Information processing systems, information processing methods, and programs

Abstract

We provide an information processing system capable of generating text tailored to job seekers. [Solution] According to one aspect of the present invention, an information processing system is provided, comprising at least one processor, the processor being configured to perform the following steps by reading a program: in a reception step, first job seeker information including information about the attributes of the job seeker; in a question information extraction step, question information corresponding to the first job seeker information is extracted from reference information, the question information including a question to the job seeker and information about an example answer corresponding to the question, the reference information including a plurality of candidate question information which are pre-generated candidate question information based on accumulated job content; in an acquisition step, the answer from the job seeker to the question information is acquired; and in a job content generation step, text about the job content of the job seeker is generated using second job seeker information, the second job seeker information including the first job seeker information and the answer acquired in the acquisition step.

Description

Technical Field

[0001] The present invention relates to an information processing system, an information processing method, and a program.

Background Art

[0002] Patent Document 1 discloses a technique for assisting in creating a sentence using a template.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] It is preferable to generate a sentence tailored to the job seeker.

[0005] In view of the above circumstances, the present invention aims to provide an information processing system and the like that can generate a sentence tailored to the job seeker.

Means for Solving the Problems

[0006] According to one aspect of the present invention, an information processing system is provided, comprising at least one processor, the processor configured to perform the following steps by reading a program: in a reception step, first job seeker information including information about the attributes of the job seeker; in a question information extraction step, question information corresponding to the first job seeker information is extracted from reference information, the question information including a question to the job seeker and information about an example answer corresponding to the question, the reference information including a plurality of candidate question information which are pre-generated candidate question information based on accumulated job content; in an acquisition step, the answer from the job seeker to the question information is acquired; and in a job content generation step, text about the job content of the job seeker is generated using second job seeker information, the second job seeker information including the first job seeker information and the answer acquired in the acquisition step.

[0007] This configuration makes it possible to provide a document creation support system that can generate documents more tailored to each job seeker.

[0008] Embodiments of the present invention will be described below. The various features shown in the embodiments below can be combined with each other. [Brief explanation of the drawing]

[0009] [Figure 1] This diagram shows the overall configuration of Information Processing System 1. [Figure 2] This diagram shows the hardware configuration of the server device 10. [Figure 3] This diagram shows the hardware configuration of the job seeker terminal 20. [Figure 4] This is a block diagram showing the functions realized by the server device 10 (control unit 11) and the job seeker terminal 20 (control unit 21). [Figure 5] This figure shows an example of the configuration of a vector database. [Figure 6] This flowchart shows an overview of the information processing performed by Information Processing System 1. [Figure 7] This is an activity diagram showing the information processing flow for building a vector database, which is performed by information processing system 1. [Figure 8] This is an explanatory diagram illustrating an example of the relationship between a specific sentence and a question information vector. [Figure 9] This is an activity diagram showing the information processing flow for building a vector database, which is performed by information processing system 1. [Figure 10] This figure shows input screen G3, which is an example of an information input screen. [Modes for carrying out the invention]

[0010] Embodiments of the present invention will be described below with reference to the drawings. The various features shown in the embodiments below can be combined with each other.

[0011] Incidentally, the program for implementing the software appearing in one embodiment may be provided as a non-transitory computer-readable medium, or it may be provided as a downloadable medium from an external server, or it may be provided so that the program is launched on an external computer and its functions are realized on a client terminal (so-called cloud computing).

[0012] Furthermore, in various information processing according to one embodiment, an input and an output corresponding to the input can be realized. Here, as long as an output is obtained as a result of the input, the form of the information referenced in such information processing (hereinafter referred to as "reference information") is not limited. The reference information may be, for example, rule-based information such as a database, a lookup table, or a predetermined function (including a decision formula such as a regression equation constructed by a statistical method), or a pre-trained model that has learned the correlation between input and output in advance, or a large-scale language model that can output a desired result by inputting a prompt.

[0013] In addition, in one embodiment, the "unit" may include, for example, hardware resources implemented by a circuit in a broad sense and information processing of software that can be specifically realized by these hardware resources. Also, in one embodiment, various types of information are handled. These information are represented, for example, by physical values of signal values representing voltage and current, the high and low of signal values as a set of binary bits composed of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculation can be executed on a circuit in a broad sense.

[0014] Furthermore, a circuit in a broad sense is a circuit realized by appropriately combining at least a circuit (Circuit), circuitry (Circuitry), a processor (Processor), a memory (Memory), etc. Also, the processor may be a general-purpose processor or a dedicated circuit. That is, it includes an application specific integrated circuit (ASIC), a programmable logic device (for example, a simple programmable logic device (SPLD), a complex programmable logic device (CPLD), and a field programmable gate array (FPGA)), etc.

[0015] 1. Hardware Configuration In this section, the hardware configuration of the information processing system according to this embodiment will be described.

[0016] FIG. 1 is a diagram showing the overall configuration of the information processing system 1. In FIG. 1, each device included in the information processing system 1 and an overview of the users who use those devices are shown. Regarding each overview, it will be described as needed while referring to other figures.

[0017] The information processing system 1 is an information processing system that executes a text generation process, which is an information processing for assisting in creating a text related to job duties included in a text related to a job seeker, such as a resume. A resume, also called a résumé, is a text in which a job seeker conveys to a potential employer his / her past work experience, skills, qualifications, etc. In this embodiment, the information processing system 1 assists the job seeker U1 in creating a text related to job duties. The text related to job duties may be any text that describes the jobs and tasks, etc. that the job seeker has experienced so far. For example, it also includes a text that summarizes the job duties (job summary). Note that, as an example of the text to be created, a resume is used for explanation, but it is not limited to this, and any text that includes a text related to job duties, such as a job summary, is not limited to this.

[0018] The information processing system 1 includes a communication line 2, a server device 10, and a job seeker terminal 20. The communication line 2 is not particularly limited, but may be constituted by the Internet network. Also, the communication line 2 may include a local area network, may include a mobile communication network, or may include a VPN (Virtual Private Network), etc. The communication line 2 mediates the exchange of data between devices connected to its own line. The server device 10 is connected to the communication line 2 by wire, and the job seeker terminal 20 is connected to the communication line 2 wirelessly. Note that the connection of each device to the communication line 2 may be wired or wireless.

[0019] The server device 10 is an information processing device that performs text generation processing while exchanging data with the job seeker terminal 20 via the communication line 2. The server device 10 stores a vocabulary database DB1 and a job seeker database DB2. The vocabulary database DB1 stores vocabulary used in the text generation processing. The job seeker database DB2 stores job seeker information used by job seeker U1 in their job search activities. Job seeker information refers to information about the job seeker and includes job seeker information entered on the registration screen and pre-registered information (registration information), as well as job seeker information entered on the input screen for creating texts about job duties (input information), and includes all information about the registered or entered job seeker. Here, the registration information and input information may be stored in separate databases. The server device 10 has artificial intelligence (AI) functionality and generates and outputs texts about job duties using the information stored in the vocabulary database DB1 and the job seeker database DB2.

[0020] The vocabulary database DB1 may store templates used in the text generation process. The vocabulary database DB1 may also store question information related to questions used to collect vocabulary corresponding to the templates. The question information includes information about the questions and their corresponding answers. The answer information is reference information for the answer and includes example answers, answer choices, etc. The answer choices are at least one option to present to the job seeker U1. The number of choices can be set as appropriate, but one example is 2 to 5 choices. The example answers and answer choices may be single words or sentences. The vocabulary database DB1 may store question information corresponding to each template. The vocabulary database DB1 may also be a vector database (vector DB) containing vector information of templates and question information.

[0021] The job seeker terminal 20 is a terminal used by job seeker U1. The job seeker terminal 20 can be a smartphone, tablet, personal computer, etc. The job seeker terminal 20 receives input necessary for text generation processing and displays the generated text related to the job description.

[0022] Figure 2 shows the hardware configuration of the server device 10. The server device 10 comprises a control unit 11, a storage unit 12, a communication unit 13, and a bus 14. The bus 14 electrically connects each part of the server device 10.

[0023] (Control Unit 11) The control unit 11 may include at least one processor. The at least one processor may consist of, for example, a Central Processing Unit (CPU), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), one or more Integrated Circuits, one or more Discrete Circuits, or a combination thereof. The control unit 11 is a computer that realizes various functions related to the information processing system 1 by reading predetermined programs stored in the storage unit 12. That is, information processing by software stored in the storage unit 12 is specifically realized by the control unit 11, which is an example of hardware, and can be executed as each functional unit included in the control unit 11. These will be described in more detail in the next section. Note that the control unit 11 is not limited to being a single unit; it may be implemented with multiple control units 11 for each function, or a combination thereof.

