Interviewer evaluation prejudice correction method and system based on real-time feedback

Abstract

The invention discloses an interviewer evaluation prejudice correction method and system based on real-time feedback, particularly relates to the technical field of human resource management, and is used for solving the problems that evaluation standards and industry characteristics are disjointed due to standardized semantic analysis in an existing recruitment system, and implicit evaluation prejudice of an interviewer is difficult to dynamically correct. Fuzzy terms in post requirements are dynamically analyzed through the industry knowledge base, industry weights and term thresholds are generated, interview interaction texts are collected in real time, and relevance between candidates and the term thresholds is analyzed; in combination with interviewer question context recognition and deviation calculation, term deviation signals and repeated question intention deviation are marked, and finally, an industry correction rule is matched to dynamically adjust a score weight; on the basis of a semantic analysis and real-time feedback linkage mechanism adaptive to industry characteristics, alignment of evaluation standards and real post requirements is realized, evaluation prejudice caused by term understanding dislocation is effectively eliminated, and objectivity and accuracy of talent selection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of human resource management, and more particularly to a method and system for correcting interviewer evaluation bias based on real-time feedback. Background Art

[0002] In existing human resource management systems, recruiters usually screen and match candidates through job requirements described in natural language, and rely on preset standardized competency tags to parse abstract terms in job descriptions to achieve automatic association between candidates and job requirements. However, due to the substantial differences in the actual competency requirements for the same term in different industries or enterprises, existing technologies use unified semantic mapping rules for demand analysis, and have not established a dynamic association mechanism between term semantics and industry characteristics.

[0003] However, standardized parsing rules are difficult to accurately reflect the contextual semantics of ambiguous terms in job requirements, resulting in systematic deviations between the matching results generated by the system and the true intentions of the employer. When interviewers evaluate candidates based on such biased data, they may form implicit evaluation biases due to the misalignment between their understanding of the terms and the actual ability standards of the position, thereby reducing the objectivity and accuracy of talent selection. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a method and system for correcting interviewer evaluation bias based on real-time feedback to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: The interviewer evaluation bias correction method based on real-time feedback includes the following steps: S1. Obtain the job requirements and industry attribute information of the target position, and collect the interactive text data between the interviewer and the candidate during the interview; S2. Extract term associations from the industry knowledge base based on industry attribute information, perform dynamic semantic analysis on fuzzy terms in job requirements text, and generate industry weights and term thresholds; S3, extracting the candidate's answer content for the fuzzy term from the interactive text data, analyzing the correlation between the answer content and the term threshold, and generating correlation analysis results; S4, real-time identification of the interviewer's actual usage context of fuzzy terms, and calculation of the initial deviation based on industry weights; S5. Correct the initial deviation based on the association analysis result, and mark it as a term deviation signal when the corrected deviation exceeds the term threshold; S6. Identify the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, generate an intention deviation result; S7. According to the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight.

[0006] In a preferred embodiment, S1 comprises: S1a, obtain the job requirement text of the target position from the recruitment platform or enterprise database, the job requirement text contains the job responsibilities and ability requirements described in natural language; S1b. Extract industry attribute information based on the enterprise registration information of the target position or the user-specified parameters. The industry attribute information includes the industry classification code and the industry standard capability definition library identifier; S1c, during the interview process, the conversation between the interviewer and the candidate is collected through real-time speech transcription technology to generate interactive text data containing timestamps; S1d. Associate and store the job requirement text, industry attribute information, and interactive text data according to the interview session number.

[0007] In a preferred embodiment, S2 includes: S2a, based on the industry classification code, access the corresponding industry characteristic data set in the preset industry knowledge base, the industry characteristic data set contains the mapping relationship between industry terms and job competency indicators; S2b, extracting term association rules in the industry knowledge base according to the industry standard capability definition library identifier, the term association rules include the core capability indicators and association strength coefficients of fuzzy terms in different industry scenarios; S2c, semantically deconstruct the fuzzy terms in the job requirements text, and generate dynamic semantic analysis results related to industry characteristics by combining core competency indicators and correlation strength coefficients; S2d. Set industry weights according to the correlation strength coefficients of the core capability indicators in the dynamic semantic analysis results, and dynamically calculate term thresholds based on the historical score distribution of the core capability indicators.

[0008] In a preferred embodiment, S3 includes: S3a, locate the candidate's answer segment from the interactive text data, the candidate's answer segment is the candidate's continuous dialogue text after the interviewer asks the fuzzy term, and the speaker is marked as the candidate; S3b, contextualize and segment the candidate's answer fragments, identify the depth of technical terms and the frequency of industry case citations related to the core competency indicators, and generate technical dimension analysis data; S3c, extracting logical structure markers and technical terminology density from the candidate's answer fragment, and generating a semantic integrity score based on the integrity assessment rules preset in the industry knowledge base. The integrity assessment rules include indicators of logical chain integrity and terminology application adaptability; S3d. Map the technical terminology depth and case citation frequency in the technical dimension analysis data to the standard reference value of the industry knowledge base, calculate the term matching deviation, and generate a qualified association analysis result when the term matching deviation is less than the term threshold and the semantic integrity score meets the standard.

[0009] In a preferred embodiment, S4 includes: S4a, extracting the context fragments of the interviewer asking the fuzzy terms from the interactive text data, and identifying the actual usage context type based on the preset context classification library; S4b, matching the weight allocation rules of the corresponding scenarios in the industry knowledge base according to the identified context type, dynamically adjusting the industry weight according to the scenario keyword coverage, and generating the context adaptation weight; S4c, analyzing the deviation direction of the core capability index of the fuzzy terms in the question context fragment, and calculating the classification deviation degree in combination with the context adaptation weight; S4d, summing up the classification deviations of different context types according to the context adaptation weights to generate an initial deviation.

[0010] In a preferred embodiment, the context classification library contains semantic feature keywords of technical scenarios, management scenarios and collaboration scenarios; the core capability indicator deviation direction is determined by comparing the focus of the question with the standard capability definition of the industry knowledge base.

