WO2020119563A1 - Applicant evaluation method and device employing neural network model - Google Patents

Abstract

The present application relates to the technical field of information-based corporate recruiting. Embodiments of the present application provide an applicant evaluation method and device employing a neural network model. The method comprises: acquiring a face image of an applicant, and extracting facial feature data in the face image; predicting, on the basis of a physiognomy-personality neural network model, a predicted personality trait type corresponding to the facial feature data of the face image of the applicant; acquiring an expected personality trait type required in a position applied for by the applicant; and comparing the expected personality trait type against the predicted personality trait type to determine a corresponding interview suggestion. Thus, artificial intelligence technology is applied to an interview process so as to eliminate tedious questionnaires, surveys, and reviews; improve the application experience for applicants; and reduce interview costs. The invention also generates a suggestion automatically of whether to give an interview invitation or not, thereby improving the efficiency of automatic resume screening and processing.

Description

Applicant evaluation method and device based on neural network model

This application requires the priority of the Chinese patent application submitted to the Chinese Patent Office on December 14, 2018 with the application number 201811536609.5 and the application name "Applicant Evaluation Method and Device Based on Neural Network Model", the entire contents of which are incorporated by reference In this application.

Technical field

This application relates to the field of artificial intelligence technology, and in particular to a method and device for evaluating candidates based on neural network models.

Background technique

In the recruitment process, companies often need to conduct psychological evaluations of job applicants to determine whether the psychological qualities of job applicants are suitable for the company and the positions they apply for. At the same time, for some important or well-paid positions, there are often too many candidates. If Interviewing one by one inevitably takes up too much time for the interviewer, and also increases the cost burden of personnel recruitment.

Therefore, there are a lot of ways to measure the personality traits of the applicants through the online and offline results of the applicants in the initial stage of recruitment, but this type of evaluation method not only takes the time of the applicant to answer the questions, but also takes up The interviewer's time review led to a further increase in personnel recruitment costs.

technical problem

The purpose of the embodiments of the present application is to provide a candidate evaluation method and device based on a neural network model to at least solve the current complicated personality evaluation process of interviewers during interviews, low user experience of applicants, and enterprise personnel costs Excessive problem.

Technical solution

In order to achieve the above objective, an embodiment of the present application provides an applicant evaluation method based on a neural network model on the one hand, including: acquiring a face image of an applicant; based on a facial-characteristic neural network model, predicting the applicant's face The predicted personality trait type corresponding to the image, wherein the face-personality neural network model is formed by training based on the corresponding training face image and the training personality trait type.

Another aspect of an embodiment of the present application provides a candidate evaluation device based on a neural network model, including: a face image acquisition unit for acquiring a face image of a candidate; a personality trait type prediction unit for a face-character-based The neural network model predicts the predicted personality trait type corresponding to the candidate's face image, wherein the face-personality neural network model is formed by training the corresponding training face image and the training personality trait type.

Another aspect of an embodiment of the present application provides a computer device, including a memory and a processor, where the memory stores a computer program, where the processor executes the computer program to implement the steps of the above-described method.

Another aspect of the embodiments of the present application provides a computer storage medium on which a computer program is stored, wherein, when the computer program is executed by a processor, the steps of the foregoing method of the present application are implemented.

Beneficial effect

Through the above technical solutions, it is proposed to use the trained face-personal neural network model to derive the personality traits corresponding to the face image, thereby applying artificial intelligence technology to the interview process, without complicated questionnaires and reviews, can Efficiently and conveniently predict the predicted personality trait type corresponding to the applicant, improve the applicant's application experience and reduce the interview cost; in addition, compare the expected personality trait type required by the candidate's application position with the predicted personality trait type ，Automatically give recommendations for recommending or not recommending interviews, which can improve the efficiency of automatic resume selection and processing.

BRIEF DESCRIPTION

FIG. 1 is a flowchart of a method for evaluating an applicant based on a neural network model according to an embodiment of the application;

FIG. 2 is a training flowchart of the face-personality neural network model used in the method shown in FIG. 1;

3 is a flowchart of a method for evaluating an applicant based on a neural network model according to another embodiment of the present application;

4 is a schematic diagram of the principle of a method for evaluating an applicant based on a neural network model according to an embodiment of the present application;

5 is a structural block diagram of an applicant evaluation device based on a neural network model according to an embodiment of the present application;

6 is a structural block diagram of an applicant evaluation device based on a neural network model according to another embodiment of the present application;

7 is a structural block diagram of a physical device of an applicant evaluation device based on a neural network model according to an embodiment of the application.