[0024] (Storage unit 12) The storage unit 12 stores various types of information as defined above. This can be implemented, for example, as a storage device such as a solid-state drive (SSD) or hard disk drive (HDD) that stores various programs related to the information processing system 1 executed by the control unit 11, or as a memory such as random access memory (RAM) that stores temporarily necessary information (arguments, arrays, etc.) related to program calculations. The storage unit 12 stores various programs and variables related to the information processing system 1 executed by the control unit 11.

[0025] (Communications Section 13) The communication unit 13 is composed of a communication module. The communication module may be a wireless communication module compliant with standards such as IEEE 802.11a / b / g / n / ac / ax, LTE, 5G, 6G, or a wired communication module compliant with standards such as IEEE 802.3. The communication unit 13 is configured to transmit various electrical signals from the server device 10 to external components. The communication unit 13 is also configured to receive various electrical signals from external components to the server device 10. More preferably, the communication unit 13 may have a network communication function, thereby enabling the communication of various information between the server device 10 and external devices via the communication line 2.

[0026] Figure 3 shows the hardware configuration of the job seeker terminal 20. The job seeker terminal 20 comprises a control unit 21, a storage unit 22, a communication unit 23, an input unit 24, an output unit 25, and a bus 26. The bus 26 electrically connects each part of the job seeker terminal 20. The control unit 21, storage unit 22, and communication unit 23 are similar hardware to the control unit 11, storage unit 12, and communication unit 13 shown in Figure 2, although their specifications and models may differ.

[0027] (Input section 24) The input unit 24 receives operation inputs made by the user. The operation inputs are transmitted as command signals to the control unit 21 via the bus 26. The control unit 21 can perform predetermined controls or calculations based on the transmitted command signals as needed. The input unit 24 may be included in the housing of the job seeker terminal 20 or it may be external. For example, the input unit 24 may be implemented as a touch panel integrated with the output unit 25. When the input unit 24 is implemented as a touch panel, the user can input tap operations, swipe operations, etc. to the input unit 24. Instead of a touch panel, the input unit 24 can be a switch button, mouse, trackpad, QWERTY keyboard, etc.

[0028] (Output section 25) The output unit 25 displays a graphical user interface (GUI) screen that can be operated by the user. The output unit 25 may be included in the housing of the job seeker terminal 20 or it may be an external component. Specifically, the output unit 25 may be implemented as a display device such as a CRT display, liquid crystal display, organic EL display, or plasma display. It is preferable that these display devices be used according to the type of job seeker terminal 20.

[0029] 2. Functional Configuration This section describes the functional configuration of this embodiment. Information processing by software stored in the memory unit 12 is specifically realized by the control unit 11, which is an example of hardware, and can be executed as each functional unit included in the control unit 11 (the processor provided by the information processing system 1).

[0030] Figure 4 is a block diagram showing the functions realized by the server device 10 (control unit 11) and the job seeker terminal 20 (control unit 21).

[0031] As shown in Figure 4A, the server device 10 (control unit 11) comprises a reception unit 111, a memory control unit 112, an artificial intelligence unit 113, a generation unit 118, an extraction unit 115, a presentation unit 116, an acquisition unit 117, a generation unit 118, a conversion unit 119, and a memory control unit 112. As shown in Figure 4B, the job seeker terminal 20 (control unit 21) comprises a user display unit 211 and an operation reception unit 212.

[0032] <Reception Desk 111> The reception unit 111 is configured to receive information from the job seeker terminal 20 or other information processing terminals. The reception unit 111 is also configured to receive various types of information by reading various types of information stored in the storage area, which is at least a part of the memory unit 12, and writing the read information to the work area, which is at least a part of the memory unit 12. The storage area is, for example, the area of ​​the memory unit 12 that functions as a storage device such as an SSD. The work area is, for example, the area that functions as memory such as RAM. For example, the reception unit 111 is configured to receive input based on terminal operations of the job seeker U1 to the job seeker terminal 20. Specifically, the reception unit 111 is configured to receive input from the job seeker U1. The reception unit 111 may also be configured to receive editing of documents by the job seeker U1. The reception unit 111 may accept editing of documents based on the job seeker's terminal operations. The reception unit 111 is configured to receive job seeker information from the job seeker U1. The reception unit 111 may receive job seeker information of job seeker U1 via the job seeker terminal 20. The reception unit 111 may also retrieve job seeker information of job seeker U1 stored in the storage unit 12 from the job seeker database.

[0033] <Memory Control Unit 112> The memory control unit 112 controls the storage unit 12 of the device and writes and reads data to and from the storage unit 12. For example, the memory control unit 112 stores the job seeker information (input information) about job seeker U1 entered on the input screen. The memory control unit 112 may also register the job seeker information (registration information) of job seeker U1, which is generated or updated in response to the input of job seeker U1, and the job seeker information (input information) entered on the input screen for creating a document about the job description, in the job seeker database.

[0034] The job seeker database registers job seeker information used by job seeker U1 in their job search activities. Job seeker information includes information about the job seeker, including pre-registered information entered on the registration screen (registration information) and information entered on the input screen for creating job description documents (input information), and encompasses all information about the job seeker that has been registered or entered. Job seeker information includes information related to resumes and work history documents. A "resume" is a document that mainly describes the job seeker's profile, current situation, educational background, work history, desired working conditions, etc. A "work history document," also called a resume, is a document in which the job seeker conveys their past work history, experience, skills, qualifications, etc. to the employer. Employers also include recruitment agencies that act as agents for organizations, mediating between job seekers and organizations. Recruitment agencies are also called headhunters or agents.

[0035] The registration information entered into the job seeker database includes basic information about job seeker U1, information about job seeker U1's work history, information about job seeker U1's skills, and a job summary (summary) which is an overview of the work history. Basic information about job seeker U1 includes, for example, job seeker U1's name, age, address, past annual income, educational background, work history, etc. Basic information about job seeker U1 may also include the industry and job type that the job seeker desires, as well as desired conditions for the organization to which they apply. Information about job content includes information about the organizations that job seeker U1 has worked in the past, and information describing the duties and tasks performed in those organizations. Job content may include, for example, the job seeker's previous organizations, department names, job titles, periods of employment, positions held, projects worked on, periods of work, the job seeker's management experience, language skills, awards, etc.

[0036] The registration information registered in the job seeker database may include the job seeker's job-seeking activity record. The job seeker's job-seeking activity record includes the job seeker U1's actions toward job postings and the job seeker U1's actions in the information processing system 1. The job seeker U1's actions toward job postings include, for example, viewing job postings, applying for job postings, adding job postings to a bookmark list, replying to scout messages sent based on job postings, and passing the screening process for job postings (document screening, interview screening, etc.). The job seeker U1's actions toward job postings consist of information indicating the job posting that was the target or starting point of the job-seeking activity and the content of that job-seeking activity. The job seeker U1's actions in the information processing system 1 include the number of times the job seeker information (registration information) of the job seeker U1 is updated, the update frequency, the number of times the job seeker information registration screen is displayed, the display frequency, the number of times messages are checked, the check frequency, etc. The job seeker U1's job-seeking activity record is recorded for each job seeker (i.e., linked to the job seeker information), for example, in the job seeker database.

[0037] The memory control unit 112 may store template candidate vectors, which are vectors of template candidates, in association with question information candidate vectors, which are vectors of question information candidates corresponding to the template candidates. Figure 5 shows an example of the configuration of the vector DB. The vector DB includes template T1, template T2, template vector TV1, template vector TV2, question information vector QAV1-1, question information vector QAV1-2, question information vector QAV2-1, and question information vector QAV2-2. Template T1 is, for example, a template predefined for generating a resume, which is an example of a document describing job duties. Template T1 may also be called a document structure or resume structure, and for example, it defines the structure and keywords of a document describing job duties. The resume structure template may define the document format and font, etc. The template includes, for example, item names such as job title, industry, job duties, and selling points, and a document template for each item, as necessary structure for a document describing job duties. Template vector TV1 is obtained by converting template T1 into a vector, and template vector TV2 is obtained by converting template T2 into a vector. Question information vectors QAV1-1 and QAV1-2 are obtained by converting the question information corresponding to template T1 into vectors, and question information vectors QAV2-1 and QAV2-2 are obtained by converting the question information corresponding to template T2 into vectors.

[0038] The vector database stores template T1 and template vector TV1, which is a vectorized version of the template, in association. The vector database may also store template vector TV1 and the corresponding question information vector QAV1-1 and question information vector QAV1-2 in association. Template T2 and template vector TV2, which is a vectorized version of the template, in association. The vector database may also store template vector TV2 and the corresponding question information vector QAV2-1 and question information vector QAV2-2 in association.

