Dialogue state analysis and auxiliary response method and system based on multi-round voice interaction

By performing speech recognition and natural language understanding on multi-turn voice interaction data, a multi-turn dialogue state representation is constructed, and response generation control parameters are generated. This solves the problem of inconsistent multi-turn dialogue state modeling in existing technologies, realizes the coherence and consistency of dialogue response, and improves the system's controllability and processing efficiency.

CN122245318APending Publication Date: 2026-06-19GUANGDONG HENGQIN SHENSHUI YUNKE DIGITAL TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
GUANGDONG HENGQIN SHENSHUI YUNKE DIGITAL TECHNOLOGY CO LTD
Filing Date
2026-03-12
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

Existing voice interaction systems lack effective state modeling and response generation control in multi-turn dialogues, resulting in inconsistent and disjointed response generation, making it difficult to adapt to complex application scenarios and affecting the stability and controllability of the interaction.

Method used

By performing speech recognition, natural language understanding, and sentiment analysis on multi-turn voice interaction data, semantic, emotional, and temporal features are extracted to construct a multi-turn dialogue state representation. Based on these features, response control parameters are generated to adjust the speech rate, tone, emotional style, and content complexity of the voice response, thereby achieving dynamic modeling of the dialogue state and technical constraints on the response.

Benefits of technology

It improves the continuity and accuracy of dialogue state analysis, enhances the controllability and consistency of the response generation process, reduces the number of invalid responses generated, and improves the system's processing efficiency and scalability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122245318A_ABST
    Figure CN122245318A_ABST
Patent Text Reader

Abstract

This invention discloses a dialogue state analysis and assisted response method based on multi-turn voice interaction. The method includes: acquiring at least one round of voice interaction data between the user and the system, and performing speech recognition processing on the voice interaction data to obtain a corresponding text sequence; based on the text sequence, extracting semantic and temporal features related to the dialogue process, constructing a multi-turn dialogue state representation, and generating dialogue state feature parameters to characterize the current dialogue stage based on the state representation; and selecting at least one response generation control parameter from a preset set of control parameters according to the dialogue state feature parameters to constrain the generation process of subsequent dialogue responses. This invention improves the coherence and consistency of responses and dialogue context, achieves dynamic adaptation and precise control of the response generation process, and ensures the scalability and maintainability of the system structure, facilitating deployment and adjustment in different application scenarios.
Need to check novelty before this filing date? Find Prior Art