Job vacancy information provision device, job vacancy information provision method, and job vacancy information provision program

Abstract

To reduce the burden on users and encourage users to apply for job offers.SOLUTION: A job vacancy information provision device 100 disclosed herein comprises an acquisition unit 122 for acquiring information about a user from text data of a conversation between a chatbot and the user, an identification unit 123 configured to identify job offers on the basis of the information about the user acquired by the acquisition unit 122, and an output unit 126 for outputting information about the job offers identified by the identification unit 123.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to a job information providing device, a job information providing method, and a job information providing program. [Background technology]

[0002] Conventionally, there are technologies that search for job information for part-time jobs, etc. For example, there are technologies that search for and introduce job opportunities based on rules based on user attributes, keywords, conditions, etc. entered by users. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2002-073868 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the conventional technology cannot reduce the burden on users and may not be able to properly encourage users to apply for job listings. For example, when searching for job listings, users must clearly define and input their desired conditions. [Means for solving the problem]

[0005] In order to solve the above-mentioned problems and achieve the objective, the job information providing device is characterized by having an acquisition unit that acquires information about the user from text data of a conversation between a chatbot and a user, an identification unit that identifies a job opportunity based on the information about the user acquired by the acquisition unit, and an output unit that outputs information about the job opportunity identified by the identification unit. Effect of the Invention

[0006] According to the present invention, it is possible to reduce the burden on users and encourage them to apply for job openings. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a diagram for explaining an overview of the process performed by a job information providing device according to an embodiment. [Diagram 2] FIG. 2 is a diagram illustrating an overview of the process performed by the job information providing device according to the embodiment. [Diagram 3] FIG. 3 is a diagram illustrating an example of the configuration of a job information providing system according to the embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of the configuration of a job information providing device according to the embodiment. [Diagram 5] FIG. 5 is a diagram showing an example of the acquisition and determination process performed by the job information providing device according to the embodiment. [Figure 6] FIG. 6 is a diagram showing an example of the acquisition and determination process performed by the job information providing device according to the embodiment. [Figure 7] FIG. 7 is a diagram showing an example of the acquisition and determination process performed by the job information providing device according to the embodiment. [Figure 8] FIG. 8 is a diagram illustrating an example of the acquisition and determination process performed by the job information providing device according to the embodiment. [Figure 9] FIG. 9 is a diagram illustrating an example of the acquisition and determination process performed by the job information providing device according to the embodiment. [Figure 10] FIG. 10 is a diagram illustrating an example of the specification process performed by the job information providing device according to the embodiment. [Figure 11] FIG. 11 is a diagram illustrating an example of a summarization process performed by the job information providing device according to the embodiment. [Figure 12] FIG. 12 is a diagram illustrating an example of an explanation process performed by the job information providing device according to the embodiment. [Figure 13] FIG. 13 is a diagram illustrating an example of an explanation process performed by the job information providing device according to the embodiment. [Figure 14] FIG. 14 is a diagram illustrating an example of an explanation process performed by the job information providing device according to the embodiment. [Figure 15]FIG. 15 is a diagram illustrating an example of an explanation process performed by the job information providing device according to the embodiment. [Figure 16] FIG. 16 is a flowchart illustrating an example of processing performed by the job information providing device according to the embodiment. [Figure 17] FIG. 17 is a diagram illustrating an example of a hardware configuration. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, with reference to the drawings, an embodiment of a job information providing device, a job information providing method, and a job information providing program according to the present application will be described in detail. Note that the present invention is not limited to the embodiment. In addition, in the description of the drawings, the same parts are denoted by the same reference numerals, and duplicated explanations will be omitted.

[0009] 〔overview〕 First, an overview of the processing performed by the job information providing device 100 will be described with reference to Figures 1 and 2. Figures 1 and 2 are diagrams for explaining an overview of the processing performed by the job information providing device 100. The job information providing device 100 according to this embodiment performs the following processing. First, when the "Answer questions to find a job" button in the lower right of Figure 1(1) is pressed from a web browser or application screen for searching for job information as shown in Figure 1(1), the job information providing device 100 transitions to a web browser or application screen for chatting with a chatbot as shown in Figure 1(2).

[0010] Next, the job information providing device 100 displays on the web browser or application screen to the user who has transitioned to the page a sentence encouraging the user to enter information, such as, "Please freely enter your preferences. AI will select the conditions and introduce you to jobs that are perfect for you. For example, among the jobs you are currently viewing, the following jobs are popular. [Job Card] Chat Examples: Are there any restaurants that offer flexible shifts and are open to people with no experience? Are there any lively jobs where the fashion is free and people in their 20s can work late after school, with a decent salary, two days a week?"

[0011] Next, as shown in Fig. 2(1), the job information providing device 100 acquires information about the user from text data of a conversation such as "I'd like to work part-time at a stylish and nice restaurant" that the user inputs in response to the information in Fig. 1(2). In response, the job information providing device 100 displays a sentence such as "I will pick out three jobs that meet your criteria. I will introduce them one by one, so please wait a moment" on the screen of the web browser or application.

[0012] Next, the job information providing device 100 identifies a job posting based on information about the user. For example, the job information providing device 100 identifies "XX Cafe Minatomirai branch", "XX Ramen Minatomirai branch", and "XX Monja Minatomirai branch" from information about the user such as "fashionable", "restaurant", and "Minato Mirai".

[0013] Next, the job information providing device 100 uses the first learning model with information about the identified job listing as input to generate sentences summarizing the job listing, such as "XX Cafe Minatomirai branch is a 5-minute walk from the station," "XX Ramen Minatomirai branch is a 7-minute walk from the station," and "XX Monjaya Minatomirai branch is a 3-minute walk from the station."

[0014] Next, the job information providing device 100 uses, for example, the acquired information about the user and the information about the specified job opportunity as input, and generates sentences explaining the reason for presenting the job opportunity to the user using a second learning model, such as "Recommended points are a solid education system and cute uniforms," ​​"Recommended points are daily pay and free meals," and "Recommended points are high hourly wage and welcoming to inexperienced people." Then, the job information providing device 100 outputs, for example, as shown in Figs. 2(1) and (2), information about the job opportunity, a sentence summarizing the job opportunity, and a sentence explaining the reason for presenting (introducing) the job opportunity.

[0015] As a result, the job information providing device 100 can reduce the burden on the user when searching for and applying for a job opportunity, and can encourage the user to apply for the job opportunity. That is, when the user searches for a job opportunity, the job information providing device 100 collects user attributes, keywords, conditions, and the like through interactions with a chatbot using a learning model, and identifies the job opportunity using the collected information, thereby reducing the burden on the user when searching for a desired job opportunity and encouraging the user to apply for the job opportunity.

[0016] Furthermore, when a user applies for a job, the job information providing device 100 can reduce the user's burden from identifying a job to applying by summarizing the job manuscript and explaining the features of the job as the reason for introducing the job to the user. In this way, the job information providing device 100 can reduce the user's burden in various steps involved in searching for and applying for a job, and can improve the user's application rate for job listings by preventing users from dropping out before applying.

[0017] [Configuration of the job information system] Next, the configuration of the job information providing system 1 will be described with reference to Fig. 3. Fig. 3 is a diagram showing an example of the configuration of the job information providing system 1 according to an embodiment. As shown in Fig. 3, the job information providing system 1 has a configuration in which a job information providing device 100 and a terminal device 200 are connected via a predetermined network NW.

[0018] The job information providing device 100 is a server that provides information on job opportunities to introduce job opportunities to users. The job information providing device 100 is a device that identifies a job opportunity from text data of a conversation between a user and a chatbot, generates a sentence summarizing the identified job opportunity and a sentence explaining the reason for presenting the job opportunity to the user, and outputs information on the job opportunity and the generated sentence.