[0039] <Artificial Intelligence Department 113> The artificial intelligence unit 113 is configured to receive input from each functional unit and return the instructed output. The artificial intelligence used by each functional unit of the server device 10 may be common to all units, or it may be individually prepared for each functional unit. In the following, the artificial intelligence unit 113 may also be referred to as the "artificial intelligence module".

[0040] The artificial intelligence unit 113 is an AI (Artificial Intelligence) that includes transformers such as GPT (Generative Pretrained Transformer, including GPT-1, GPT-2, GPT-3, and GPT-4), BERT (Bidirectional Encoder Representations from Transformers), BART (Bidirectional and Auto-regressive Transformer), and language models such as recurrent neural networks (RNNs), and includes generative AI.

[0041] The language model is an example of a learning model using a machine learning algorithm. Specific machine learning algorithms include nearest neighbors, naive Bayes, decision trees, support vector machines, and deep learning using neural networks. The artificial intelligence unit 113 can apply the above algorithms as appropriate.

[0042] The artificial intelligence unit 113 has a trained model constructed by a learning method such as supervised learning, unsupervised learning, or self-supervised learning. In supervised learning, machine learning is performed using training data. Training data consists of pairs of input data and output data (correct answer data) for training. Furthermore, the language model may not only be one trained for a specific task, but also a general-purpose model that can be used universally for a wide range of tasks.

[0043] The artificial intelligence unit 113 includes a general-purpose natural language processing learning model, such as a large-scale language model (LLM) that has learned from a large amount of data. Such a general-purpose learning model includes language models that can handle various tasks without fine-tuning using methods such as one-shot learning and few-shot learning. Furthermore, the general-purpose learning model can also handle various tasks using zero-shot learning. The artificial intelligence used in each functional unit of the control unit 11 may be a separate learning model, or it may be a common general-purpose learning model.

[0044] The learning model included in the artificial intelligence unit 113 can undergo additional learning. For example, the artificial intelligence unit 113 learns whether the text describing the job duties of job seeker U1 that was created has been modified by job seeker U1 or others. In other words, as feedback to the text describing the job duties of job seeker U1 created by the learning model, the artificial intelligence unit 113 uses the actual modified text describing the job duties by job seeker U1 as training data to perform additional learning and fine-tune the model. This optimizes the text describing the job duties of job seeker U1 output from the learning model.

[0045] The artificial intelligence unit 113 may include a component extraction model, a template candidate generation model, a question information candidate generation model, a question information extraction model, a template extraction model, a job content generation model, and the like.

[0046] The component extraction model may be a large-scale language model, or it may be a trained model that takes job description-related texts as input and outputs common components in those job description-related texts. If the template candidate generation model is a large-scale language model, the generation unit 118 may, as a component extraction step, input instructions to the model to take job description-related texts as input and output common components in the input job description-related texts.

[0047] The template candidate generation model may be a large-scale language model, or a learning model trained to take job description-related sentences as input and output template candidates for job description-related sentences. If the template candidate generation model is a large-scale language model, the generation unit 118 may, as a template candidate generation step, input an instruction to the model to take job description-related sentences as input and output template candidates. Alternatively, the template candidate generation model may be a learning model trained to take specific sentences that satisfy predetermined conditions for learning as input and output template candidates for job description-related sentences. If the template candidate generation model is a large-scale language model, the generation unit 118 may, as a template candidate generation step, input an instruction to the model to take specific sentences that satisfy predetermined conditions as input and output template candidates. Alternatively, the template candidate generation model may be a learning model trained to take common components for learning as input and output template candidates for job description-related sentences. If the template candidate generation model is a large-scale language model, the generation unit 118 may, as a template candidate generation step, input an instruction to the model to take common components as input and output template candidates.

[0048] The component extraction model and the template candidate generation model may be the same model, or the template candidate generation model may execute the component extraction step and the template candidate generation step. In that case, the template candidate generation model may be a trained model that takes job description texts as input, extracts common components, and outputs template candidates based on the extracted components. In the case of a large-scale language model, the model may be instructed to take job description texts as input, extract common components in the input job description texts, and output template candidates based on the extracted components.

[0049] The question information candidate generation model may be a large-scale language model, or it may be a learning model that takes a text about job content for training as input and outputs question information candidates for generating text about job content. If the question information candidate generation model is a large-scale language model, the generation unit 118 may, as a question information candidate generation step, input an instruction to the model to take a text about job content as input and output question information candidates for generating text about job content. Alternatively, the question information candidate generation model may be a learning model that takes a specific text that satisfies predetermined conditions for training as input and outputs question information candidates for generating text about job content. If the question information candidate generation model is a large-scale language model, the generation unit 118 may, as a question information candidate generation step, input an instruction to the model to take a specific text that satisfies predetermined conditions as input and output question information candidates for generating text about job content. Alternatively, the question information candidate generation model may be a learning model that takes a template candidate for text about job content as input and outputs question information candidates for generating text about job content. If the question information candidate generation model is a large-scale language model, the generation unit 118 may, as a question information candidate generation step, input an instruction to the model to take a template candidate for a sentence about job duties as input and output a question information candidate for generating a sentence about job duties. Alternatively, the question information candidate generation model may be a learning model that has been trained to generate the question information candidates necessary to generate a sentence about job duties that conforms to the template candidate, or, if it is a large-scale language model, it may be instructed to output the question information necessary to generate a sentence about job duties that conforms to the input template candidate.

[0050] The template extraction model may be a learned model that has been trained (machine learning or fine-tuning) to take either the first or second job seeker information as input and output a template of a sentence related to the job description. For example, the template extraction model is a learned model that has been trained using the first or second job seeker information and its corresponding template as training data. The extraction unit 115 inputs the first job seeker information and template candidates into the template extraction model of the artificial intelligence unit 113, causing the template extraction model to output a template from among the template candidates. Alternatively, the template extraction model may be a generative AI that includes a large-scale language model. In this case, the extraction unit 115 inputs an instruction to extract a template based on the first job seeker information and a prompt with the first job seeker information inserted into it into the template extraction model, causing the template extraction model to output a template. Furthermore, the extraction unit 115 may input a prompt to the template extraction model that, in addition to the template extraction instruction and the first job seeker information, includes, for example, one or more samples of the first job seeker information and one or more corresponding template samples.

[0051] The question information extraction model may be a trained model that has been trained (machine learning or fine-tuning) to take either the first or second job seeker information as input and output question information. For example, the question information extraction model is a trained model that has been trained using the first or second job seeker information and the corresponding question information as training data. The extraction unit 115 inputs the first job seeker information to the question information extraction model of the artificial intelligence unit 113 and causes the question information extraction model to output the first question information. Alternatively, the question information extraction model may be a generative AI that includes a large-scale language model. In this case, the extraction unit 115 inputs an instruction to extract the first question information based on the first job seeker information and a prompt with the first job seeker information inserted into it to the question information extraction model and causes the question information extraction model to output the first question information. Furthermore, the extraction unit 115 may input a prompt to the question information extraction model that includes, for example, one or more samples of the first job seeker information and one or more corresponding samples of the first question information, in addition to the instruction to extract the first question information and the first job seeker information. When the extraction unit 115 obtains an answer to the first question, the memory control unit 112 saves the obtained answer as job seeker information in the job seeker database DB2. The extraction unit 115 inputs the second job seeker information to the question information extraction model of the artificial intelligence unit 113, causing the question information extraction model to output the second question information. The question information extraction model may also be a generative AI including a large-scale language model. In this case, the extraction unit 115 inputs an instruction to extract the second question information based on the second job seeker information and a prompt that includes the second job seeker information, causing the question information extraction model to output the second question information. Furthermore, the extraction unit 115 may input to the question information extraction model a prompt that, in addition to the second question information extraction instruction and the second job seeker information, includes, for example, one or more samples of the second job seeker information and one or more corresponding samples of the second question information.

[0052] In this embodiment, the artificial intelligence module may include a learning model. The learning model is a model trained to take a large-scale language model or second job seeker information as input and output text about the job seeker's job description. If the learning model is a large-scale language model, the learning model inputs the second job seeker information and an instruction to output text about the job seeker's job description. The generation of text about job description is performed by referring to a job description generation model that includes the correlation between the first or second job seeker information and the text about job description. As the job description generation model, for example, a job description generation model is used which is a learning model trained to take the first or second job seeker information as input and output text about job description. The generation unit 118 inputs the first or second job seeker information to the job description generation model of the artificial intelligence unit 113 and causes the job description generation model to output text about job description. The job description generation model may also be a generation AI that includes a large-scale language model. In this case, the generation unit 118 inputs an instruction to generate a document describing the job duties based on the first or second job seeker information, and a prompt inserting the first or second job seeker information, into the job duties generation model, causing the job duties document to output. In addition, the generation unit 118 may input a prompt to the job duties generation model that, in addition to the instruction to generate a document describing the job duties and the first or second job seeker information, inserts, for example, one or more samples of the first or second job seeker information and one or more corresponding samples of the job duties document. When inputting into the job duties generation model, the generation unit 118 may also input templates extracted based on the job seeker information into the learning model as reference information.

