Interview method, device and equipment and computer readable storage medium

Abstract

The embodiment of the invention provides an interview method, device and equipment and a computer readable storage medium. The method comprises the following steps: generating an interview topic for a specified picture based on a Touye argumentation model; wherein the interview questions comprise a question provided according to and assertion, a question provided according to arguments, a question provided according to counter examples and a question asserted to be corrected; according to a preset rule, outputting the interview questions to a receiving terminal of the job hunting user in sequence; and receiving the answer content of the job hunting user for each interview question, and scoring the answer content based on a preset evaluation index to obtain an interview score of the job hunting user. According to the method, comprehensive evaluation of the auditing thinking ability of the job hunting user is realized to a certain extent, the interview process is effectively simplified through online interview, time and labor cost are saved, meanwhile, the interview screening standard is not influenced by the subjective consciousness of the recruitment user, and the method has high efficiency and objectivity.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data analysis technology, and in particular to an interview method, apparatus, device, and computer-readable storage medium. Background Art

[0002] Currently, job openings across various industries are typically conducted manually through interviews. For positions with high turnover and high hiring demand, HR departments must expend significant effort and resources conducting frequent interviews, resulting in high recruitment costs and a significant drain on human resources. Furthermore, manual interviews are subject to the subjective opinions of interviewers, making it difficult to effectively achieve uniform screening criteria for new hires. Summary of the Invention

[0003] The embodiments of the present application provide an interview method, apparatus, device, and computer-readable storage medium, which aim to address the shortcomings of traditional interview methods, such as low efficiency, high cost, and inability to achieve unified screening standards.

[0004] In a first aspect, an embodiment of the present application provides an interview method, comprising:

[0005] Generate interview questions for a specified image based on the Toulmin argument model; wherein the interview questions include: questions on providing basis and claim, questions on providing evidence, questions on providing counterexamples, and questions on revising claim;

[0006] According to the preset rules, each interview question is output to the job seeker's receiving terminal in sequence;

[0007] The job seeker's answers to the interview questions are received, and the answers are scored based on preset evaluation indicators to obtain the job seeker's interview score.

[0008] In a second aspect, an embodiment of the present application provides an interview device, comprising:

[0009] A question generation module is used to generate interview questions for a specified image based on the Toulmin argument model; wherein the interview questions include: questions on providing basis and claim, questions on providing evidence, questions on providing counterexamples, and questions on revising claim;

[0010] The question output module is used to output each interview question in sequence to the job seeker's receiving terminal according to the preset rules;

[0011] The interview scoring module is used to receive the job seeker's answers to each interview question and score the answers based on preset evaluation indicators to obtain the job seeker's interview score.

[0012] In a third aspect, an embodiment of the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the interview method provided in the first aspect of the embodiment of the present application are implemented.

[0013] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the interview method provided in the first aspect of the embodiment of the present application are implemented.

[0014] The technical solution provided in the embodiment of the present application is that during the interview process, the system automatically generates interview questions for specified pictures based on the Toulmin argument model. The interview questions based on the Toulmin argument model completely cover the test of key critical thinking abilities such as analysis, evaluation, refutation, and correction, and the job seeker's answers are scored based on preset evaluation indicators to obtain the job seeker's interview score, thereby achieving a comprehensive assessment of the job seeker's critical thinking ability to a certain extent. The online interview effectively simplifies the interview process, saves time and labor costs, and the interview screening criteria are not affected by the subjective consciousness of the recruiting user, and are efficient and objective. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart of an interview method provided in an embodiment of the present application;

[0016] Figure 2 A flowchart of the interview score calculation process provided in an embodiment of the present application;

[0017] Figure 3 A schematic diagram of the structure of an interview device provided in an embodiment of the present application;

[0018] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the technical solutions in the embodiments of this application are further described in detail through the following embodiments and in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application. Those skilled in the art can make adjustments to them as needed to suit specific application scenarios. It should also be noted that, for ease of description, the drawings only show parts related to this application rather than all structures.

[0020] As an example, in the recruitment process, the interview method often used is: the recruiting user and the job-seeking user conduct offline interviews or online face-to-face video interviews, etc. However, this interview method requires the recruiting user to spend a lot of time and manpower costs, and the interview process is easily affected by the personal subjective consciousness of the recruiting user (or interviewer). It is difficult to effectively achieve the uniformity of the newcomer screening standards, and it is difficult to comprehensively test the job-seeking user's critical thinking ability, resulting in low interview efficiency and accuracy of interview evaluation.

[0021] In this regard, in an embodiment of the present application, the critical thinking ability of job seekers is comprehensively tested through an online AI (Artificial Intelligence) interview. Specifically, during the online interview process, the system automatically generates interview questions for the specified picture based on the Toulmin argument model. The interview questions include basis and proposition providing questions, argument providing questions, counterexample providing questions and proposition correction questions, and outputs the generated interview questions to the job seeker's receiving terminal in sequence according to the preset rules. Through these subjective discussion questions, the job seeker's critical thinking ability can be tested in all aspects. For the job seeker, he can answer in the order of the output of the interview questions. The system scores the job seeker's answer content based on the preset evaluation indicators to obtain the job seeker's interview score, which realizes a comprehensive evaluation of the job seeker's critical thinking ability to a certain extent, simplifies the interview process, is efficient and objective, and also improves the accuracy of the interview evaluation.

