Interview simulation method, system and equipment based on knowledge graph and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and particularly provides a knowledge graph-based interview simulation method, system and device and a storage medium, and the method comprises the steps: obtaining a resume document, and extracting key information from the resume document through a keyword extraction technology; retrieving a local graph from a pre-constructed knowledge graph based on the key information; generating a problem based on entities and relationships in the local atlas; receiving answer data for the question, and extracting entities and relationships from the answer data; and matching the extracted entities and relationships with the entities and relationships corresponding to the questions, and generating evaluation data based on a matching result. According to the method, the personalized resume of the interviewer can be deeply analyzed, the specific demand of a target post can be accurately met, and a series of interview problems which are in accordance with reality and are rich in pertinence can be intelligently generated. And the answer quality of the interviewer can be evaluated instantly and objectively.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and specifically relates to an interview simulation method, system, device and storage medium based on knowledge graph. Background Art

[0002] For many job seekers, lack of actual interview experience is a major obstacle they have to face in the job search process. In the fiercely competitive job market, an excellent interview performance can often determine whether a person can successfully obtain the desired job opportunity. However, traditional interview preparation methods often have certain limitations.

[0003] Many job seekers rely on past question banks to prepare and improve their interview skills through repeated practice. Although this method can help them become familiar with the interview process and master some basic answering skills to a certain extent, the questions in the question bank are often limited and it is difficult to cover all possible interview scenarios and questions. In addition, over-reliance on question banks may also cause job seekers to fall into a mechanical answering mode during the interview, lacking flexibility and innovation.

[0004] In addition to question banks, some job seekers also prepare for the interview by learning from other people's interview experiences. Although these experiences may have certain reference value, each person's interview situation is unique, and other people's experiences may not be completely applicable to their own. Moreover, this preparation method lacks an immediate feedback mechanism, so job seekers cannot understand their preparation status in time, let alone make targeted improvements to their shortcomings.

[0005] Therefore, it is particularly important for job seekers to find a more comprehensive, diverse and immediate feedback method for interview preparation. Only in this way can they fully demonstrate their strengths in the interview and win the favor of the employer. Summary of the invention

[0006] In view of the above-mentioned deficiencies in the prior art, the present invention provides an interview simulation method, system, device and storage medium based on a knowledge graph to solve the above-mentioned technical problems.

[0007] In a first aspect, the present invention provides an interview simulation method based on a knowledge graph, comprising: Obtain resume documents and use keyword extraction technology to extract key information from the resume documents; Retrieving a local graph from a pre-built knowledge graph based on the key information; Generate problems based on entities and relations in local graphs; receiving answer data for the question, and extracting entities and relationships from the answer data; The extracted entities and relations are matched with the entities and relations corresponding to the question, and evaluation data is generated based on the matching results.

[0008] In an optional implementation, a resume document is obtained, and key information is extracted from the resume document using a keyword extraction technique, including: Use regular expressions to remove non-alphabetic characters from the text and convert them to lowercase. Define the key information categories to be extracted, including education background, work experience, skills and expertise, and project experience; Use TF-IDF to vectorize the resume text; Analyze the words with high TF-IDF scores and classify them into corresponding key information categories.

[0009] In an optional implementation, the method for constructing the knowledge graph includes: collecting interview data, the interview data comprising questions and answers; Performing clustering processing on the interview data to obtain multiple data sets; Extract entities and relationships from the interview data in the dataset, and clean the extracted entities and relationships; Build a knowledge graph based on the cleaned entities and relationships.

[0010] In an optional implementation, retrieving a local graph from a pre-built knowledge graph based on the key information includes: If the key information includes multiple key entities, semantic similarity recognition is performed on the multiple key entities, and a representative entity is screened out from the multiple key entities based on the recognition result; Build a graph query statement containing all represented entities; Submit the graph query statement to the query engine of the knowledge graph to obtain a local graph.

[0011] In an optional implementation, generating questions based on entities and relationships in a local graph includes: Set up question templates; Substitute some entities and relationships in the local graph into the variables of the question template to form a question.

[0012] In a second aspect, the present invention provides an interview simulation system based on a knowledge graph, comprising: Resume processing module, used to obtain resume documents and extract key information from resume documents using keyword extraction technology; A graph retrieval module, used to retrieve a local graph from a pre-built knowledge graph based on the key information; The question generation module is used to generate questions based on entities and relations in the local graph; An answer analysis module, configured to receive answer data for the question and extract entities and relationships from the answer data; The evaluation generation module is used to match the extracted entities and relations with the entities and relations corresponding to the question, and generate evaluation data based on the matching results.