[0053] This learning model is a pre-trained model that has been machine-trained using job seeker information for training. The job description documents for training are the job description documents used as training data for training the learning model. The job description documents for training are information about job seekers other than job seeker U1, who is a user of information processing system 1, but may also include the registration information and input information of job seeker U1.

[0054] Furthermore, this learning model may also be a model generated by performing so-called supervised learning machine learning, where, for example, the model is trained using training data that associates the job seeker information of job seekers who have previously engaged in job-seeking activities with training data (explanatory variables), and training data (dependent variables) that associates the job description documents, such as resumes submitted by those job seekers to the companies where they actually found employment (i.e., documents describing effective job descriptions), as training data.

[0055] Furthermore, the learning model may be generated by training it on a corpus (a database of structured and large-scale collections of natural language sentences), and job seeker information may be included as one of the large corpora. The artificial intelligence unit 113 uses this learning model to create a sentence about the job description derived from the input information about the job seeker, and outputs the created sentence. The display control unit 114 also functions as an example of a display unit that displays the sentence about the job description output by the artificial intelligence unit 113 based on instructions from the generation unit 118.

[0056] <Display Control Unit 114> The display control unit 114 performs processing to display system screens related to the information processing system 1 on each terminal. For example, the display control unit 114 performs processing such as generating and sending HTML (Hyper Text Markup Language) files to display a web page showing the system screen on the job seeker terminal 20. The display control unit 114 may also perform processing such as generating and sending display data for applications that use the information processing system 1. For example, the display control unit 114 can display the registration screen, input screen, and reception screen of job seeker information related to job seeker U1 on the job seeker terminal.

[0057] <Extraction part 115> The extraction unit 115 is configured to extract various types of information from reference information. For example, as a specific document extraction step, the extraction unit 115 can extract specific documents that meet predetermined conditions from stored documents related to job descriptions. As a component extraction step, the extraction unit 115 can extract common components common to multiple specific documents from multiple specific documents. As a template extraction step, the extraction unit 115 can extract templates corresponding to the first job seeker information from the reference information. As a question information extraction step, the extraction unit 115 extracts question information corresponding to the first job seeker information from the reference information.

[0058] <Presentation part 116> The presentation unit 116 performs processing to display system screens related to the information processing system 1 to each terminal. The presentation unit 116 is configured to be able to display the question information extracted by the extraction unit 115 to the job seeker U1.

[0059] <Acquisition part 117> The acquisition unit 117 is configured to acquire various types of information. The acquisition unit 117 can acquire information entered into the job seeker terminal 20 as an answer.

[0060] <Generation unit 118> The generation unit 118 is configured to generate various types of text based on job seeker information relating to the job-seeking activities of job seeker U1.

[0061] The generation unit 118 may be configured to generate a document relating to the job description of job seeker U1 as part of the job description generation step. For example, as part of the job description generation step, the generation unit 118 may instruct a learning model to generate a document relating to the job description based on the job seeker information, causing the learning model to generate the document relating to the job description. Here, the document relating to the job description may be, for example, a document describing the job description of job seeker U1, or a document summarizing the job description of job seeker U1.

[0062] The generation unit 118 controls the input to the artificial intelligence unit 113. The generation unit 118 is configured to generate prompts that instruct the system to output information related to the job duties of job seeker U1, using as input information information entered by job seeker U1 on the job seeker information input screen and information entered and registered on the registration screen. The generation unit 118 may also be configured to generate prompts that instruct the system to output text related to the job duties, using as input job seeker information of job seeker U1, for example. Alternatively, the generation unit 118 may generate prompts by inserting as input job seeker information of job seeker U1, for example, into a pre-set prompt template.

[0063] Furthermore, the generation unit 118 is configured to instruct the artificial intelligence unit 113 to create a document describing the job duties of a job seeker based on the job seeker information, which includes registered information and input information. The job seeker information referred to here includes the input information that has been entered and the registered information that has been registered in advance.

[0064] <Conversion Unit 119> The conversion unit 119 can convert the generated candidate question information into vector data. The conversion unit 119 can also convert template candidates and job seeker information into vector data. Vector data is a vector value obtained by quantifying text data using known methods such as natural language processing by morphological analysis and encoding of categorical variables. Vector data may be dense vectors or sparse vectors.

[0065] <User display unit 211> The user display unit 211 of the job seeker terminal 20 displays the screen indicated by the screen data transmitted from the server device 10. The user display unit 211 of the job seeker terminal 20 displays the system screen related to the information processing system 1 indicated by the screen data transmitted from the server device 10.

[0066] <Operation reception unit 212> The operation reception unit 212 of the job seeker terminal 20 is configured to accept operations from users (job seekers) using the job seeker terminal 20. When job seeker U1 uses the information processing system 1 by operating the job seeker terminal 20 as a user, they log in by entering their user ID (Identification) and password. As a result, the user ID is associated with the information generated or updated on the job seeker terminal 20, making it possible to identify which user the information belongs to.

[0067] 3. Overview of Information Processing Methods This section describes the information processing method of the server device 10. In this information processing method, each part of the server device 10 is executed by a computer as a step. As shown below, the information processing method comprises each step executed by the information processing system. The program of this embodiment causes the computer to execute each step of the information processing system 1. The order of processing can be changed as appropriate, multiple processes may be executed simultaneously, and some processes may be omitted.

[0068] 3.1 Overview Figure 6 is a flowchart illustrating the overview of the information processing performed by the information processing system 1. In this process, first, the reception unit 111 receives first job seeker information, which includes information about the job seeker's attributes (step S001). Next, the extraction unit 115 extracts question information from the reference information corresponding to the first job seeker information as a question information extraction step (step S002). Here, the question information includes questions for the job seeker and information about example answers corresponding to the questions, and the reference information includes multiple candidate question information, which are pre-generated candidate question information based on stored job description documents. Next, the acquisition unit 117 acquires answers from the job seeker to the question information (step S003). Next, the generation unit 118 generates a document about the job description of the job seeker using the second job seeker information as a job description generation step (step S004). The second job seeker information includes the first job seeker information and the answers acquired by the acquisition unit 117.

[0069] In summary, according to one embodiment, the information processing system comprises at least one processor. The processor includes the following parts by reading a program. The reception unit 111 receives first job seeker information, which includes information about the attributes of the job seeker. The extraction unit 115 extracts question information from reference information corresponding to the first job seeker information as a question information extraction step. The question information includes questions for the job seeker and information about example answers corresponding to the questions. The reference information includes a plurality of candidate question information, which are pre-generated candidate question information based on stored job description documents. The acquisition unit 117 acquires answers from the job seeker to the question information. The generation unit 118 generates a document about the job description of the job seeker using second job seeker information as a job description generation step. The second job seeker information includes the first job seeker information and the answers acquired by the acquisition unit 117. In this manner, questions corresponding to the job seeker information can be presented.

[0070] 4. Details of the information processing method This section will provide a more detailed explanation of the information processing in this embodiment, which involves causing a computer controlling the information processing system 1 to execute a program. The information processing system 1 includes at least one processor. The processor is configured to perform the steps shown in Figures 7, 9, etc., by reading a program.

[0071] Figure 7 is an activity diagram showing the flow of information processing for constructing a vector database executed by the information processing system 1. An example of the flow may be included within the scope defined in the overview described above. Figure 8 is an explanatory diagram illustrating an example of the relationship between a specific sentence and a question information vector. The explanatory diagram includes a specific sentence, constituent elements, a template, question information, and a question information vector. Hereafter, the explanation will follow each activity in this activity diagram using the explanatory diagram. Note that the information processing may include any exception handling not shown. Exception handling includes interrupting the information processing or omitting each process. The selection or input performed in the information processing may be based on operations by the job seeker U1, or it may be performed automatically without operations by the job seeker U1.

[0072] First, the acquisition unit 117 acquires the accumulated job description documents (activity A001). The accumulated job description documents are the job description documents of multiple job seekers. The acquisition unit 117 can acquire the job description documents accumulated in the job seeker database DB2.

[0073] Next, the extraction unit 115 performs a specific document extraction step, which involves extracting specific documents that meet predetermined conditions from the stored documents related to job descriptions (Activity A002). The stored documents related to job descriptions may be those stored in the job seeker database DB2. The predetermined conditions may be conditions that allow for the extraction of high-quality specific documents from the stored documents related to job descriptions.