Implementation of this application

The specific implementation manners of the embodiments of the present application will be described in detail below with reference to the drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application.

As shown in FIG. 1, a method for evaluating an applicant based on a neural network model according to an embodiment of the present application includes:

S11. Acquire the face image of the candidate and extract the face feature data in the face image.

With regard to the implementation subject of the method of the embodiments of the present application, on the one hand, it may be a dedicated integrated component, a dedicated server, or a dedicated terminal dedicated to the evaluation of the personality traits of the applicant; on the other hand, it may also be a general-purpose server, where the general-purpose server may It is a server for managing recruitment software, and the recruitment software can be used by applicants on each user terminal device, so that the applicant can interact with the server through the user terminal device, so as to realize the evaluation of the personality characteristics of the corresponding applicant , And all of the above are within the scope of protection of this application.

The face feature data in this embodiment may be face texture feature data, face contour feature point data and other data used to reflect face information, which should fall within the protection scope of the present application; regarding face features The data extraction method can be implemented based on various methods such as deep neural network DNN, convolutional neural network CNN or support vector machine SVM, and all fall within the protection scope of the present application.

S12. Based on the face-characteristic neural network model, predict the predicted personality trait type corresponding to the face feature data of the candidate’s face image, where the face-characteristic neural network model is based on multiple pre-collected face image samples Face feature data and personality trait type samples are formed through training.

In this embodiment, the face-characteristic neural network model trained by "training face image-training personality trait type" is applied to predict the personality trait type corresponding to the candidate's face image based on artificial intelligence technology.

Regarding the training process of the facial-characteristic neural network model, in some embodiments, it may be implemented through the process shown in FIG. 2 as follows:

S21. Send a search request to the network search engine server, where the search request includes keywords about the face image.

S22. Acquire a training facial image in response to the search request from the search engine server, and extract facial feature data in the training facial image.

The keywords may be various keywords related to faces, such as "face" and "avatar". In addition, in addition to collecting training facial images in S21 and S22, training facial images can be supplemented based on manual collection. For the face feature data and the extraction method thereof, reference may be made to the description of the foregoing embodiment, and details are not described here.

S23. Statistical training personality trait types corresponding to the training facial image, wherein the training personality trait types include manually marked information.

Among them, each training facial image can be manually marked with a corresponding training personality trait type, which can be marked by industry experts based on experience; in addition, it can also be a statistical personality analysis report corresponding to the training facial image, Therefore, based on the personality analysis report, the corresponding personality trait types are obtained to label each training facial image.

S24. Input the statistical training personality trait types and corresponding training personality trait types into the facial-characteristic neural network model to train the facial-characteristic neural network model.

The specific training process of the neural network should not be limited here, it can be determined according to the network type of the applied neural network model, such as a convolutional neural network or a deep neural network, etc., thereby realizing the use of face images The sample's face feature data and personality trait type samples are used to train the face-personality neural network model.

With regard to personality trait types, it refers to a parameter that companies need to investigate during the recruitment process, which can reflect to a certain extent whether the personality of the applicant is suitable for a specific job position. As an example, the personality trait types may include one or more of the following: eagle personality, bee personality, wolf personality, cat personality, fox personality, dog personality, monkey personality, lion personality, and Panda personality.

The following will describe and explain the details and differences of the nine types of personality traits:

1. Perfect type: Patience, perseverance, keeping promises, implementing consistently, loving family and family, obeying the law, influential leaders, like to control, and being upright.

The most obvious personality characteristics of perfect personality: on principle, self-control is strong, and dominance is strong.

The perfect type corresponds to animals: eagle, eagle type personality is perfectionist, demanding, careful, cruel, principled. Just about.

2. Full love type: longing for the love or good relationship of others, willing to accommodate others. People-oriented, very concerned about the feelings and needs of others, very enthusiastic, willing to give love to others.

The most prominent personality traits of the whole love personality are: happy giving, happy group high,

The full-love type corresponds to animals: bees, bee-type personalities are unknown, caring, and sacrifice-oriented. Love will.

3. Achievement type: Strong sense of competence, like authority, often compare with others, measure one's own value highland by achievements, hope to get everyone's affirmation.

The most notable personality traits of achievement personality: self-confidence, love comparison, change, smooth and pragmatic.

Achievement-type corresponding animals: wolves, wolf-type personalities pursue achievements, have a strong desire for self-realization, like to be the focus person, like to express themselves, and be brave.

4. Artistic: emotional, romantic, afraid of being rejected, easy to be depressed, jealous, self-reflection, self-pursuit.

The most prominent personality characteristics of artistic personality: emotional instability and sensitivity.