[0013] In an optional implementation, the resume processing module includes: A regular expression processing unit, used to remove non-alphabetic characters in the text using regular expressions and convert them to lowercase; A category definition unit, used to define the key information categories to be extracted, wherein the key information categories include educational background, work experience, skills and expertise, and project experience; Vector conversion unit, used to vectorize resume text using TF-IDF; The vocabulary analysis unit is used to analyze words with high TF-IDF scores and classify them into corresponding key information categories.

[0014] In an optional implementation, the method for constructing the knowledge graph includes: collecting interview data, the interview data comprising questions and answers; Performing clustering processing on the interview data to obtain multiple data sets; Extract entities and relationships from the interview data in the dataset, and clean the extracted entities and relationships; Build a knowledge graph based on the cleaned entities and relationships.

[0015] In a third aspect, a device is provided, comprising: A memory, used for storing an interview simulation program based on a knowledge graph; A processor is used to implement the steps of the knowledge graph-based interview simulation method provided in the first aspect when executing the knowledge graph-based interview simulation program.

[0016] In a fourth aspect, a computer-readable storage medium is provided, on which a knowledge graph-based interview simulation program is stored. When the knowledge graph-based interview simulation program is executed by a processor, the steps of the knowledge graph-based interview simulation method provided in the first aspect are implemented.

[0017] The beneficial effect of the present invention is that the interview simulation method, system, device and storage medium based on the knowledge graph provided by the present invention can deeply analyze the personalized resume of the interviewee, accurately match the specific needs of the target position, and intelligently generate a series of practical and targeted interview questions. It can also evaluate the quality of the interviewee's answer instantly and objectively. Through advanced natural language processing technology and big data analysis, the system can carefully analyze the interviewee's knowledge reserve, technical experience, language expression, adaptability and other dimensions, and provide instant feedback and constructive opinions. After multiple rounds of simulated questions and answers, the interviewee can not only be familiar with the interview process, but also continuously optimize his answer strategy in actual combat, and improve his interview skills and expressiveness.

[0018] In addition, the invention has a reliable design principle, a simple structure and a very broad application prospect. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 is a schematic flow chart of a method according to an embodiment of the present invention.

[0021] Figure 2 is another schematic flow chart of a method according to an embodiment of the present invention.

[0022] Figure 3 is a schematic block diagram of a system according to an embodiment of the present invention.

[0023] Figure 4 A schematic diagram of the structure of a device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0026] The knowledge graph-based interview simulation method provided in an embodiment of the present invention is executed by a computer device, and accordingly, the knowledge graph-based interview simulation system runs in the computer device.

[0027] Figure 1 is a schematic flow chart of a method according to an embodiment of the present invention. Figure 1 The execution subject can be an interview simulation system based on knowledge graph. According to different requirements, the order of the steps in the flowchart can be changed, and some can be omitted.

[0028] like Figure 1 As shown, the method includes: S1. Obtain a resume document and use keyword extraction technology to extract key information from the resume document.

[0029] First, we need to obtain resume documents from job seekers, which are usually submitted in the form of electronic documents (such as Word, PDF, etc.). Next, the system uses advanced keyword extraction technologies, which may include natural language processing (NLP) algorithms, machine learning models, or deep learning methods, to conduct in-depth analysis of resume documents to extract key information related to job seekers' education background, work experience, skills, project participation, etc. This information will be used in subsequent steps to more accurately retrieve relevant information in the knowledge graph.

[0030] S2. Retrieve a local graph from a pre-built knowledge graph based on the key information.

[0031] Using the key information extracted from the resume, search in a pre-built large knowledge graph. This knowledge graph may contain a lot of industry knowledge, company background, job information, skill requirements, etc. The system retrieves a local graph that is highly relevant to the job seeker's resume by matching the key information with the entities and relationships in the knowledge graph. This local graph will serve as the basis for subsequent question generation.

[0032] S3. Entity and relation generation problem based on local graph.

[0033] Based on the entities and relationships in the retrieved local graph, a series of structured questions are automatically generated. These questions may involve the applicant's specific work experience, skills, project contributions, etc. The question generation process may rely on some preset templates or rules, or it may be dynamically generated through a machine learning model. The purpose of these questions is to help the employer gain a deeper understanding of the applicant's abilities and background.

[0034] S4. Receive answer data for the question, and extract entities and relationships from the answer data.

[0035] Receive the job seeker's answers to the generated questions and extract entities and relationships from them. These entities and relationships may include the job seeker's specific work content, project experience, skill application, etc. The extraction process also relies on natural language processing technology and machine learning models to ensure accuracy and efficiency.