[0019] The terminal device 200 is a terminal such as a personal computer, a smartphone, or a tablet, and displays the output result by the job information providing device 100. For example, the terminal device 200 is a smartphone that displays the answer output by the job information providing device 100 on a web browser or application screen. Also, for example, the terminal device 200 is a smartphone that displays the job offer information output by the job information providing device 100 and the generated text on a web browser or application screen.

[0020] [Configuration of the job information providing device] Next, the configuration of the job information providing device 100 will be described with reference to Fig. 4. Fig. 4 is a diagram showing an example of the configuration of the job information providing device 100 according to the embodiment. As shown in Fig. 4, the job information providing device 100 has a communication unit 110, a control unit 120, and a storage unit 130. Note that each of these units may be held in a distributed manner in a plurality of devices. The processing of each of these units will be described below.

[0021] The communication unit 110 is realized by a NIC (Network Interface Card) or the like, and enables communication between an external device and the control unit 120 via an electric communication line such as a LAN (Local Area Network) or the Internet. For example, the communication unit 110 enables communication between the external device and the control unit 120.

[0022] The storage unit 130 is realized by a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 130 stores, for example, information about the user, information about the job offer, default search conditions, recommendation reasons, various machine learning algorithms, learning data for machine learning, learned models, prompts, generated information (including text, images, voice, etc.), information that associates information about the user with search conditions, application rate, hiring rate, retention rate (turnover rate), evaluation (work performance, achievements), happiness (satisfaction), and other information required for identifying the job offer, generating a text summarizing the job offer, generating a text explaining the reason for presenting the job offer to the user, and the like. Here, the information about the user includes information such as a user ID, a user name, user attributes, user input information (including information analyzed by voice recognition), search history (including search conditions, keywords, etc.), features of the job offer that the user considers important, browsing history, and application and interview history.

[0023] The information on the job listing includes information such as the job listing ID, the job listing name, the job listing manuscript, and the job listing characteristics. The default search conditions are search conditions that include keywords used to determine whether or not a job listing is related to the job listing, as described below, and to identify the job listing. New search conditions can be added to the default search conditions as appropriate. The recommendation reason is a feature of the job listing that is used to explain the reason for presenting the job listing to the user.

[0024] The information may be information that has been analyzed by natural language processing. For example, information expressed by characters may be subjected to morphological analysis, and vectorized by a method using frequency such as Term Frequency Inverse Document Frequency (TF-IDF) or a method using distributed representation such as Word2Vec (Word to Vector). The vectorization method is not limited to the above, and any known method according to the purpose may be used. The information stored in the storage unit 130 is not limited to the above.

[0025] The control unit 120 is realized using a CPU (Central Processing Unit), an NP (Network Processor), an FPGA (Field Programmable Gate Array), etc., and executes a processing program stored in a memory. As shown in Fig. 4, the control unit 120 has a determination unit 121, an acquisition unit 122, an identification unit 123, a summarization unit 124, an explanation unit 125, and an output unit 126. Each unit of the control unit 120 will be described below.

[0026] The determination unit 121 determines whether or not the text data of the conversation input by the user is related to identifying a job posting based on the relationship between the vectorized text data of the conversation input by the user and the default search conditions. For example, the determination unit 121 vectorizes the text data of the conversation input by the user and the default search conditions using an existing method such as TF-IDF or Word2Vec, and compares the cosine similarity to determine whether or not the text data of the conversation input by the user is related to identifying a job posting.

[0027] More specifically, when the cosine similarity between the text data of the conversation input by the user and the default search condition is equal to or greater than a predetermined threshold, the determination unit 121 determines that the text data of the conversation input by the user is related to identifying a job listing. As another example, the determination unit 121 sorts the text data of the conversation input by the user in descending order of cosine similarity with the default search condition, and determines that the text data of the conversation input by a top predetermined number of users is related to identifying a job listing.

[0028] The acquisition unit 122 acquires information about the user. For example, the acquisition unit 122 acquires information about the user from text data of a conversation between a chatbot and a user. For example, the acquisition unit 122 acquires information about the user from text data of a conversation between a chatbot and a user using an existing learning model, an example of which is a large-scale language model. Note that the text data of a conversation between a chatbot and a user includes not only a text conversation, but also speech or the like converted into text using a speech recognition technology.

[0029] Furthermore, when the determination unit 121 determines that the text data of the conversation input by the user is related to the identification of a job application, the acquisition unit 122 acquires the information as information about the user. More specifically, when the determination unit 121 determines that the sentence "It would be nice to have a part-time job at a stylish and nice restaurant" input by the user is related to the identification of a job application, the acquisition unit 122 acquires the information "It would be nice to have a part-time job at a stylish and nice restaurant".

[0030] At this time, when the determination unit 121 determines that the text data of the conversation input by the user is related to identifying a job application, the acquisition unit 122 may acquire only words in the sentence input by the user whose cosine similarity with the default search condition satisfies a predetermined condition as information about the user. More specifically, the acquisition unit 122 acquires, as information about the user, the words "stylish" and "restaurant" whose cosine similarity exceeds a predetermined threshold among the words included in the sentence input by the user, "I'd like to work part-time at a stylish and nice restaurant."

[0031] Furthermore, when the determination unit 121 determines that the input information is not related to identifying a job opportunity, the acquisition unit 122 outputs information that prompts the user to input information together with a response to the text data of the conversation input by the user. For example, when the user inputs a keyword that is not related to identifying a job opportunity, the acquisition unit 122 outputs a sentence that prompts the user to input information that has not been acquired by the job information providing device 100 together with the response to the user's input.

[0032] Here, the text output by the acquisition unit 122 for urging the user to input information that the job information providing device 100 has not yet acquired may be, for example, a question, an answer, a request, or the like by a chatbot that prompts the user to input information about users that the job information providing device 100 has not acquired. In addition, for example, the text for urging the user to input information that the job information providing device 100 has not acquired may be a fixed phrase that prompts the user to input information about users that the job information providing device 100 has not acquired.

[0033] More specifically, when a user inputs a sentence such as "I can't get a girlfriend," the acquisition unit 122 displays a sentence prompting the user to input age information, such as "If you tell us your age, we can introduce you to jobs where there are people of a similar age," along with the chatbot's response such as "Take care of yourself and value your own charms and wonderful qualities to build a comfortable relationship."

[0034] The acquisition unit 122 may change the sentences depending on the acquired information about the user when outputting the sentences. For example, the acquisition unit 122 outputs sentences changed to a dialect of the user's hometown depending on the information about the user's hometown. Also, for example, the acquisition unit 122 outputs sentences changed to a positive expression or an expression that makes the user feel a sense of crisis based on the information about the user's interview situation.

[0035] Furthermore, when the determination unit 121 determines that the text data of the conversation input by the user is not related to identifying a job application, the acquisition unit 122 inputs a prompt that gives the chatbot a predetermined role according to the keyword into the learning model. More specifically, when the keywords "job hunting" and "part-time job" are input by the user, the acquisition unit 122 inputs a prompt that gives the chatbot a role as a "career advisor" into the learning model. That is, the acquisition unit 122 gives the chatbot a role as an expert in a specific field, and inputs a prompt to the chatbot to respond to the input from the user as an expert in the specific field.

[0036] The identifying unit 123 identifies a job posting based on information about a user acquired by the acquiring unit 122. For example, the identifying unit 123 identifies a job posting from a search result based on a search condition identified from the information about a user acquired by the acquiring unit 122. More specifically, when the acquiring unit 122 acquires information about a user, "occupation: student", the identifying unit 123 identifies a search condition, "students welcome", that corresponds to "occupation: student", and identifies a job posting from a search result based on the identified search condition, "students welcome". That is, for example, when the search condition is "students welcome", the identifying unit 123 identifies a job posting whose information (manuscript, characteristics, etc.) about the job posting includes "students welcome".