[0011] In a preferred embodiment, S5 includes: S5a, verify the validity of the technical term depth and case citation frequency in the association analysis results, and generate a correction factor when the technical term depth is lower than the standard level of the industry knowledge base and the case citation frequency does not meet the standard; S5b. Proportional adjustment is made to the initial deviation according to the correction factor, which is dynamically set based on the ratio of the industry weight to the deviation from the mean of the historical scenario; S5c, weighting the adjusted deviation and the semantic integrity score to generate a composite deviation; S5d. When the composite deviation exceeds the term threshold, it is marked as a term deviation signal and associated with the corresponding core capability indicator and context type.

[0012] In a preferred embodiment, S6 includes: S6a. According to the core competency indicators and context types associated with the term deviation signal, the repeated question tolerance threshold of the corresponding indicator is extracted from the industry knowledge base. The repeated question tolerance threshold is dynamically set based on the deviation risk coefficient of the industry weight and historical question behavior statistics; S6b. Count the number of questions asked by the interviewer on the same core competency indicator and the interval between questions from the interactive text data. When the number of questions exceeds the tolerance threshold for repeated questions and the interval between questions is less than the reasonable interval preset in the industry knowledge base, mark it as repeated questioning behavior. S6c, verify whether the technical term depth and semantic completeness scores in the association analysis results meet the term threshold. When the term threshold is met at the same time and there is repeated questioning behavior, generate an intention deviation result; S6d. Bind the intention deviation results with the industry weights and context types associated with the term deviation signals to generate structured deviation records.

[0013] In a preferred embodiment, S7 includes: S7a, matching correction rules from the industry knowledge base according to the term deviation signal type and the associated context type, the correction rules including the weight adjustment ratio and the feedback trigger condition, the feedback trigger condition is set based on the difference between the number of repeated questions in the intention deviation result and the deviation tolerance threshold; S7b, generating a dynamic adjustment coefficient according to the difference ratio between the corrected deviation degree and the deviation tolerance threshold value in the intention deviation result, and calculating the real-time score weight correction value in combination with the weight adjustment ratio; S7c, judging whether to generate a real-time feedback instruction according to the feedback triggering condition, and triggering a feedback instruction including a weight correction value and a deviation type prompt when the difference in the number of repeated questions exceeds the preset intervention threshold of the industry knowledge base; S7d. Add the real-time score weight correction value to the industry weight to generate an updated score weight and store it in the interview evaluation decision record.

[0014] In another aspect, the present invention provides an interviewer evaluation bias correction system based on real-time feedback, comprising: Data collection module: obtains the job requirements and industry attribute information of the target position, and collects the interactive text data between the interviewer and the candidate during the interview process; Semantic parsing module: extracts term associations from the industry knowledge base based on industry attribute information, performs dynamic semantic parsing on fuzzy terms in job requirements text, and generates industry weights and term thresholds; Relevance analysis module: extracts the candidate's answer content for fuzzy terms from the interactive text data, analyzes the correlation between the answer content and the term threshold, and generates the relevance analysis results; Deviation calculation module: real-time identification of the interviewer's actual usage context of fuzzy terms, and calculation of initial deviation based on industry weights; Deviation marking module: corrects the initial deviation based on the association analysis results, and marks it as a term deviation signal when the corrected deviation exceeds the term threshold; Behavior recognition module: identifies the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, an intention deviation result is generated; Feedback generation module: Based on the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By building an evaluation and correction mechanism that combines dynamic semantic analysis with real-time feedback, the objectivity and industry adaptability of interview evaluations are effectively improved, and industry characteristics are deeply embedded in the entire process of term analysis and evaluation monitoring: the term association rules based on the industry knowledge base dynamically analyze the fuzzy terms in job requirements, and combine industry weights and term thresholds to generate multi-dimensional evaluation indicators, which can accurately reflect the differences in actual requirements for the same competency item in different industries; through the context recognition of real-time interactive texts and the correlation verification of candidates' answers, the potential deviations of interviewers in question focus and term usage habits can be dynamically captured, and term deviation signals can be marked in a timely manner based on the calculation of composite deviation, so as to avoid the implicit evaluation bias caused by standard misalignment from continuously affecting decision-making; 2. The synergy of intention deviation recognition and dynamic score adjustment significantly improves the self-correction ability of the evaluation process. While real-time monitoring of interviewers' repeated questions, context deviation and other behaviors, the system combines the compliance status of candidates' answers with the correction rules of the industry knowledge base to generate targeted feedback instructions and dynamically adjust the score weights to ensure that the evaluation standards are always aligned with the real needs of the position. This two-way verification mechanism not only solves the semantic distortion problem caused by traditional one-way analysis, but also enhances the adaptability of the evaluation model to complex scenarios through dynamic adaptation of industry weights and real-time data, thereby achieving a better balance between the accuracy and fairness of talent selection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flow chart of the interviewer evaluation bias correction method based on real-time feedback of the present invention; Figure 2 It is a structural schematic diagram of the interviewer evaluation bias correction system based on real-time feedback of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0018] Embodiment 1: Figure 1 The present invention provides an interviewer evaluation bias correction method based on real-time feedback, which includes the following steps: S1. Obtain the job requirements and industry attribute information of the target position, and collect the interactive text data between the interviewer and the candidate during the interview; S2. Extract term associations from the industry knowledge base based on industry attribute information, perform dynamic semantic analysis on fuzzy terms in job requirements text, and generate industry weights and term thresholds; S3, extracting the candidate's answer content for the fuzzy term from the interactive text data, analyzing the correlation between the answer content and the term threshold, and generating correlation analysis results; S4, real-time identification of the interviewer's actual usage context of fuzzy terms, and calculation of the initial deviation based on industry weights; S5. Correct the initial deviation based on the association analysis result, and mark it as a term deviation signal when the corrected deviation exceeds the term threshold; S6. Identify the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, generate an intention deviation result; S7. According to the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight.