[0036] S5. Match the extracted entities and relations with the entities and relations corresponding to the question, and generate evaluation data based on the matching results.

[0037] The extracted entities and relations are matched with the entities and relations in the local graph used when generating questions. By comparing the answers of job seekers with the information in the knowledge graph, the system can generate a series of evaluation data, such as the skill matching degree of job seekers, the relevance of project experience, etc. These evaluation data will provide valuable references for recruiters and help them make more informed recruitment decisions.

[0038] The entire process is a highly automated and intelligent process that uses advanced natural language processing technology and machine learning models to extract key information from resumes, generate structured questions, receive and process job seekers' answers, and ultimately generate evaluation data to provide recruiters with comprehensive job seeker assessments.

[0039] In an embodiment of the present invention, based on step S1, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner.

[0040] S101. Obtain resume document Receive resume documents submitted by job seekers and support common document formats such as Word, PDF, TXT, etc.

[0041] Pre-process the document, including removing unnecessary headers, footers, watermarks, etc.

[0042] S102. Text cleaning and preprocessing Use regular expressions to remove non-alphabetic characters (such as numbers, punctuation marks, special symbols, etc.) from the text and retain the plain text information.

[0043] The text is lowercase to ensure uniformity and accuracy in subsequent processing.

[0044] S103. Define key information categories Identify the key information categories that need to be extracted, including educational background, work experience, skills expertise, project experience, etc.

[0045] For each category, detailed extraction rules and feature identification are formulated.

[0046] S104. TF-IDF vectorized resume text Use the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm to vectorize the resume text.

[0047] The TF-IDF algorithm can evaluate the importance of a word to a document set or a document in a corpus.

[0048] By calculating the TF-IDF value of each word, we get the vector representation of the resume text.

[0049] S105. Analyze words with high TF-IDF scores Sort the TF-IDF vectors and find the words with higher scores.

[0050] These words are usually closely related to the key information in the resume and are important clues for extracting key information.

[0051] S106. Classify words into corresponding key information categories According to the predefined key information categories and extraction rules, the words with high scores are classified into corresponding categories.

[0052] For each category, all words and phrases related to the category are extracted to form structured key information.

[0053] S107. Result output and verification: Output the extracted key information in a structured format, such as JSON, XML, etc.

[0054] Verify the extraction results to ensure the accuracy and completeness of the information.

[0055] If any anomalies or missing information are found, manual corrections or re-extraction can be carried out.

[0056] In an embodiment of the present invention, based on step S2, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner.

[0057] First, we need to build a knowledge graph. The specific construction methods include: 1. Interview data collection Data sources: Questions and answers recorded during the interview process, including text, audio, video and other forms. Resumes and other relevant materials submitted by job seekers.

[0058] Data collection method: During the interview process, the interviewer or the system records the questions and answers in real time. The resumes and related materials submitted by job seekers are extracted from the recruitment management system or database.

[0059] Data preprocessing: Clean the collected data to remove duplicate, invalid or redundant information. Unify the data format to facilitate subsequent processing and analysis.

[0060] 2. Interview data clustering processing Clustering method: Use unsupervised learning algorithms (such as K-means, DBSCAN, etc.) to cluster interview data. Clustering is performed based on the topic of the question, the keywords of the answer, or the background information of the job seeker.

[0061] Clustering result evaluation: clustering effect is evaluated by calculating the similarity within clusters and the difference between clusters. Clustering algorithm parameters are adjusted according to the evaluation results to optimize clustering results.

[0062] Dataset division: Based on the clustering results, the interview data is divided into multiple datasets, each of which contains information on similar topics or backgrounds.

[0063] 3. Entity and relationship extraction and cleaning Entity and relationship extraction: Use natural language processing (NLP) technology to extract entities and relationships from the interview data in each dataset. Entities include applicants’ names, positions, skills, projects, etc.; relationships include applicants’ matching degree with positions, skill mastery, project experience, etc.

[0064] Cleaning: Clean the extracted entities and relationships to remove noise, redundancy, and erroneous information. Standardize entities, such as unifying naming conventions and correcting spelling errors. Verify and correct relationships to ensure their accuracy and rationality.

[0065] 4. Knowledge graph construction Graph design: Design the nodes and edges of the knowledge graph based on the extracted entities and relationships. Determine the graph's hierarchy and relationship types, such as job seeker-position, job seeker-skill, job seeker-project, etc.

