Human resources assignment system, human resources assignment method, and human resources assignment program

Abstract

To provide a personnel work allocation system or the like capable of acquiring atypical work information and atypical personnel information, acquiring information required for allocating personnel to work, and reducing labor for inputting the information.SOLUTION: A human resource work assignment system or the like according to the present disclosure acquires work skill information related to work skills required of personnel from atypical work information, and acquires holding skill information related to skills held by the personnel from atypical human resource information for each personnel; Based on the comparison between the holding skill information and the work skill information for each personnel, candidates of personnel engaged in the work are selected from the introduced personnel, level information indicating the level of the work skill of the candidate is acquired for each candidate from the holding skill information of the selected candidate, weight information indicating the importance of the work skill is acquired based on the work information, and the total score for each candidate is calculated based on the weight information and the level information for each work skill.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a human resource task assignment system, a human resource task assignment method, and a human resource task assignment program, and more particularly to a human resource task assignment system, a human resource task assignment method, and a human resource task assignment program that are suitable for assigning human resources suitable for tasks. [Background technology]

[0002] An assignment is the process of selecting and allocating personnel to participate in a project, business plan, matter, project, job, etc. that is currently being implemented or is planned to be implemented. The person in charge of assignments is required to select and assign the most suitable personnel for each task from a large number of candidates within a limited time. In addition, there are cases where the abilities and character of a candidate must be judged from documents such as resumes and curriculum vitae. Not only do these documents come in a variety of file formats and appearances, but their content also includes non-standard elements such as free-form writing, so the person in charge of assignments must carefully read and interpret each one. Furthermore, assignments often required taking multiple conditions into consideration, which meant that it took a huge amount of time to assign the right personnel for each job, and there was sometimes a risk of mismatch. For this reason, there was a need for a human resource task allocation system that could automatically select and assign suitable personnel from multiple personnel to each task.

[0003] Therefore, an allocation system capable of matching job recruiters (employing companies) with job seekers has been proposed (see, for example, Patent Document 1). The allocation system disclosed in Patent Document 1 claims that even if job seekers and employers (employing companies) are unable to clearly prioritize the types of companies or personnel they are looking for, they can accurately extract and match companies or personnel that meet the conditions simply by selecting and entering registration information.

[0004] However, the allocation system disclosed in Patent Document 1 stores job seeker value information selected and input by job seekers from a number of options provided for each of a number of common items in a job seeker database, and also stores employer value information selected and input by employers (employing companies) from a number of options provided for each of a number of common items in a employer database. In other words, the allocation system disclosed in Patent Document 1 only allows the input of job seeker value information using pre-prepared, standardized items and standardized options, which forces job seekers and employers (employing companies) to input job seeker value information and employer value information in a standardized manner, which is cumbersome and burdensome for job seekers and employers (employing companies).In addition, it is not possible to input information such as resumes that include non-standardized elements. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2006-59195 Summary of the Invention [Problem to be solved by the invention]

[0006] Therefore, the present invention aims to provide a human resources task assignment system, a human resources task assignment method, and a human resources task assignment program that can acquire non-standard task information related to tasks provided by those who wish to introduce personnel to work on the tasks, and non-standard personnel information related to personnel to be introduced as potential personnel, and can acquire information necessary to assign personnel to tasks based on this task information and personnel information, thereby reducing the effort required to input this information. [Means for solving the problem]

[0007] That is, the human resources task allocation system according to the first aspect is characterized by comprising: a task skill information acquisition unit that acquires task skill information regarding the task skills required of personnel engaged in the task from non-standard task information regarding the task provided by a client who wishes to introduce personnel to engage in the task; a possessed skill information acquisition unit that acquires, for each personnel, possessed skill information regarding the skills possessed by the personnel from non-standard personnel information regarding the personnel introduced by a recruitment agency; a candidate selection unit that selects candidates for personnel to engage in the task from the introduced personnel based on a comparison between the possessed skill information and the task skill information; a level information acquisition unit that acquires, for each candidate, level information indicating the level of the candidate's task skill from the possessed skill information of the selected candidate; a weight information acquisition unit that acquires weight information indicating the importance of the task skill based on the task information; and a total score calculation unit that calculates a total score indicating the suitability of each candidate for the task based on the weight information and level information for each task skill.

