Method for processing information, program, and information processor

Abstract

To provide a method for processing information which allows reduction of labor to set an evaluation item according to the type of a work.SOLUTION: A computer acquires work type information on the type of a work according to a program. The computer inputs the acquired work type information into a language processing model and acquires information on a plurality of evaluation items for the type of the work.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to an information processing method, a program, and an information processing device. [Background technology]

[0002] Appropriately evaluating the skills of employees in a workplace such as a company, and appropriately evaluating the skills of job seekers at the recruitment site, are important for proper personnel evaluation and recruitment. Patent Document 1 discloses a system for evaluating employees fairly and fairly by making a final evaluation taking into consideration the employee's own evaluation of himself, human factors such as the employee's skills and attitude, and objective numerical values ​​such as the employee's goal achievement rate. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2023-104852 A Summary of the Invention [Problem to be solved by the invention]

[0004] Each company sets in advance the skills required for its employees to perform their jobs, and evaluates and hires employees and job seekers by determining whether they have each skill and the skill level of each skill. However, it is not easy to determine the evaluation items (skills and skill levels required for employees and job seekers) to appropriately evaluate employees and job seekers. In addition, since the skills required differ depending on the type of job, it is necessary to set appropriate evaluation items for each type of job, which places a heavy burden on human resources personnel and recruiters.

[0005] In one aspect, an object of the present invention is to provide an information processing method etc. that makes it possible to reduce the burden of the task of setting evaluation items according to job type. [Means for solving the problem]

[0006] In one aspect, an information processing method includes a process in which a computer executes a process of acquiring job type information related to a job type, inputting the acquired job type information into a language processing model, and acquiring information on a plurality of evaluation items corresponding to the job type. Effect of the Invention

[0007] On the one hand, it is possible to reduce the burden of setting evaluation items according to job type. [Brief description of the drawings]

[0008] [Figure 1] 1 is a block diagram showing an example of the configuration of an information processing device; [Diagram 2] 13 is a flowchart illustrating an example of a procedure for generating evaluation information. [Diagram 3] FIG. 13 is an explanatory diagram showing an example of a screen. [Figure 4] 13 is a flowchart showing an example of a procedure for generating an evaluation score. [Diagram 5] FIG. 13 is an explanatory diagram showing an example of a screen. [Figure 6] 13 is a flowchart showing an example of a procedure for generating a final evaluation score. [Figure 7] FIG. 13 is an explanatory diagram showing an example of a screen. [Figure 8] 13 is a flowchart showing an example of a procedure for generating a final evaluation score according to the second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an information processing method, a program, and an information processing device according to the present disclosure will be described in detail with reference to the drawings showing embodiments thereof.

[0010] (Embodiment 1) In this embodiment, an information processing device is described that uses general-purpose large-scale language models (LLMs) to generate evaluation information for evaluating evaluation targets such as employees and job seekers in various occupations, and evaluates the evaluation targets based on the generated evaluation information. In this embodiment, an example is described in which Japanese is used as the language input to the large-scale language model, but other languages ​​may also be used. Fig. 1 is a block diagram showing an example of the configuration of an information processing device.

[0011] The information processing device 10 is a computer capable of various information processing and transmitting and receiving information, and is a server computer, a personal computer, a tablet terminal, or the like. The information processing device 10 includes a control unit 11, a storage unit 12, a communication unit 13, an input unit 14, a display unit 15, a reading unit 16, and the like, and these units are connected to each other via a bus. The control unit 11 is configured using one or more processors such as a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-purpose computing on graphics processing units), a TPU (Tensor Processing Unit), or an AI chip (semiconductor for AI). The control unit 11 executes various information processing and control processing to be performed by the information processing device 10 by appropriately executing a program P stored in the storage unit 12. Note that when the control unit 11 includes multiple processors, the control unit 11 may execute each process by a different processor.