[0019] S1. Obtain the job requirements and industry attribute information of the target position, and collect the interactive text data between the interviewer and the candidate during the interview, including: S1a, obtain the job requirement text of the target position from the recruitment platform or enterprise database, the job requirement text contains the job responsibilities and ability requirements described in natural language; S1b. Extract industry attribute information based on the enterprise registration information of the target position or the user-specified parameters. The industry attribute information includes the industry classification code and the industry standard capability definition library identifier; S1c, during the interview process, the conversation between the interviewer and the candidate is collected through real-time speech transcription technology to generate interactive text data containing timestamps; S1d. Associate and store the job requirement text, industry attribute information, and interactive text data according to the interview session number.

[0020] In step S1a, the specific method of obtaining the job requirement text of the target position is: calling the job requirement text of the target position through the public interface of the recruitment platform or the enterprise database, and the job requirement text contains the job responsibilities and ability requirements described in natural language. For example, when the target position is "Java development engineer", the job requirement text may contain ability requirements described in natural language such as "proficient in Spring framework" and "experience in distributed system design". The job requirement text obtained in step S1a will be used as input data for dynamic semantic parsing in subsequent steps.

[0021] In step S1b, the method for extracting industry attribute information is: determine the industry attribute information according to the industry classification code registered in the registration information of the enterprise to which the target position belongs in the industrial and commercial department, or determine the industry attribute information according to the industry classification parameters manually specified by the user in the recruitment system. The industry classification code adopts the four-digit code in the national standard "National Economic Industry Classification", for example, "software development" corresponds to the industry classification code "6513". The industry standard capability definition library identifier is used to associate the capability definition data in the pre-built industry knowledge base, for example, the "Internet Industry Capability Definition Library" corresponds to the identifier "IT_SKILL_LIB_2023".

[0022] In step S1c, the process of generating interactive text data is as follows: during the interview, the conversation between the interviewer and the candidate is converted into text data through real-time speech transcription technology, which is implemented by a speech recognition engine based on an acoustic model and a language model. The converted text data is marked with a timestamp in the order of the conversation, for example, the start time of each conversation text is marked as "2023-08-20 14:30:05". The timestamp is used for time series analysis in the subsequent steps to identify the actual use context of the interviewer's use of fuzzy terms in real time.

[0023] In step S1d, the specific method of data association storage is: assign a unique interview session number to each interview, such as "INTERVIEW_20230820_001", and associate and store the position requirement text obtained in step S1a, the industry attribute information extracted in step S1b, and the interactive text data generated in step S1c according to the interview session number. When storing, the position requirement text and the industry attribute information are associated in the form of key-value pairs. For example, the key "POSITION_REQ" corresponds to the position requirement text content, and the key "INDUSTRY_ATTR" corresponds to the industry classification code and the industry standard capability definition library identifier. The interactive text data is stored in a structured data table in timestamp order, and the data table contains the fields "timestamp", "speaker identity", and "conversation content".

[0024] S2. Extract term associations from the industry knowledge base based on industry attribute information, perform dynamic semantic analysis on fuzzy terms in job requirements text, and generate industry weights and term thresholds, including: S2a, based on the industry classification code, access the corresponding industry characteristic data set in the preset industry knowledge base, the industry characteristic data set contains the mapping relationship between industry terms and job competency indicators; S2b, extracting term association rules in the industry knowledge base according to the industry standard capability definition library identifier, the term association rules include the core capability indicators and association strength coefficients of fuzzy terms in different industry scenarios; S2c, semantically deconstruct the fuzzy terms in the job requirements text, and generate dynamic semantic analysis results related to industry characteristics by combining core competency indicators and correlation strength coefficients; S2d. Set industry weights according to the correlation strength coefficients of the core capability indicators in the dynamic semantic analysis results, and dynamically calculate term thresholds based on the historical score distribution of the core capability indicators.

[0025] In step S2a, the specific method of accessing the corresponding industry characteristic data set in the preset industry knowledge base based on the industry classification code is: according to the industry classification code extracted in step S1b, the corresponding industry characteristic data set is matched from the industry knowledge base. The industry classification code adopts the four-digit code in the national standard "National Economic Industry Classification", and its industry characteristic data set contains the mapping relationship between the core terms of the industry and the job competency indicators. In the industry characteristic data set of the software development industry, "code ability" is mapped to job competency indicators such as "code review pass rate" and "unit test coverage". The industry characteristic data set is constructed through the job competency white paper issued by the industry standardization organization and stored in the industry knowledge base.

[0026] In step S2b, the specific method for extracting term association rules is: according to the industry standard capability definition library identifier in step S1b, load the corresponding capability definition library from the industry knowledge base, and parse the term association rules therein. The term association rules are stored in the form of a structured data table, and each rule contains fuzzy terms, core capability indicators and association strength coefficients. For example, for the fuzzy term "communication ability", in the term association rules of the sales industry, the core capability indicators are "customer demand conversion rate" and "cross-departmental collaboration frequency", and the association strength coefficients are 0.8 and 0.6 respectively, where the association strength coefficient is pre-set by industry experts based on the importance of job capabilities. In the term association rules of the technical industry, the core capability indicators are "clarity of technical solution explanation" and "normativeness of document writing", and the association strength coefficients are 0.7 and 0.5 respectively.

[0027] In step S2c, the specific process of dynamic semantic analysis is: semantically deconstruct the fuzzy terms in the job requirement text obtained in step S1a, and the deconstruction process is based on the semantic rule base preset in the industry knowledge base, which is formulated by industry experts based on the usage habits of terms in different scenarios. For example, when the fuzzy term "leadership" appears in the job requirement text and the industry attribute information is "Internet industry", it is deconstructed into "technical decision-making participation" (association strength coefficient 0.75) and "team project leadership times" (association strength coefficient 0.7) according to the semantic rule base; if the industry attribute information is "financial industry", it is deconstructed into "risk decision accuracy" (association strength coefficient 0.85) and "team resource coordination efficiency" (association strength coefficient 0.65). The dynamic semantic analysis results are stored in the form of key-value pairs, with the key being the fuzzy term and the value being a combination of the core capability indicator and its association strength coefficient.