[0008] In a second aspect, the human resources task assignment system according to the first aspect may further include a ranking unit that assigns a recommendation ranking to the candidates for the tasks based on the total score.

[0009] In a third aspect, the human resource task allocation system according to the second aspect may further include an output unit that outputs a candidate list in which the candidates are arranged in association with a recommendation ranking of the candidates.

[0010] In a fourth aspect, in the human resource task allocation system according to the first aspect, the task skill information acquisition unit may use a large-scale language model to acquire task skill information from non-standard task information and store it in the memory unit as key-value format information.

[0011] In a fifth aspect, in the human resource task allocation system according to the first aspect, the possessed skill information acquisition unit may use a large-scale language model to acquire possessed skill information from non-standard human resource information and store the information in the memory unit as key-value format information.

[0012] In a sixth aspect, in the human resource task assignment system according to the first aspect, the candidate selection unit may select as candidates human resources whose possessed skill information has a similarity to the business skill information that is equal to or greater than a threshold value through a semantic search based on the business skill information and the possessed skill information.

[0013] In a seventh aspect, in the human resources job assignment system according to the first aspect, the level information acquisition unit may use a large-scale language model to acquire level information indicating the level of the candidate's job skills from the candidate's skill information, and store the information in a memory unit as key-value format information combining the job skill information and level information for each candidate.

[0014] In an eighth aspect, in the human resources task assignment system according to the first aspect, task skills include required skills, which are required skills imposed on personnel, and satisfactory skills, which are skills that are different from the required skills and that personnel are desired to possess, and the level information acquisition unit acquires, for each candidate, required level information indicating the level of the candidate's required skills and satisfactory level information indicating the level of the candidate's satisfactory skills from the skill information possessed by the selected candidate, the weight information acquisition unit acquires, based on the task information, required skill weight information and satisfactory skill weight information indicating the importance of the required skills and satisfactory skills, respectively, and the total score calculation unit calculates a total score for each candidate based on the required level information, satisfactory level information, required skill weight information, and satisfactory skill weight information.

[0015] In a ninth aspect, in the human resources task allocation system according to the first aspect, the system further includes a table information acquisition unit that acquires the contents of the table information expressed in table format using a large-scale language model and an image recognition model, and the possessed skill information acquisition unit may acquire the contents of the table information using the table information acquisition unit when the possessed skill information includes table information.

[0016] A tenth aspect of the human resource task allocation method is characterized in that a computer executes the following steps: a task skill information acquisition step for acquiring task skill information related to the task skills required of personnel engaged in the task from non-standard task information related to the task provided by a client who wishes to introduce personnel to engage in the task; a possessed skill information acquisition step for acquiring, for each personnel, possessed skill information related to the skills possessed by the personnel from non-standard personnel information related to the personnel introduced by a recruitment agency; a candidate selection step for selecting candidates for personnel to engage in the task from the introduced personnel based on a comparison between the possessed skill information and the task skill information; a level information acquisition step for acquiring, for each candidate, level information indicating the level of the candidate's task skill from the possessed skill information of the selected candidate; a weight information acquisition step for acquiring weight information indicating the importance of the task skill based on the task information; and a total score calculation step for calculating a total score indicating the suitability of each candidate for the task based on the weight information and level information for each task skill.

[0017] The human resources task allocation program of the eleventh aspect is characterized in that it implements on a computer the following: a task skill information acquisition function that acquires task skill information regarding the task skills required of personnel engaged in the task from non-standard task information regarding the task provided by a client who wishes to introduce personnel to engage in the task; a possessed skill information acquisition function that acquires, for each personnel, possessed skill information regarding the skills possessed by the personnel from non-standard personnel information regarding the personnel introduced by a recruitment agency; a candidate selection function that selects candidates for personnel to engage in the task from the introduced personnel based on a comparison between the possessed skill information and the task skill information; a level information acquisition function that acquires, for each candidate, level information indicating the level of the candidate's task skill from the possessed skill information of the selected candidate; a weight information acquisition function that acquires weight information indicating the importance of the task skill based on the task information; and a total score calculation function that calculates a total score indicating the suitability of each candidate for the task based on the weight information and level information for each task skill. [Effects of the Invention]