[0037] At this time, for example, the identification unit 123 may identify the information about the user itself as the search condition, or may identify the search condition from the information about the user by referring to a database that defines the correspondence between the information about the user and the search condition.

[0038] Furthermore, for example, the identifying unit 123 identifies a job offer according to the application rate of users who have the user-related information acquired by the acquiring unit 122. For example, the identifying unit 123 identifies a job offer with a high application rate from users who have the user-related information acquired by the acquiring unit 122, using a model that has learned the relationship between user attributes and the application rate for each company, industry, and job type.

[0039] That is, first, the identification unit 123 learns the relationship between information about a user and the user's application rate for each company, industry, and job type, and obtains an output of the application rate for each company, industry, and job type by a user having information about a specified user by inputting information about the user into a learning model that outputs the user's application rate for each company, industry, and job type by inputting information about the user.

[0040] Then, the identifying unit 123 identifies a job posting according to the level of the user's application rate. For example, the identifying unit 123 may identify a job posting of a company, industry, or occupation for which the user's application rate is equal to or higher than a predetermined threshold, or may identify a job posting of a company, industry, or occupation for which the user's application rate is in the top predetermined number.

[0041] More specifically, when the acquisition unit 122 acquires information about a user, "occupation: student," the identification unit 123 identifies a job posting with a high application rate from users with the user attribute "occupation: student." When identifying a job posting according to a high user application rate, the information on the company, industry, and occupation may be used independently or in appropriate combination. For example, the identification unit 123 may identify a job posting from a company with a high user application rate, or may identify a job posting from a company and an industry with a high user application rate.

[0042] Furthermore, for example, the identifying unit 123 identifies a job offer according to the hiring rate of a user who has information about the user acquired by the acquiring unit 122. For example, the identifying unit 123 identifies a job offer with a high hiring rate of a user who has information about the user acquired by the acquiring unit 122, using a model that has learned the relationship between user attributes and the hiring rate of the user in each company, each industry, and each job type.

[0043] That is, first, the identification unit 123 learns the relationship between information about a user and the hiring rate of the user in each company, industry, and job type, and obtains an output of the hiring rate of a user in each company, industry, and job type for a user having information about a specified user by inputting information about the user into a learning model that outputs the hiring rate of the user in each company, industry, and job type for a user having information about the specified user.

[0044] The identifying unit 123 then identifies job listings according to the hiring rate of the user in each company, industry, and job type. For example, the identifying unit 123 may identify job listings of companies, industries, and job types in which the hiring rate of the user is equal to or higher than a predetermined threshold, or may identify job listings of companies, industries, and job types in which the hiring rate of the user is in the top predetermined number.

[0045] More specifically, when the acquisition unit 122 acquires information about a user such as "Working conditions: available for night shifts", the identification unit 123 identifies job openings with a high hiring rate for users with the user attribute "Working conditions: available for night shifts". When identifying job openings according to the high hiring rate of a user, the information on the company, industry, and occupation may be used independently or in appropriate combination. For example, the identification unit 123 may identify job openings from companies with a high hiring rate for users, or may identify job openings from companies with a high hiring rate for users and in industries with a high hiring rate for users.

[0046] Also, for example, the identifying unit 123 identifies a job offer according to the retention rate of a user who has information about the user acquired by the acquiring unit 122. For example, the identifying unit 123 identifies a job offer with a high retention rate of a user who has information about the user acquired by the acquiring unit 122, using a model that has learned the relationship between user attributes and the retention rate of the user in each company, each industry, and each job type.

[0047] That is, first, the identification unit 123 learns the relationship between information about a user and the user retention rate in each company, industry, and job type, and obtains an output of the retention rate in each company, industry, and job type of a user having information about a specified user by inputting information about the user into a learning model that outputs the user retention rate in each company, industry, and job type by inputting information about the user.

[0048] Then, the identifying unit 123 identifies job listings according to the level of user retention rate in each company, industry, and occupation. For example, the identifying unit 123 may identify job listings of companies, industries, and occupations in which the user retention rate is equal to or higher than a predetermined threshold, or may identify job listings of companies, industries, and occupations in which the user retention rate is in the top predetermined number.

[0049] More specifically, when the acquisition unit 122 acquires information about a user, such as "career: sales experience", the identification unit 123 identifies a job opportunity with a high retention rate for users with the user attribute "career: sales experience". When identifying a job opportunity according to the high retention rate of a user, the information on the company, industry, and occupation may be used independently or in appropriate combination. For example, the identification unit 123 may identify a job opportunity from a company with a high user retention rate, or may identify a job opportunity from a company with a high user retention rate and in an industry with a high user retention rate.

[0050] Also, for example, the identifying unit 123 identifies a job opening according to the evaluation of a user who has information about the user acquired by the acquiring unit 122. For example, the identifying unit 123 identifies a job opening that is highly rated by a user who has information about the user acquired by the acquiring unit 122, using a model that has learned the relationship between user attributes and user evaluations in each company, industry, and job type.

[0051] That is, first, the identification unit 123 learns the relationship between information about a user and the user's evaluation in each company, industry, and job type, and obtains an output of the evaluation in each company, industry, and job type of a user having information about a specified user by inputting information about the user into a learning model that outputs the user's evaluation in each company, industry, and job type.

[0052] Then, the identification unit 123 identifies job listings according to the level of the user's evaluation in each company, industry, and job type. For example, the identification unit 123 may identify job listings in companies, industries, and job types for which the user's evaluation is equal to or higher than a predetermined threshold, or may identify job listings in companies, industries, and job types for which the user's evaluation is in the top predetermined number of positions.

[0053] More specifically, when the acquisition unit 122 acquires information about a user such as "Qualification: Eiken Grade Pre-2", the identification unit 123 identifies job listings that are highly rated by users with the user attribute "Qualification: Eiken Grade Pre-2". When identifying job listings according to the level of a user's rating, the information on the company, industry, and job type may be used independently or in appropriate combination. For example, the identification unit 123 may identify job listings of companies highly rated by the user, or may identify job listings of companies highly rated by the user and in industries highly rated by the user.

[0054] Also, for example, the identifying unit 123 identifies a job opportunity according to the happiness level of a user who has information about the user acquired by the acquiring unit 122. For example, the identifying unit 123 identifies a job opportunity in which the happiness level of a user who has information about the user acquired by the acquiring unit 122 is high, using a model that has learned the relationship between user attributes and the happiness level of the user in each company, each industry, and each occupation.

[0055] That is, first, the identification unit 123 learns the relationship between information about a user and the user's happiness level in each company, industry, and job type, and obtains an output of the happiness level in each company, industry, and job type of a user who has information about a specified user by inputting information about the user into a learning model that outputs the user's happiness level in each company, industry, and job type.

[0056] The identifying unit 123 then identifies job openings according to the level of the user's happiness in each company, industry, and occupation. For example, the identifying unit 123 may identify job openings in companies, industries, and occupations where the user's happiness is equal to or higher than a predetermined threshold, or may identify job openings in companies, industries, and occupations where the user's happiness is in the top predetermined number. Note that the happiness (satisfaction) level can be determined using a well-being index calculated from responses to a questionnaire including questions about work.

[0057] More specifically, when the acquisition unit 122 acquires information about the user such as "personality: logical" or "personality: passionate," the identification unit 123 identifies job listings in which the user with the user attributes "personality: logical" or "personality: passionate" has a high level of happiness. When identifying job listings in accordance with the level of the user's happiness, the information on the company, industry, and occupation may be used independently or in appropriate combination. For example, the identification unit 123 may identify job listings in companies in which the user has a high level of happiness, or may identify job listings in companies in which the user has a high level of happiness and in an industry in which the user has a high level of happiness.