[0066] Graph construction: Use graph databases (such as Neo4j, JanusGraph, etc.) or knowledge graph platforms (such as ApacheJena, Dgraph, etc.) to build knowledge graphs. Import the cleaned entities and relationships into the graph database to form a complete knowledge graph.

[0067] Graph optimization: Optimize the performance of the knowledge graph, such as adding indexes, optimizing query paths, etc. Further verify and correct the entities and relationships in the graph to ensure the accuracy and completeness of the graph.

[0068] Furthermore, the local graph related to the key information is retrieved from the pre-built knowledge graph, including the following process: 1. Key information preprocessing Key information extraction: Extract key information from input data, which may contain multiple key entities.

[0069] Semantic similarity identification: If the key information contains multiple key entities, it is necessary to identify the semantic similarity of these entities. Use natural language processing (NLP) technology, such as word embedding, semantic role labeling, etc., to calculate the semantic similarity between entities. According to the similarity threshold or sorting results, select representative entities from multiple key entities. The representative entity should be able to better summarize or represent the main information of other entities. The specific screening method is to select one from the phrases whose similarity reaches the threshold as the representative entity and discard other words.

[0070] 2. Graph query statement construction Determine the query target: According to business needs, clarify the query target, that is, the specific content of the local graph that needs to be retrieved.

[0071] Construct query statements: Use graph query languages ​​(such as SPARQL, Cypher, etc.) to construct graph query statements that contain all represented entities. The query statement should be able to accurately describe the entities to be retrieved and the relationships between them, as well as the attribute information that may be involved.

[0072] 3. Knowledge Graph Retrieval Query engine selection: Select an appropriate query engine (such as Neo4j's Cypher query engine, Apache Jena's SPARQL query engine, etc.) based on the knowledge graph's storage and query technology.

[0073] Submit query statement: Submit the constructed graph query statement to the query engine of the knowledge graph.

[0074] Obtaining a local graph: The query engine retrieves relevant information from the knowledge graph based on the query statement and returns a local graph containing all representative entities and their relationships.

[0075] 4. Result processing and feedback Result parsing: Parse the query results to extract the required entity, relationship and attribute information.

[0076] Result verification: Verify the parsed results to ensure the accuracy and completeness of the information.

[0077] Result feedback: The verified results are fed back to the business demanders in an appropriate format (such as JSON, XML, charts, etc.) for subsequent analysis and decision-making.

[0078] 5. Solution optimization and improvement Performance optimization: Based on actual application situations, we optimize the query statement construction, query engine selection, and result processing to improve retrieval efficiency and accuracy.

[0079] Functional expansion: Expand the functionality of the solution based on changes in business needs, such as adding query support for complex relationships, introducing intelligent recommendation algorithms, etc.

[0080] Continuous maintenance: Regularly update and maintain the knowledge graph to ensure its timeliness and accuracy. At the same time, continuously monitor and improve the solution to adapt to the ever-changing needs and technical environment.

[0081] Through the implementation of the above solutions, accurate retrieval of key information and efficient extraction of local maps can be achieved, providing strong guarantees for subsequent decision support, data analysis, etc.

[0082] In an embodiment of the present invention, based on step S3, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner.

[0083] 1. Question template settings Template design: Design a series of problem templates based on business needs and application scenarios. These templates should be able to cover all types of entities and relationships that may be involved in the local graph. The template should contain variable placeholders so that specific entities and relationships in the local graph can be substituted later.

[0084] Template classification: Categorize templates into different categories, such as basic information, skills and experience, project experience, etc. This will help you choose the right template based on your specific needs.

[0085] Template optimization: Continuously optimize and improve templates based on actual application conditions, for example, adjust the language style of templates, add new template types, etc.

[0086] 2. Local graph information extraction Entity recognition: Identify all relevant entities from the local graph, which may include job applicant name, position, skills, projects, etc.

[0087] Relationship combing: Sort out the relationships between entities, clarify the position of each entity in the graph and its connection with other entities.

[0088] Information screening: Based on business needs, filter out key entities and relationships needed to generate questions.

[0089] 3. Question generation Template matching: Select a suitable template from the preset template library based on the filtered key entities and relationships.

[0090] Variable substitution: Substitute the selected entities and relationships into the variable placeholders in the template to form specific questions. During the substitution process, pay attention to maintaining the grammatical correctness and semantic clarity of the question.

[0091] Question verification: Verify the generated questions to ensure their accuracy and rationality. If any ambiguity or error is found in the questions, correct them in a timely manner.

[0092] 4. Problem optimization and adjustment Language style adjustment: According to the target audience and application scenarios, adjust the language style of the questions to make them more in line with the audience's acceptance habits and context.