[0018] The human resource task allocation system according to the present invention includes a task skill information acquisition unit that acquires task skill information related to task skills required of personnel engaged in the task from non-routine task information related to the task provided by a client who desires the introduction of personnel engaged in the task, a possessed skill information acquisition unit that acquires, for each personnel, possessed skill information related to skills possessed by the personnel from the non-routine personnel information related to personnel introduced by a recruitment agent, a candidate selection unit that selects candidates for personnel to engage in the task from the introduced personnel based on a comparison between the possessed skill information and the task skill information, and a level indicator that indicates the level of the candidate's task skill from the possessed skill information of the selected candidate. The system is characterized by comprising a level information acquisition unit that acquires information for each candidate, a weight information acquisition unit that acquires weight information indicating the importance of business skills based on business information, and a total score calculation unit that calculates a total score for each candidate based on the weight information and level information for each business skill.As a result, it is possible to acquire business information provided by a client who wishes to introduce personnel to work, and non-standard human resources information regarding personnel introduced by a recruitment agent who is a candidate for personnel, and to acquire information required to assign personnel to work based on this non-standard business information and human resources information, thereby reducing the effort required to input this information. Furthermore, the human resource task assignment method and human resource task assignment program of the present invention, like the human resource task assignment system of the present invention, can acquire non-standard task information provided by a client who wishes to be introduced to personnel to work on the task, and non-standard personnel information regarding personnel introduced by a recruitment agent who is a candidate for personnel, and can acquire information required to assign personnel to tasks based on this non-standard task information and non-standard personnel information, thereby reducing the effort required to input this information. [Brief explanation of the drawings]

[0019] [Figure 1] FIG. 1 is a diagram for explaining an overview of a human resource task assignment system according to this embodiment. [Figure 2] FIG. 2 is a diagram showing an example of a business information list according to this embodiment. [Figure 3] FIG. 3 is a block diagram illustrating an example of the hardware configuration of the human resource task assignment system according to this embodiment. [Figure 4] FIG. 4 is a block diagram illustrating an example of the functional configuration of the human resource task assignment system according to this embodiment. [Figure 5] FIG. 5 is a diagram showing an example of the body of an email sent from a recruitment agent according to this embodiment. [Figure 6] FIG. 6 is a diagram showing an excerpt from a first example of a resume of a person introduced by a recruitment agent according to this embodiment. [Figure 7] FIG. 7 is a diagram showing an excerpt from a second example of a resume of a person introduced by a recruitment agent according to this embodiment. [Figure 8] FIG. 8 is a diagram for explaining an example of an output screen of the human resource task assignment system according to this embodiment. [Figure 9] FIG. 9 is an example of a block diagram for explaining the structure of a human resource task assignment program according to this embodiment. [Figure 10] FIG. 10 is an example of a flowchart of the human resource task assignment program according to this embodiment. [Figure 11] FIG. 11 is an example of a flowchart of a human resource task assignment program according to another embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0020] (Overview of Human Resources Task Allocation System 10) A human resource task assignment system 10 according to an embodiment of the present disclosure will be described with reference to Figures 1 to 9. First, an overview of the human resource task assignment system 10 according to this embodiment will be described with reference to Figure 1. The human resource task assignment system 10 matches tasks with human resources based on non-standard task information acquired from a client 1 and non-standard human resource information acquired from a recruitment agency 2, and provides the resulting assignment candidates 13 to an aggregator 3. Users of the human resource task assignment system 10, the human resource task assignment method, and the human resource task assignment program may be the requester 1, the recruitment agent 2, and the aggregator 3, as well as the human resource introduced by the recruitment agent 2. Also, users of the human resource task assignment system 10, the human resource task assignment method, and the human resource task assignment program may be the requester 1, the recruitment agent 2, and the aggregator 3, or those who operate on their behalf. Matching refers to creating a suitable combination by linking mutually compatible elements (in this disclosure, business and human resources). Matching in this disclosure is the matching of human resources introduced by recruitment agency 2 with the business of client 1, and an appropriate employment relationship is established for both parties when the human resources' skills and desired conditions match the skills and recruitment conditions required by the business. The assignment candidates 13 refer to the results of the human resource task allocation system 10 matching tasks with human resources and are presented to the aggregator 3, and the assignment candidate information refers to information describing the assignment candidates. The assignment candidates 13 are presented to the aggregator 3 with priorities assigned to the candidates based on predetermined conditions. The predetermined conditions are criteria for evaluating the priorities, and specifically, priorities are assigned to the candidates based on the evaluation criteria such as skill match rate, profit margin, partner reliability, and training slots. The assignment candidates 13 may be called matching candidates in the operation of the human resource task allocation system 10, and the assignment candidate information may be called matching candidate information.

