically Insufficient adaptability requires manual intervention to complete data updates. Some solutions seem to implement multiple rounds of dialogue, but the actual recognition process does not recognize the logical relationship between each round of dialogue, only a simple stack of single-round dialogue

Images