[0022] Figure 1 A flow chart of the interview method provided in the embodiment of the present application. Figure 1 As shown, the method may include:

[0023] S101. Generate interview questions for specified pictures based on the Toulmin argument model.

[0024] Among them, the interview questions include: questions about providing basis and claims, questions about providing arguments, questions about providing counterexamples, and questions about revising claims.

[0025] The designated images serve as the observation objects in the interview questions and serve as visual vehicles for structured argumentation. By using the concrete scenes depicted in the designated images, job seekers are encouraged to demonstrate their critical thinking skills. Based on the job position, images with a relevance exceeding a preset threshold can be selected from the image database as designated images. For example, if the position is project management, images depicting teamwork scenarios can be selected as designated images.

[0026] The Toulmin argument model consists of six key elements: claim, data, warrant, support, refutation, and qualification. During an interview, you can use the Toulmin argument model and the visual information of a designated image to design a series of interview questions that comprehensively cover the six key elements of the Toulmin argument model. These questions can systematically assess a job applicant's critical thinking ability. Questions that provide evidence and evidence cover the Toulmin argument model elements of "data" and "claim," questions that provide evidence cover the Toulmin argument model elements of "warrant" and "support," questions that provide counterexamples cover the Toulmin argument model elements of "refutation" / "exceptions," and questions that revise a claim cover the Toulmin argument model elements of "qualification" and "conclusion."

[0027] As an example, the interview system can generate interview questions that cover Toulmin elements according to a template based on the key visual elements in a specified image and the six major model elements included in the Toulmin argument model, or instruct the large language model to use the specified image and Toulmin argument model to generate more flexible interview questions that cover Toulmin elements.

[0028] For example, assuming the specified image is a picture of a mountain climbing scene, the following interview questions can be generated using this image and the Toulmin argument model:

[0029] 1. Provide the topic based on the basis and proposition

[0030] “Please observe the pictures shown, describe their content and talk about the inspiration you get from them.”

[0031] 2. Provide the topic with evidence

[0032] "Please provide corresponding arguments to support the point you just made based on information from personal experience, life events, knowledge data, etc."

[0033] 3. Counterexamples provide questions

[0034] "Under what circumstances do you think the point you just made may become invalid and unreasonable."

[0035] 4. Suggest revising the title

[0036] "Please restate your opinion based on your complete description of this image."

[0037] S102: Output each interview question to the job seeker's receiving terminal in sequence according to preset rules.

[0038] After obtaining the interview questions, each interview question can be output to the job seeker's receiving terminal in sequence according to the preset question output order. For example, the interview questions can be pushed to the job seeker's receiving terminal in the order of claim-argument-counterexample-correction.

[0039] Corresponding applications, applets, browser pages, etc. can be run in the job seeker's receiving terminal, and corresponding online job search functions can be provided to the job seeker through applications, applets, browser pages, etc. Therefore, each interview question can be output to the interview page of the receiving terminal in sequence according to preset rules, and the job seeker can answer each interview question in sequence according to the output order of the questions.

[0040] As an example, an AI interviewer's AI image can be displayed on the interview page of the receiving terminal to simulate an interview with a real interviewer, and the AI ​​interviewer can ask the job seeker the above interview questions in turn. During the job seeker's answering process, the job seeker's facial data can be obtained through the image acquisition device of the receiving terminal, the job seeker's expression can be judged by analyzing the facial data, and the job seeker's voice emotion can be obtained based on the voice information fed back by the job seeker during the answering process. The speed of the AI ​​interviewer's questioning can be dynamically adjusted according to the job seeker's expression and voice emotion. For example, when the job seeker is detected to be nervous based on the job seeker's expression and voice emotion, the speed of the questioning can be slowed down or prompt information can be provided to the job seeker.

[0041] Optionally, during the process of asking a job seeker questions through an interview question, the job seeker can answer the interview question. During the answering process, the answer may be incomplete or deviate significantly from the question. In this case, the interview system can determine whether the job seeker's answer contains key information required for the interview question. If it is determined that the answer lacks any key information required for the interview question, the interview system can generate follow-up questions corresponding to the interview question based on the interview question and the job seeker's answer, and output the follow-up questions to the job seeker's receiving terminal to guide the job seeker to provide the key information required for the interview question. Optionally, the interview system can provide the interview question and the job seeker's answer to a target macro model and provide prompt information to the target macro model. The prompt information is used to prompt the target macro model to ask follow-up questions based on the interview question and answer content. The prompt information includes at least one of an interview question example, an answer content example, and a corresponding follow-up question example. The target macro model generates follow-up questions for the interview question based on the follow-up question generation knowledge learned from the prompt information. For example, if a question about providing evidence and claims only describes the content of an image but doesn't provide any insights or claims drawn from it, the interview system can generate follow-up questions based on the missing key information and the interview question, such as "Please provide your insights based on your description." Another example is if a question about providing evidence requires three pieces of evidence but the applicant only provides one, the interview system can generate follow-up questions based on the missing key information and the interview question, such as "Please provide two more pieces of evidence." In other words, when generating follow-up questions for an interview question, only follow-up questions are asked based on the missing key information, eliminating the need to repeat previous questions. This not only reduces the burden of generating follow-up questions and improves interview efficiency, but also avoids repeatedly asking the same questions to the applicant, thereby avoiding user frustration and effectively improving the interview experience.