[0012] The storage unit 12 includes a RAM (Random Access Memory), a flash memory, a hard disk, an SSD (Solid State Drive), and the like. The storage unit 12 stores a program P (program product) executed by the control unit 11 and various data required for executing the program P. The storage unit 12 also stores an application (hereinafter referred to as a skill generation application AP1) for realizing a process of generating evaluation information when evaluating an evaluation target person, an application (hereinafter referred to as a document evaluation application AP2) for realizing a process of evaluating an evaluation target person based on documents such as a resume and a CV of the evaluation target person, and an application (hereinafter referred to as an interview evaluation application AP3) for realizing a process of evaluating an evaluation target person based on answers to questions during an interview. The storage unit 12 also temporarily stores data and the like generated when the control unit 11 executes the program P and each of the applications AP1 to AP3. The storage unit 12 further stores a language generation model M (language processing model). The language generation model M is a general-purpose large-scale language model constructed by performing pre-learning using a large-scale set of sentences, and the like. The language generation model M is trained to perform a calculation to generate output data according to the contents of the input data based on the input data when input data including, for example, text data written in a natural language and / or image data is input, and to output the generated output data. The language generation model M can be configured using algorithms such as GPT (Generative Pre-trained Transformer)-3, GPT-3.5, GPT-4, etc., and may be configured by combining a plurality of algorithms. Note that the language generation model M is not limited to the above-mentioned transformer-based model. The language generation model M is expected to be used as a program module constituting artificial intelligence software. The language generation model M performs a predetermined calculation on the input data and outputs the calculation result, and data such as coefficients and thresholds of functions that define the calculation are stored in the storage unit 12 as the language generation model M.Instead of storing the language generation model M in the storage unit 12, the information processing device 10 may access a language processing server that stores the language generation model M and read it out.

[0013] The communication unit 13 is a communication module for performing processes related to wired or wireless communication, and transmits and receives information to and from other devices via a network. The network may be the Internet or a public telephone line network, or may be a LAN (Local Area Network) constructed in a facility where the information processing device 10 is installed. The input unit 14 accepts an operation input by a user, and sends a control signal corresponding to the operation content to the control unit 11. The input unit 14 includes, for example, a keyboard and a mouse, and a microphone for voice input. The display unit 15 is a liquid crystal display or an organic EL (electroluminescence) display, and displays various information according to instructions from the control unit 11. A part of the input unit 14 and the display unit 15 may be a touch panel configured as an integrated unit.

[0014] The reading unit 16 reads information stored in a portable storage medium 10a including a CD (Compact Disc), a DVD (Digital Versatile Disc), a USB (Universal Serial Bus) memory, an SD (Secure Digital) card, etc. The program P and each of the applications AP1 to AP3 stored in the storage unit 12 may be read by the control unit 11 from the portable storage medium 10a via the reading unit 16 and stored in the storage unit 12. In addition, the program P and each of the applications AP1 to AP3 may be written in the storage unit 12 during the manufacturing stage of the information processing device 10, or the control unit 11 may download them from another device via the communication unit 13 and store them in the storage unit 12.

[0015] In this embodiment, the information processing device 10 may be a multi-computer consisting of multiple computers, or may be a virtual machine virtually constructed by software in one device. In addition, when the information processing device 10 is configured as a server computer, the information processing device 10 may be a local server installed in a facility where the information processing device 10 is installed, or may be a cloud server connected to the information processing device 10 via a network such as the Internet. In the following, the information processing device 10 is described as being one computer. In addition, the program P and each of the applications AP1 to AP3 may be arranged on a single computer or at one site, or may be distributed across multiple sites and distributed and executed on multiple computers connected to each other via a network. Furthermore, the input unit 14 and the display unit 15 are not essential for the information processing device 10, and the information processing device 10 may be configured to accept operations through a connected computer, or may be configured to output information to be displayed to an external display device. In addition, the storage unit 12 of the information processing device 10 may be configured by multiple storage devices, or may include an external storage device connected to the information processing device 10.

[0016] When generating evaluation information for evaluating the evaluation target for various occupations, the information processing device 10 having the above-mentioned configuration receives information on the occupation for which the evaluation information is to be generated as text data described in a natural language, such as text data input via the input unit 14 and text data converted from voice data uttered in a natural language. In addition, when the information processing device 10 has an OCR (Optical Character Recognition) function, the information processing device 10 may receive information on the occupation as text data generated from an image by the OCR function. Then, the information processing device 10 executes a process of generating evaluation information for the occupation from the received information on the occupation, using the language generation model M. In addition, the information processing device 10 executes a process of generating evaluation scores for the evaluation target for each evaluation item included in the evaluation information and questions for the evaluation target from the evaluation information generated for the occupation and documents such as the resume and CV of the evaluation target in the occupation, using the language generation model M. Furthermore, the information processing device 10 executes a process of generating a final evaluation score for each evaluation item of the subject from the evaluation score for each evaluation item and the answer to the question generated for the subject, using the language generation model M. In this way, the information processing device 10 of this embodiment uses the language generation model M when generating evaluation information for each occupation from information about each occupation, when generating evaluation scores and questions for the subject from evaluation information for each occupation and the document of the subject in each occupation, and when generating a final evaluation score for the subject from the evaluation score for each evaluation item of the subject and the answer to the question.