[0074] Specifically, the extraction unit 115 may, as a specific document extraction step, extract specific documents that meet predetermined conditions from the accumulated documents relating to job content and the results of job-seeking activities. The results of job-seeking activities may be stored in the job seeker database DB2, associated with documents relating to job content. Examples of job-seeking activity results include the number of scout emails received, whether or not scout emails were received, the number of scout emails replied to, whether or not scout emails were replied to, the number of document screenings passed, whether or not document screenings were passed, the number of job offers received, whether or not a job offer was received, and whether or not a decision was made to join the company. The predetermined conditions may include the number of scout emails received being greater than or equal to a predetermined number.

[0075] Furthermore, the specified conditions may also include the inclusion of specified components in the text. These specified components are, for example, the 5W1H. The 5W1H are "Who," "What," "When," "Where," "Why," and "How." A text describing job duties that includes the 5W1H can be considered a high-quality text.

[0076] In Activity A002, the extraction unit 115 may, as a specific document extraction unit, extract specific documents from the stored documents related to job descriptions for each attribute of the job seeker. Specifically, as a specific document extraction step, the extraction unit 115 may extract specific documents from the stored documents related to job descriptions for each attribute of the job seeker, such as by industry, job type, age, place of residence, or place of work. If the extraction unit 115 extracts specific documents for each attribute, the storage control unit 112 may store the extracted specific documents in association with the attributes of the job seeker used for extraction.

[0077] Next, the extraction unit 115, in the component extraction step, uses an artificial intelligence module to extract common components that are common to multiple specific documents (activity A003). In the example in the explanatory diagram, the extraction unit 115, as part of the component extraction step, uses the artificial intelligence module M1 (e.g., a component extraction model) to extract components common to specific documents (common components), such as component A, component B, component C, etc. This allows for the extraction of components that are commonly included in documents related to high-quality job descriptions (e.g., selling points, management experience, specific results, quantitative achievements, etc.). In other words, the common components relate to at least one of the job seeker's achievements, skills, qualifications, and selling points.

[0078] Here, the extraction unit 115 may, as a component extraction step, extract common components common to each attribute of the job seeker. For example, as a component extraction step, the extraction unit 115 may extract common components common to the accumulated job description documents for each attribute of the job seeker, such as by industry, job type, age, place of residence, or place of work. Specifically, as a component extraction step, the extraction unit 115 may extract common components common to specific documents extracted for each attribute. If the extraction unit 115 extracts common components for each attribute, the storage control unit 112 may store the extracted common components in association with the attributes.

[0079] For example, the generation unit 118 generates template candidates that include common components as a template candidate generation step (activity A004). In the example in the explanatory diagram, the extraction unit 115 can generate template candidates for job description documents based on common components using the artificial intelligence module M2 (e.g., template candidate generation model) as a template candidate generation step. A template candidate may also be called a set of components consisting of multiple common components. Multiple template candidates may be generated (e.g., template T1, template T2, etc.). This makes it possible to generate template candidates that include components commonly found in high-quality job description documents.

[0080] Here, the generation unit 118 may generate template candidates for each attribute of the job seeker as a template candidate generation step. For example, the generation unit 118 may generate template candidates for each attribute of the job seeker, such as by industry, job type, age, place of residence, or place of work, based on the accumulated texts related to job content. Specifically, the generation unit 118 may generate template candidates based on specific texts extracted for each attribute as a template candidate generation step. Alternatively, the generation unit 118 may generate template candidates that include common components extracted for each attribute as a template candidate generation step. This makes it possible to generate template candidates that correspond to the attributes of the job seeker. When the generation unit 118 generates template candidates for each attribute, the storage control unit 112 may store the template candidates in association with the attributes.

[0081] Next, the generation unit 118 generates candidate question information using an artificial intelligence module based on the stored job description texts as a question information candidate generation step (activity A005). Specifically, the extraction unit 115 may generate candidate question information using an artificial intelligence module as a question information candidate generation step to generate texts about the job description of job seekers corresponding to template candidates. Here, the generation unit 118 may generate candidate question information for each attribute of the job seeker as a question information candidate generation step. For example, the generation unit 118 may generate candidate question information for each industry, each job type, each age, each place of residence, and each place of work as a question information candidate generation step. Specifically, the generation unit 118 may generate candidate question information corresponding to the template candidates generated for each attribute. If candidate question information is generated for each attribute, the storage control unit 112 may store the candidate question information in association with the attribute. This makes it possible to generate candidate question information according to the attributes of the job seeker. In the example diagram, the extraction unit 115 can generate candidate questions based on a template using the artificial intelligence module M3 (e.g., a candidate question generation model) as a question information candidate generation step. In other words, it generates candidate questions that job seekers will answer in order to obtain the information of job seekers necessary to generate text according to the template. The extraction unit 115 may generate multiple candidate questions for each template as a question information candidate generation step. For example, the extraction unit 115 can generate candidate questions QA1-1 and QA1-2 for template T1, and candidate questions QA2-1 and QA2-2 for template T2. The candidate questions may include information about one question and multiple answer choices corresponding to that question. Here, the answer choices are candidate answers to the corresponding question.For example, candidate question information QA1-1 may include question Q1-1 and corresponding answers A1-1-1 and A1-1-2; candidate question information QA1-2 may include question Q1-2 and corresponding answers A1-2-1 and A1-2-2; candidate question information QA2-1 may include question Q2-1 and corresponding answers A2-1-1 and A2-1-2; and candidate question information QA2-2 may include question Q2-2 and corresponding answers A2-2-1 and A2-2-2.

[0082] Next, the conversion unit 119 converts the generated candidate question information into vectors (activity A006). In the example in the explanatory diagram, the conversion unit 119 converts template T1 into template vector TV1 and template T2 into template vector TV2. The conversion unit 119 also converts candidate question information QA1-1 into question information vector QAV1-1, candidate question information QA1-2 into question information vector QAV1-2, candidate question information QA2-1 into question information vector QAV2-1, and candidate question information QA2-2 into question information vector QAV2-2. Furthermore, if the generation unit 118 generates candidate templates and candidate question information for each attribute of the job seeker, the conversion unit 119 may also convert the associated attributes of the job seeker into vectors along with the candidate templates and candidate question information.

[0083] Next, the memory control unit 112 stores the candidate question information in the vector DB (activity A007). Specifically, the memory control unit 112 can store a template candidate vector, which is a vectorized template candidate, and a question information candidate vector, which is a vector of a question information candidate corresponding to the template candidate, in association with each other. The reference information (e.g., the vector DB) includes the question information candidate and the question information candidate vector, which is a vector of the question information candidate. This constructs a vector DB in which the template candidate and the question information candidate are stored in association with each other. Furthermore, if the conversion unit 119 converts the associated attributes along with the template candidate and the question information candidate into vectors, the memory control unit 112 may store a vector corresponding to the attributes of the job seeker corresponding to the template candidate or the question information candidate, in association with the template candidate vector and the question information candidate vector.

[0084] Next, we will explain the information processing that generates text related to job descriptions using Figures 9 and 10. Figure 9 is an activity diagram showing the flow of information processing for constructing a vector DB executed by the information processing system 1. Figure 10 is a diagram showing input screen G3, which is an example of an information input screen. Input screen G3 includes areas 31 to 40. Areas 31, 33, 35, 37, and 39 are areas where the extraction unit 115 presents questions generated based on the extracted question information as a question information extraction step. Areas 32, 34, 36, 38, and 40 are areas where the job seeker U1's answers to the questions are displayed.

[0085] First, the server device 10 receives first job seeker information via the reception unit 111 (Activity A101). The first job seeker information includes information about the job seeker's attributes. Specifically, the job seeker terminal 20 may receive the first job seeker information from job seeker U1 via the operation reception unit 212, and the reception unit 111 may receive the first job seeker information from the job seeker terminal 20. Alternatively, the reception unit 111 may receive the first job seeker information from the job seeker database DB2 stored in the storage unit 12.

[0086] Next, the server device 10 converts the first job seeker information into a vector (job seeker information vector) using the conversion unit 119 (Activity A102). For example, the conversion unit 119 can convert the first job seeker information into an embedding vector.

[0087] Furthermore, the conversion unit 119 may vectorize the job seeker information using a vectorization API. A vectorization API is an API for converting data into a vector (an array of numbers). The conversion unit 119 may use a vectorization API located outside the information processing system 1 as the API.

[0088] Next, the extraction unit 115 searches the vector database (activity A103) and extracts templates corresponding to the first job seeker information (activity A104). For example, the extraction unit 115 can extract template vectors from the vector database using a search query based on the first job seeker information, and then extract templates corresponding to those template vectors.

[0089] Specifically, the extraction unit 115, as a template extraction step, can extract templates corresponding to the first job seeker information based on the similarity between the first job seeker information vector, which is a vectorized version of the first job seeker information, and the template candidate vector, which is a vectorized version of the template candidates. For example, the extraction unit 115 can extract question information with high similarity by calculating the similarity between the first job seeker information vector and each template vector in the vector DB. Cosine similarity or Euclidean distance can be used as indicators when calculating similarity. This makes it possible to extract templates that match the job seeker information.