[0042] Through the above process, based on the job seeker's answer to the interview question, it is determined whether to trigger a follow-up question. After determining that the follow-up question is triggered, a follow-up question is generated. The job seeker is further questioned through the follow-up question to guide the job seeker to answer in a targeted manner, thereby ensuring the effectiveness of the answer content, thereby improving the effectiveness of the interview and increasing the fun and interactivity of the interview; in addition, the question speed can be dynamically adjusted according to the job seeker's expression and voice emotion to further enhance the interactivity of the interview.

[0043] S103: Receive the job seeker's answers to the interview questions, and score the answers based on preset evaluation indicators to obtain the job seeker's interview score.

[0044] Specifically, the preset evaluation indicators are pre-set standardized measurement dimensions for evaluating the quality of the answer content. They can be designed based on the assessment dimensions of critical thinking ability. For example, the assessment indicators for critical thinking ability may include: logic, criticality, relevance, clarity, fairness, and breadth. Among them, logic is used to examine the rigor of the reasoning of the answer content, criticality is used to examine whether the counterexample information proposed effectively reveals potential loopholes or limitations of the claim, relevance is used to examine the degree of connection between the answer content, clarity is used to examine whether the answer content is concise and easy to understand and will not cause confusion or ambiguity, fairness is used to examine whether the answer content is honest, fair, and unbiased, and breadth is used to examine whether the argument dimensions of the answer content are diverse.

[0045] The interview system obtains the job seeker's responses to each interview question based on the voice information input by the job seeker. The system then scores the job seeker's responses using the aforementioned pre-defined evaluation criteria, thereby obtaining the job seeker's interview score. Furthermore, the job seeker's interview score can be associated with the job seeker's personal information and sent to the recruiter's receiving terminal for reference in making hiring decisions.

[0046] The interview method provided in the embodiment of the present application is such that during the interview process, the system automatically generates interview questions for the specified pictures based on the Toulmin argument model. The interview questions based on the Toulmin argument model completely cover the test of key critical thinking abilities such as analysis, evaluation, refutation, and correction, and the job seeker's answers are scored based on preset evaluation indicators to obtain the job seeker's interview score, thereby achieving a comprehensive assessment of the job seeker's critical thinking ability to a certain extent. The online interview effectively simplifies the interview process, saves time and labor costs, and the interview screening criteria are not affected by the subjective consciousness of the recruiting user, and are efficient and objective.

[0047] Optionally, at least one preset evaluation indicator can be set for each interview question. For example, the preset evaluation indicators set for questions providing basis and claims may include at least one of the following: consistency / accuracy, relevance, fairness, and clarity; the preset evaluation indicators set for questions providing arguments may include at least one of the following: relevance, logic, breadth, and clarity; the preset evaluation indicators set for questions providing counterexamples may include at least one of the following: criticality, relevance, and clarity, etc.; the preset evaluation indicators set for questions providing critical claim corrections may include at least one of the following: logic, fairness, and clarity, etc.

[0048] Based on this, after obtaining the job seeker's answers to each interview question, Figure 2 As shown, the process of the interview system scoring the answer content based on the preset evaluation indicators to obtain the interview score of the job seeker may include:

[0049] S201. For each interview question, the answer to the interview question may be scored based on at least one preset evaluation indicator to obtain an indicator score of the at least one preset evaluation indicator.

[0050] Specifically, for each interview question, the answer to the interview question is scored using at least one preset evaluation indicator set for the interview question, thereby obtaining an indicator score for the at least one preset evaluation indicator.

[0051] For example, taking the topic of providing basis and proposition as an example, assuming that the preset evaluation indicators set for the interview question include consistency / accuracy, relevance, fairness and clarity, the job seeker's answer content will be scored from multiple indicator dimensions such as consistency / accuracy, relevance, fairness and clarity to obtain accuracy score, relevance score, fairness score and clarity score, etc.

[0052] As an optional implementation method, for each preset evaluation indicator under each interview question, the interview system can perform text mining and semantic recognition on the job seeker's answer to the interview question to extract key semantic information in the answer content, determine the degree of deviation between the key semantic information and the preset reference information of the interview question, and calculate the indicator score of the preset evaluation indicator under the interview question based on the degree of deviation and the preset standard score corresponding to the preset reference information.

[0053] As another optional implementation method, for each interview question, the interview system can construct the answer content of the interview question and the scoring rules corresponding to at least one preset evaluation indicator as input data of the target big model, and obtain prompt information. The prompt information is used to prompt the target big model to score the answer content according to the scoring rules of at least one preset evaluation indicator in the input data. The prompt information also includes at least: an answer content instance, a scoring rule instance, and a corresponding indicator score instance. The input data and the prompt information are then input into the target big model for indicator scoring to generate an indicator score for at least one preset evaluation indicator.