[0093] Question sequence optimization: Sort the generated questions according to their logical relationship and importance to form a reasonable question sequence.

[0094] Question supplement: Supplement and improve the generated questions according to business needs to ensure that the questions can fully cover the key information in the local map.

[0095] 5. Problem output and application Question output: Output the generated questions in an appropriate format (such as text file, database record, etc.) for subsequent use.

[0096] In an embodiment of the present invention, based on step S4, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner.

[0097] 1. Receive answer data Data collection: During the interview process, when job applicants are asked questions, their answers need to be collected. This data may be in the form of text, audio, or video. For answers in text form, they can be directly recorded and stored in the database. For answers in audio or video form, they need to be converted into text form with the help of speech recognition or video analysis technology for subsequent processing.

[0098] Data preprocessing: The collected answer data is preprocessed, including removing irrelevant information (such as redundant polite expressions, insignificant details, etc.), word segmentation, part-of-speech tagging, etc. For the text converted by speech recognition or video analysis, additional verification and correction are required to ensure its accuracy and readability.

[0099] 2. Entity and Relation Extraction Entity recognition: Identify entities related to the question in the preprocessed answer data. These entities may include specific skills mentioned by the job seeker, project names, company names, job titles, etc.

[0100] Use natural language processing techniques such as named entity recognition to automatically identify and extract these entities.

[0101] Relationship extraction: After identifying entities, further analyze the relationships between entities. These relationships may include the association between a job seeker and a project, the degree of match between the skills a job seeker possesses and the job requirements, etc. Use relationship extraction techniques (such as rule-based methods, machine learning-based methods, etc.) to automatically identify and extract these relationships.

[0102] Data integration: Integrate the extracted entities and relationships with the original local graph. If the extracted entities or relationships already exist in the local graph, they will be updated or supplemented; if not, they will be added to the local graph.

[0103] Through this step, the content of the local map can be continuously improved and enriched to provide more comprehensive information support for subsequent analysis and decision-making.

[0104] Data verification: Verify the integrated data to ensure the accuracy and consistency of entities and relationships. This can be achieved through manual review or the use of data verification tools. If errors or inconsistencies are found in the data, make corrections and adjustments in a timely manner.

[0105] By adding a processing step for answer data, the information collected during the interview process can be used more comprehensively to further improve the content of the local map. This will help to conduct a more accurate and comprehensive evaluation of job seekers in the future, improve the efficiency and accuracy of recruitment, and provide a richer and more reliable information basis for subsequent data analysis and mining.

[0106] In an embodiment of the present invention, based on step S5, a possible embodiment is given below to illustrate its specific implementation in a non-limiting manner.

[0107] 1. Entity and relationship matching Matching strategy formulation: Formulate matching strategies for entities and relationships based on business needs. These strategies may include exact matching, fuzzy matching, semantic similarity matching, etc. Exact matching requires that the extracted entities and relationships are completely consistent with the entities and relationships in the question; fuzzy matching allows a certain degree of difference, such as spelling errors in names, synonym replacements, etc.; semantic similarity matching matches based on the semantic features of entities and relationships, which does not require exact consistency, but requires semantic similarity or relevance.

[0108] Matching process implementation: Match the extracted entities and relations with the corresponding entities and relations in the question. This can be achieved by writing a matching algorithm or using an existing matching tool. In the matching process, according to the formulated matching strategy, the extracted entities and relations are compared one by one to determine whether they match the entities and relations in the question.

[0109] Matching result record: record the matching results, including the matched entities and relationships, the matching type (exact match, fuzzy match, semantic similarity match, etc.), the success or failure of the match, etc. These records can be used for subsequent evaluation data generation and problem analysis.

[0110] 2. Evaluation data generation Determine the evaluation dimensions: Determine the dimensions of the evaluation data based on business needs. These dimensions may include matching accuracy, completeness, timeliness, etc. Matching accuracy refers to the correct degree of matching between the extracted entities and relationships and the corresponding entities and relationships in the problem; completeness refers to whether the extracted entities and relationships fully cover the entities and relationships involved in the problem; timeliness refers to whether the extracted and matched entities and relationships reflect the actual situation of the problem in a timely manner.

[0111] Evaluation data calculation: Based on the matching results, the evaluation data of each dimension is calculated. This can be achieved by counting the number of successfully matched entities and relationships, and calculating matching accuracy, completeness, and other indicators. For semantic similarity matching, the semantic similarity score can also be introduced as part of the evaluation data.