[0017] In this embodiment, the same language generation model M is used in the process of generating evaluation information from job type information, the process of generating evaluation scores and questions for each job type of the evaluation target, and the process of generating the final evaluation score of the evaluation target, but this configuration is not limited to this. For example, a language generation model M for each process may be prepared, and a different language generation model M may be used for each process. In addition, the language generation model M may be configured to use a general-purpose large-scale language model as is, or may be fine-tuned for each process. For example, the language generation model M for the process of generating evaluation information from job type information may be fine-tuned using the work content, project information, recruitment requirements, and hiring requirements for each job type published on a company guide site or a job site. In addition, information that can be used to evaluate the evaluation target may be scraped from this information by keyword search, and fine-tuned using the obtained information.

[0018] In addition, the language generation model M for generating the evaluation score and question of the evaluation target of each job may be a model fine-tuned using training data in which information (documents) about the evaluation target and evaluation information about the job the evaluation target applies for are associated with the evaluation score and question of the evaluation target. The language generation model M here learns to output the correct evaluation score and question in the training data when the document and evaluation information included in the training data are input. Specifically, the language generation model M performs calculations based on the input data and generates an evaluation score and question for each evaluation item according to the content of the input data. The language generation model M then compares the generated evaluation score and question with the correct evaluation score and question, and optimizes various parameters such as the weights (coupling coefficients) between neurons and the coefficients of the activation function using the steepest descent method, the backpropagation method, etc. so that the two are close to each other. As a result, when the document of the evaluation target is input, a language generation model M that outputs an evaluation score and question according to the content of the input document is obtained. Furthermore, the language generation model M for the process of generating the final evaluation score of the person to be evaluated may be a model fine-tuned using training data in which the evaluation scores for each evaluation item determined from the document of the person to be evaluated and the answers (text data) of the person to be evaluated to questions related to each evaluation item are associated with the final evaluation score of the person to be evaluated identified from the answers. Fine-tuning of the language generation model M may be performed by the information processing device 10 or by another learning device. The language generation model M fine-tuned by the other learning device is downloaded from the learning device to the information processing device 10 via a network or a portable storage medium 10a, for example, and stored in the storage unit 12.

[0019] The process of generating evaluation information related to a specified job type by the information processing device 10 of this embodiment will be described below. Fig. 2 is a flowchart showing an example of the evaluation information generation process procedure, and Fig. 3 is an explanatory diagram showing an example screen. The following process is executed by the control unit 11 of the information processing device 10 in accordance with the program P and skill generation application AP1 stored in the storage unit 12.

[0020] A user who generates evaluation information regarding an arbitrary occupation instructs the information processing device 10 to start the skill generation application AP1. When the control unit 11 of the information processing device 10 receives the instruction to start the skill generation application AP1 via the input unit 14, the control unit 11 starts the skill generation application AP1 and executes the following process.

[0021] The control unit 11 of the information processing device 10 displays a job type input screen as shown in FIG. 3A on the display unit 15 (S11). The screen shown in FIG. 3A has input fields for each information of job type name, skill level, company information, and tag, and accepts input of job type information for specifying the target for generating evaluation information. Each input field may be configured to allow any information to be input, or may be configured to have a pull-down menu for selecting any one of multiple options. In the example of FIG. 3A, the input field for tag is configured to allow any information to be input, and the input fields for job type name, skill level, and company information are configured to have pull-down menus, but are not limited to this configuration. For example, the input field for skill level is provided with a pull-down menu for selecting beginner, intermediate, or advanced. Information such as the company's business content and project content, and the recruitment and hiring requirements for each job type is input into the input field for company information. This information may be input directly, or the URL (Uniform Resource Locator) of a website on which this information is posted may be input, or the file name of a file containing this information may be input. The file may be in any file format, such as a pdf (Portable Document Format) file, a word file, a text file, etc. The tag input field is used to input information about the job content and required skills of the person being evaluated. The user inputs at least the job title via the job title input screen.

[0022] The control unit 11 accepts input of each piece of information for each input field of the occupation type input screen via the input unit 14, and displays each piece of accepted information in the corresponding input field (S12). The control unit 11 judges whether or not the Create button in the occupation type input screen has been operated (S13), and if it judges that it has not been operated (S13: NO), returns to the process of step S12 and continues accepting input of information related to the occupation. If it judges that the Create button has been operated (S13: YES), the control unit 11 creates a prompt for creating evaluation information for the occupation based on the information related to the entered occupation (occupation information) (S14). For example, the control unit 11 generates text data in one natural language by associating the information entered in each input field with the items (occupation name, skill level, company information, tag) of each input field. The control unit 11 then adds a message to the beginning of the generated text data to control the output result from the language generation model M, such as "Based on the information of this company, please generate skill evaluation information including the skills necessary to work at this company, the reasons why each skill is important, and the evaluation criteria," to create a prompt. The control unit 11 may also add messages such as "Please list 15 skills necessary to work at this company," or "Please list the skills necessary to work at this company, divided into hard skills related to knowledge and technology, and soft skills related to personality and character," to the text data based on the information input via the job type input screen. The message to be added to the text data may be generated in advance and stored in the storage unit 12. The control unit 11 may be configured to input the text data directly into the language generation model M without adding a message to the text data to acquire evaluation information.