[0054] Among them, after the scoring rules of the preset evaluation indicators, the job seeker's answer content and the prompt information are input into the target large model, the target large model can generate the indicator score of the preset evaluation indicator based on the indicator scoring knowledge learned from the reference example in the prompt information, the answer content of the job seeker and the scoring rules of the preset evaluation indicators. In an optional implementation method, the step of the target large model generating the indicator score of the preset evaluation indicator may include: the target large model identifies the logical elements in the answer content through the self-attention mechanism, compares the preset reference information of the interview question, determines the matching elements and missing elements in the answer content, and generates the indicator score of the preset evaluation indicator based on the matching elements and missing elements obtained based on the scoring knowledge learned from the reference example. For example, if the preset reference information contains 3 scoring points and the job seeker's answer content only contains 1 scoring point, the indicator score of the preset evaluation indicator can be generated based on the preset score corresponding to the 1 scoring point.

[0055] S202. For each preset evaluation indicator, determine the relevant interview questions of the preset evaluation indicator in all interview questions, and determine the comprehensive score of the preset evaluation indicator based on the indicator scores of the preset evaluation indicator in the relevant interview questions.

[0056] In actual applications, the preset evaluation indicators set for some interview questions may overlap. For example, for each interview question, the job seeker's answer will be examined to see whether it is clear and unambiguous. Therefore, after obtaining the indicator score of at least one preset evaluation indicator corresponding to each interview question, for each preset evaluation indicator, the associated interview questions of the preset evaluation indicator are determined from all interview questions, and the comprehensive score of the preset evaluation indicator is determined based on the indicator score of the preset evaluation indicator in each associated interview question. Among them, the interview questions scored by the preset evaluation indicator can be called the associated interview questions of the preset evaluation indicator. For example, assuming that the basis and proposition providing questions, argument providing questions, counterexample providing questions, and proposition correction questions are all set with the same preset evaluation indicator, then these four interview questions can all be called the associated interview questions of the preset evaluation indicator.

[0057] Furthermore, if a preset evaluation indicator is associated with only one interview question, the score of the preset evaluation indicator in the associated interview question is determined as the comprehensive score of the preset evaluation indicator. If a preset evaluation indicator is associated with multiple interview questions, the scores of the preset evaluation indicator in each associated interview question are weighted and summed, and the weighted sum result is determined as the comprehensive score of the preset evaluation indicator.

[0058] S203: Determine the interview score of the job seeker based on the comprehensive score of each preset evaluation indicator and the preset weight value of each preset evaluation indicator.

[0059] Logic and critical thinking can better evaluate critical thinking ability and are the core indicators for assessing critical thinking ability. Therefore, the preset weight values ​​corresponding to logic and critical thinking are higher, and the preset weight values ​​corresponding to the other preset evaluation indicators are relatively low.

[0060] After obtaining the comprehensive scores of the preset evaluation indicators, the comprehensive scores of the preset evaluation indicators are weighted and summed according to the preset weight values ​​of the preset evaluation indicators to obtain the interview score of the job seeker.

[0061] In this embodiment, by independently calculating the score for each preset evaluation indicator, the critical thinking ability of job seekers can be examined from multiple dimensions, thereby improving the accuracy of identifying the critical thinking ability of job seekers, and unifying the scoring rules to avoid subjective bias in manual interviews and improve the objectivity of interview evaluations.

[0062] Optionally, when the interview question is a basis and claim provision question, the preset evaluation indicators corresponding to the basis and claim provision question include at least one of the following: consistency between the basis information provided by the job seeker and the reference content description information of the designated image; relevance between the claim information provided by the job seeker and the reference meaning interpretation information of the designated image; fairness of the claim information; and clarity of the basis information and claim information. Therefore, for the basis and claim provision question, the above S201 may include at least one of the following:

[0063] S2011a. Calculate the consistency score between the basis information in the answer content of the basis and claim provision question and the reference content description information of the designated image.

[0064] Specifically, reference content description information and reference meaning interpretation information are extracted from a given image. The reference content description information provides a reference description of the objective elements in the given image, which serves as a benchmark for evaluating the job seeker's observation ability. The reference meaning interpretation information provides a reference answer to the symbolic meaning and thematic implications of the given image, which serves as a benchmark for evaluating the job seeker's logical extraction and critical thinking.

[0065] Semantic extraction is performed on the answers provided by job seekers regarding the basis and claim questions. This extraction process includes the following steps: obtaining the text data corresponding to the answers and extracting multiple specific entities contained in the text data as semantic information for the answers. Specific entities are meaningful, interview-related concepts and relationships within the text data. Specific entities include, but are not limited to, tools, people, actions, environments, results, and claims. Furthermore, the semantic information is segmented into basis information and claim information.

[0066] Next, the text similarity between the basis information and the reference content description information is calculated using a text similarity algorithm to obtain a consistency score. The text similarity algorithm may include, but is not limited to, cosine similarity, word embedding algorithm, and edit distance.