[0028] In step S2d, the method for generating the industry weight and term threshold is as follows: according to the correlation strength coefficient of the core capability indicator of the dynamic semantic analysis result in step S2c, the industry weight is set according to the absolute value of the correlation strength coefficient. For example, if the correlation strength coefficient of the core capability indicator "technical decision-making participation" is 0.75, then its industry weight is 0.75.

[0029] The dynamic calculation method of the term threshold is: based on the historical score distribution data of the same core competency indicator in the industry standard competency definition library, the upper quartile of the historical score data is taken as the threshold. For example, in the historical score data of "Technical Decision Participation" in the Internet industry, the upper quartile is 85 points, so the term threshold is set to 85 points. The historical score distribution data is extracted from the enterprise's historical interview database and stored by industry classification code and core competency indicator.

[0030] S3. Extract the candidate's answer content for the fuzzy term from the interactive text data, analyze the correlation between the answer content and the term threshold, and generate correlation analysis results, including: S3a, locate the candidate's answer segment from the interactive text data, the candidate's answer segment is the candidate's continuous dialogue text after the interviewer asks the fuzzy term, and the speaker is marked as the candidate; S3b, contextualize and segment the candidate's answer fragments, identify the depth of technical terms and the frequency of industry case citations related to the core competency indicators, and generate technical dimension analysis data; S3c, extracting logical structure markers and technical terminology density from the candidate's answer fragment, and generating a semantic integrity score based on the integrity assessment rules preset in the industry knowledge base. The integrity assessment rules include indicators of logical chain integrity and terminology application adaptability; S3d. Map the technical terminology depth and case citation frequency in the technical dimension analysis data to the standard reference value of the industry knowledge base, calculate the term matching deviation, and generate a qualified association analysis result when the term matching deviation is less than the term threshold and the semantic integrity score meets the standard.

[0031] In step S3a, the specific method for locating the candidate's answer fragment is to filter out the conversation content in which the speaker identity is marked as the candidate from the interactive text data generated in step S1c, and intercept the continuous conversation text after the interviewer asks the fuzzy term as the candidate's answer fragment. For example, when the interviewer asks "Please describe how you demonstrate leadership in the project", the text fragment in which the candidate answers "I led the cross-departmental collaboration project and coordinated 5 teams to complete the microservice architecture upgrade" is marked as the answer fragment. The speaker identity tagging is achieved by extracting the "speaker identity" field in the structured storage table of the interactive text data. This field is automatically annotated during speech transcription through existing voiceprint recognition technology (such as Azure Speaker Recognition), without the need for custom algorithms.

[0032] In step S3b, the process of generating contextual semantic segmentation and technical dimension analysis data is as follows: the candidate's answer fragment is divided into multiple semantic blocks according to natural paragraphs or semantic transition words (such as "first" and "secondly"). For each semantic block, the depth of the technical terms contained therein is identified. The depth of the technical terms is determined based on the pre-set technical terminology hierarchy library in the industry knowledge base. The hierarchy library is constructed based on the skill classification in industry standard documents (such as the "Software Engineer Competency Standards"). For example, "Microservice architecture" belongs to the "Architecture Design" category, and the depth value is 3. At the same time, the frequency of industry case citations is counted. For example, the mention of "Financial Industry Risk Control System Migration Case" in the answer fragment is counted as 1 case citation. The case citation standard is based on the pre-defined industry case list in the industry knowledge base.

[0033] In step S3c, the semantic completeness score is calculated by extracting logical structure marker words (such as "therefore" and "in summary") from the candidate's answer fragment, counting their number and calculating the logical chain completeness score. For example, if the number of logical connecting words in the answer fragment containing the complete "problem description-solution-implementation effect" logical chain is ≥3, the logical chain completeness score is 90 points. The technical term density is calculated by the ratio of the number of technical terms to the total number of words in the answer fragment. For example, if the proportion of technical terms exceeds 15%, the term density score is A. Combined with the integrity assessment rules formulated by industry experts in the industry knowledge base (such as a logical chain completeness weight of 60% and a term application adaptability weight of 40%), a comprehensive semantic completeness score is generated.

[0034] The adaptability of term application is determined by comparing the matching ratio of the technical terms used by the candidate with the mainstream technical term list of the current position in the industry knowledge base. For example, the matching ratio of "Spring Cloud" in the candidate's answer and "Kubernetes" in the mainstream term list is 70%.

[0035] In step S3d, the calculation method of term matching deviation and the generation method of correlation analysis results are as follows: the technical term depth and case citation frequency obtained in step S3b are mapped to the standard reference value of the same core capability indicator in the industry knowledge base, and the standard reference value is obtained based on historical interview data statistics. For example, the standard technical term depth of "architecture design capability" in the Internet industry is level 3 and the case citation frequency standard is 2 times. If the technical term depth in the candidate's answer is level 2 and the case citation frequency is 1 time, the term matching deviation is calculated as (1 / 3 + 0 / 2) × 100% = 33%.

[0036] When the term matching deviation is less than the term threshold generated in step S2d (for example, the threshold for the Internet industry is 40%) and the semantic integrity score of step S3c meets the standard (for example, ≥80 points), it is marked as a qualified association analysis result. The qualified association analysis result is stored in the form of structured data, including the technical term depth deviation value, case citation frequency deviation value and semantic integrity score.