[0023] The control unit 11 inputs the generated prompt into the language generation model M, and generates evaluation information according to the contents of the prompt by the language generation model M (S15). Here, the control unit 11 acquires evaluation information (text data) in which a plurality of evaluation items (skill types) used in the evaluation of the evaluation subject and evaluation criteria for each evaluation item are described in natural language from the language generation model M. Then, the control unit 11 displays the generated evaluation information on the display unit 15 (S16). The control unit 11 displays a screen as shown in FIG. 3B. In the screen shown in FIG. 3B, the evaluation information generated for the data scientist is displayed, and for example, evaluation criteria for each skill (evaluation item) such as programming, statistical analysis, and machine learning are displayed as Hard Skill, and evaluation criteria for each skill (evaluation item) such as communication ability, problem-solving ability, and adaptability are displayed as Soft Skill.

[0024] The control unit 11 assigns a file name to the evaluation information displayed on the screen of FIG. 3B and stores it in the storage unit 12, and when an instruction to end the above-mentioned process is received, the display of the screen of FIG. 3B is ended. The above-mentioned process makes it possible to generate evaluation information that can be used when evaluating the evaluation target in the job type from the job type information input via the job type input screen. In addition, by including company information in the input information, it is possible to generate evaluation information for each company by taking into account the information of each company. The criteria for evaluating the evaluation target may vary depending on the experience or knowledge of the person in charge, but in this embodiment, the evaluation criteria are determined using the language generation model M, so that appropriate evaluation criteria can be generated. Therefore, the burden on the person in charge of setting the evaluation items according to each job type can be reduced, and even a person without specialized knowledge of each job type can set evaluation information that can appropriately evaluate the evaluation target in each job type. Furthermore, in this embodiment, when generating evaluation information from input job information, a prompt is created by adding a message to the job information, and evaluation information is generated from the created prompt. Therefore, the person inputting the job information only needs to input the job name, skill level, company information, etc., and desired evaluation information can be generated without increasing the operational burden.

[0025] In the above-described processing, the processing of creating a prompt from input job type information and / or the processing of generating evaluation information from the prompt using the language generation model M are not limited to being configured to be performed locally by the information processing device 10. For example, these processing may be configured to be performed by a server that provides the language generation model M. For example, the information processing device 10 can be configured to transmit input job type information to the server, and receive evaluation information generated by the server based on the prompt created from the job type information.

[0026] Next, a process for evaluating an evaluation target based on the evaluation information corresponding to each job type generated by the above-mentioned process and the evaluation target's information will be described. Fig. 4 is a flowchart showing an example of the evaluation score generation process procedure, and Fig. 5 is an explanatory diagram showing an example screen. The following process is executed by the control unit 11 of the information processing device 10 in accordance with the program P and document evaluation application AP2 stored in the storage unit 12.

[0027] A user who evaluates an evaluation target such as an employee or a job seeker for a job type for which evaluation information has been generated by the above-mentioned process instructs the information processing device 10 to launch the document evaluation application AP2. When the control unit 11 of the information processing device 10 receives an instruction to launch the document evaluation application AP2 via the input unit 14, the control unit 11 launches the document evaluation application AP2 and executes the following process.

[0028] The control unit 11 of the information processing device 10 displays a document input screen as shown in FIG. 5 on the display unit 15 (S21). The document input screen here is the upper area of ​​the screen shown in FIG. 5, and has input fields for each information of resume, company information, evaluation information, and free text, and accepts input of information about the person to be evaluated (subject information). Each input field may be configured to allow any information to be input, or may be configured to have a pull-down menu for selecting any one of multiple options. In the example of FIG. 5, the input field for free text is configured to allow any information to be input, and the input fields for resume, company information, and evaluation information are configured to have pull-down menus, but are not limited to this configuration. In the input field for resume, the contents of the resume of the person to be evaluated may be directly input, or a file name in which the resume is written may be input. In the input field for company information, information similar to that in the input field for company information in the screen of FIG. 3 may be input, or a company profile in which the company information is summarized may be input. Here again, in the input field for company information, company information or a company profile may be directly input, the URL of a website on which the company information or the company profile is posted may be input, or the file name of a file on which the company information or the company profile is written may be input. In the input field for evaluation information, the file name of the evaluation information generated by the process shown in FIG. 4 is input. In the input field for free text, the blog of the person to be evaluated, information posted on a social networking service (SNS), program code created in the past by the person to be evaluated, etc. may be directly input, the URL of a website on which such information is posted may be input, or the file name of a file on which such information is written may be input. The file format of the file input through each input field is not limited. The user inputs at least a resume and evaluation information through a document input screen.