[0067] S2012a. Calculate the correlation score between the claim information and the reference meaning interpretation information.

[0068] The correlation here is mainly used to evaluate the degree of association between the claim information and the reference meaning interpretation information. The lower the degree of association between the claim information and the reference meaning interpretation information, the lower the correlation score. Conversely, the higher the degree of association between the claim information and the reference meaning interpretation information, the higher the correlation score. In an optional implementation, the interview system can use the target macro model to calculate the correlation score between the claim information and the reference meaning interpretation information. The interview system can input the claim information, the reference meaning interpretation information and the corresponding prompt information into the target macro model. The prompt information is used to instruct the target macro model to perform a correlation calculation on the claim information and the reference meaning interpretation information. The target macro model learns knowledge based on the prompt information and calculates the correlation score between the claim information and the reference meaning interpretation information. The specific calculation process can refer to the description of the critical score in the following embodiment, and this embodiment will not be described in detail here.

[0069] S2013a. Calculate the fairness score of the claim information.

[0070] Fairness is primarily used to assess whether claims are honest, fair, and free of personal bias. Optionally, the target macro model and prompt information related to the fairness score can be used to calculate the fairness score of the claim. The specific calculation process can be found in the description of the critical score in the following example and will not be detailed here.

[0071] S2014a. Calculate the clarity scores of the evidence and claim information.

[0072] Clarity is used to assess whether the evidence and claim information are concise, understandable, and unambiguous. Optionally, clarity scores for the evidence and claim information can be calculated using the target macro model and prompts related to the clarity score. The specific calculation process can be found in the description of the criticality score in the following example and will not be detailed here.

[0073] Optionally, when the interview question is an argument provision question, the preset evaluation indicators for the argument provision question include at least one of the following: the relevance between the argument information provided by the job seeker and the assertion information; the logic between the argument information and the assertion information; the breadth of the argument information's discussion dimensions; and the clarity of the argument information. Therefore, for the argument provision question, the above S201 may include at least one of the following:

[0074] S2011b. Calculate the correlation score between the argument information and the claim information in the answer content of the argument provision question.

[0075] The system performs semantic extraction on the answer to the argument-providing question to obtain semantic information. This semantic information is then segmented to obtain at least one piece of argument information. The claim information provided by the job seeker is then retrieved, and a correlation score is calculated between the at least one piece of argument information and the claim information. Correlation here assesses the degree of correlation between the argument information and the claim information. The lower the correlation, the lower the correlation score; conversely, the higher the correlation, the higher the correlation score.

[0076] Optionally, the target macro model and the prompt information related to the relevance score can be used to calculate the relevance score between the argument information and the claim information. The specific calculation process can refer to the description of the critical score in the following embodiment, and this embodiment will not be repeated in detail here.

[0077] S2012b. Calculate the logical score between the argument information and the claim information.

[0078] In addition, the logic here is used to evaluate the degree to which the argument information supports the claim information. The higher the degree to which the argument information supports the claim information, the higher the logic score. Conversely, if the argument information is difficult or unable to support the main information, the logic score will be lower.

[0079] Optionally, the target macro model and the prompt information related to the logic score can be used to calculate the logic score between the argument information and the claim information. The specific calculation process can refer to the description of the critical score in the following embodiment, and this embodiment will not be repeated in detail here.

[0080] S2013b. Calculate the breadth score based on the amount of evidence and the discussion dimensions of the evidence.

[0081] Among them, breadth is used to evaluate whether the argument information supports the claim information from different dimensions. If the amount of argument information is greater and each argument information supports the claim information from different dimensions, the breadth score will be higher. Conversely, the fewer the amount of argument information is and the discussion dimensions of each argument information are similar, the breadth score will be lower.

[0082] S2014b. Calculate the clarity score of the argument information.

[0083] Clarity here is used to assess whether the argument information is concise, understandable, and unambiguous. Optionally, the target macro model and clarity score-related prompts can be used to calculate the clarity score of the argument information. The specific calculation process can be referred to the description of the critical score in the following embodiment and will not be detailed here.

[0084] Optionally, in the case where the interview question is a counterexample provision question, the preset evaluation indicators for the counterexample provision question include at least one of the following: the relevance between the counterexample information provided by the job seeker and the assertion information; the criticality of the counterexample information with respect to the assertion information; and the clarity of the counterexample information. Therefore, for the counterexample provision question, the above S201 may include at least one of the following:

[0085] S2011c. Calculate the correlation score between the counterexample information and the assertion information in the answer content of the counterexample provision question.

[0086] We perform semantic analysis on the responses to the counterexample questions to obtain the counterexample information provided by the job seeker, as well as the claim information provided by the job seeker. We then calculate the correlation score between the counterexample and claim information. This correlation measures the degree of association between the counterexample and claim information. The lower the correlation, the lower the correlation score; conversely, the higher the correlation, the higher the correlation score.

[0087] Optionally, the target large model and the hint information related to the relevance score can be used to calculate the relevance score of the counterexample information and the claim information. The specific calculation process can refer to the description of the critical score in the following embodiment, and this embodiment will not be repeated in detail here.