[0037] S4. Real-time identification of the interviewer’s actual usage context of fuzzy terms and calculation of initial deviation based on industry weights, including: S4a, extracting the context fragments of the interviewer's questions about fuzzy terms from the interactive text data, and identifying the actual usage context type based on the preset context classification library, which contains semantic feature keywords of technical scenarios, management scenarios, and collaboration scenarios; S4b, matching the weight allocation rules of the corresponding scenarios in the industry knowledge base according to the identified context type, dynamically adjusting the industry weight according to the scenario keyword coverage, and generating the context adaptation weight; S4c, analyzing the core competency index deviation direction of fuzzy terms in the question context fragment, and calculating the classification deviation degree in combination with the context adaptation weight. The core competency index deviation direction is determined by comparing the question focus with the standard competency definition of the industry knowledge base; S4d, summing up the classification deviations of different context types according to the context adaptation weights to generate an initial deviation.

[0038] In step S4a, the process of real-time identification of the interviewer's actual usage context of fuzzy terms is as follows: extract the conversation content before and after the question position of the fuzzy term from the interactive text data generated in step S1c as context fragments. For example, the two sentences after the interviewer asked the question "Please explain how you demonstrate leadership in the project" "This project involves multiple teams, and I am responsible for technical decisions..." are captured as context fragments.

[0039] The context type of the context fragment is identified based on the preset context classification library. The context classification library is constructed by industry experts based on the job capability scenario division, and contains semantic feature keywords for technical scenarios, management scenarios, and collaboration scenarios. For example, the keywords for technical scenarios include "architecture design" and "code performance", the keywords for management scenarios include "team coordination" and "project planning", and the keywords for collaboration scenarios include "cross-department communication" and "resource integration". The method for identifying the context type is to count the number of matches with the keywords of each scenario in the context fragment, and the scenario with the largest number of matches is determined as the current context type.

[0040] In step S4b, the specific implementation of dynamically adjusting the industry weight is as follows: according to the context type determined in step S4a, the weight ratio adjustment coefficient of the corresponding scenario is read from the industry knowledge base. The ratio coefficient is obtained by industry experts based on the analysis of the importance of the scenarios in the historical interview data. For example, the adjustment coefficient for the technical scenario is +20%, the management scenario is -15%, and the collaboration scenario is -5%. Based on the original industry weight (the "technical decision-making participation" weight of 0.75 from step S2d), if the current scenario is a technical scenario, the adjusted context adaptation weight is 0.75×1.2=0.9; if it is a management scenario, it is adjusted to 0.75×0.85=0.64. The adjustment coefficient is set based on the statistical results of the correlation between the recruitment success of the ability indicators in different scenarios in the same industry.

[0041] In step S4c, the specific method for judging the deviation direction of the core competency indicators and calculating the classification deviation is: compare the core competency indicators mentioned in the context of the interviewer's question with the standard competency definition of the industry knowledge base. The standard competency definition comes from the term association rules in step S2b. For example, the standard definition of "architecture design capability" in the Internet industry is "technology selection" (weight 40%), "performance optimization" (weight 35%), and "code specification" (weight 25%). If 80% of the interviewer's questions focus on "technology selection", the deviation direction is excessive emphasis on "technology selection", and the deviation is calculated as the target proportion difference (80%-40%=40%) multiplied by the context adaptation weight (0.9), resulting in a classification deviation of 36%.

[0042] In step S4d, the logic for generating the initial deviation is: weighted accumulation of classification deviations in multiple context scenarios according to the context adaptation weights. For example, in a certain interview, the technical scenario classification deviation is 36% (context adaptation weight 0.9) and the management scenario classification deviation is 10% (weight 0.64). If the context distribution is 70% for technical scenarios and 30% for management scenarios, then the initial deviation = 36% × 0.7 + 10% × 0.3 = 27%. The weighted weight comes from the context adaptation weight generated in step S4b, ensuring that the deviation of high-weight scenarios has a greater impact on the final result.

[0043] S5. Correct the initial deviation based on the association analysis result, and mark it as a term deviation signal when the corrected deviation exceeds the term threshold, including: S5a, verify the validity of the technical term depth and case citation frequency in the association analysis results, and generate a correction factor when the technical term depth is lower than the standard level of the industry knowledge base and the case citation frequency does not meet the standard; S5b. Proportional adjustment is made to the initial deviation according to the correction factor, which is dynamically set based on the ratio of the industry weight to the deviation from the mean of the historical scenario; S5c, weighting the adjusted deviation and the semantic integrity score to generate a composite deviation; S5d. When the composite deviation exceeds the term threshold, it is marked as a term deviation signal and associated with the corresponding core capability indicator and context type.

[0044] In step S5a, the specific process of verifying the validity of the association analysis results is as follows: extract the technical terminology depth and case citation frequency from the technical dimension analysis data generated in step S3b, and compare them with the standard level preset in the industry knowledge base. The standard level is defined by the job competency level specification issued by the industry standardization organization. For example, the "architecture design capability" in the Internet industry requires the technical terminology depth to reach level 3 (covering sub-items such as architecture mode and performance optimization) and the case citation frequency ≥ 2 times. If the technical terminology depth in the candidate's answer is level 2 and the case citation frequency is 1 time, it is determined to be unsatisfactory and a correction factor is generated. The correction factor is calculated as the weighted sum of the technical terminology depth deviation value (level 1) and the case citation frequency deviation value (1 time), for example, the correction factor = (1 / 3 + 1 / 2) × 0.5 = 0.42, where the weight 0.5 is determined by the statistical results of the deviation of historical scenarios from the mean in the industry knowledge base.

[0045] In step S5b, the correction factor is dynamically set and the initial deviation is adjusted by calculating the correction factor based on the ratio of the industry weight generated in step S2d to the deviation from the mean of historical scenarios. The deviation from the mean of historical scenarios is extracted from the industry knowledge base. For example, the average deviation of the technical terminology depth of "architecture design ability" in the Internet industry in historical interviews is 0.3 levels, and the average deviation of case citation frequency is 0.5 times. Then the correction factor = industry weight (0.75) × (deviation value / historical mean) = 0.75 × (0.42 / 0.4) = 0.79. Multiply the correction factor by the initial deviation (12.5% ​​from step S4d) to obtain the adjusted deviation = 12.5% ​​× 0.79 ≈ 9.88%.