[0029] The control unit 11 accepts input of each piece of information (document) for each input field of the document input screen via the input unit 14, and displays each piece of accepted information (such as the file name of each document) in each input field (S22). The control unit 11 judges whether or not the evaluation button in the document evaluation screen has been operated (S23), and if it judges that it has not been operated (S23: NO), it returns to the process of step S22 and continues to accept input of documents and the like containing information about the person to be evaluated. If it judges that the evaluation button has been operated (S23: YES), the control unit 11 creates a prompt for generating (calculating) an evaluation score for the person to be evaluated based on the input document and the like (S24). Here, the control unit 11 generates one piece of text data including the description of the document entered in each input field. The control unit 11 then adds a message to the beginning of the generated text data to control the output result from the language generation model M, such as "Based on the resume of the person to be evaluated and the evaluation criteria, please rate the skill level of the person to be evaluated on a 5-point scale for each evaluation item", "Based on the evaluation results of the person to be evaluated, please create questions to be asked of the person to be evaluated. Also, please create reasons for asking the questions and predicted answers to those questions", or "Please create two questions for each evaluation item", to create a prompt. The messages here may also be generated in advance and stored in the storage unit 12.

[0030] The control unit 11 inputs the created prompt to the language generation model M, and generates the evaluation score of the evaluation target person for each evaluation item, a question to be asked to the evaluation target person (deep-dive question), the reason for asking the question, and a predicted answer by the language generation model M (S25). Here, the control unit 11 acquires document evaluation information (text data) in which the above-mentioned information is described in natural language from the language generation model M. The control unit 11 generates at least the evaluation score of the evaluation target person and the deep-dive question by the language generation model M. The control unit 11 may also be configured to first generate an evaluation score for each evaluation item based on the created prompt, and then input a part or all of the created prompt and the generated evaluation score to the language generation model M to generate the deep-dive question, etc. Then, the control unit 11 displays the generated document evaluation information on the display unit 15 (S26). The control unit 11 displays, for example, a screen as shown in FIG. 5. The screen shown in Figure 5 displays document evaluation information generated for a data scientist. For example, for the evaluation items of feature engineering and communication skills, the evaluation score, probing questions, reasons for the questions, and predicted answers are displayed.

[0031] The control unit 11 assigns a file name to the document evaluation information displayed on the screen of FIG. 5 and stores it in the storage unit 12, and when an instruction to end the above-mentioned process is received, the display of the screen of FIG. 5 is ended. Through the above-mentioned process, an evaluation score of the person to be evaluated and in-depth questions to the person to be evaluated can be generated from the evaluation information (evaluation criteria) for each job type and the document in which each information of the person to be evaluated is described. In personnel evaluation, there is a possibility that the evaluation result varies depending on the subjectivity of the person in charge of evaluation. However, in this embodiment, the evaluation score of the person to be evaluated is determined using the language generation model M, so that a fair evaluation can be performed. Therefore, the burden on the person in charge of evaluating employees and job seekers can be reduced. In addition, in this embodiment, when generating document evaluation information from an input document, a prompt is generated by adding a message to the document (text data), and the document evaluation information is generated from the generated prompt. Therefore, the person in charge only needs to specify the file of each information, and can perform appropriate evaluation processing without increasing the operational burden.

[0032] In the above-described processing, the processing of creating a prompt from an input document and / or the processing of generating an evaluation score, a question, and the like from the prompt using the language generation model M are not limited to being configured to be performed locally by the information processing device 10. For example, these processing may be performed by a server that provides the language generation model M. For example, the information processing device 10 can be configured to transmit an input document to the server and receive document evaluation information generated on the basis of a prompt created from the document by the server.

[0033] Next, a process for determining a final evaluation of a person to be evaluated based on the document evaluation information generated by the above-mentioned process and the person's answers to questions will be described. Fig. 6 is a flowchart showing an example of a process for generating a final evaluation score, and Fig. 7 is an explanatory diagram showing an example screen. The following process is executed by the control unit 11 of the information processing device 10 in accordance with the program P and the interview evaluation application AP3 stored in the memory unit 12.

[0034] Using the document evaluation information generated for the person to be evaluated by the above-mentioned process, a user who conducts an interview with the person to be evaluated instructs the information processing device 10 to launch the interview evaluation app AP3. When the control unit 11 of the information processing device 10 receives an instruction to launch the interview evaluation app AP3 via the input unit 14, it launches the interview evaluation app AP3 and executes the following process.