[0021] The human resource task assignment system 10 is a so-called computer, and can be realized by general-purpose devices such as a server, a personal computer (hereinafter referred to as a PC), a desktop PC, a notebook PC, a tablet PC, a smartphone, etc. The human resource task assignment system 10 may also be constructed as a dedicated computer system. The requester 1 is a person who needs personnel to work on a business and wants to be introduced to personnel, and may be an individual or an organization, or even a person involved in the operation of a business may request personnel to work on that business using the name of the business. An organization may include not only a corporation or organization, but also an organization such as a department that forms part of a corporation or organization. The recruitment agency 2 refers to a department or company that provides personnel. The aggregator 3 refers to an assignor or a department in charge of assignment that matches tasks with personnel and assigns suitable personnel to the tasks. The department refers to, for example, an assignment office 3a established within a company that assigns personnel to tasks such as projects, business plans, cases, plans, and jobs that are currently being implemented or are scheduled to be implemented.

[0022] (Regarding Client 1) As shown in Figure 1, a project 1a and an end user 1b, etc., act as requesters 1 and provide non-routine task information to a human resource task assignment system 10 to make an assignment request (staff request) 12. The end user 1b refers to the person or organization that actually uses the human resource task assignment system 10. An assignment request (personnel request) 12 refers to a request for personnel required for a task such as a project, and may be made by providing task information including detailed task information to the human resources task allocation system 10. The task information is information about a task provided by a requester 1 who wishes to be introduced to personnel who will be engaged in the task, and includes information such as the status of the task, the number of personnel required for the task, task skill information, the personality profile of the personnel, the task budget, etc. Furthermore, the assignment request may be called a personnel request when operating the human resource task assignment system 10. Job skill information refers to information about the job skills required of personnel engaged in a job. Job skills may include required skills, which are skills that are required for personnel, and preferred skills, which are skills that are different from required skills and that personnel are desired to possess, as described below. Therefore, job skill information may include required skill information, which is information about required skills, and preferred skill information, which is information about preferred skills.

[0023] (Regarding Recruiter 2) As shown in Figure 1, departments 2a, partner companies 2b, and sole proprietors / freelancers 2c that function as part of a business entity act as recruitment agents 2 and provide non-standard personnel information to a personnel task allocation system 10 to make personnel proposals (personnel responses) 11. The department 2a that functions as part of the business operator refers to the business operator's human resources department, general affairs department, and labor department, etc., which are responsible for the recruitment, training, evaluation, operation, and labor management of personnel within the business operator, but is not limited to these and may also be the technology department, manufacturing department, development department, etc. Partner company 2b refers to another business entity with which the company has a cooperative relationship, conducts business jointly, or cooperates toward a specific goal. The sole proprietor / freelancer 2c does not belong to any organization such as a business, and acts as a recruitment agent 2, providing non-standard personnel information about himself / herself to the personnel task allocation system 10 and making personnel proposals (personnel responses) 11. The personnel proposal (personnel response) 11 refers to proposing personnel who can be assigned as personnel for a task, and may be performed by providing personnel information including detailed information about the personnel to the personnel task assignment system 10. The human resources information is information about human resources provided by the recruitment agency 2 and introduced by the recruitment agency 2, and is standardized information, or non-standardized information, or a mixture of standardized and non-standardized information, including information such as the work history of the human resources, skill information regarding the skills possessed by the human resources, personality profile of the human resources, desired unit price of the human resources, and affiliation of the human resources. The personnel proposals 11 may be referred to as personnel responses in the operation of the human resource task assignment system 10. Furthermore, the personnel information may be referred to as proposal information in the operation of the human resource task assignment system 10.