[0088] S2012c. Calculate the critical score of the counterexample information against the claim information.

[0089] Criticality is used to evaluate whether the counterexample information provided by the job seeker effectively reveals potential loopholes or limitations in the claim. Optionally, the criticality scores of the counterexample information and the claim information can be calculated using the target large model and prompt information related to the criticality score.

[0090] In some examples, the critical scoring-related prompts provided to the target large model may include:

[0091] "You are a rigorous and professional talent assessment expert who can score the criticality of the counterexample information provided by job seekers based on the semantics of the claims provided by job seekers."

[0092] Requirements: 1. Criticality refers to whether the counterexample effectively reveals potential loopholes or limitations in the claim. That is, the counterexample directly targets the core logic of the claim rather than marginal details, has the possibility of real occurrence or existing evidence to support it, and can reveal the underlying mechanism that invalidates the claim.

[0093] 2. Scoring Rules: The criticality level ranges from low to high, corresponding to a score of 1-100. [1,50) represents extremely low criticality, that is, no valid counterexample is provided or the counterexample is irrelevant to the claim, and is based on emotional expressions (such as "obviously wrong" or "unreliable") without specific analysis; [50,65) represents low criticality, that is, the counterexample relies on extreme assumptions or fictional scenarios, does not touch the core of the claim, and only attacks the peripheral details of the claim rather than the core logic; [65,75) represents moderate criticality, that is, the counterexample can be cited, but it is common sense, and the refutation logic remains superficial without in-depth analysis of the mechanism; [75,90) represents high criticality, that is, the counterexample is supported by data, research, etc., directly addresses the implicit assumptions of the claim, and can explain the relationship between the counterexample and the conditions for the claim to be valid; [90,100] represents extremely high criticality, that is, the counterexample dismantles the claim from multiple dimensions (such as the contradiction between theory and reality, the contradiction between different evidence), the logical chain is complete, and the evidence can be independently verified.

[0094] 3. If the counterexample information is empty, it is impossible to make a critical evaluation and a low score can be given.

[0095] 4. Please score only from the perspective of criticality and do not consider other factors when scoring.

[0096] 5. When grading, only grade based on the given content and do not over-interpret or over-extend it.

[0097] Please output the corresponding indicator score based on the following case examples and requirements.

[0098] “Example claim message: Electric cars are environmentally friendly.

[0099] Counterexample information example 1: Electric cars are ugly.

[0100] Critical Scoring Example 1: The score is extremely low, such as 20 points.

[0101] Counterexample 2: A Swedish research institute pointed out that carbon emissions from battery production are overestimated, but the decarbonization of the power grid in the future will enhance the advantages of electric vehicles. Therefore, environmental protection needs to be evaluated in stages and coordinated with power grid reform.

[0102] Critical Scoring Example 2: An extremely high score, such as 98.

[0103] On this basis, the prompt information related to critical scoring (including the definition of criticality, critical scoring requirements, scoring rules, and critical examples, etc.), counterexample information and claim information provided by job seekers are provided to the target big model, so as to use the target big model to calculate the critical score of the counterexample information against the claim information.

[0104] S2013c. Calculate the clarity score of the counterexample information.

[0105] Clarity is used to assess whether the counterexample information is concise, understandable, and unambiguous. Optionally, the clarity score of the counterexample information can be calculated using the target macro model and related clarity score information. The specific calculation process can be found in the description of the criticality score above and will not be further described in this embodiment.

[0106] Optionally, if the interview question is a proposition revision question, the preset evaluation indicators for the proposition revision question include at least one of the following: the logic between the conclusion information provided by the job seeker and the proposition information and counterexample information; the fairness of the conclusion information; and the clarity of the conclusion information. Therefore, for the proposition revision question, the above S201 may include at least one of the following:

[0107] S2011d. Combine the claim information and counterexample information to calculate the logic score of the conclusion information in the answer content of the claim correction question.

[0108] Perform semantic analysis on the answers to the claim modification questions to obtain the conclusion information provided by the job seeker, as well as the claim information and counterexample information provided by the job seeker. Analyze whether the conclusion information adds restrictive conditions based on the counterexample information compared to the claim information, and analyze whether the meaning of the conclusion information is consistent with that of the claim information. Determine the logic score of the conclusion information based on the analysis results of the above dimensions.

[0109] S2012d. Calculate the fairness score of the conclusion information.

[0110] Fairness is primarily used to assess whether the conclusion information is honest, fair, and free of personal bias. Optionally, the fairness score of the conclusion information can be calculated using the target macro model and prompt information related to the fairness score. The specific calculation process can be found in the description of the critical score above and will not be repeated in this embodiment.

[0111] S2013d. Calculate the clarity score of the conclusion information.

[0112] Clarity is used to assess whether the conclusion information is concise, understandable, and unambiguous. Optionally, the clarity score of the conclusion information can be calculated using the target macro model and prompt information related to the clarity score. The specific calculation process can be referred to the description of the critical score above and will not be repeated in this embodiment.