[0035] The control unit 11 of the information processing device 10 displays an interview screen as shown in FIG. 7 on the display unit 15 (S31). The control unit 11 reads the evaluation criteria (evaluation information) and document evaluation information generated for the evaluation target who will be interviewed as the information of the evaluation target. The interview screen here is a video display field R1 on the upper left side of the screen shown in FIG. 7, and displays an image (video) of the face area or upper body of the evaluation target. The captured image of the evaluation target may be image data captured by a camera of a terminal (not shown) used by the evaluation target, or may be image data captured by a camera (not shown) connected to the information processing device 10. The control unit 11 of the information processing device 10 acquires the speech uttered by the evaluation target and outputs the speech from a speaker or earphones (not shown). The speech of the evaluation target may be audio data acquired by a microphone of the evaluation target's terminal, or may be audio data acquired by a microphone connected to the information processing device 10. In the case of a face-to-face interview, the control unit 11 does not need to output the speech of the evaluation target from a speaker or the like.

[0036] The control unit 11 acquires a probing question for one evaluation item from the document evaluation information, and displays it in the question display field R2 on the upper right side of the screen in Fig. 7 (S32). The interviewer asks the subject the probing question displayed in the question display field R2, and the control unit 11 acquires voice data (speech) of the subject's response to the probing question (S33). The control unit 11 converts the acquired voice data into text data, and displays the acquired text data in the question display field R2 on the screen (S34).

[0037] Based on the answer (text data) of the person to be evaluated, the control unit 11 creates a prompt for generating the evaluation score of the person to be evaluated that is updated by the answer and the question to be asked next to the person to be evaluated (S35). Here, the control unit 11 generates one piece of text data including the evaluation criteria (evaluation information) and document evaluation information for the person to be evaluated, the probing questions asked to the person to be evaluated, and the answers to the probing questions. Then, the control unit 11 creates a prompt by adding a message for controlling the output result from the language generation model M to the beginning of the generated text data, such as "Please rate the skill level of the person to be evaluated on a 5-point scale for each evaluation item based on the answers to the questions. If the skill level has changed, please explain the reason." The message here may also be generated in advance and stored in the storage unit 12.

[0038] The control unit 11 inputs the created prompt to the language generation model M, and generates the evaluation score (updated evaluation score) of the evaluation target person for each evaluation item and a question to be asked next to the evaluation target person by the language generation model M (S36). At this time, the control unit 11 may generate a reason for asking the next question and a predicted answer to the next question in addition to the evaluation score and the next question. Here, the control unit 11 acquires final evaluation information (text data) in which each of the above-mentioned information is described in natural language from the language generation model M. Then, the control unit 11 displays the generated final evaluation information on the display unit 15 (S37). The control unit 11 displays a screen as shown in FIG. 7, for example. In the screen shown in FIG. 7, the answer of the evaluation target person is displayed in the question display field R2, and the document evaluation score, the final evaluation score, and the reason for the final evaluation score are displayed in the evaluation display field R3 below for the evaluation item corresponding to the probing question in the question.

[0039] The control unit 11 judges whether there is a next question generated in step S36 (S38), and when it judges that there is a next question (S38: YES), it displays the generated next question in the question display field R2 on the screen (S39). Then, the control unit 11 returns to the process of step S33. If the next question is displayed, the interviewer asks the next question to the subject, and the control unit 11 executes the processes of steps S33 to S37 for the next question. On the other hand, if there is no next question, the interviewer operates the "Next evaluation skill" button. Therefore, when it judges that there is no next question (S38: NO), the control unit 11 judges whether the "Next evaluation skill" button has been operated (S40), and when it judges that it has been operated (S40: YES), it returns to the process of step S32. Then, the control unit 11 repeats the processes of steps S32 to S39 for the evaluation items for which a question has not yet been asked. In addition, if the interviewer wishes to end the questions for the evaluation item even when the next question is displayed, the interviewer may operate the ``Go to next evaluation skill'' button.In this case, the control unit 11 determines that the ``Go to next evaluation skill'' button has been operated (S40: YES) and proceeds to step S32.

[0040] When the control unit 11 determines that the "Next Evaluation Skill" button has not been operated (S40: NO), it determines whether or not an instruction to end the above-mentioned process has been received (S41). For example, the control unit 11 determines whether or not an instruction to end the ongoing interview has been received via the input unit 14. When the control unit 11 determines that an instruction to end has not been received (S41: NO), it returns to the process of step S40 and waits until the "Next Evaluation Skill" button is operated or an instruction to end is received. When it determines that an instruction to end has been received (S41: YES), the control unit 11 ends a series of processes. At this time, the control unit 11 assigns a file name to the information displayed on the interview screen during the interview and stores it in the storage unit 12. This allows the video of the person being evaluated during the interview, the questions asked to the person being evaluated, and their answers to be checked at a later date.