[0046] In step S5c, the composite deviation is generated by weighting the deviation adjusted in step S5b and the semantic integrity score in step S3c. The weighting coefficient is determined by the integrity weight rule preset in the industry knowledge base. For example, when the integrity score is ≥80 points, the deviation weight is 70% and the integrity weight is 30%. Assuming that the adjusted deviation is 9.88% and the semantic integrity score is 85 points, the composite deviation = 9.88% × 0.7 + (100-85)% × 0.3 = 6.92% + 4.5% = 11.42%. The integrity weight rule is set based on the evaluation results of industry experts on the impact of answer quality.

[0047] In step S5d, the marking logic of the term deviation signal is: compare the composite deviation with the term threshold generated in step S2d. For example, the term threshold of "architecture design capability" in the Internet industry is 10%. If the composite deviation is 11.42%>10%, it is marked as a term deviation signal. When marking, the corresponding core capability indicators (such as "architecture design capability") and context types (such as technical scenarios) are associated and stored in a structured data format, such as the fields "deviation signal type: technical scenario-architecture design capability" and "deviation value: 11.42%".

[0048] S6. Identify the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, generate the intention deviation result, including: S6a. According to the core competency indicators and context types associated with the term deviation signal, the repeated question tolerance threshold of the corresponding indicator is extracted from the industry knowledge base. The repeated question tolerance threshold is dynamically set based on the deviation risk coefficient of the industry weight and historical question behavior statistics; S6b. Count the number of questions asked by the interviewer on the same core competency indicator and the interval between questions from the interactive text data. When the number of questions exceeds the tolerance threshold for repeated questions and the interval between questions is less than the reasonable interval preset in the industry knowledge base, mark it as repeated questioning behavior. S6c, verify whether the technical term depth and semantic completeness scores in the association analysis results meet the term threshold. When the term threshold is met at the same time and there is repeated questioning behavior, generate an intention deviation result; S6d. Bind the intention deviation results with the industry weights and context types associated with the term deviation signals to generate structured deviation records.

[0049] In step S6a, the method for dynamically setting the tolerance threshold for repeated questions is: according to the core competency indicator and context type associated with the term deviation signal marked in step S5d, the historical question behavior statistics of the indicator are extracted from the industry knowledge base. The historical question behavior statistics include the average frequency of questions asked by the interviewer on the same competency indicator in historical interviews for the same position and the deviation risk coefficient. The deviation risk coefficient is calculated based on the proportion of the candidate's score reduction due to repeated questions. For example, the historical average question frequency of "architecture design ability" in the Internet industry is 2 times, and the deviation risk coefficient is 0.6 (indicating that 60% of candidates have their scores reduced due to repeated questions).

[0050] The calculation formula of tolerance threshold is: tolerance threshold = historical frequency average × (1-deviation risk coefficient) × industry weight. Assuming the industry weight is 0.75, tolerance threshold = 2 × (1-0.6) × 0.75 = 0.6 times, rounded up to 1 time. The deviation risk coefficient is generated by the historical data analysis model preset in the industry knowledge base. The model is based on linear regression analysis of the relationship between question frequency and score changes in historical interview data.

[0051] In step S6b, the identification logic of repeated questioning behavior is: extract the interviewer's question record on the same core competency indicator from the interactive text data generated in step S1c, and count the number of questions and the time interval between adjacent questions. The reasonable interval preset in the industry knowledge base is set according to the position type. For example, the reasonable interval for technical positions is 15 minutes, and the reasonable interval for management positions is 10 minutes. If the interviewer asks about "architecture design ability" 3 times within 10 minutes (exceeding the tolerance threshold 1 time) and the interval time is less than the reasonable interval, it is marked as repeated questioning behavior. The question record is realized by matching the sentences in the interactive text data where the speaker is the interviewer and contains the core competency indicator keywords.

[0052] In step S6c, the verification process of the association analysis result is to check whether the technical terminology depth and semantic integrity score in the qualified association analysis result generated in step S3d meet the terminology threshold. For example, the technical terminology depth must reach the industry knowledge base standard level 3, and the semantic integrity score must be ≥80 points. If the candidate's answer meets the terminology threshold at the same time and there is repeated questioning behavior marked in step S6b, the intention deviation result is generated. The intention deviation result includes the number of repeated questions, the deviation from the tolerance threshold, and the associated context type (such as technical scenario).

[0053] In step S6d, the method for generating structured deviation records is: bind the intention deviation result with the industry weight (from step S2d) and context type (from step S4a) associated with the term deviation signal to form structured data containing the fields "core capability indicator", "industry weight", "context type", "number of repeated questions", and "deviation tolerance threshold". For example, the structured data record is "core capability indicator: architecture design capability; industry weight: 0.75; context type: technical scenario; number of repeated questions: 3 times; deviation tolerance threshold: 2 times". The structured data is stored in the historical deviation library of the industry knowledge base for correction rule matching and feedback instruction generation in the subsequent step S7.

[0054] S7. Based on the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight, including: S7a, matching correction rules from the industry knowledge base according to the term deviation signal type and the associated context type, the correction rules including the weight adjustment ratio and the feedback trigger condition, the feedback trigger condition is set based on the difference between the number of repeated questions in the intention deviation result and the deviation tolerance threshold; S7b, generating a dynamic adjustment coefficient according to the difference ratio between the corrected deviation degree and the deviation tolerance threshold value in the intention deviation result, and calculating the real-time score weight correction value in combination with the weight adjustment ratio; S7c, judging whether to generate a real-time feedback instruction according to the feedback triggering condition, and triggering a feedback instruction including a weight correction value and a deviation type prompt when the difference in the number of repeated questions exceeds the preset intervention threshold of the industry knowledge base; S7d. Add the real-time score weight correction value to the industry weight to generate an updated score weight and store it in the interview evaluation decision record.

