Intelligent simulation interview method and device, computer equipment and storage medium

By using voice acquisition, real-time speech recognition and speaker separation, combined with intent recognition and question classification, structured or open-ended answers are generated. The system is optimized using user feedback, which solves the problems of insufficient dynamic response and low speech recognition accuracy in existing intelligent mock interview systems, and achieves an efficient and personalized mock interview experience.

CN121144459APending Publication Date: 2025-12-16SHENZHEN KINGSUN SCIENCE & TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511274952.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-08
Publication Date
2025-12-16

AI Technical Summary

Technical Problem

Existing intelligent mock interview systems lack dynamic response capabilities, are unable to conduct multiple rounds of in-depth follow-up questions, and have low speech recognition accuracy, failing to distinguish the speaker. This results in a large gap between the simulated environment and the real interview scenario, leading to a poor user experience.

Method used

By collecting voice data, performing real-time speech recognition and speaker separation, and combining intent recognition and question classification, the system selects a knowledge graph or a large language model to generate answers, optimizes the system based on user feedback, and generates an interview performance evaluation report.

Benefits of technology

It enhances the realism and interaction efficiency of mock interviews, continuously optimizes system performance through a data-driven feedback mechanism, and provides a personalized interview experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121144459A_ABST
    Figure CN121144459A_ABST
Patent Text Reader

Abstract

The invention discloses an intelligent simulation interview method and device, computer equipment and a storage medium. The method comprises the steps that voice data of a virtual interviewer and voice data of a job seeker are collected, speaker separation is carried out, and first text data corresponding to the virtual interviewer and second text data corresponding to the job seeker are obtained; performing intention recognition and question classification on the first text data, and distinguishing technical questions from non-technical questions; if the question type of the technical question is a structured question, retrieving a knowledge graph to return a reference answer with a code example, and if the question type of the technical question is an open question, calling a large language model to generate a reference answer conforming to an STAR principle; and returning the reference answer and an interview performance evaluation report generated based on the second text data of the job seeker to the client. According to the invention, a plurality of links such as voice acquisition, speaker separation, intention recognition, question classification, multi-channel answer generation and real-time display are combined, a closed-loop intelligent interview simulation system is constructed, and the sense of reality of simulation interview is improved.
Need to check novelty before this filing date? Find Prior Art