[0041] The above-mentioned process can generate the final evaluation score of the person to be evaluated and the next question to the person to be evaluated from the evaluation information (evaluation criteria) and document evaluation information for the person to be evaluated, and the answers to the questions asked by the person to be evaluated during the interview. Therefore, the next questions to the person to be evaluated are displayed one after another during the interview, so that the interviewer can ask probing questions in order, and it is possible to prevent questions to the person to be evaluated from being missed. In addition, although there is a possibility that the evaluation results will vary depending on the subjectivity of the person in charge of personnel evaluation, in this embodiment, the final evaluation score of the person to be evaluated is determined using the language generation model M, so that it is possible to perform a fair evaluation. Therefore, the burden on the interviewer can be reduced. In addition, in this embodiment, when generating the final evaluation information from the text data of the answer of the person to be evaluated, a prompt is created by adding a message to the text data, and the final evaluation information is generated from the created prompt, so that the person in charge can concentrate on the question and answer session with the person to be evaluated, and the evaluation process can be performed appropriately without increasing the burden on the interviewer.

[0042] In the above-described processing, the processing of creating a prompt from the answer to the probing question of the evaluation subject and / or the processing of generating a final evaluation score and a next question, etc. from the prompt using the language generation model M are not limited to being configured to be performed locally by the information processing device 10. For example, these processing may be performed by a server that provides the language generation model M. For example, the information processing device 10 can be configured to transmit text data of the answer of the evaluation subject to the server and receive final evaluation information generated based on the prompt created by the server from the text data.

[0043] In this embodiment, for example, a configuration is assumed in which a person in charge of setting evaluation information (evaluation criteria) according to a job type executes the process of FIG. 2, a person in charge of evaluating an evaluation target based on the evaluation target's document (work history, etc.) executes the process of FIG. 4, and an interviewer executes the process of FIG. 6, but the configuration is not limited to this. For example, the process of FIG. 4 may be executed after the process of FIG. 2 is executed. In this case, for example, the processes of FIG. 2 and FIG. 4 may be executed for each evaluation target. In addition, each process provided by the information processing device 10 is assumed to be used not only by personnel personnel and recruiters who evaluate evaluation targets, such as human resources agents and job change agents, but also by the evaluation target himself, such as a job seeker. For example, the evaluation target himself can generate evaluation information for a desired job type by inputting job type information via the screen of FIG. 3A, and can grasp the skills required for the desired job type. In addition, the evaluation target himself can grasp his own evaluation score for the skills (each evaluation item) required for the desired job type by inputting his own document, such as his work history, via the screen of FIG. 5. Furthermore, by simulating an interview, the person being evaluated can understand the final evaluation as a result of the interview by answering the probing questions displayed on the screen in Figure 7. By using it in this way, job seekers can understand what skills they lack before actually applying for a job or before actually attending an interview, and can take measures early on.

[0044] (Embodiment 2) In the above-mentioned first embodiment, an information processing device is described that estimates the emotion of the person to be evaluated when answering a question during an interview, in addition to the answer given by the person to be evaluated, and generates a final evaluation score and the next question by taking the estimated emotion into consideration. The information processing device of this embodiment has a configuration similar to that of the information processing device 10 of the first embodiment shown in FIG. 1, and therefore a description of the configuration is omitted.

[0045] FIG. 8 is a flowchart showing an example of a process procedure for generating a final evaluation score in the second embodiment. The process shown in FIG. 8 is the process shown in FIG. 6 with step S51 added between steps S34 and S35. The same steps as in FIG. 6 will not be described. In this embodiment, the control unit 11 of the information processing device 10 estimates the emotion of the evaluation subject based on the captured image and / or the spoken voice of the evaluation subject after the process of step S34 (S51). For example, the control unit 11 estimates the emotion of the evaluation subject in the captured image using an application program that detects the emotion of the subject from the captured image, such as Microsoft's Emotion API. In addition, the control unit 11 extracts information such as tone and strength of voice and speaking speed from the spoken voice of the evaluation subject, and estimates the emotion of the evaluation subject based on the extracted information. The emotion of the evaluation subject includes, for example, tension, relaxation, joy, sadness, anger, surprise, etc., and each emotion is expressed by a numerical value, for example, out of 10 points (10 points being the highest and 0 points being the lowest). In addition, when the information processing device 10 can acquire vital data in addition to the captured image and speech of the evaluation subject, the control unit 11 may estimate the emotion by analyzing the vital data of the evaluation subject. For example, the control unit 11 may extract the heart rate or breathing state from the vital data of the evaluation subject and detect the emotion based on the extracted information. In addition, the control unit 11 may estimate the emotion of the evaluation subject using a neural network. For example, a learning model configured with a CNN (Convolution Neural Network) model and trained to output information on the emotion of the evaluation subject when the captured image and / or speech of the evaluation subject are input may be used. In this case, the control unit 11 inputs the captured image and / or speech of the evaluation subject sequentially acquired during the interview into the learning model, and can estimate the emotion of the evaluation subject based on the output information from the learning model. In addition, a learning model trained to output information on the emotion of the evaluation subject when the vital data is input in addition to the captured image and / or speech of the evaluation subject may be used.