[0058] As another example, the identification unit 123 may identify search conditions from information about a user by using a learning model. For example, the identification unit 123 may identify search conditions by inputting information about a user to a learning model that learns the relationship between information about a user and search conditions and outputs search conditions by inputting information about a user.

[0059] As another example, the identification unit 123 may use a learning model to identify a job request from information about a user. For example, the identification unit 123 uses a model that has learned the relationship between information about a user and a job request to identify a job request from information about a user acquired by the acquisition unit 122. That is, the identification unit 123 identifies a job request by inputting information about a user into a learning model that learns the relationship between information about a user and a job request and outputs a job request by inputting information about a user.

[0060] The summarization unit 124 receives information about the job request identified by the identification unit 123 as input, and generates a sentence summarizing the job request using the first learning model. For example, the summarization unit 124 receives a manuscript of the job request identified by the identification unit 123 as input, and generates a sentence summarizing the job request using the first learning model.

[0061] The summarizing unit 124 also inputs information about the user, and generates a sentence summarizing the job posting using a first learning model to which a prompt for summarizing the job posting manuscript has been input in accordance with the information about the user. More specifically, the summarizing unit 124 inputs a prompt "Please summarize the job posting manuscript in XX characters or less, including the location conditions" to the learning model to summarize the job posting manuscript so as to include the "location conditions" characteristic of the job posting that the user considers important, which is included in the information about the user, and generates a sentence summarizing the job posting. The first learning model used by the summarizing unit 124 is an existing learning model, an example of which is a large-scale language model, and is not particularly limited. The first learning model can also be processed using a second learning model.

[0062] The explanation unit 125 uses the information about the user acquired by the acquisition unit 122 and the information about the job request identified by the identification unit 123 as input, and generates a sentence explaining the reason for presenting the job request to the user using a second learning model. For example, the explanation unit 125 generates a sentence explaining the reason for presenting the job request to the user using a second learning model in which a predetermined role is assigned according to the information about the user and a prompt for explaining the features of the job request is input.

[0063] More specifically, the explanation unit 125 assigns the chatbot the role of "advertising copy recommendation expert" according to the information about the user "college student" and inputs a prompt to the second learning model to have the chatbot explain the feature of the job posting "students welcome". That is, the explanation unit 125 assigns the chatbot the role of an expert in a specific field and inputs a prompt to the chatbot to respond to input from the user as an expert in the specific field according to user attributes, etc. Note that the second learning model used by the explanation unit 125 is an existing learning model, an example of which is a large-scale language model, and is not particularly limited. Also, the second learning model can perform processing using the first learning model.

[0064] The output unit 126 outputs information related to the job request identified by the identification unit 123, the summarized text generated by the summarization unit 124, and the explanatory text generated by the explanation unit 125. For example, the output unit 126 outputs on a timeline an image of the job request identified by the identification unit 123, the headline of the job request, the URL of the job request, the text summarizing the draft of the job request generated by the summarization unit 124, and the text generated by the explanation unit 125 explaining the reason for presenting the job request to the user.

[0065] [Acquisition and determination process] Next, an example of the acquisition and determination process performed by the job information providing device 100 according to the embodiment will be described with reference to Fig. 5 to Fig. 9. Fig. 5 to Fig. 9 are diagrams for explaining an example of the acquisition and determination process performed by the job information providing device 100 according to the embodiment. First, a case where a search history does not exist and a case where a search history exists will be described.

[0066] Here, search history refers to, for example, information on job search results conducted by the user in the past (search keywords, conditions, job feature, etc.), search keywords used when transitioning from a search result by a search engine to a service for searching job results, etc. The above is an example of a search history, and the search history also includes conditions necessary to display to the user job results with a high number of applications under certain conditions.

[0067] The job information providing device 100 determines that a search history exists when, for example, information about a user who has transitioned to a web browser or application screen for chatting with a chatbot as shown in Figure 1 (2) exists, such as user attributes, user input information, search history, features of job listings that the user considers important, browsing history, and application / interview history.

[0068] If there is no search history, the acquisition unit 122 outputs information prompting the user to input area information, as shown in Fig. 5. For example, the acquisition unit 122 outputs a sentence such as "We will help you find a job. First, please tell us the location of the job you are looking for below" and buttons such as "Search by area" and "Search by station" as information prompting the user to input area information.

[0069] Here, when the user presses the [Search by area] button, for example, the acquisition unit 122 outputs a sentence such as "Please select one prefecture where you would like to work" as shown in Fig. 5. At this time, for example, the acquisition unit 122 allows the user to select a prefecture by a drum roll. Similarly, the acquisition unit 122 may allow the user to select a city, ward, town, village, etc. by a drum roll.

[0070] On the other hand, when the user presses the [Search by station] button, for example, the acquisition unit 122 outputs a sentence such as, "Please tell us the name of the station where you would like to work. For example, I would like to work around Shinjuku Station! Maybe Shimbashi on the Ginza Line! I would like to search around Ueno," as shown in FIG. 6 (1).

[0071] In response to this, as shown in Fig. 6(2), if the user inputs a sentence such as "Ueno, please," the acquisition unit 122 outputs, for example, a sentence such as "Ueno, right! Please choose one desired station from the candidates!", radio buttons indicating the candidate stations [Ueno], [Ueno Okachimachi], and [Keisei Ueno], and an [OK] button. That is, when acquiring information on the desired workplace from the station name entered by the user, if there are multiple stations including the station name entered by the user (for example, "Ueno") in their names, the acquisition unit 122 acquires information on the desired workplace by presenting station candidates including the station name entered by the user in their names and accepting a selection from the user.

[0072] As another example of a case where the user presses the [Search by station] button, when the user inputs a sentence such as "Mitaka, please," as shown in Fig. 7(1), the acquisition unit 122 outputs a sentence such as "Mitaka station, right! Can I also display jobs at nearby stations that are often applied for at the same time?", check buttons for [Kichijoji], [Inokashira Park station], and [Musashisakai] as stations around Mitaka station, and an [OK] button. That is, the acquisition unit 122 acquires information on the desired work location by accepting a selection to include stations located around the station (e.g., "Mitaka") input by the user in the conditions.

[0073] The acquisition unit 122 may acquire the desired work location by accepting a selection by line from the user, as shown in FIG. 7(2). For example, when the user inputs a sentence such as "Hibiya Line, please," the acquisition unit 122 outputs, for example, a sentence such as "Tokyo Metro Hibiya Line, right! Please select a station," to prompt the user to select a station on the selected line. Here, any input format according to the purpose, such as free input or checkboxes, can be used to select a station on the selected line. In this way, the acquisition unit 122 can reduce the burden of information input by the user. After that, the acquisition unit 122 performs the same process as when a search history exists.

[0074] On the other hand, if the search history exists, the acquisition unit 122 displays job offers with many applications under the specified conditions from the information of the search history. For example, as shown in FIG. 8, the acquisition unit 122 displays information on the screen of a web browser or application that prompts the user to input information, such as "Please freely enter your request. AI will select the conditions and introduce the job that is perfect for you. For example, among the job offers you are currently viewing, the following jobs are popular. [Job offer information] Chat example: Are there any restaurants with flexible shifts that I can work at even if I have no experience? Are there any lively jobs where you can dress casually and are mainly for people in their 20s? Are there any jobs where you can work late at night after school, with a fairly high salary, twice a week?"

[0075] In response to this, the user inputs a sentence such as "I'd like to work part-time at a stylish and nice restaurant" as information about the part-time job he / she desires. Then, the acquiring unit 122 acquires the text data of the conversation input by the user as information about the user.

[0076] Next, the process performed by the determination unit 121 will be described. When the user inputs information as described above, the determination unit 121 determines whether or not the text data of the conversation input by the user is related to identifying a job posting. For example, the determination unit 121 vectorizes the text input by the user and the default search criteria by using an existing method such as TF-IDF or Word2Vec.