Artistic type corresponds to animals: cats, cat-type personalities are casual, do things by feelings, emotional, sensitive and willful. Spiritual.

5. Intellectual type: Leng Yan looks at the world, is lonely and likes to stay away from the crowd. Like to think and analyze, but lack of mobility.

The most prominent personality traits of rational personality: high IQ, low music group, low dare, low pragmatism.

The rational type corresponds to animals: fox, fox type has high personality IQ, and likes to have its own space, clean and intelligent.

6. Loyalty: Do things carefully and cautiously, don't trust others easily, do your best and trust authority. The team is very conscious.

The most prominent personality characteristics of loyal personality are: high stability, pragmatism, and suspicious.

Loyalty corresponds to animals: dogs, dog-type personality loyal, pragmatic, and vigilant. Loyalty.

7. Active type: optimistic, fresh, catch up with the trend, do not like to withstand pressure, afraid of negative emotions; want to live a happy life, want to innovate, entertain people, eager to live a more enjoyable life, make the world unhappy Vanished.

The most prominent personality characteristics of active personality are: optimistic, lively, and emotional.

Active type corresponds to animals: monkey, monkey type personality is optimistic, lively, happy.

8. Leadership: Pursue power, emphasize strength, do not rely on others, and have a sense of justice. If you want to talk about things (say), like to do big things; it is an absolute activist, and take action to solve problems as soon as you encounter them. Want to be independent, everything depends on yourself.

The most prominent personality characteristics of a leadership personality: strong dominance, pragmatism, and independence.

Leader-type corresponding animals: lion, lion-type personality strong leadership, strong strength, strong execution, the general.

9. Peaceful: Very gentle, do not like conflicts with others, do not boast, do not like the limelight, and have a weak personality. Want to live in harmony with people

The most prominent personality traits of a peaceful personality: high peace, high safety, low dominance.

Peace-type corresponding animals: panda-type, panda-type personality is indisputable, lazy and gentle, invincible and victorious, benevolent.

It can be seen from the above description that different personality trait types can be competent for different jobs, that is, different job positions will require different types of expected personality traits. For example, Internet product managers may be talents who need fox-type personality.

S13. Obtain the type of expected personality traits required by the candidate's application position.

S14. Compare the expected personality trait type with the predicted personality trait type to determine the corresponding interview suggestion, where the interview suggestion indicates that the candidate is recommended or not recommended to participate in the interview.

Specifically, it may be that when the expected personality trait type is the same as the predicted personality trait type, a recommendation is given to recommend the candidate to participate in the interview, and when the expected personality trait type is different from the predicted personality trait type, the candidate is not recommended to participate in the interview automatically Suggestions.

As shown in FIG. 3, a method for evaluating an applicant based on a neural network model according to an embodiment of the present application includes:

S31. Acquire the face image of the candidate, and extract face feature data in the face image.

S32. Based on the face-characteristic neural network model, predict the predicted personality trait types corresponding to the face feature data of the candidate’s face image, where the face-characteristic neural network model is a person who has collected multiple pre-collected face image samples Face feature data and personality trait type samples are formed through training.

S33. Obtain the type of expected personality traits required by the candidate's application position.

For details of S31-S33, reference may be made to the related description in the embodiment of FIG. 1, and details are not described herein.

S34. Based on the investigation item reference table, determine the interview question corresponding to the expected personality trait type, wherein the investigation item reference table stores the mapping relationship between the expected personality trait type and the interview question.

Various personality trait evaluation systems used in current related technologies, such as 16PF personality test, Holland professional interest test, mbti professional personality test, etc., generally explore the deep-level personality characteristics of people through the form of questions; However, companies also need to spend time by exploring the differences between the various personality evaluation systems to select the appropriate evaluation system, which is very inconvenient and not smart enough. In contrast, S34 in the method of the embodiment of the present application can better solve this problem, and can intelligently present candidates with a level of interview questions, without the need for corporate personnel to repeatedly think about what evaluation should be based on the inquiry System to ask interview questions.

In addition, in some application scenarios, if the interviewer has doubts about the results derived from the face-personal neural network model, or if it is necessary to further verify the personality type of the interviewer, targeted interview questions can be automatically generated for the interview For reference or use by officials or candidates.

In a more preferred embodiment, the evaluation method for the personality traits of the applicant may further include the following operations:

S35. Obtain a feedback option selected by the user for multiple options under the interview question, wherein the multiple options are used to indicate different personality trait types, respectively.

The reference table for the investigation item may be a storage mapping relationship "expected personality trait type-interview question-options-indicated personality trait type", so that when the interview question is pushed, multiple options under the question can also be pushed. As an example, a problem and multiple options under the problem will be displayed on the user terminal, and the user may select the option on the terminal to upload the feedback option to the server.