[0055] In step S7a, the matching method of the correction rule is: according to the term deviation signal type (such as "technical scenario-architecture design ability deviation") and the associated context type (such as technical scenario), the preset correction rule is retrieved from the industry knowledge base. The correction rule is formulated by industry experts based on the degree of influence of different deviation types in historical interview data on the accuracy of scoring. For example, the correction rule for the deviation of "architecture design ability" in the technical scenario is: weight adjustment ratio -10% (that is, the weight of the indicator is reduced by 10%), and the feedback trigger condition is set to "the difference in the number of repeated questions ≥ 2 times". The difference in the number of repeated questions is calculated by subtracting the tolerance threshold of step S6a from the actual number of questions counted in step S6b, for example, the actual number of questions asked is 3 times - the tolerance threshold is 1 time = the difference is 2 times.

[0056] In step S7b, the logic for generating the real-time score weight correction value is: generate a dynamic adjustment coefficient based on the proportional relationship between the corrected deviation (the composite deviation of 11.42% from step S5c) and the deviation tolerance threshold difference (such as 2 times) in the intention deviation result. Dynamic adjustment coefficient = corrected deviation / (deviation tolerance threshold difference × preset unit impact factor), where the unit impact factor is set by industry experts based on the impact of each deviation on the score in historical data (for example, the Internet industry is preset to 5%).

[0057] In step S7c, the judgment basis for the feedback trigger condition is: the intervention threshold preset in the industry knowledge base is set according to the position type, for example, the intervention threshold for technical positions is the difference in the number of repeated questions ≥ 1.5 times. If the difference in step S6b is 2 times (≥ 1.5 times), the feedback instruction is triggered. The feedback instruction content includes the weight correction value (0.6645) and the deviation type prompt (such as "Technical scenario-architecture design ability weight is too high"), and the instruction format is "Adjustment indicator: architecture design ability; Corrected weight: 0.6645; Deviation reason: Repeated questioning exceeds the threshold".

[0058] In step S7d, the specific operation of updating the scoring weight is: replacing the original industry weight (0.75) generated in step S2d with the real-time scoring weight correction value (0.6645) to generate an updated scoring weight of 0.6645. The updated weight is associated with the interview session number, core competency indicator, and deviation type and stored in the interview assessment decision record. The record storage format includes the fields "session number: INTERVIEW_20230820_001; indicator name: architecture design capability; original weight: 0.75; revised weight: 0.6645; deviation type: technical scenario-repeated questioning".

[0059] Embodiment 2: Figure 2 A structural schematic diagram of the interviewer evaluation bias correction system based on real-time feedback of the present invention is given. The interviewer evaluation bias correction system based on real-time feedback includes: Data collection module: obtains the job requirements and industry attribute information of the target position, and collects the interactive text data between the interviewer and the candidate during the interview process; Semantic parsing module: extracts term associations from the industry knowledge base based on industry attribute information, performs dynamic semantic parsing on fuzzy terms in job requirements text, and generates industry weights and term thresholds; Relevance analysis module: extracts candidates' answers to fuzzy terms from interactive text data, analyzes the relevance between the answers and term thresholds, and generates relevance analysis results; Deviation calculation module: real-time identification of the interviewer's actual usage context of fuzzy terms, and calculation of initial deviation based on industry weights; Deviation marking module: corrects the initial deviation based on the association analysis results, and marks it as a term deviation signal when the corrected deviation exceeds the term threshold; Behavior recognition module: identifies the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, an intention deviation result is generated; Feedback generation module: Based on the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight.

[0060] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0061] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0062] In addition, each functional module in each embodiment of the present application 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.

[0063] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0064] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for correcting interviewer evaluation bias based on real-time feedback, characterized in that: The steps include: S1. Obtain the job requirements and industry attribute information of the target position, and collect the interactive text data between the interviewer and the candidate during the interview; S2. Extract term associations from the industry knowledge base based on industry attribute information, perform dynamic semantic analysis on fuzzy terms in job requirements text, and generate industry weights and term thresholds; S3, extracting the candidate's answer content for the fuzzy term from the interactive text data, analyzing the correlation between the answer content and the term threshold, and generating correlation analysis results; S4, real-time identification of the interviewer's actual usage context of fuzzy terms, and calculation of the initial deviation based on industry weights; S5. Correct the initial deviation based on the association analysis result, and mark it as a term deviation signal when the corrected deviation exceeds the term threshold; S6. Identify the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, generate an intention deviation result; S7. According to the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight.

2. The interviewer evaluation bias correction method based on real-time feedback according to claim 1 is characterized in that: S1 includes: S1a, obtain the job requirement text of the target position from the recruitment platform or enterprise database, the job requirement text contains the job responsibilities and ability requirements described in natural language; S1b. Extract industry attribute information based on the enterprise registration information of the target position or the user-specified parameters. The industry attribute information includes the industry classification code and the industry standard capability definition library identifier; S1c, during the interview process, the conversation between the interviewer and the candidate is collected through real-time speech transcription technology to generate interactive text data containing timestamps; S1d. Associate and store the job requirement text, industry attribute information, and interactive text data according to the interview session number.

3. The interviewer evaluation bias correction method based on real-time feedback according to claim 1, characterized in that S2 include: S2a, based on the industry classification code, access the corresponding industry characteristic data set in the preset industry knowledge base, the industry characteristic data set contains the mapping relationship between industry terms and job competency indicators; S2b, extracting term association rules in the industry knowledge base according to the industry standard capability definition library identifier, the term association rules include the core capability indicators and association strength coefficients of fuzzy terms in different industry scenarios; S2c, semantically deconstruct the fuzzy terms in the job requirements text, and generate dynamic semantic analysis results related to industry characteristics by combining core competency indicators and correlation strength coefficients; S2d. Set industry weights according to the correlation strength coefficients of the core capability indicators in the dynamic semantic analysis results, and dynamically calculate term thresholds based on the historical score distribution of the core capability indicators.