[0077] Then, when the cosine similarity between the vectorized text input by the user and the default search conditions is equal to or greater than a predetermined threshold, the determination unit 121 determines that the text input by the user is related to identifying a job request. At this time, the determination unit 121 may rearrange the text data of the conversation input by the user in descending order of cosine similarity with the default search conditions, and determine that a top predetermined number of text data are related to identifying a job request.

[0078] The above example in the [Acquisition / Determination Process] is an example in which the determination unit 121 determines that the text data of the conversation input by the user is related to identifying a job listing. Below, a description will be given of a case in which the determination unit 121 determines that the text data of the conversation input by the user is not related to identifying a job listing.

[0079] When the process performed by the determination unit 121 determines that the text entered by the user is not related to identifying a job opportunity, the acquisition unit 122 responds by combining the chatbot's response with a text that prompts the user to enter information about the user. For example, Figs. 9(1) to (3) are display screens when a keyword not related to identifying a job opportunity is entered. For example, as shown in Fig. 9(1), when the user enters a text such as "I can't get a girlfriend," the acquisition unit 122 displays a standard text such as "If you tell us your age, we can introduce you to jobs where there are people of a similar age" along with the chatbot's response such as "Take care of yourself and value your charm and greatness to build a comfortable relationship."

[0080] As a result, the job information providing device 100 can respond to the conversation text data entered by the user while prompting the user to enter information about the user that has not yet been obtained, thereby satisfying the user's needs and collecting information to identify the job openings that the user desires.

[0081] As another example, when the determination unit 121 determines that the text data of the conversation input by the user is not related to identifying a job application, the acquisition unit 122 may input a prompt that gives a predetermined role according to the text data of the conversation input by the user to the learning model used for the chatbot. For example, as shown in FIG. 9(2), when the user inputs a sentence such as "Please tell me some tips for finding a part-time job," the acquisition unit 122 inputs a prompt such as "Please answer as a career advisor" according to the keyword "part-time job" input by the user to the learning model used for the chatbot.

[0082] Then, the acquisition unit 122 displays a standard phrase such as "If you tell us about your current occupation, we can introduce you to a job that suits you better" along with the chatbot's answer such as "Tips for finding a part-time job: · Clarify your goals and decide on a working style · Search for job information from online or real information sources · Carefully create a resume and curriculum vitae · Prepare by researching the company and the industry before the interview · Check the reputation of the part-time job and check the environment and treatment · Highlight your strengths and experience · Create a self-promotional pitch · Actively apply and find a part-time job that suits you · Use your network to get information and introductions. Put these tips into practice and find a part-time job that suits you."

[0083] As another example, as shown in FIG. 9(3), when a user inputs a sentence such as "Please tell me about part-time jobs that will be useful in job hunting," the acquisition unit 122 inputs a prompt such as "Please answer as a career advisor" into the learning model used by the chatbot in response to the keyword "job hunting" input by the user.

[0084] The acquisition unit 122 then displays a standard phrase such as "The following part-time jobs are useful in job hunting: 1. Part-time jobs related to companies: Part-time work experience at a company related to your desired industry or occupation is advantageous in job hunting. 2. Customer service or sales part-time jobs: You can acquire communication skills and how to interact with people, which leads to improved business skills. 3. Project manager assistant: Experience in project progress management and task management is useful for project work at a company. 4. Marketing or advertising-related part-time jobs: Experience in marketing strategy and advertising production is advantageous for finding employment in the marketing department of a company. 5. Event staff: Experience in planning and managing events is useful for finding employment in the event industry or as a project manager. These part-time jobs can give you an advantage in job hunting through practical experience and the acquisition of business skills. In addition, part-time work experience related to the industry or occupation will be useful for determining your own interests and aptitude." along with the chatbot's answer such as "If you tell us the occupation or job you want to work in in your job hunting, we may be able to introduce you to jobs that are useful in your job hunting."

[0085] As a result, when a user inputs information unrelated to identifying a job listing, the job information providing device 100 inputs a prompt that assigns a specified role to the chatbot, thereby providing a professional response to the user's requests, etc., and providing the information the user is looking for, thereby reducing the effort required for the user to search for information themselves and reducing the burden on the user.

[0086] [Specific Processing] Next, an example of the specification process performed by the job information providing device 100 according to the embodiment will be described with reference to Fig. 10. Fig. 10 is a diagram illustrating an example of the specification process performed by the job information providing device 100 according to the embodiment. In the example of Fig. 10, information such as "gender: female", "occupation: university student", "desired place of work: Tokyo", and "desired occupation: food / restaurant" is acquired by the acquisition unit 122 as information related to the user.

[0087] 10 is an example of a job offer identified by the identifying unit 123 when the user's gender is female, occupation is university student, desired place of work is Tokyo, and desired occupation type is food / restaurants. For example, the identifying unit 123 identifies a search condition "students welcome" for the information about the user acquired by the acquiring unit 122, "occupation: university student."

[0088] In addition, for example, the identification unit 123 identifies the search criteria "occupation: hall staff (serving food), place of work: cafe, coffee, coffee shop, place of work: sweets (ice cream shop, cake shop, creperie)" for the information about the user acquired by the acquisition unit 122, "gender: female" and "desired job type: food / restaurant."

[0089] Here, the information about the user and the search conditions do not have to have a one-to-one relationship. That is, the identification unit 123 may identify multiple search conditions from information about one user, or may identify one search condition from information about multiple users. In addition, the information about the user may be identified as a search condition as it is, such as a geographical condition such as a desired work area.

[0090] Then, the identification unit 123 identifies a job posting such as "No. 01234567:" as shown in FIG. 10 based on search conditions such as "students welcome, hall staff (serving food), cafe, coffee, coffee shop" that correspond to the information about the user such as "gender: female," "occupation: university student," "desired place of work: Tokyo," and "desired job type: food and beverage."

[0091] Next, a process for identifying a job posting based on a keyword in a sentence will be described. For example, when a sentence such as "I want to work at a ramen shop in Shibuya with free meals" is acquired as information about a user by the acquisition unit 122, the identification unit 123 identifies a geographical condition from the keyword "Shibuya", identifies a feature of meal subsidies from the keyword "free meals", and identifies a place to work (type of shop) from the keyword "ramen shop".

[0092] Identification of search conditions by keywords included in the text may be performed by comparing the scores of the vectorized keywords and existing search conditions. For example, in the above example, the vectorized keyword "includes meals" is compared in cosine similarity with the vectorized existing search conditions, and "subsidized meals" with a high cosine similarity is identified as the search condition. The identification unit 123 identifies search conditions corresponding to keywords included in the text acquired by the acquisition unit 122, and searches for job listings using the identified search conditions, thereby identifying job listings.

[0093] In this way, the job information providing device 100 identifies job listings based on conditions specified from keywords contained in natural text, making it easier for users to identify desired job listings and reducing the burden on users.

[0094] [Summarization process] Next, an example of summarization processing performed by the job information providing device 100 according to the embodiment will be described with reference to Fig. 11. Fig. 11 is a diagram illustrating an example of summarization processing performed by the job information providing device 100 according to the embodiment. The summarization unit 124 summarizes the job request identified by the identification unit 123. For example, the summarization unit 124 inputs a prompt such as "Please summarize the manuscript of the job request No. ****** in 300 characters" to a first learning model used to generate a sentence summarizing the job request, and generates a sentence summarizing the manuscript of the job request.

[0095] At this time, the summarizing unit 124 can generate a sentence summarizing the draft of the job request so as to include the feature of the job request that the user considers important. For example, the identifying unit 123 inputs a prompt such as "Please summarize the draft of the job request No. ****** in 300 characters. At that time, please include the feature of the job request that the user considers important, "location condition." into the first learning model used to generate a sentence summarizing the job request, and generates a sentence summarizing the draft of the job request. For example, the summarizing unit 124 generates a sentence summarizing the draft of the job request identified by the identifying unit 123 (FIG. 11(1)) so as to include the feature of the job request that the user considers important, "location condition" (FIG. 11(2)).