S36. Based on the feedback options, verify the predicted personality trait types.

S37. If the personality trait type indicated by the feedback option matches the predicted personality trait type, determine that the predicted personality trait type is correct.

As an example, for the position of "Internet product manager", the expected personality trait type is the fox personality type trait. When the question asked is "Whether you like to go out with friends or play games at home on weekends." The corresponding options can be "going out activities" and "playing games at home". If the candidate chooses "going out activities", it means that the applicant has good communication skills and meets the needs of the job. Further, if the face attributes predicted by the neural network also correspond to the fox-type personality traits, it proves that the personality traits derived by the neural network model are correct.

S38. If the personality trait type indicated by the feedback option does not match the predicted personality trait type, it is determined that the predicted personality trait type is incorrect.

Among them, if the personality trait type deduced by the neural network model is the fox personality type, and the personality trait type indicated by the option of the candidate's feedback through the interview question is the eagle personality, it may be considered that the feedback of the interview question result is more The true point, and it can be determined that the predicted personality trait type corresponding to the neural network model is incorrect.

S39. Associate and store the personality trait type indicated by the feedback option and the candidate's face image to supplement the training data for training the facial-personal neural network model.

In this embodiment, the candidate's feedback option for the interview question may be used to use the personality trait type indicated by the feedback to train the neural network model for second optimization, so as to improve and calibrate the function of the neural network model so as to be able to The next time he applies the neural network model, he can derive personality trait types that are more accurate or more suitable for facial images.

Therefore, combining the results predicted by the neural network model and the reference table of the investigation items, the interview questions corresponding to the key investigation items are designed to ensure that no talents are missed and the efficiency of personality trait recognition can be improved.

In the applicant evaluation method based on the neural network model provided in the embodiments of the present application, the interviewer's picture is identified based on the neural network, so as to derive the corresponding personality type; it can analyze the personality of the recruiter qualitatively and can The interviewer or the applicant provides corresponding investigation suggestions, and can also obtain a more accurate personality type based on the feedback of the applicant to assist the neural network model to avoid sifting out talents who meet the position.

As shown in FIG. 4, the execution of the method of the embodiment of the present application mainly includes the training phase of the neural network model and the application phase of the neural network model:

1) Training stage of neural network model

First, collect the face pictures and the corresponding character report points.

The collection method of facial images should not be limited here. As an example, on the one hand, it can be artificially collected, and the corresponding personality report points can be obtained by expert analysis, where the personality report points can include facial attributes, which can be It is the nine personality types (eagle personality, bee personality, wolf personality, cat personality, fox personality, dog personality, monkey personality, lion personality, panda personality) including the applicant's characteristics and their Corresponding personality type; on the other hand, it can also collect face pictures through website crawler technology, for example, it can be to input keywords corresponding to the face picture to the search engine (it may preferably be a search engine maintained by a talent recruitment website) To obtain the face picture (for example, resume picture) obtained by the search engine server in response to the keyword, and then people can obtain the key points of the personality report (that is, personality type) corresponding to the face picture based on expert knowledge.

Then, input the collected data as a training data source to the neural network model to train the neural network model; that is, input the mapping relationship of personality type-face image to the neural network model to complete the training of the neural network model .

2) Application stage of neural network model

The neural network model is mainly used in the recruitment interview process, for example, it can be configured in the server. And, after obtaining the frontal image of the candidate, the server inputs it to the neural network model, and then obtains the personality type for the interviewer from the neural network model, and transmits the personality type to the interviewer. After the interviewer obtains the personality type, there will be a comparison between the desired personality type required by the position and the personality type output by the neural network model.

First, the server can obtain the avatar of the candidate. Among them, the server may be directly an interview operation server that operates recruitment software, or may be an additional server that communicates with the interview operation server, thereby obtaining the front image of the interviewer from the interview operation server.

Then, the frontal image is input to the neural network model, and the corresponding personality type is derived from the neural network model.

It should be noted that different job types require different personality types, for example, Internet product manager positions require "fox-type personality"; the interviewer obtains the personality type required by the job type (such as fox-type personality), and Compare with the personality type derived from the neural network model.

Preferably, a job requirement table may be stored in the server, and a mapping relationship of “job type-expected personality type” may be stored in the job requirement table, from which the server obtains the corresponding job type by extracting keywords in the job description Then, the server determines the corresponding expected personality type according to the job demand table and the extracted job type.