4. The interviewer evaluation bias correction method based on real-time feedback according to claim 1, characterized in that S3 include: S3a, locate the candidate's answer segment from the interactive text data, the candidate's answer segment is the candidate's continuous dialogue text after the interviewer asks the fuzzy term, and the speaker is marked as the candidate; S3b, contextualize and segment the candidate's answer fragments, identify the depth of technical terms and the frequency of industry case citations related to the core competency indicators, and generate technical dimension analysis data; S3c, extracting logical structure markers and technical terminology density from the candidate's answer fragment, and generating a semantic integrity score based on the integrity assessment rules preset in the industry knowledge base. The integrity assessment rules include indicators of logical chain integrity and terminology application adaptability; S3d. Map the technical terminology depth and case citation frequency in the technical dimension analysis data to the standard reference value of the industry knowledge base, calculate the term matching deviation, and generate a qualified association analysis result when the term matching deviation is less than the term threshold and the semantic integrity score meets the standard.

5. The interviewer evaluation bias correction method based on real-time feedback according to claim 1, characterized in that S4 include: S4a, extracting the context fragments of the interviewer asking the fuzzy terms from the interactive text data, and identifying the actual usage context type based on the preset context classification library; S4b, matching the weight allocation rules of the corresponding scenarios in the industry knowledge base according to the identified context type, dynamically adjusting the industry weight according to the scenario keyword coverage, and generating the context adaptation weight; S4c, analyzing the deviation direction of the core capability index of the fuzzy terms in the question context fragment, and calculating the classification deviation degree in combination with the context adaptation weight; S4d, summing up the classification deviations of different context types according to the context adaptation weights to generate an initial deviation.

6. The interviewer evaluation bias correction method based on real-time feedback according to claim 5, characterized in that: The context classification library contains semantic feature keywords for technical scenarios, management scenarios, and collaboration scenarios; the deviation direction of the core capability indicators is determined by comparing the focus of the questions with the standard capability definitions of the industry knowledge base.

7. The interviewer evaluation bias correction method based on real-time feedback according to claim 1, characterized in that S5 include: S5a, verify the validity of the technical term depth and case citation frequency in the association analysis results, and generate a correction factor when the technical term depth is lower than the standard level of the industry knowledge base and the case citation frequency does not meet the standard; S5b. Proportional adjustment is made to the initial deviation according to the correction factor, which is dynamically set based on the ratio of the industry weight to the deviation from the mean of the historical scenario; S5c, weighting the adjusted deviation and the semantic integrity score to generate a composite deviation; S5d. When the composite deviation exceeds the term threshold, it is marked as a term deviation signal and associated with the corresponding core capability indicator and context type.

8. The interviewer evaluation bias correction method based on real-time feedback according to claim 1, characterized in that S6 include: S6a. According to the core competency indicators and context types associated with the term deviation signal, the repeated question tolerance threshold of the corresponding indicator is extracted from the industry knowledge base. The repeated question tolerance threshold is dynamically set based on the deviation risk coefficient of the industry weight and historical question behavior statistics; S6b. Count the number of questions asked by the interviewer on the same core competency indicator and the interval between questions from the interactive text data. When the number of questions exceeds the tolerance threshold for repeated questions and the interval between questions is less than the reasonable interval preset in the industry knowledge base, mark it as repeated questioning behavior. S6c, verify whether the technical term depth and semantic completeness scores in the association analysis results meet the term threshold. When the term threshold is met at the same time and repeated questioning behavior exists, generate an intention deviation result; S6d. Bind the intention deviation results with the industry weights and context types associated with the term deviation signals to generate structured deviation records.

9. The interviewer evaluation bias correction method based on real-time feedback according to claim 1, characterized in that: S7 includes: S7a, matching correction rules from the industry knowledge base according to the term deviation signal type and the associated context type, the correction rules including the weight adjustment ratio and the feedback trigger condition, the feedback trigger condition is set based on the difference between the number of repeated questions in the intention deviation result and the deviation tolerance threshold; S7b, generating a dynamic adjustment coefficient according to the difference ratio between the corrected deviation degree and the deviation tolerance threshold value in the intention deviation result, and calculating the real-time score weight correction value in combination with the weight adjustment ratio; S7c, judging whether to generate a real-time feedback instruction according to the feedback triggering condition, and triggering a feedback instruction including a weight correction value and a deviation type prompt when the difference in the number of repeated questions exceeds the preset intervention threshold of the industry knowledge base; S7d. Add the real-time scoring weight correction value to the industry weight to generate an updated scoring weight and store it in the interview evaluation decision record.

10. A system for correcting bias in interviewer evaluation based on real-time feedback, used to implement the method for correcting bias in interviewer evaluation based on real-time feedback as described in any one of claims 1 to 9, characterized in that: include: Data collection module: obtains the job requirements and industry attribute information of the target position, and collects the interactive text data between the interviewer and the candidate during the interview process; Semantic parsing module: extracts term associations from the industry knowledge base based on industry attribute information, performs dynamic semantic parsing on fuzzy terms in job requirements text, and generates industry weights and term thresholds; Relevance analysis module: extracts candidates' answers to fuzzy terms from interactive text data, analyzes the relevance between the answers and term thresholds, and generates relevance analysis results; Deviation calculation module: real-time identification of the interviewer's actual usage context of fuzzy terms, and calculation of initial deviation based on industry weights; Deviation marking module: corrects the initial deviation based on the association analysis results, and marks it as a term deviation signal when the corrected deviation exceeds the term threshold; Behavior recognition module: identifies the interviewer's repeated questioning behavior associated with the term deviation signal. When the association analysis result meets the term threshold and there is repeated questioning behavior, an intention deviation result is generated; Feedback generation module: Based on the term deviation signal type, the corrected deviation degree and the intention deviation result, the correction rules of the industry knowledge base are matched to generate feedback instructions and adjust the interviewer's real-time scoring weight.

Images