[0096] In this way, the job information providing device 100 can summarize job requests including the features of the job requests that the user considers important, thereby reducing the burden on the user of carefully examining the information on each job request included in the search results for job requests.

[0097] [Explanation Processing] Next, an example of the explanation process performed by the job information providing device 100 according to the embodiment will be described with reference to Figs. 12 to 14. Figs. 12 to 14 are diagrams for explaining an example of the explanation process performed by the job information providing device 100 according to the embodiment. First, the recommendation reason will be described. The recommendation reason is a feature of a job listing used to explain the reason for presenting the job listing to a user.

[0098] More specifically, for a user having user-related information such as "gender: female, occupation: university student," the job information providing device 100 identifies "solid educational system, cute uniform" as the reason for recommendation from among the job posting's characteristics, "5 minutes' walk from station, meal subsidy, willing to make friends, solid educational system, cute uniform."

[0099] The job information providing device 100 may identify the recommendation reason using a learning model. For example, the job information providing device 100 learns the relationship between information about the user, features of the job, and the recommendation reason, and identifies the recommendation reason by inputting information about the user and features of the job to a learning model that outputs a recommendation reason using information about the user and features of the job. Note that the processing performed using the learning model described in this specification may be performed by an external system, and in that case, the job information providing device 100 can receive the processing result performed by the external system.

[0100] The explanation unit 125 receives as input the information on the user acquired by the acquisition unit 122 and the information on the job listing identified by the identification unit 123, and uses the second learning model to generate a sentence explaining the reason for presenting the job listing to the user. For example, the explanation unit 125 receives as input the user attributes and the like and the characteristics of the job listing (including the reason for recommendation), and uses the second learning model to generate a sentence explaining the reason for presenting the job listing to the user.

[0101] At this time, the explanation unit 125 may assign a predetermined role according to information about the user. The explanation unit 125 may also input a prompt that restricts the contents of the output to the second learning model. More specifically, as shown in FIG. 12(1), the explanation unit 125 first inputs a prompt such as "Presentation You are a recommendation sentence creator for a recruitment company. Your job is to create recommendation sentences based on the facts written in the recruitment content. Please write an introduction sentence that recommends female college students using the recruitment article as a keyword. #Constraint Please do not add conditions other than the content I provide. Also, please do not use bullet points that list keywords. #Recruitment article Short-term★Donut sales! OK for both academic studies and work #Keywords Job type: Deli, prepared food sales, sweets sales Workplace: Chuo-ku, Osaka City, Osaka Prefecture Work facility: Shopping mall, outlet Recruitment characteristics: 5 minutes walk from the station Recruitment characteristics: Limited time" to the second learning model.

[0102] Then, as shown in FIG. 12(2), the explanation unit 125 generates a sentence explaining the reason for presenting the job opportunity to the user, such as "Attention female college students, we are recruiting short-term staff members selling donuts at a shopping mall with excellent access in Chuo Ward, Osaka City. This is an environment where you can work with peace of mind even during your busy daily life, as you can balance your work with your studies. It is also a convenient 5-minute walk from the station, making it easy to commute! What's more, this job is for a limited period of time. It is the perfect job opportunity for students. It is a great opportunity to gain experience in deli and prepared food sales, and sweets sales. Why not gain new experience and some pocket money at the same time? We look forward to receiving your application!"

[0103] As another example, as shown in FIG. 13, the explanation unit 125 may send the following message to the user: "You are a world-class expert in writing recommendation text for job advertisements. Please write the content in Japanese of 300 characters or less. The ideal answer for the input format is shown below. #Job title: Are you in financial trouble? Why not rebuild your life at a factory? #Reasons for recommendation: Reason 1 for recommendation: Work place: Factory, Reason 2 for recommendation: High salary, Reason 3 for recommendation: City, ward, town, or village: Chuo-ku, Osaka City, Osaka Prefecture, Reason 4 for recommendation: Middle-aged people working, Reason 5 for recommendation: Commuting by car is OK #Ideal answer Good news for those who are seriously broke! There is a chance to get back on track with life support, same-day dormitory entry, and daily pay. Since it is a factory job, the place of work is stable. Moreover, you can expect a high income, so you can be financially secure. The work location is in Chuo-ku, Osaka City, Osaka Prefecture, so it is easy to access and convenient to commute. In addition, since it is a workplace where middle-aged people are also active, you can grab the opportunity regardless of experience or age. Furthermore, you can commute by car, so you will not have to worry about commuting. This job offer is a perfect opportunity for those who are serious about working. If you want to solve your money problems and make a new start, please apply. Now, while following the rules I have specified, please be careful and write a recommendation for a job advertisement based on the job title and recommendation reasons below, referring to the ideal answers shown above. The target of the recommendation is a female housewife. #Job title No resume required! Production-related office work (experienced people welcome) #Reason for recommendation Prompts such as "Reason for recommendation 1: Housewives welcome, Reason for recommendation 2: Small job type: General office work, Reason for recommendation 3: Daytime work: Available, Reason for recommendation 4: Middle-aged workers active, Reason for recommendation 5: Commuting by car OK" are input into the second learning model.

[0104] Then, as shown in FIG. 14, the explanation unit 125 generates a sentence explaining the reason for presenting the job listing to the user, such as "We have job listings perfect for housewives! No resume is required, and first-time applicants as well as experienced applicants are all welcome. The job is a small-scale general office job related to production, and you can work during the day, so you can easily balance it with housework and child-rearing. You can work with peace of mind without worrying about experience or age in a work environment where middle-level employees are also active. Commuting by car is also OK! This is a job listing recommended for housewives who want to work at their own pace without being bound by commuting time or method. Why not start your challenge at a new workplace here?"

[0105] In addition, in the text explaining the reason for presenting the job listing to the user, the order in which elements such as the job type and the appealing points of the job are described and the order in which the reasons for recommendation are explained can be specified by a prompt so that they are in a predetermined order according to the purpose. In this way, the job information providing device 100 can eliminate doubts that the user has about the presented job listing by explaining the reason for recommending the job listing, and reduce the burden on the user. In addition, the job information providing device 100 can encourage users to apply for the job listing by explaining the reason for recommendation according to user attributes, etc., and can improve the application rate.

[0106] [Output processing] Next, an example of output processing performed by the job information providing device 100 according to the embodiment will be described with reference to FIG. 15. FIG. 15 is a diagram for explaining an example of specific processing performed by the job information providing device 100 according to the embodiment. For example, as shown in FIG. 15(1)(2), the output unit 126 outputs information on the job opportunity (including the title, URL, etc.), a sentence summarizing the job opportunity (for example, "XX Cafe Minatomirai branch is a 5-minute walk from the station," "XX Ramen Minatomirai branch is a 7-minute walk from the station," "XX Monja Minatomirai branch is a 3-minute walk from the station"), and a sentence explaining the reason for presenting (introducing) the job opportunity (for example, "Recommended points are a solid education system and cute uniforms," ​​"Recommended points are daily pay and free meals," "Recommended points are high hourly wage and welcoming to inexperienced people.").

[0107] Furthermore, when outputting information on a job offer, information summarizing the job offer, and text explaining the reason for presenting the job offer, the output unit 126 can integrate each piece of information and output the information by changing the display order according to the purpose. That is, the output unit 126 can change the display order when outputting according to existing know-how. For example, when it is desired to appeal for a job offer, the output unit 126 outputs the information so that a description regarding the merits of the job offer is displayed first.

[0108] In this way, the job information providing device 100 identifies job openings using user information collected through chat as search criteria, summarizes the identified job openings, and outputs a sentence explaining the reason for presenting them to the user, thereby reducing the burden on the user.