In order to avoid missing talents, in a preferred embodiment, the server will also push questions and key survey items corresponding to the deduced personality type. The key survey items may be the specific options corresponding to the question, and according to the interviewer's Question the feedback of each option to verify whether the deduced personality type is biased. Among them, the server can pre-configure the survey item reference table, and store the mapping relationships "eagle personality-question 1-option A", "fox personality-question 2-option B" and "monkey personality-question" in the survey item reference table 1-Option C” etc.

Specifically, the server will construct a question based on the expected personality type corresponding to the position and/or the personality type derived from the neural network model, for example, it may be based on the expected personality type to propose key investigation items to further verify whether the interviewer corresponds to the expected personality type Or, according to the deduced personality type, a key investigation item is proposed to verify whether the personality type deduced by the neural network model is accurate, that is, to verify the reliability of the neural network.

More preferably, in the reference table of the investigation items above, the corresponding questions and options may be determined according to the expected personality type and the deduced personality type, for example, "expected personality type-derived personality type-problem m-option n", at this time, It can be based on the expected personality trait type and predicted personality trait type query query item reference table to determine the corresponding interview questions; that is, when both the "expected personality type" and "derived personality type" are determined, then look up the table Only the corresponding questions will be output.

Further, the server may send the question to the interviewer, so that the interviewer can present the interviewer to the interviewer in the process of actually interviewing the interviewer, and obtain the feedback of the interviewer's options for the question (that is, key investigation items).

Preferably, the key investigation item can be regarded as further feedback or calibration of the personality estimated by the neural network. For example, when the neural network infers that the person interviewed is a fox-type personality based on the face, but in order to avoid the neural network’s personality The possible deviation of the output results in the wrong screening of talents. At this time, it is necessary to construct questions in conjunction with the key investigation items, and further confirm or calibrate the deduced personality based on the interviewers' answers. As an example, the key investigation item may be a question customized to a specific personality and/or specific options under it, for example, the personality type of the interviewer derived by the neural network is the fox personality, and the key investigation item corresponding to the fox personality in the database may be Is the question "Which people like to play with friends on weekends or stay at home to play games?" Option A "Weekends like to play with friends on weekends" is more sure that the interviewer belongs to the fox personality. When choosing other options, it may point to other personalities. In this case, it may be necessary to infer the type of personality corresponding to the interviewee again, for example, a comprehensive evaluation by the interviewer in combination with the results of other interview questions; more preferably , When the interviewer’s feedback results deviate from the results derived by the neural network, it can be combined with the interviewer’s feedback results for multiple interview questions to determine the interviewer’s personality type (which can be artificially judged), and then, The determined personality type and the interviewer's image are input to the neural network as a new training data source for training, thereby achieving calibration and compensation of the neural network derivation function.

In an application scenario, if Interviewer A is applying for the position of an Internet product manager, the position requires talents of fox personality, and the model predicts that the interviewer is an eagle personality, and the interviewer can know the interviewer and the position The degree of compliance is average. In order to further verify the prediction results of the model to avoid screening talents, the model also designed related questions for the interviewer according to the applicant's personality (for example, eagle type) and job requirements (fox type), such as generally watching movies or friends with friends on weekends Play games at home. If the applicant likes to work with friends, it shows that the applicant has good communication skills and meets the needs of the job.

It should be noted that the training phase and the application phase of the neural network model in the embodiments of the present application do not necessarily need to set a necessary sequence. As described above, the neural network model may also be generated for the secondary Train the training data of the neural network model to ensure the high reliability of the neural network model.

As shown in FIG. 5, a candidate evaluation device based on a neural network model according to an embodiment of the present application includes:

The face image obtaining unit 501 is used to obtain a face image of an applicant and extract face feature data in the face image;

The personality trait type prediction unit 502 is used to predict the predicted personality trait type corresponding to the face feature data of the candidate’s face image, wherein the face-personality neural network model is a plurality of pre-collected face images The sample's face feature data and personality trait type samples are formed through training;

The desired personality trait type obtaining unit 503 is used to obtain the desired personality trait type required by the applied position;

The interview suggestion determination unit 504 is used to compare the expected personality trait type and the predicted personality trait type to determine the corresponding interview suggestion, wherein the interview suggestion indicates whether to recommend or not recommend the candidate to participate in the interview .