[0090] Here, the extraction unit 115 may, as a template extraction step, extract templates corresponding to the attributes of the job seeker included in the first job seeker information. That is, the extraction unit 115 may, as a template extraction step, extract templates corresponding to the industry, occupation, age, place of residence, place of work, etc. included in the first job seeker information. Specifically, the extraction unit 115 may extract templates corresponding to the first job seeker information based on the similarity between the vectors of the first job seeker information vector and a vector obtained by combining the template candidate and the attributes associated with the template candidate. Alternatively, the extraction unit 115 may first determine whether an attribute included in the first job seeker information corresponds to a template candidate with similar attributes, and then extract templates based on the similarity between the template candidate vector of the template candidate that corresponds to the attribute included in the first job seeker information and the first job seeker information vector. This makes it possible to extract templates corresponding to the attributes of the job seeker.

[0091] Next, the extraction unit 115 performs a question information extraction step, which involves extracting question information from the reference information corresponding to the first job seeker information (Activity A105).

[0092] For example, the extraction unit 115 may, as a question information extraction step, extract first question information corresponding to the first job seeker information based on the similarity between a first job seeker information vector, which is a vectorized representation of the first job seeker information, and a question information candidate vector, which is a vectorized representation of the question information candidates. For example, the extraction unit 115 can extract question information with high similarity by calculating the similarity between the first job seeker information vector and each question information candidate vector in the vector DB. Cosine similarity or Euclidean distance can be used as indicators when calculating similarity.

[0093] The extraction unit 115 may extract question information corresponding to a template as a question information extraction step. The extraction unit 115 may extract question information corresponding to the extracted template from reference information. The extraction unit 115 may extract question information stored in association with the extracted template from the word database DB1, and the word database DB1 may be a vector database.

[0094] Furthermore, the extraction unit 115 may, as a question information extraction step, extract question information corresponding to the attributes of the job seeker included in the first job seeker information. If, in activity A104, the extraction unit 115 extracts a template corresponding to the attributes of the job seeker as a template extraction step, then in activity A105, the extraction unit 115 may, as a question information extraction step, extract question information corresponding to the attributes of the job seeker.

[0095] Next, the presentation unit 116 presents the extracted question information to the user display unit 211 of the job seeker terminal 20 (Activity A106). The extraction unit 115 may have the extracted question information presented to the presentation unit 116 as is, or it may edit the extracted question information to suit the job seeker U1 before presenting it to the presentation unit 116. For example, the extraction unit 115 may extract the question information using a learning model, and the learning model may be a large-scale language model. Based on the extracted question information, the extraction unit 115 may use a large-scale language model to edit it to suit the job seeker U1 before having the presentation unit 116 present it.

[0096] The display unit 116 transmits the extracted question information data to the job seeker terminal 20. The job seeker terminal 20 displays an input screen G3 containing the transmitted question information data via the user display unit 211. The input screen G3 includes input fields where the user can enter answers to the question information. The first question information is displayed in area 31 of the input screen G3. Area 31 contains the question: "We will automatically generate job details from 3-10 questions. First, please tell us what kind of work you have done at which companies in the following format. Please summarize each job on one line. Employment period Company name Job title example 2020-2021 Company A Job title B"

[0097] Next, job seeker U1 inputs their answers to the questions presented in area 31 into area 32. The job seeker terminal 20 receives the input from job seeker U1 via the operation reception unit 212. The job seeker terminal 20 transmits the answer data to the server device 10. The acquisition unit 117 acquires the answers from the job seeker terminal 20 (activity A107).

[0098] Next, the control unit 11 adds the response to the first job seeker information to create the second job seeker information (Activity A108). Specifically, the memory control unit 112 stores the response obtained from job seeker U1 in the job seeker database, associating it with the first job seeker information of job seeker U1.

[0099] Next, the control unit 11 determines whether additional questions are necessary based on the second job seeker information for which answers were added in activity A108 (activity A109).

[0100] If it is determined in Activity A109 that additional questions are needed, return to Activity A105.

[0101] The extraction unit 115 extracts second question information corresponding to the second job seeker information. In other words, as a question information extraction step, the extraction unit 115 can extract first question information and second question information. The second question information may be extracted from reference information according to the second job seeker information after obtaining the answer to the first question information. For example, if the reference information includes a vector DB, the extraction unit 115 may, as a question information extraction step, extract second question information corresponding to the second job seeker information based on the similarity between the vectors of the second job seeker information vector, which is obtained by vectorizing the second job seeker information, and the candidate question information vector.

[0102] Next, in Activity A106, the presentation unit 116 presents the second question information to the job seeker terminal 20. The control unit 11 may use the artificial intelligence module to determine whether the answer entered by the job seeker U1 in area 32 satisfies the requirements of the question in area 31. For example, area 32 contains the answer "2015-2023 Company C". The control unit 11 determines that the answer in area 32 does not contain the information of "job type" and causes it to be presented as the next question in area 33. Area 33 contains the question, "Information is missing. We apologize for the inconvenience, but please fill in "Employment Period Company Name Job Type" in the following format. Example: 2020-2021 Company A Job Type B".

[0103] Area 34 includes the answer "2015-2023 Company C Store Manager". Next, the control unit 11 determines whether additional questions are needed based on the second job seeker information to which the answers entered in Area 34 have been added. If the control unit 11 determines that additional questions are needed after the input of the job title, the extraction unit 115 extracts the following question information, and the generation unit 118 generates questions based on the extracted question information. Area 35 includes the question "Question: How many people were on the team? Example answer: 6 people". Area 36 includes the answer "80 people". Area 37 presents three example answers along with the question "Question: What role did you play?". Area 38 includes the answer "I did XX.". Area 39 presents three example answers along with the question "Question: What were the specific results?". Area 40 includes the answer "I increased XX.". Thus, the control unit 11 may repeatedly extract and present question information in order to generate a document about the job description. The question information extracted includes information necessary to generate a document about the job description, such as the job seeker's job type, work area, specific roles, and achievements.

[0104] The presentation unit 116 may present question information to the job seeker U1 via the chat system and receive responses from the job seeker U1. The chat system is a system that responds in real time, with the presentation unit 116 displaying the question information on the user display unit 211, and when it receives a response from the job seeker U1, the extraction unit 115 extracts question information corresponding to the received response. Specifically, when the presentation unit 116 presents question information and receives a response from the job seeker U1, the extraction unit 115 uses natural language processing technology to extract the next question information corresponding to the second job seeker information, and the presentation unit 116 presents it. Here, the presentation unit 116 may display a list of messages sent and received with the job seeker U1.

[0105] If the control unit 11 determines that no additional questions are needed to generate a document about the job description, the presentation unit 116 may present the job seeker U1 with a question requesting instructions to generate a document about the job description. Subsequently, the acquisition unit 117 acquires instructions to generate a document about the job description from the job seeker terminal 20. Subsequently, as a job description generation step, the generation unit 118 can use the artificial intelligence module to generate a document about the job description of the job seeker, taking the second job seeker information as input (activity A110). In this manner, by extracting question information that matches the job seeker information, personalized questions can be presented to each job seeker, and by generating a document based on these questions, a more personalized document can be generated.

[0106] In the above embodiment, in activity A002, the extraction unit 115 extracts specific sentences that meet predetermined conditions from the accumulated sentences related to job content, and the generation unit 118 generates template candidates and question information candidates based on the specific sentences. By using the template candidates and question information candidates generated in this way based on the specific sentences, the generation unit 118 can generate high-quality sentences (sentences that are highly likely to meet predetermined conditions).

[0107] Furthermore, the generation unit 118 may, as a job content generation step, generate a document describing the job content of the job seeker based on a specific document using an artificial intelligence module. The artificial intelligence module may be a learning model, and the learning model may be a model that has been trained to output a document describing the job content of the job seeker, taking a specific document and second job seeker information as input. Alternatively, the artificial intelligence module may be a large-scale language model, and the large-scale language model may be input with a specific document, second job seeker information, and an instruction to output a document describing the job content of the job seeker based on the specific document and the second job seeker information. In this manner, a document describing the job content of job seeker U1 can be generated by referring to a document describing job content that satisfies predetermined conditions.

[0108] Furthermore, the generation unit 118 may, as a job content generation step, generate text about the job seeker's job content using an artificial intelligence module based on the job seeker U1's answers to the questions. The artificial intelligence module may be a learning model, and the learning model may be a model that has been trained to take the job seeker U1's answers as second job seeker information as input and output text about the job seeker's job content. Alternatively, the artificial intelligence module may be a learning model, and the learning model may be a large-scale language model, and the job seeker U1's answers to the questions and an instruction to output text about the job seeker's job content based on the answers may be input to the large-scale language model. In this manner, text about the job content of job seeker U1 can be generated based on the job seeker U1's answers to the questions.