[0109] 〔flowchart〕 Next, the flow of processing by the job information providing device 100 will be described with reference to Fig. 16. Fig. 16 is a flowchart showing an example of the flow of processing according to this embodiment. The steps below may be executed in a different order, and some processing may be omitted.

[0110] First, the job information providing device 100 determines whether the user has a search history (step S101). If the job information providing device 100 determines that the user does not have a search history (step S101: No), the acquisition unit 122 acquires information necessary to specify the area (step S102). For example, the acquisition unit 122 specifies the area where the user will do the desired work by having the user input information about the area where the user will do the desired work in the form of a button.

[0111] Next, the acquisition unit 122 presents to the user jobs with many applications under the acquired conditions (step S103). For example, the job information providing device 100 presents to the user jobs with many applications in the area based on information on the work area desired by the user specified by the acquisition unit 122.

[0112] On the other hand, if the job information providing device 100 determines that a search history exists (step S101: Yes), the process proceeds to step S103. For example, the acquisition unit 122 presents to the user job listings with many applications among job listings searched for using keywords in the user's past job searches.

[0113] Next, the acquisition unit 122 acquires text data of the conversation input by the user (step S104). For example, the acquisition unit 122 acquires information about the user from the text data of the conversation between the chatbot and the user. Next, the determination unit 121 determines whether the text data of the conversation input by the user is related to the introduction of a job offer (step S105).

[0114] Here, if the determination unit 121 determines that the text data of the conversation input by the user is not related to the introduction of a job listing (step S105: No), the acquisition unit 122 responds to the text data of the conversation input by the user (step S106). For example, if the user inputs a keyword that is not related to identifying a job listing, the acquisition unit 122 responds to the text data of the conversation input by the user.

[0115] Next, the acquisition unit 122 responds with a sentence that prompts the user to input information about the user (step S107). For example, the acquisition unit 122 outputs a standard phrase that prompts the user to input information together with the response to the user's input. Then, the process returns to step S104.

[0116] On the other hand, if the determination unit 121 determines that the text data of the conversation input by the user is related to the introduction of job information (step S105: Yes), the identification unit 123 identifies search conditions from information about the user (step S108). For example, the identification unit 123 identifies corresponding search conditions from information about the user by referring to a database that defines the correspondence between the acquired information about the user and the search conditions.

[0117] Next, the identifying unit 123 identifies a job request based on the identified search condition (step S109). For example, the identifying unit 123 searches for a job request using the identified search condition as a filter, thereby identifying a job request that satisfies the condition.

[0118] Next, the summarizing unit 124 receives information about the job request identified by the identifying unit 123 as an input and generates a sentence summarizing the job request using the first learning model (step S110). For example, the summarizing unit 124 receives a manuscript of the job request identified by the identifying unit 123 as an input and generates a sentence summarizing the job request using the first learning model.

[0119] Next, the explanation unit 125 uses the information about the user acquired by the acquisition unit 122 and the information about the job request identified by the identification unit 123 as input, and generates a sentence explaining the reason for presenting the job request to the user using the second learning model (step S111). For example, the explanation unit 125 generates a sentence explaining the reason for presenting the job request to the user using the second learning model in which a predetermined role is assigned according to the information about the user and a prompt for explaining the features of the job request is input.

[0120] Next, the output unit 126 outputs information related to the job request identified by the identification unit 123, the summarized text generated by the summarization unit 124, and the explanatory text generated by the explanation unit 125 (step S112). For example, the output unit 126 outputs on the timeline an image of the job request identified by the identification unit 123, the headline of the job request, the URL of the job request, the text summarizing the draft of the job request generated by the summarization unit 124, and the text generated by the explanation unit 125 explaining the reason for presenting the job request to the user.

[0121] 〔effect〕 Thus, the job information providing device 100 of the embodiment has an acquisition unit 122 that acquires information about the user from text data of the conversation between the chatbot and the user, an identification unit 123 that identifies a job request based on the information about the user acquired by the acquisition unit 122, and an output unit 126 that outputs information about the job request identified by the identification unit 123.

[0122] As a result, the job information providing device 100 can identify job openings using user information collected through chat as search conditions and present them to the user, thereby reducing the burden on the user and encouraging the user to apply for job openings.

[0123] In other words, when a user searches for a job, the job information providing device 100 collects user attributes, keywords, conditions, etc. through interactions with a chatbot using a learning model, and identifies the job using the collected information, thereby reducing the burden on the user when searching for a desired job and encouraging the user to apply for a job.

[0124] That is, in the conventional technology, there were cases where it was not possible to reduce the burden on the user when searching for and applying for job listings, but the job information providing device 100 can solve problems such as the following: For example, in the conventional technology, in order to find a desired job listing, it is necessary to search for job listings by inputting information such as user attributes, keywords, and conditions one by one, which places a heavy burden on the user.

[0125] Furthermore, for example, in conventional technology, if a user has not clearly articulated what type of job opportunity they are looking for, the user must spend time considering search conditions and keywords, which places a heavy burden on the user.

[0126] In addition, the job information providing device 100 according to the embodiment further has a determination unit 121 that determines whether or not the text data of the conversation input by the user is related to the identification of a job listing based on the relationship between the vectorized text data of the conversation input by the user and default search conditions, and when the determination unit 121 determines that the text data of the conversation input by the user is related to the identification of a job listing, the acquisition unit 122 acquires the information as information about the user.

[0127] As a result, the job information providing device 100 makes a judgment based on the cosine similarity between the information collected via chat and the default search conditions, and acquires the information about the user only if the specified conditions are met, thereby saving the user the trouble of inputting search conditions, etc., reducing the burden on the user, and encouraging users to apply for job listings.

[0128] In addition, when the determination unit 121 determines that the text data of the conversation input by the user is not related to identifying a job listing, the acquisition unit 122 of the job information providing device 100 according to the embodiment outputs a sentence prompting the input of information about the user along with a response to the text data of the conversation input by the user.

[0129] As a result, when a user inputs information unrelated to identifying a job opening, the job information providing device 100 provides a normal response while also outputting a sentence encouraging the user to obtain information about the user that has not yet been obtained, thereby satisfying the user's needs and collecting information needed to identify a job opening that the user is expected to apply for.

[0130] In addition, when the determination unit 121 determines that the text data of the conversation input by the user is not related to identifying a job listing, the acquisition unit 122 of the job information providing device 100 according to the embodiment inputs a prompt that assigns a predetermined role according to the text data of the conversation input by the user into a learning model used by the chatbot.

[0131] As a result, when a user inputs information unrelated to identifying a job listing, the job information providing device 100 inputs a prompt that gives the chatbot a specified role, thereby providing a professional response to the user's requests, etc., and providing the information the user is looking for, thereby reducing the effort required for the user to search for information themselves, reducing the burden on the user, and encouraging users to apply for job listings.

[0132] Furthermore, the identifying unit 123 of the job information providing device 100 according to the embodiment identifies a job opportunity from a search result based on search conditions identified from the information about the user acquired by the acquiring unit 122. As a result, the job information providing device 100 identifies a job opportunity using search conditions according to the information about the user, thereby reducing the effort required for the user to identify a job opportunity by considering search conditions on his / her own, reducing the burden on the user, and encouraging the user to apply for the job opportunity.

[0133] Furthermore, the identifying unit 123 of the job information providing device 100 according to the embodiment identifies a job request from the information about the user acquired by the acquiring unit 122, using a model that has learned the relationship between information about the user and the job request. In this way, the job information providing device 100 identifies a job request according to the information about the user, thereby reducing the effort required for the user to identify a job request by considering search conditions, reducing the burden on the user, and encouraging the user to apply for the job request.