[0112] Evaluation data output: The calculated evaluation data is output in an appropriate format (such as tables, charts, etc.) for subsequent analysis and use. These evaluation data can be used to evaluate the effects of the extraction and matching algorithms and guide the improvement and optimization of subsequent algorithms.

[0113] Evaluation data application: Evaluation data is applied to various scenarios such as interview evaluation, data analysis, knowledge mining, etc. For example, in interview evaluation, the answers of job applicants can be quantitatively scored based on evaluation data to improve the objectivity and accuracy of the evaluation; in data analysis, evaluation data can be used to evaluate the quality of extracted entities and relationships to ensure the reliability and validity of the data.

[0114] By adding entity and relationship matching and evaluation data generation, we can more comprehensively evaluate the matching degree between the extracted entities and relationships and the corresponding entities and relationships in the question, and provide more accurate and reliable information support for subsequent interview evaluation, data analysis, etc. At the same time, it also provides strong data support for algorithm optimization and improvement.

[0115] Please refer to Figure 2 In another embodiment of the present invention, the following steps are included: 1. Interviewer information input In order to ensure the accuracy and relevance of the interview evaluation system, you first need to enter the interviewee's resume and job requirements information. The system supports multiple ways to import this information: a. Upload resume and job requirements documents Interviewers can upload their resume and job description files through the interface provided by the system. Supported file formats include but are not limited to PDF, DOC, DOCX and other common document formats. After uploading, the system will automatically extract and parse the key information in the file.

[0116] b. Manually enter resume and job requirements For interviewees who are unable to provide electronic documents, the system provides a dialog box that allows interviewees to manually enter resume information and job requirements.

[0117] c. Import resumes and job requirements into other systems Considering that the interviewee may have maintained his resume and job information in other recruitment platforms or human resource management systems, this system supports easily importing resumes and job requirement information into the AI ​​simulation interview system through API interface, reducing the workload of repeated input.

[0118] 2. Intelligent generation problem Based on the job requirements and the interviewee's resume, the system will accurately design several interview questions through the big model behind it, aiming to fully cover the resume content while closely matching the job requirements. The prompt design principles of the big model are as follows: Ask interview questions based on job requirements and user resumes. The principles to be followed are as follows: It is forbidden to ask questions that exceed the scope of the resume; the output must be strictly in accordance with the output format; the questions should cover the resume as comprehensively as possible. Job requirements: {input_demands}. Interviewee resume: {input_resume}.

[0119] The system will format and parse the questions returned by the large model and store them in the interview question queue of the database for use in the subsequent interview process.

[0120] 3. Deal with and evaluate the interviewer’s answers In the process of AI simulation interview, a highly structured question-and-answer dialogue mode is adopted to create a professional and efficient interview environment. The system will automatically receive and analyze the interviewee's response: if the interviewee expresses unfamiliarity with a certain question or requests to change the question, the system will change the next round of interview questions; if the interviewee gives an answer that is not relevant to the question, the system will reject this answer and repeat the current question; if the interviewee's answer is related to the interview question, the big model will conduct a detailed analysis of the question and answer, and output a concise and professional evaluation. The prompt design principles of the big model are as follows: provide feedback based on the interview questions and answers provided. If the answer deviates from the topic of the question, please reply directly "Do not give irrelevant answers"; if the respondent requests to change the question or indicates that he cannot answer, reply "The question has been changed for you". If the answer gives specific content for the question, provide a brief evaluation. Question: {output_question} Answer: {input_question} 4. Generate interview summary evaluation After several rounds of interview questions, the interview is evaluated based on the interviewee's resume, job requirements, and all Q&A records. The prompt design principles are as follows: give an overall interview evaluation based on job requirements, the interviewee's resume, and each round of questions and answers. The principles that need to be followed are as follows: the evaluation must be conducted from two perspectives: the fit between the resume and the job requirements and the main body of the user's resume; cases can be added to enrich the question evaluation; and some personalized suggestions are made for interview questions and answers to facilitate the interviewee's improvement. Job requirements {intput_demands}. User resume: {input_resume}. Questions and answers: {questions&answers} 5. Export interview report After completing the interview, the interviewee can review the entire interview process and deeply analyze his / her performance. For each question encountered in the interview, the system will use the analysis and processing capabilities of the large-scale language model behind it to generate detailed reference answers, providing the interviewee with valuable learning resources and opportunities for self-improvement. This function can help the interviewee better understand the interviewer's questioning intentions and learn how to construct more powerful answers.