In a specific application scenario, as shown in FIG. 6, the device further includes an interview question determination unit 505, a personality trait type verification unit 506, a training data supplement unit 507, and a training unit 508;

The interview question determination unit 505 may be used to determine the interview question corresponding to the expected personality trait type based on the investigation item reference table after acquiring the desired personality trait type required by the candidate position, wherein the investigation item reference table Stored the mapping relationship between the types of expected personality traits and interview questions;

The interview question determination unit 505 may also be used to query the investigation item reference table according to the expected personality trait type and the predicted personality trait type to determine the corresponding interview question, wherein the investigation item reference table The mapping relationship includes the mapping relationship between expected personality trait types, predicted personality trait types and interview questions;

The personality trait type verification unit 506 may be used to obtain a feedback option selected by the user for multiple options under the interview question after determining the interview question corresponding to the desired personality trait type based on the reference table of the investigation item , Wherein the plurality of options are used to indicate different personality trait types, based on the feedback option, verify the predicted personality trait type, and, if the personality trait type indicated by the feedback option and the predicted personality trait If the types match, it is determined that the predicted personality trait type is correct;

The training data supplementing unit 507 is configured to determine the prediction if the personality trait type indicated by the feedback option does not match the predicted personality trait type after verifying the predicted personality trait type based on the feedback option The personality trait type is incorrect, and the personality trait type indicated by the feedback option and the candidate’s face image are stored in association to supplement the training data for training the facial-personal neural network model in;

The training unit 508 is configured to send a search request to a network search engine server, where the search request includes keywords about a facial image, obtain a training facial image in response to the search request from the search engine server, and count the training A training personality trait type corresponding to the facial image, wherein the training personality trait type includes a manually marked personality trait type, and the statistically trained personality trait type and the corresponding training personality trait type are input to the facial phase -Personality neural network model to train the face-personal neural network model.

It should be noted that, for other corresponding descriptions of each functional unit involved in an applicant evaluation device based on a neural network model provided by an embodiment of the present application, reference may be made to the corresponding descriptions in FIGS. 1-4, and details are not described herein again.

Based on the above method shown in FIGS. 1-4, correspondingly, an embodiment of the present application further provides a storage device on which a computer program is stored. Appraisal method of candidates based on neural network model.

Based on the above method shown in FIGS. 1-4 and the embodiment of the virtual device shown in FIGS. 5 and 6, in order to achieve the above purpose, as shown in FIG. 7, an embodiment of the present application also provides a neural network-based model The physical device 70 of the applicant appraisal device includes a storage device 701 and a processor 702; the storage device 701 is used to store a computer program; the processor 702 is used to execute the computer program to implement the above Appraisal method of candidates based on neural network model shown in Figure 1-4.

By applying the technical scheme of this application, artificial intelligence technology can be applied to the interview process, without complicated questionnaires and reviews, can effectively and conveniently predict the predicted personality trait type corresponding to the applicant, and improve the applicant's Application experience and reduced interview costs.

Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by hardware, or by software plus a necessary general hardware platform. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, U disk, mobile hard disk, etc.), including several The instruction is used to enable a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in each implementation scenario of this application.

An embodiment of the present application further provides a computer device, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, when the processor executes the computer-readable instructions Implement the candidate evaluation method based on the neural network model shown in Figure 1-4.

Exemplarily, the computer-readable instructions may be divided into one or more modules/units, and the one or more modules/units are stored in the memory and executed by the processor to complete the present Application. The one or more modules/units may be a series of computer-readable instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer-readable instructions in the computer device.

The processor may be a central processing unit (Central Processing Unit, CPU), or may be other general-purpose processors, digital signal processors (Digital signal processors) Signal Processor (DSP), Application Specific Integrated Circuit (ASIC), Field-Programmable Array (Field-Programmable) Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory may be an internal storage unit of the computer device, such as a hard disk or a memory of the computer device. The memory may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a smart memory card (Smart Media Card, SMC), Secure Digital, SD) card, flash card (Flash Card), etc. Further, the memory may also include both an internal storage unit of the computer device and an external storage device. The memory is used to store the computer-readable instructions and other instructions and data required by the computer device. The memory can also be used to temporarily store data that has been or will be output.

The functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or software functional unit.

If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer non-volatile readable storage medium. Based on this understanding, the technical solution of the present application essentially or part of the contribution to the existing technology or all or part of the technical solution can be embodied in the form of a software product, the computer software product is stored in a storage medium Includes several computer-readable instructions to enable a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store computer-readable instructions.