[0134] In addition, the job information providing device 100 of the embodiment further has a summarization unit 124 that receives information about the job opportunity identified by the identification unit 123 as input and generates a sentence summarizing the job opportunity using a first learning model, and an explanation unit 125 that receives information about the user acquired by the acquisition unit 122 and information about the job opportunity identified by the identification unit 123 as input and generates a sentence explaining the reason for presenting the job opportunity to the user using a second learning model, and the output unit 126 outputs the information about the job opportunity identified by the identification unit 123, the summarized sentence generated by the summarization unit 124, and the explanatory sentence generated by the explanation unit 125.

[0135] Furthermore, when a user applies for a job, the job information providing device 100 can reduce the burden on the user from identifying a job to applying by summarizing the job manuscript and explaining the features of the job as the reason for introducing the job to the user. In this way, the job information providing device 100 can reduce the burden on various users when searching for and applying for a job, and can promote users' applications for job listings and improve the application rate by preventing users from dropping out before applying.

[0136] Furthermore, with conventional technology, even when a search could be performed without any problems, job listings contained a large amount of information, and the user had to carefully examine the information for each listing to determine whether or not it was a job listing they wished to apply for, which placed a heavy burden on the user.

[0137] Furthermore, for example, in conventional technology, when a job listing is presented (introduced), the user has a psychological resistance (burden) such as a question, such as "Why was this job listing introduced?" In conventional technology, the user has a burden as described above, and it is not rare for the user to drop out before applying for the job listing. The job information providing device 100 can solve such problems, reduce the burden on the user, and encourage the user to apply for the job listing.

[0138] In addition, the summarization unit 124 of the job information providing device 100 according to the embodiment further inputs information about the user and generates a sentence summarizing the job posting using a first learning model that inputs a prompt to summarize the manuscript of the job posting in accordance with the information about the user.

[0139] As a result, the job information providing device 100 can summarize the draft of a job listing so as to include the features of the job listing that the user considers important, thereby reducing the burden on the user of carefully examining the information on each job listing and encouraging users to apply for the job listing.

[0140] In addition, the explanation unit 125 of the job information providing apparatus 100 according to the embodiment uses a second learning model that inputs a prompt for giving a predetermined role and explaining the characteristics of a job opening according to the information about the user, and generates a sentence explaining the reason for presenting the job opening to the user.

[0141] As a result, the job information providing apparatus 100 can give the role of an expert and explain the reason for recommending the job opening, thereby dispelling the doubts of the user about the job opening presented to the user and reducing the burden on the user. In addition, the job information providing apparatus 100 can promote the user's application for the job opening and improve the application rate by explaining the reason for recommendation according to the user attributes and the like.

[0142] 〔Hardware Configuration〕 In addition, the job information providing apparatus 100 according to the above-described embodiment is realized by, for example, a computer 1000 having a configuration as shown in FIG. 17. FIG. 17 is a hardware configuration diagram showing an example of a computer that realizes the functions of the job information providing apparatus 100. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM 1300, an HDD 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F) 1700.

[0143] The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400, and controls each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

[0144] The HDD 1400 stores a program executed by the CPU 1100 and data used by such a program. The communication interface 1500 receives data from other devices via a predetermined communication network and sends it to the CPU 1100, and sends the data generated by the CPU 1100 to other devices via a predetermined communication network.

[0145] The CPU 1100 controls output devices such as a display and a printer, and input devices such as a keyboard and a mouse, via the input / output interface 1600. The CPU 1100 acquires data from the input devices via the input / output interface 1600. The CPU 1100 also outputs generated data to the output devices via the input / output interface 1600.

[0146] The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads the program from the recording medium 1800 onto the RAM 1200 via the media interface 1700 and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or a PD (Phase change rewritable Disc), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory.

[0147] For example, when the computer 1000 functions as the job information providing device 100 according to the embodiment, the CPU 1100 of the computer 1000 executes programs loaded onto the RAM 1200 to realize the functions of the control unit 120. The CPU 1100 of the computer 1000 reads and executes these programs from the recording medium 1800, but as another example, the CPU 1100 may obtain these programs from another device via a predetermined communication network.

[0148] 〔others〕 Although various embodiments have been described in detail herein with reference to the drawings, these embodiments are merely examples and are not intended to limit the present invention. In other words, the present invention is capable of learning, constructing, and updating various multi-value classification models and estimating a desired probability by changing the feature amount used as an input. The features described herein can be realized by various methods, including various modifications and improvements based on the knowledge of those skilled in the art.

[0149] Moreover, the above-mentioned "section" can be read as a unit, a means, a circuit, etc. For example, the communication section, the control section, and the storage section can be read as a communication unit, a control unit, and a storage unit, respectively. [Explanation of symbols]

[0150] 100 Recruitment information providing device 110 Communications Department 120 Control section 121 Judgment section 122 Acquisition Department 123 Specific part 124 Summary 125 Explanation Section 126 Output section 130 Storage section

Claims

1. An acquisition unit that acquires information about the user from text data of a conversation between the chatbot and the user; an identification unit that identifies a job offer based on the information about the user acquired by the acquisition unit; an output unit that outputs information regarding the job offer identified by the identification unit; A job information providing device comprising:

2. The method further includes a determination unit that determines whether the text data of the conversation input by the user is related to the identification of the job offer based on a relationship between the vectorized text data of the conversation input by the user and a preset search condition, When the determination unit determines that the text data of the conversation input by the user is related to the identification of the job offer, the acquisition unit acquires the information as information about the user.

2. The job information providing device according to claim 1.

3. The acquisition unit is When the determination unit determines that the text data of the conversation input by the user is not related to the identification of the job application, a sentence prompting the input of information about the user is output together with a response to the text data of the conversation input by the user.

3. The job information providing device according to claim 2.

4. The acquisition unit is When the determination unit determines that the text data of the conversation input by the user is not related to the identification of the job application, a prompt that assigns a predetermined role according to the text data of the conversation input by the user is input to a learning model used by the chatbot.

3. The job information providing device according to claim 2.

5. The identification unit is Identifying the job posting from a search result based on search criteria identified from the information about the user acquired by the acquisition unit 2. The job information providing device according to claim 1.

6. The identification unit is Using a model that has learned the relationship between information about the user and the job listing, the job listing is identified from the information about the user acquired by the acquisition unit.

2. The job information providing device according to claim 1.

7. a summarizing unit that uses information about the job offer identified by the identifying unit as an input and generates a sentence summarizing the job offer using a first learning model; an explanation unit that uses as input the information about the user acquired by the acquisition unit and the information about the job offer identified by the identification unit, and uses a second learning model to generate a sentence explaining the reason for presenting the job offer to the user; and The output unit is Outputting information related to the job offer identified by the identification unit, the summarized text generated by the summarizing unit, and the explanatory text generated by the explanatory unit.

2. The job information providing device according to claim 1.

8. The summary portion comprises: Furthermore, by inputting information about the user, the first learning model is used to generate a sentence summarizing the job posting in response to the information about the user, the first learning model being input with a prompt to summarize the job posting.

8. The job information providing device according to claim 7.

9. The explanation section includes: A sentence explaining the reason for presenting the job listing to the user is generated using the second learning model to which a predetermined role is assigned and a prompt for explaining the characteristics of the job listing is input according to information about the user.

8. The job information providing device according to claim 7.

10. A job information providing method executed by a job information providing device, An acquisition step of acquiring information about the user from text data of a conversation between the chatbot and the user; a specifying step of specifying a job offer based on the information about the user acquired by the acquiring step; an output step of outputting information regarding the job offer identified by the identification step; A method for providing job information, comprising:

11. An acquisition step of acquiring information about the user from text data of a conversation between the chatbot and the user; A step of identifying a job offer based on the information about the user acquired by the acquisition step; an output unit that outputs information regarding the job offer identified by the identifying step; A job information providing program that causes a computer to execute the above steps.

Images