[0109] Furthermore, the generation unit 118 can generate text about the job seeker's job description using an artificial intelligence module as part of the job description generation step, based on the second job seeker information and the template extracted in the template extraction step. The artificial intelligence module may be a learning model, which is a model that has been trained to take the extracted template and the second job seeker information as input and output text about the job seeker's job description. Alternatively, the artificial intelligence module may be a large-scale language model, which may input the template, the second job seeker information, and instructions to output text about the job seeker's job description based on the template and the second job seeker information. In this manner, text about the job description of job seeker U1 can be generated based on a template. In other words, for example, text about the job description of job seeker U1 can be generated using a template generated based on high-quality text about job description, with a configuration that conforms to a template similar to the job seeker information.

[0110] Furthermore, the generation unit 118 may input a template as reference information into the learning model, and as a job content generation step, the generation unit 118 may use an artificial intelligence module to generate text about the job content of the job seeker, using second job seeker information as input. Here, the generation unit 118 may use an artificial intelligence module that includes a learning model as part of the job content generation step. The learning model may be a large-scale language model or a model trained to take second job seeker information as input and output text about the job content of the job seeker. By inputting reference information in this way, the learning model can generate text about the job content of job seeker U1 based on the reference information.

[0111] Furthermore, in Activity A104, if the template extraction step extracts a template corresponding to the job seeker's attributes, the generation unit 118 can, as a job content generation step, generate text about the job seeker's job content using an artificial intelligence module based on the second job seeker information and the template extracted according to the attributes. This makes it possible to generate text about job content that is more closely matched to the job seeker.

[0112] Although embodiments of the present invention have been described above, the present invention is not limited thereto and can be modified as appropriate without departing from the technical spirit of the invention.

[0113] 5. Others In the above embodiment, the server device 10 performed various storage and control functions, but instead of the server device 10, multiple external devices may be used. That is, various information and programs may be stored in a distributed manner across multiple external devices using blockchain technology or the like. In particular, the artificial intelligence unit 113 may be an external component of the server device 10. In that case, the external artificial intelligence unit 113 may be provided by, for example, an artificial intelligence service server, and is configured to receive input from each functional unit of the server device 10, receive requests to execute artificial intelligence services, and return the instructed output as a processing result to the server device 10. The artificial intelligence service server may be a server that provides services using a language model as a learning model, or a server that executes language processing tasks using a language model. The artificial intelligence service server may be constructed using a large-scale language model. The artificial intelligence service server receives prompt input in the form of text, images, audio, etc., and generates and responds with answers to the prompts.

[0114] In the above embodiment, after extracting common components from specific sentences, a template was generated based on these common components. However, the generation unit 118 may, as a template candidate generation step, generate candidate sentence structure templates based on specific sentences using an artificial intelligence module. In this case, the artificial intelligence module may include a learning model. The learning model is a large-scale language model or a model trained to take a specific sentence as input and output a template of sentences related to the job description of a job seeker. If the learning model is a large-scale language model, the learning model inputs the specific sentence and an instruction to output a sentence structure template based on the specific sentence.

[0115] <Variations in configuration> The configuration shown in Figure 1, etc., is just one example, and other configurations are possible as long as they do not cause inconvenience in implementation. For example, one device may be distributed among two or more devices, or it may be replaced by a cloud computing system. Also, the functions of one device may be distributed among two or more devices, or the functions of two or more devices may be concentrated in one device. Furthermore, the operations performed by one function may be distributed among two or more functions, or two or more functions may be integrated into one function. In short, as long as each necessary function is realized in the overall information processing system 1, the devices that realize those functions may be configured in any way. In particular, the artificial intelligence unit 113 may be an external configuration of the server device 10. In that case, the external artificial intelligence unit 113 may be provided by, for example, an artificial intelligence service server, and is configured to receive input from each functional unit of the server device 10, receive requests to execute artificial intelligence services, and return the instructed output as a processing result to the server device 10. The artificial intelligence service server may be a server that provides services using a language model as a learning model, or a server that executes language processing tasks using a language model. The artificial intelligence service server may be constructed using an LLM. The artificial intelligence service server accepts prompt input in the form of text, images, audio, etc., and generates and responds to the prompt.

[0116] <Other variations> The output destination for information or data (hereinafter referred to as "information, etc.") may be other devices, displays, storage units (including built-in and external storage units), etc. Acquisition of information, etc. includes not only acquiring information, etc. transmitted from other devices, but also acquiring information, etc. generated by the device itself. The table associating parameters is not limited to the illustrated table; the number of parameters may be reduced or increased. Furthermore, information, etc. corresponding to parameters may be obtained using mathematical formulas or conditional expressions, etc., without using a table.

[0117] <Note> Furthermore, they may be provided in the following embodiments.

[0118] (1) An information processing system comprising at least one processor, wherein the processor is configured to perform the following steps by reading a program, the system comprising: reception step, reception of first job seeker information including information about the attributes of the job seeker; question information extraction step, extraction of question information corresponding to the first job seeker information from reference information, wherein the question information includes a question to the job seeker and information about an example answer corresponding to the question, wherein the reference information includes a plurality of candidate question information which are pre-generated candidate question information based on accumulated job content; acquisition step, acquisition of the job seeker's answer to the question information; and job content generation step, generation of a document relating to the job content of the job seeker using second job seeker information, wherein the second job seeker information includes the first job seeker information and the answer acquired in the acquisition step.

[0119] In this manner, questions tailored to the job seeker's information can be presented.

[0120] (2) An information processing system as described in (1) above, wherein in the question information candidate generation step, the artificial intelligence module generates the question information candidates based on the stored text relating to the job content, and in the conversion step, the generated question information candidates are converted into vectors, and the reference information includes the question information candidates and the question information candidate vectors which are vectors of the question information candidates.

[0121] In this manner, candidate question information can be accumulated based on the accumulated texts related to job descriptions.

[0122] (3) An information processing system as described in (1) or (2) above, wherein the question information extraction step extracts first question information and second question information, wherein the second question information is extracted from the reference information in accordance with the second job seeker information after obtaining the answer to the first question information.

[0123] In this manner, job seeker information and question information corresponding to their answers can be extracted.

[0124] (4) An information processing system as described in (3) above, wherein in the question information extraction step, the system extracts the first question information corresponding to the first job seeker information based on the similarity between the vectors of a first job seeker information vector obtained by vectorizing the first job seeker information and a question information candidate vector obtained by vectorizing the question information candidate, and extracts the second question information corresponding to the second job seeker information based on the similarity between the vectors of a second job seeker information vector obtained by vectorizing the second job seeker information and the question information candidate vector.

[0125] In this manner, it is possible to extract question information that is similar in vector to job seeker information.

[0126] (5) An information processing system according to any one of (1) to (4) above, wherein in the job content generation step, the artificial intelligence module generates a document relating to the job content of the job seeker, wherein the artificial intelligence module includes a learning model, the learning model is a large-scale language model or a model that has been trained to take the second job seeker information as input and output a document relating to the job content of the job seeker, and if the learning model is the large-scale language model, the second job seeker information and an instruction to output a document relating to the job content of the job seeker are input to the large-scale language model.

[0127] In this manner, a large-scale language model can be used to generate text describing the job duties of job seekers.

[0128] (6) An information processing system according to any one of (1) to (5) above, wherein in the specific text extraction step, specific texts that satisfy predetermined conditions are extracted from the stored texts relating to job descriptions, and in the job description generation step, texts relating to the job descriptions of the job seeker are generated by an artificial intelligence module based on the specific texts.

[0129] In this manner, high-quality documents can be extracted from the accumulated documents related to job descriptions, and these high-quality documents can be used to generate documents related to the job descriptions of job seekers.

[0130] (7) An information processing system as described in (6) above, wherein the predetermined condition includes the number of scouts received being equal to or greater than a predetermined number.

[0131] In this manner, it is possible to extract from the accumulated job description documents documents for which the number of scouts received exceeds a predetermined number, and to use these documents to generate job description documents for job seekers.

[0132] (8) An information processing system as described in (6) or (7) above, wherein the predetermined condition includes that the document contains predetermined components.

[0133] In this manner, it is possible to extract documents containing predetermined components from the accumulated documents related to job descriptions, and to use those documents to generate documents related to the job descriptions of job seekers.

[0134] (9) An information processing system according to any one of (6) to (8) above, wherein in the template candidate generation step, the artificial intelligence module generates a template candidate for a sentence structure based on the specific sentence, and in the question information candidate generation step, the artificial intelligence module generates a question information candidate for generating a sentence relating to the job description of the job seeker according to the template candidate.