Those of ordinary skill in the art may understand that all or part of the processes in the method of the above embodiments may be completed by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions may be stored in a computer non-volatile In the readable storage medium, when the computer-readable instructions are executed, they may include the processes of the foregoing method embodiments. Wherein, any reference to the memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still implement the foregoing The technical solutions described in the examples are modified, or some of the technical features are equivalently replaced; and these modifications or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims (20)

1. A candidate evaluation method based on neural network model, which is characterized by:

   Acquiring the face image of the candidate, and extracting face feature data in the face image;

   Based on the face-characteristic neural network model, predict the predicted personality trait type corresponding to the face feature data of the candidate's face image, wherein the face-characteristic neural network model is a plurality of pre-collected face image samples The face feature data and personality trait type samples are formed by training;

   Obtain the type of expected personality traits required by the candidate's application position;

   The expected personality trait type is compared with the predicted personality trait type to determine a corresponding interview suggestion, where the interview suggestion indicates whether or not to recommend the candidate to participate in the interview.
2. The method according to claim 1, characterized in that after obtaining the desired personality trait type required by the candidate for the job position, the method further comprises:

   Based on the investigation item reference table, the interview question corresponding to the expected personality trait type is determined, wherein the investigation item reference table includes a mapping relationship between the expected personality trait type and the interview question.
3. The method according to claim 2, wherein the interview question includes a plurality of options, the plurality of options are used to indicate different personality trait types, based on the investigation item reference table, to determine the expected personality After the interview question corresponding to the trait type, the method further includes:

   Obtaining feedback options of multiple options selected by the user for the interview question;

   Verify the predicted personality trait type based on the feedback option;

   If the personality trait type indicated by the feedback option matches the predicted personality trait type, it is determined that the predicted personality trait type is correct.
4. The method of claim 3, wherein after verifying the predicted personality trait type based on the feedback option, the method further comprises:

   If the personality trait type indicated by the feedback option does not match the predicted personality trait type, it is determined that the predicted personality trait type is incorrect;

   The personality trait type indicated by the feedback option and the face image of the candidate are stored in association to supplement to the training data for training the facial-personal neural network model.
5. The method according to claim 1, wherein the method further comprises a training process for the facial-personal neural network model, specifically including:

   Send a search request to a network search engine server, where the search request includes keywords about the face image;

   Acquiring training facial images in response to the search request from the search engine server, and extracting facial feature data in the training facial images;

   Statistic the training personality trait types corresponding to the training facial image, wherein the training personality trait types include manually marked personality trait types;

   The statistical personality trait types and the corresponding training facial image are input to the facial-personal neural network model to train the facial-personal neural network model.
6. The method according to claim 3, wherein the mapping relationship in the investigation item reference table includes a mapping relationship between expected personality trait types, predicted personality trait types and interview questions, wherein the The investigation item reference table determines that the interview questions corresponding to the expected personality trait types include:

   Query the investigation item reference table according to the expected personality trait type and the predicted personality trait type to determine the corresponding interview question.
7. A candidate evaluation device based on neural network model, including:

   A face image obtaining unit, used to obtain a face image of a candidate, and extract face feature data in the face image;

   The personality trait type prediction unit is used to predict the predicted personality trait type corresponding to the face feature data of the candidate's face image based on the face-personal neural network model, wherein the face-personal neural network model is based on Face feature data and personality trait type samples of a pre-collected face image sample are formed through training;

   An expected personality trait type acquisition unit, used to acquire the expected personality trait type required by the candidate's application position;

   The interview suggestion determination unit is used for comparing the expected personality trait type and the predicted personality trait type to determine the corresponding interview suggestion, wherein the interview suggestion indicates whether or not to recommend the applicant to participate in the interview.
8. The device according to claim 7, wherein the device further comprises:

   The interview question determination unit is used to determine the interview question corresponding to the expected personality trait type based on the investigation item reference table after obtaining the desired personality trait type required by the candidate, wherein the investigation item reference table includes the expected personality The mapping relationship between trait types and interview questions.
9. The device according to claim 8, wherein the device further comprises:

   The personality trait type verification unit is used to obtain the feedback option selected by the user for the multiple options under the interview question after determining the interview question corresponding to the desired personality trait type based on the reference table of the investigation item. The plurality of options are respectively used to indicate different personality trait types, based on the feedback option, the predicted personality trait type is verified, and, if the personality trait type indicated by the feedback option is different from the predicted personality trait type If it matches, it is determined that the predicted personality trait type is correct.
10. The device according to claim 9, wherein the device further comprises:

    The training data supplementary unit is configured to determine the predicted personality if the predicted personality trait type does not match the predicted personality trait type after verifying the predicted personality trait type based on the feedback option The trait type is incorrect, and the personality trait type indicated by the feedback option and the candidate’s face image are stored in association to supplement to the training data used to train the facial-personal neural network model .
11. The device according to claim 7, wherein the device further comprises:

    The training unit is configured to send a search request to a network search engine server, where the search request includes keywords about a facial image, obtain a training facial image in response to the search request from the search engine server, and count the training facial The training personality trait type corresponding to the image, wherein the training personality trait type includes a manually marked personality trait type, and the statistically trained personality trait type and the corresponding training personality trait type are input to the face- Personality neural network model to train the face-personal neural network model.
12. A computer nonvolatile readable storage medium, the computer nonvolatile readable storage medium stores computer readable instructions, characterized in that, when the computer readable instructions are executed by a processor, the following steps are realized:

    Acquiring the face image of the candidate, and extracting face feature data in the face image;

    Based on the face-characteristic neural network model, predict the predicted personality trait type corresponding to the face feature data of the candidate's face image, wherein the face-characteristic neural network model is a plurality of pre-collected face image samples The face feature data and personality trait type samples are formed by training;

    Obtain the type of expected personality traits required by the candidate's application position;

    The expected personality trait type is compared with the predicted personality trait type to determine a corresponding interview suggestion, where the interview suggestion indicates whether or not to recommend the candidate to participate in the interview.
13. The computer non-volatile storage medium according to claim 12, wherein the computer-readable instructions are further implemented as follows when executed by the processor:

    After obtaining the type of expected personality traits required by the candidate's application position, based on the investigation item reference table, determine the interview questions corresponding to the expected personality trait type, wherein the investigation item reference table includes the expected personality traits The mapping between types and interview questions.
14. The computer non-volatile storage medium according to claim 13, wherein the interview question includes multiple options, the multiple options are respectively used to indicate different personality trait types, and the computer is readable When the instruction is executed by the processor, the following steps are also implemented:

    After determining the interview question corresponding to the expected personality trait type based on the investigation item reference table, obtain feedback options for the multiple options for the interview question selected by the user;

    Verify the predicted personality trait type based on the feedback option;

    If the personality trait type indicated by the feedback option matches the predicted personality trait type, it is determined that the predicted personality trait type is correct.
15. The computer non-volatile storage medium according to claim 14, wherein the computer-readable instructions are further implemented as follows when executed by the processor:

    If the personality trait type indicated by the feedback option does not match the predicted personality trait type, it is determined that the predicted personality trait type is incorrect;

    The personality trait type indicated by the feedback option and the face image of the candidate are stored in association to supplement to the training data for training the facial-personal neural network model.
16. The computer non-volatile storage medium according to claim 12, wherein the computer-readable instructions are further implemented as follows when executed by the processor:

    Send a search request to a network search engine server, where the search request includes keywords about the face image;

    Acquiring training facial images in response to the search request from the search engine server, and extracting facial feature data in the training facial images;

    Statistic the training personality trait types corresponding to the training facial image, wherein the training personality trait types include manually marked personality trait types;

    The statistical personality trait types and the corresponding training facial image are input to the facial-personal neural network model to train the facial-personal neural network model.
17. A computer device, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, characterized in that, when the processor executes the computer-readable instructions, it is implemented as follows step:

    Acquiring the face image of the candidate, and extracting face feature data in the face image;

    Based on the face-characteristic neural network model, predict the predicted personality trait type corresponding to the face feature data of the candidate's face image, wherein the face-characteristic neural network model is a plurality of pre-collected face image samples The face feature data and personality trait type samples are formed by training;

    Obtain the type of expected personality traits required by the candidate's application position;

    The expected personality trait type is compared with the predicted personality trait type to determine a corresponding interview suggestion, where the interview suggestion indicates whether or not to recommend the candidate to participate in the interview.
18. The computer device according to claim 17, wherein the processor further implements the following steps when executing the computer-readable instructions:

    After obtaining the type of expected personality traits required by the candidate's application position, based on the investigation item reference table, determine the interview questions corresponding to the expected personality trait type, wherein the investigation item reference table includes the expected personality traits The mapping between types and interview questions.
19. The computer device according to claim 18, wherein the interview question includes a plurality of options, the plurality of options are respectively used to indicate different personality trait types, and when the processor executes the computer-readable instructions The following steps are also achieved:

    After determining the interview question corresponding to the expected personality trait type based on the investigation item reference table, obtain feedback options for the multiple options for the interview question selected by the user;

    Verify the predicted personality trait type based on the feedback option;

    If the personality trait type indicated by the feedback option matches the predicted personality trait type, it is determined that the predicted personality trait type is correct.
20. The computer device of claim 19, wherein the processor further implements the following steps when executing the computer-readable instructions:

    If the personality trait type indicated by the feedback option does not match the predicted personality trait type, it is determined that the predicted personality trait type is incorrect;

    The personality trait type indicated by the feedback option and the face image of the candidate are stored in association to supplement to the training data for training the facial-personal neural network model.