[0046] The process of step S34 and the process of step S51 may be executed in parallel. After that, the control unit 11 executes the processes of step S35 and after. In this embodiment, in step S35, the control unit 11 generates text data including the evaluation criteria (evaluation information) and document evaluation information for the evaluation target, the probing questions asked to the evaluation target, and the answers to the probing questions, as well as the emotion of the evaluation target estimated in step S51, and creates a prompt from the generated text data.

[0047] By the above-mentioned process, in this embodiment, the final evaluation score of the person to be evaluated and the next question to be asked to the person to be evaluated can be generated by taking into consideration not only the content of the answer to the question asked by the person to be evaluated during the interview, but also the emotion of the person to be evaluated when answering. Therefore, since an evaluation score more suitable for the person to be evaluated can be specified, a more appropriate evaluation can be performed. In addition, since a next question more suitable for the person to be evaluated can be generated, the interview can be conducted more smoothly.

[0048] In this embodiment, the process is the same as that of the first embodiment except for the process of generating the final evaluation score and the next question of the person to be evaluated while taking into account the emotions of the person to be evaluated during the interview, and the same effects as those of the first embodiment can be obtained. Also in this embodiment, the modified examples described as appropriate in the first embodiment can be applied.

[0049] The matters described in each of the above embodiments can be combined with each other. In addition, the independent claims and dependent claims described in the claims can be combined with each other in any and all combinations regardless of the citation format. Furthermore, the claims use a format in which a claim cites two or more other claims (multiple claim format), but this is not limited to this. A multiple claim (multi-multi claim) that cites at least one multiple claim may also be used.

[0050] The embodiments disclosed herein are illustrative in all respects and should not be considered as limiting. The scope of the present invention is defined by the claims, not by the above meaning, and is intended to include all modifications within the scope and meaning equivalent to the claims. [Explanation of symbols]

[0051] 10. Information processing device 11 Control section 12 Storage section 13. Communications Department 14 Input section 15 Display M Language Generation Model

Claims

1. Obtain job type information about the job type; The acquired job type information is input into a language processing model to acquire information on a plurality of evaluation items corresponding to the job type. An information processing method in which processing is performed by a computer.

2. The job type information and the skill level in the job type are input to the language processing model to obtain information on a plurality of evaluation items corresponding to the skill level of the job type. The information processing method according to claim 1 , wherein the processing is executed by the computer.

3. Get company information about the company, The acquired job type information and company information are input into the language processing model to acquire information on a plurality of evaluation items corresponding to the job type of the company.

3. The information processing method according to claim 1, wherein the processing is executed by the computer.

4. The information on the evaluation items includes the type of each evaluation item and the evaluation criteria for each evaluation item.

3. The information processing method according to claim 1 or 2.

5. Obtain subject information regarding the person being evaluated; Inputting information about the plurality of evaluation items and the subject information into a language processing model to obtain evaluation information for each of the plurality of evaluation items.

3. The information processing method according to claim 1, wherein the processing is executed by the computer.

6. Inputting information about the plurality of evaluation items and the subject information into a language processing model to obtain questions about the plurality of evaluation items. The information processing method according to claim 5 , wherein the processing is executed by the computer.

7. Obtaining answers to questions regarding the evaluation items; The answer to the question is input into a language processing model to obtain the next question; and Based on the answers to each question, a final evaluation for the evaluation item is determined. The information processing method according to claim 6 , wherein the processing is executed by the computer.

8. Obtain job type information about the job type; The acquired job type information is input into a language processing model to acquire information on a plurality of evaluation items corresponding to the job type. A program that causes a computer to carry out processing.

9. An information processing device having a control unit, The control unit is Obtain job type information about the job type; The acquired job type information is input into a language processing model to acquire information on a plurality of evaluation items corresponding to the job type. Information processing device.

Images