[0135] (10) An information processing system as described in (9) above, wherein in the component extraction step, the artificial intelligence module extracts common components common to a plurality of specific documents, and in the template candidate generation step, the template candidate including the common components is generated.

[0136] (11) In the information processing system described in (10) above, the common component is an information processing system relating to at least one of the job seeker's achievements, skills, qualifications, and selling points.

[0137] (12) An information processing system according to any one of (9) to (11) above, wherein in the specific text extraction step, the specific text is extracted for each industry or job type, and in the template candidate generation step, the template candidate is generated for each industry or job type.

[0138] (13) An information processing system according to any one of (9) to (12) above, wherein in the saving step, the system further saves a template candidate vector obtained by vectorizing the template candidate and a question information candidate vector which is a vector of the question information candidate corresponding to the template candidate, in association with each other.

[0139] (14) An information processing system according to any one of (9) to (13) above, wherein in the template extraction step, a template corresponding to the first job seeker information is extracted based on the similarity between a first job seeker information vector obtained by vectorizing the first job seeker information and a template candidate vector obtained by vectorizing the template candidate, and in the question information extraction step, the question information corresponding to the template is extracted.

[0140] (15) An information processing system as described in (14) above, wherein in the job content generation step, the artificial intelligence module generates text relating to the job content of the job seeker based on the second job seeker information and the template extracted in the template extraction step.

[0141] (16) An information processing system as described in (15) above, wherein in the job content generation step, the artificial intelligence module generates a document relating to the job content of the job seeker, wherein the artificial intelligence module includes a learning model, the learning model is a large-scale language model or a model that has been trained to take the second job seeker information as input and output a document relating to the job content of the job seeker, and the template is input to the learning model as reference information.

[0142] (17) An information processing method comprising each step performed by the information processing system described in any one of (1) to (16) above.

[0143] (18) A program that causes a computer to perform each step of the information processing system described in any one of (1) to (16) above. Of course, this is not always the case. Furthermore, the embodiments and modifications described above may be implemented in any combination.

[0144] Finally, various embodiments of the present invention have been described, but these are presented as examples only and are not intended to limit the scope of the invention. Novel embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. Embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of symbols]

[0145] 1: Information Processing System 2: Communication lines 10: Server device 11: Control Unit 111: Reception Department 112: Memory Control Unit 113: Artificial Intelligence Department 114: Display Control Unit 115:Extraction part 116:Presentation part 117: Acquisition Department 118 :Generation part 119: Conversion section 12: Storage section 13: Communications Department 14: Bus 2: Communication lines 20: Job seeker terminal 21: Control Unit 211: User display section 212: Operation Reception Section 22: Storage section 23: Communications Department 24: Input section 25: Output section 26: Bus 31 :Area 32 :Area 33: area 34 :Area 35: area 36: area 37 :Area 38: area 39: area 40: area A1-1-1: Answer A1-1-2: Answer A1-2-1: Answer A1-2-2: Answer A2-1-1: Answer A2-1-2: Answer A2-2-1: Answer A2-2-2: Answer DB: Vector DB1: Vocabulary Database DB2: Job Seeker Database G3: Input screen M1: Artificial Intelligence Module M2: Artificial Intelligence Module M3: Artificial Intelligence Module Q1-1: Question Q1-2: Question Q2-1: Question Q2-2: Question QA1-1: Candidate Question Information QA1-2: Candidate Question Information QA2-1: Candidate Question Information QA2-2: Candidate Question Information QAV1-1: Question Information Vector QAV1-2: Question Information Vector QAV2-1: Question Information Vector QAV2-2: Question Information Vector T1: Template T2: Template TV1: Template Vector TV2: Template Vector U1: Job seeker

Claims

1. An information processing system, Equipped with at least one processor, The aforementioned processor is configured to perform the following steps by reading a program: In the application step, the first job seeker information, including information about the job seeker's attributes, is received. In the question information extraction step, question information corresponding to the first job seeker information is extracted from the reference information. The aforementioned question information includes questions to the job seeker and information regarding example answers corresponding to those questions. The aforementioned reference information includes multiple candidate question information, which are pre-generated candidate question information based on accumulated job content. In the acquisition step, the answers from the job seeker to the aforementioned question information are obtained. In the job description generation step, the information processing system generates text relating to the job description of the job seeker using second job seeker information, wherein the second job seeker information includes the first job seeker information and the response obtained in the acquisition step.

2. In the information processing system described in claim 1, Furthermore, in the question information candidate generation step, the artificial intelligence module generates the question information candidates based on the accumulated texts related to job descriptions. Furthermore, in the conversion step, the generated candidate question information is converted into a vector, The reference information includes the candidate question information and the candidate question information vector, which is a vector of the candidate question information, in an information processing system.

3. In the information processing system described in claim 1, In the aforementioned question information extraction step, the first question information and the second question information are extracted, and here, The information processing system extracts the second question information from the reference information in accordance with the second job seeker information after obtaining the answer to the first question information.

4. In the information processing system described in claim 3, In the aforementioned question information extraction step, Based on the similarity between the first job seeker information vector, which is obtained by vectorizing the first job seeker information, and the question information candidate vector, which is obtained by vectorizing the question information candidate, the first question information corresponding to the first job seeker information is extracted. An information processing system that extracts the second question information corresponding to the second job seeker information based on the similarity between the vectors of the second job seeker information vector, which is obtained by vectorizing the second job seeker information, and the question information candidate vector.

5. In the information processing system described in claim 1, In the job description generation step, the artificial intelligence module uses the second job seeker information as input to generate text about the job description of the job seeker, and here, The artificial intelligence module includes a learning model. The aforementioned learning model is a large-scale language model or a model trained to take the second job seeker information as input and output text related to the job description of the job seeker. If the learning model is the large-scale language model, the second job seeker information and an instruction to output text relating to the job description of the job seeker are input to the large-scale language model. Information processing system.

6. In the information processing system described in claim 1, Furthermore, in the specific text extraction step, specific texts that meet predetermined conditions are extracted from the accumulated texts related to job duties. In the job description generation step, an information processing system generates a document relating to the job description of the job seeker using an artificial intelligence module, based on the specified document.

7. In the information processing system described in claim 6, The aforementioned predetermined conditions include the fact that the number of scouts received is greater than or equal to a predetermined number, in this information processing system.

8. In the information processing system described in claim 6, An information processing system in which the aforementioned predetermined conditions include the fact that a document contains predetermined components.

9. In the information processing system described in claim 6, Furthermore, in the template candidate generation step, the artificial intelligence module generates template candidates for the sentence structure based on the specified sentence. Furthermore, in the question information candidate generation step, the information processing system generates the question information candidates using the artificial intelligence module to generate text about the job description of the job seeker according to the template candidate.

10. In the information processing system described in claim 9, Furthermore, in the component extraction step, the artificial intelligence module extracts common components that are common to multiple specified sentences, The information processing system generates the template candidate that includes the common components in the template candidate generation step.

11. In the information processing system according to claim 10, The aforementioned common component is an information processing system relating to at least one of the job seeker's achievements, skills, qualifications, and selling points.

12. In the information processing system described in claim 9, In the aforementioned specific text extraction step, the specific text is extracted for each industry or job type. An information processing system that generates template candidates for each industry or each job type in the template candidate generation step.

13. In the information processing system described in claim 9, Furthermore, in the saving step, the information processing system saves the template candidate vector, which is a vector of the template candidate obtained by vectorizing the template candidate, and the question information candidate vector, which is a vector of the question information candidate corresponding to the template candidate, in association with each other.

14. In the information processing system described in claim 9, Furthermore, in the template extraction step, a template corresponding to the first job seeker information is extracted based on the similarity between the first job seeker information vector, which is obtained by vectorizing the first job seeker information, and the template candidate vector, which is obtained by vectorizing the template candidates. An information processing system that extracts the question information corresponding to the template in the question information extraction step.

15. In the information processing system described in claim 14, In the job description generation step, the information processing system generates text relating to the job description of the job seeker using the artificial intelligence module, based on the second job seeker information and the template extracted in the template extraction step.

16. In the information processing system described in claim 15, In the job description generation step, the artificial intelligence module uses the second job seeker information as input to generate text relating to the job seeker's job description, and here, The artificial intelligence module includes a learning model. The aforementioned learning model is a large-scale language model or a model trained to take the second job seeker information as input and output text related to the job description of the job seeker. An information processing system in which the template is input as reference information to the learning model.

17. Information processing method, An information processing method comprising each step performed by the information processing system according to any one of claims 1 to 16.

18. It is a program, A program that causes a computer to perform each step of the information processing system described in any one of claims 1 to 16.

Images