[0113] Through the above process, the index scores of consistency / accuracy, relevance, fairness, clarity, logic, breadth and criticality in each interview question can be obtained. For each preset evaluation indicator, the average index score of the preset evaluation indicator in each interview question is calculated as the comprehensive score of the preset evaluation indicator, such as the comprehensive score of consistency / accuracy, relevance, fairness, clarity, logic, breadth and criticality. Based on the comprehensive score of consistency / accuracy, relevance, fairness, clarity, logic, breadth and criticality and their preset weight values, the total score of the job seeker's critical thinking is calculated.

[0114] In this embodiment, the job seeker's answer content is evaluated based on accuracy, relevance, fairness, clarity, logic, breadth and criticality. Each evaluation index comprehensively covers the critical thinking assessment dimension, thereby improving the accuracy of identifying the job seeker's critical thinking ability.

[0115] Figure 3 A schematic diagram of the structure of the interview device provided in the embodiment of the present application. Figure 3 As shown, the device may include: a question generating module 301, a question output module 302 and an interview scoring module 303.

[0116] Specifically, the question generation module 301 is used to generate interview questions for a specified picture based on the Toulmin argument model; wherein the interview questions include: basis and claim provision questions, argument provision questions, counterexample provision questions, and claim correction questions;

[0117] The question output module 302 is used to output each interview question in sequence to the job seeker's receiving terminal according to preset rules;

[0118] The interview scoring module 303 is used to receive the job seeker's answers to each interview question and score the answers based on preset evaluation indicators to obtain the job seeker's interview score.

[0119] On the basis of the above embodiment, optionally, the interview scoring module 303 is also used to score the answer content of each interview question based on at least one preset evaluation indicator to obtain an indicator score of at least one preset evaluation indicator; for each preset evaluation indicator, determine the associated interview questions of the preset evaluation indicator in all interview questions, and determine the comprehensive score of the preset evaluation indicator based on the indicator score of the preset evaluation indicator in the associated interview questions; determine the interview score of the job seeker based on the comprehensive score of each preset evaluation indicator and the preset weight value of each preset evaluation indicator.

[0120] On the basis of the above embodiment, optionally, the interview scoring module 303 is also used to construct, for each interview question, the answer content of the interview question and the scoring rules corresponding to at least one preset evaluation indicator as input data of the target big model; obtain prompt information; the prompt information is used to prompt the target big model to score the answer content according to the scoring rules in the input data, and the prompt information also includes at least: an answer content instance, a scoring rule instance, and a corresponding indicator score instance; the input data and the prompt information are input into the target big model for indicator scoring to generate an indicator score for at least one preset evaluation indicator.

[0121] Based on the above embodiment, optionally, the question generation module 301 is also used to determine whether the answer content provided by the job seeker contains key information required for the interview question; if it is determined that the answer content lacks any key information required for the interview question, then generate follow-up questions corresponding to the interview question based on the interview question and the answer content; and output the follow-up questions to the job seeker's receiving terminal to guide the job seeker to provide the key information required for the interview question.

[0122] Based on the above embodiment, optionally, the preset evaluation indicators corresponding to the topic provided by the claim include at least one of the following: consistency between the basis information provided by the job seeker and the reference content description information of the specified picture; relevance between the claim information provided by the job seeker and the reference meaning interpretation information of the specified picture; fairness of the claim information; clarity of the basis information and the claim information.

[0123] Based on the above embodiment, optionally, the preset evaluation indicators corresponding to the argument provision topic include at least one of the following: the relevance between the argument information and the claim information provided by the job seeker; the logic between the argument information and the claim information; the breadth of the discussion dimension of the argument information; and the clarity of the argument information.

[0124] Based on the above embodiment, optionally, the preset evaluation indicators corresponding to the counterexample provision topic include at least one of the following: the relevance between the counterexample information provided by the job seeker and the assertion information; the criticality of the counterexample information against the assertion information; and the clarity of the counterexample information.

[0125] Based on the above embodiment, optionally, the preset evaluation indicators corresponding to the proposed revised topic include at least one of the following: the logic between the conclusion information provided by the job seeker and the proposed information and counterexample information; the fairness of the conclusion information; and the clarity of the conclusion information.

[0126] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of the present application is shown as follows: Figure 4As shown, the computer device includes a processor 40, a memory 41, an input device 42 and an output device 43; the number of processors 40 in the computer device can be one or more. Figure 4 In the example, a processor 40 is used; the processor 40, memory 41, input device 42 and output device 43 in the computer device can be connected by a bus or other means. Figure 4 The bus connection is taken as an example.

[0127] Memory 41, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the interview method in the embodiments of the present application (e.g., question generation module 301, question generation module 302, and interview scoring module 303). Processor 40 executes the software programs, instructions, and modules stored in memory 41 to execute various functional applications and data processing of the device, thereby implementing the interview method provided in any of the above embodiments.

[0128] Memory 41 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function; the data storage area may store data used in the interview method provided in the above-described embodiment, among other things. Furthermore, memory 41 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state memory device. In some instances, memory 41 may further include memory remotely located relative to processor 40, and such remote memory may be connected to the device / terminal / server via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0129] The input device 42 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the computer device. The output device 43 may include a display device such as a display screen.