[0121] In order to facilitate the interviewee's in-depth analysis and self-reflection, the system generates an AI simulation interview report based on the questions in the interview process, the interviewee's personal answers, and the reference answers generated by the system, for the interviewee to export or download as a whole. The generated AI simulation interview report, as the essence of this process, its core value lies in helping users identify the strengths and weaknesses in the interview, so as to formulate more accurate improvement strategies. By comparing personal answers with reference answers, users can deeply understand the interviewer's expectations and evaluation criteria, and then show more mature and professional performance in future interviews. This innovative service undoubtedly provides strong support and guarantee for improving users' interview capabilities.

[0122] This scheme has the following beneficial effects: 1. Improve interview competitiveness: By simulating real interview scenarios, it helps interviewees better prepare for interviews, enhance their self-confidence, and improve their interview skills and expressiveness.

[0123] 2. Personalized feedback: The system can intelligently generate interview questions based on the interviewee’s resume and job requirements, and provide instant feedback and constructive suggestions, so that the interviewee can improve his or her answers in a targeted manner.

[0124] 3. Ease of use: Interviewers can export or download interview reports as needed, so they can review and reflect on the interview process at any time.

[0125] In some embodiments, the interview simulation system based on knowledge graph may include multiple functional modules composed of computer program segments. The computer program of each program segment in the interview simulation system based on knowledge graph may be stored in the memory of a computer device and executed by at least one processor to perform (see Figure 1 Description) Functionality of interview simulation based on knowledge graph.

[0126] In this embodiment, the interview simulation system based on knowledge graph can be divided into multiple functional modules according to the functions it performs, such as Figure 3As shown. The functional modules of the system 300 may include: a resume processing module 310, a graph retrieval module 320, a question generation module 330, an answer analysis module 340 and an evaluation generation module 350. The module referred to in the present invention refers to a series of computer program segments that can be executed by at least one processor and can complete fixed functions, which are stored in a memory. In this embodiment, the functions of each module will be described in detail in subsequent embodiments.

[0127] Resume processing module, used to obtain resume documents and extract key information from resume documents using keyword extraction technology; A graph retrieval module, used to retrieve a local graph from a pre-built knowledge graph based on the key information; The question generation module is used to generate questions based on entities and relations in the local graph; An answer analysis module, configured to receive answer data for the question and extract entities and relationships from the answer data; The evaluation generation module is used to match the extracted entities and relations with the entities and relations corresponding to the question, and generate evaluation data based on the matching results.

[0128] Optionally, as an embodiment of the present invention, the resume processing module includes: A regular expression processing unit, used to remove non-alphabetic characters in the text using regular expressions and convert them to lowercase; A category definition unit, used to define the key information categories to be extracted, wherein the key information categories include educational background, work experience, skills and expertise, and project experience; Vector conversion unit, used to vectorize resume text using TF-IDF; The vocabulary analysis unit is used to analyze words with high TF-IDF scores and classify them into corresponding key information categories.

[0129] Optionally, as an embodiment of the present invention, the method for constructing the knowledge graph includes: collecting interview data, the interview data comprising questions and answers; Performing clustering processing on the interview data to obtain multiple data sets; Extract entities and relationships from the interview data in the dataset, and clean the extracted entities and relationships; Build a knowledge graph based on the cleaned entities and relationships.

[0130] Figure 4The interview simulation method based on knowledge graph provided for the embodiment of the present application can be applied to the device. It can be understood by those skilled in the art that the device structure involved in the embodiment of the present invention does not constitute a limitation on the device, and the device may include more or fewer components than shown, or combine certain components, or arrange different components. In an embodiment of the present invention, the device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0131] The device 400 may include: a processor 410, a memory 420 and a communication unit 430. These components communicate via one or more buses. Those skilled in the art will appreciate that the server structure shown in the figure does not limit the present invention, and it may be a bus structure or a star structure, and may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently.

[0132] The memory 420 may be used to store the execution instructions of the processor 410, and the memory 420 may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. When the execution instructions in the memory 420 are executed by the processor 410, the device 400 is enabled to perform some or all of the steps in the following method embodiments.

[0133] The processor 410 is the control center of the storage device, and uses various interfaces and lines to connect various parts of the entire electronic device. It runs or executes software programs and / or modules stored in the memory 420, and calls data stored in the memory to perform various functions of the electronic device and / or process data. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of a plurality of packaged ICs with the same or different functions. For example, the processor 410 can include only a central processing unit (CPU). In an embodiment of the present invention, the CPU can be a single computing core or multiple computing cores.

[0134] The communication unit 430 is used to establish a communication channel so that the storage device can communicate with other devices, receive user data sent by other devices or send user data to other devices.