[0130] In one embodiment, a computer-readable storage medium is further provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the interview method provided in any of the above embodiments are implemented.

[0131] In one embodiment, a computer program product is also provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the interview method provided in any of the above embodiments are implemented.

[0132] The interview apparatus, computer device, computer-readable storage medium, and computer program product provided in the above embodiments can execute the interview method provided in any embodiment of this application, and have the corresponding functional modules and beneficial effects of executing the method. For technical details not fully described in the above embodiments, please refer to the interview method provided in any embodiment of this application.

[0133] Through the above description of the implementation methods, those skilled in the art can clearly understand that the present application can be implemented with the help of software and necessary general-purpose hardware, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as a computer's floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0134] It is worth noting that the various units and modules included in the above embodiments are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application.

[0135] Note that the above are only preferred embodiments of the present application and the technical principles employed. Those skilled in the art will understand that the present application is not limited to the specific embodiments herein, and that various obvious changes, readjustments, and substitutions can be made by those skilled in the art without departing from the scope of protection of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments and may include many other equivalent embodiments without departing from the scope of the present application. The scope of the present application is determined by the scope of the appended claims.

Claims

1. An interview method, characterized in that: include: Generate interview questions for a specified image based on the Toulmin argument model; wherein the interview questions include: questions on providing basis and claim, questions on providing evidence, questions on providing counterexamples, and questions on revising claim; According to the preset rules, each interview question is output to the job seeker's receiving terminal in sequence; The job seeker's answers to the interview questions are received, and the answers are scored based on preset evaluation indicators to obtain the job seeker's interview score.

2. The method according to claim 1, characterized in that Each interview question corresponds to at least one preset evaluation indicator. Scoring the answer based on the preset evaluation indicator to obtain the job seeker's interview score includes: For each interview question, scoring the answer to the interview question based on the at least one preset evaluation indicator to obtain an indicator score for the at least one preset evaluation indicator; For each preset evaluation indicator, determine the associated interview questions of the preset evaluation indicator among all interview questions, and determine the comprehensive score of the preset evaluation indicator based on the indicator scores of the preset evaluation indicator in the associated interview questions; The interview score of the job seeker is determined based on the comprehensive score of each preset evaluation indicator and the preset weight value of each preset evaluation indicator.

3. The method according to claim 2, characterized in that For each interview question, scoring the answer to the interview question based on the at least one preset evaluation indicator to obtain an indicator score for the at least one preset evaluation indicator includes: For each interview question, the answer content of the interview question and the scoring rule corresponding to the at least one preset evaluation indicator are constructed as input data of the target large model; Obtain prompt information; the prompt information is used to prompt the target large model to score the answer content according to the scoring rule in the input data, and the prompt information also includes at least one of: an answer content instance, a scoring rule instance, and a corresponding indicator score instance; The input data and the prompt information are input into the target macro model for index scoring to generate an index score of the at least one preset evaluation index.

4. The method according to claim 1, wherein After receiving the job seeker's answer to any interview question, the process further includes: Determining whether the answer provided by the job seeker contains key information required for the interview question; If it is determined that the answer lacks any key information required for the interview question, generating follow-up questions corresponding to the interview question based on the interview question and the answer; The follow-up questions are output to a receiving terminal of the job seeker to guide the job seeker to provide key information required for the interview questions.

5. The method according to claim 2, characterized in that The preset evaluation indicators corresponding to the topic provided by the claim include at least one of the following: The consistency between the basis information provided by the job seeker and the reference content description information of the designated image; the correlation between the claim information provided by the job seeker and the reference meaning interpretation information of the designated image; the fairness of the information claimed; The clarity of the supporting information and the claimed information.

6. The method according to claim 2, characterized in that The preset evaluation indicators corresponding to the argument provision topic include at least one of the following: the relevance between the argument information and the claim information provided by the job seeker; The logic between the argument information and the claim information; The breadth of the discursive dimension of the evidence; The clarity of the information presented.

7. The method according to claim 2, characterized in that The preset evaluation indicators corresponding to the counterexample provided by the question include at least one of the following: the correlation between the counterexample information and the claim information provided by the job seeker; The criticality of the counterexample information with respect to the assertion information; The clarity of the counterexample information.

8. The method according to claim 2, characterized in that The preset evaluation indicators corresponding to the proposed revision topic include at least one of the following: The logic between the conclusion information, assertion information and counterexample information provided by the job seeker; The fairness of the conclusion information; Clarity of the conclusion message.

9. An interview device, characterized in that: include: A question generation module is used to generate interview questions for a specified image based on the Toulmin argument model; wherein the interview questions include: questions on providing basis and claim, questions on providing evidence, questions on providing counterexamples, and questions on revising claim; The question output module is used to output each interview question in sequence to the job seeker's receiving terminal according to the preset rules; The interview scoring module is used to receive the job seeker's answers to each interview question and score the answers based on preset evaluation indicators to obtain the job seeker's interview score.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images