[0135] The present invention also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, the program may include some or all of the steps in each embodiment provided by the present invention. The storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0136] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes, including several instructions for enabling a computer device (which can be a personal computer, a server, or a second device, a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0137] In this specification, the same or similar parts between the various embodiments can be referred to each other. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiment.

[0138] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of systems or modules, which can be electrical, mechanical or other forms.

[0139] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0140] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0141] Although the present invention has been described in detail with reference to the accompanying drawings and in combination with the preferred embodiments, the present invention is not limited thereto. Without departing from the spirit and essence of the present invention, a person of ordinary skill in the art may make various equivalent modifications or substitutions to the embodiments of the present invention, and these modifications or substitutions shall be within the scope of the present invention. Any person of ordinary skill in the art may easily think of changes or substitutions within the technical scope disclosed by the present invention, and these shall be within the scope of protection of the present invention.

Claims

1. An interview simulation method based on knowledge graph, characterized in that: include: Obtain resume documents and use keyword extraction technology to extract key information from the resume documents; Retrieving a local graph from a pre-built knowledge graph based on the key information; Generate problems based on entities and relations in local graphs; receiving answer data for the question, and extracting entities and relationships from the answer data; The extracted entities and relations are matched with the entities and relations corresponding to the question, and evaluation data is generated based on the matching results.

2. The method according to claim 1, characterized in that Obtain resume documents and use keyword extraction technology to extract key information from resume documents, including: Use regular expressions to remove non-alphabetic characters from the text and convert them to lowercase. Define the key information categories to be extracted, including education background, work experience, skills and expertise, and project experience; Use TF-IDF to vectorize the resume text; Analyze the words with high TF-IDF scores and classify them into corresponding key information categories.

3. The method according to claim 2, characterized in that The method for constructing the knowledge graph includes: collecting interview data, the interview data comprising questions and answers; Performing clustering processing on the interview data to obtain multiple data sets; Extract entities and relationships from the interview data in the dataset, and clean the extracted entities and relationships; Build a knowledge graph based on the cleaned entities and relationships.

4. The method according to claim 1, characterized in that Retrieving a local graph from a pre-built knowledge graph based on the key information includes: If the key information includes multiple key entities, semantic similarity recognition is performed on the multiple key entities, and a representative entity is screened out from the multiple key entities based on the recognition result; Build a graph query statement containing all represented entities; Submit the graph query statement to the query engine of the knowledge graph to obtain a local graph.

5. The method according to claim 1, characterized in that Generate problems based on entities and relations in local graphs, including: Set up question templates; Substitute some entities and relationships in the local graph into the variables of the question template to form a question.

6. An interview simulation system based on knowledge graph, characterized in that: include: Resume processing module, used to obtain resume documents and extract key information from resume documents using keyword extraction technology; A graph retrieval module, used to retrieve a local graph from a pre-built knowledge graph based on the key information; The question generation module is used to generate questions based on entities and relations in the local graph; An answer analysis module, configured to receive answer data for the question and extract entities and relationships from the answer data; The evaluation generation module is used to match the extracted entities and relations with the entities and relations corresponding to the question, and generate evaluation data based on the matching results.

7. The system according to claim 6, characterized in that The resume processing module includes: A regular expression processing unit, used to remove non-alphabetic characters in the text using regular expressions and convert them to lowercase; A category definition unit, used to define the key information categories to be extracted, wherein the key information categories include educational background, work experience, skills and expertise, and project experience; Vector conversion unit, used to vectorize resume text using TF-IDF; The vocabulary analysis unit is used to analyze words with high TF-IDF scores and classify them into corresponding key information categories.

8. The system according to claim 7, characterized in that The method for constructing the knowledge graph includes: collecting interview data, the interview data comprising questions and answers; Performing clustering processing on the interview data to obtain multiple data sets; Extract entities and relationships from the interview data in the dataset, and clean the extracted entities and relationships; Build a knowledge graph based on the cleaned entities and relationships.

9. A device, characterized in that: include: A memory, used for storing an interview simulation program based on a knowledge graph; A processor, used to implement the steps of the knowledge graph-based interview simulation method as described in any one of claims 1 to 5 when executing the knowledge graph-based interview simulation program.

10. A computer-readable storage medium storing a computer program, characterized in that: The readable storage medium stores an interview simulation program based on a knowledge graph, and when the interview simulation program based on the knowledge graph is executed by a processor, the steps of the interview simulation method based on the knowledge graph as described in any one of claims 1 to 5 are implemented.

Images