[0024] (Business and personnel information) The above-mentioned business information and human resource information may be provided to the human resource task assignment system 10 as electronic data in unstructured formats, such as files created by various applications, such as word processing application software, spreadsheet software, and presentation software; email text documents; documents written in markup languages, such as HTML (registered trademark); and electronic document files, such as PDF (registered trademark). The above-mentioned business information and human resource information may include, as part of a structured format, such as a CSV (Comma Separated Values) file, a JSON (JavaScript Object Notation) file, or a YAML (YAML Ain't Markup Language) file, and may be a mixture of structured and unstructured content. The above-mentioned business information and human resource information may also include, as part of a structured format, such as a table or object in a database (DB) system, and may be a mixture of structured and unstructured content. In the following, the terms "business information" and "personnel information" are used in the above sense. In addition, the business information and personnel information may be printed on paper, in which case the business information and personnel information may be converted into the above-mentioned electronic data by the human resources job assignment system 10 or other equipment before being provided to the human resources job assignment system 10. In addition, business information and human resources information are non-standard documents, and there are no specific specifications regarding the format, such as type, format, style, and appearance, and the format of the business information and human resources information can be selected at the discretion of the creator of the business information and human resources information. As described above, business information and human resources information are non-standard documents, but when acquired by the human resources task assignment system 10, the contents of the business information are stored in the memory unit 10d described below as key-value format data in which the contents of the business information are organized into pairs of key and value. The key-value format is the simplest format for storing data in a database or other database as key-value pairs, where each key is associated with exactly one value from a collection of values.

[0025] (About Aggregator 3) By introducing the human resource task allocation system 10, the aggregator 3 can reduce its burden and improve the accuracy of matching tasks and personnel. Hereinafter, an overview of various processes executed by the human resource task assignment system 10 will be described with reference to FIG. The human resource task allocation system 10 executes processes such as personnel proposal reception 15, assignment request reception 16, assignment DB 17, assignment matching 20, assignment result 21, assignment proposal 22, and assignment candidate presentation 23. The assignment DB 17 refers to a database that stores personnel information, business information, assignment candidate information, and the like in a predetermined format (for example, key-value format).

[0026] In the personnel proposal reception 15 , the personnel task allocation system 10 receives a personnel proposal (personnel response) 11 by acquiring personnel information provided by the recruitment agent 2 . In the assignment request reception 16 , the human resource task allocation system 10 receives an assignment request (personnel request) 12 by acquiring task information provided by the requester 1 . As described above, the human resource information and task information are not set to a specific format, but are created using various formats. The human resource task allocation system 10 reads the content of the task information and human resource information received at the personnel proposal reception 15 and the assignment request reception 16 using a large language model (LLM) as described below, stylizes it into a predetermined format (for example, key-value format), and stores it in the storage unit 10d of the assignment DB 17 described below. In the assignment DB 17, the human resource task allocation system 10 stores in the memory unit 10d the human resource information acquired in the personnel proposal reception 15, the task information acquired in the assignment request reception 16, and the assignment candidate information generated by the assignment matching 20 described below. The assignment candidate information is information about candidates for personnel to engage in the work selected by the candidate selection unit 43 described below, and specifically refers to the candidate's personal information, possessed skill information, total score Zk, etc.

[0027] In assignment matching 20, the human resource task allocation system 10 matches tasks with human resources based on the human resource information acquired in personnel proposal reception 15 and the task information acquired in assignment request reception 16, and assigns personnel to the tasks. In assignment matching 20, the human resource task allocation system 10 generates assignment candidates by matching tasks and personnel based on the task information and human resource information stored in the storage unit 10d in the assignment DB 17. The human resource task allocation system 10 evaluates the assignment candidates based on predetermined criteria, and presents the assignment candidates in descending order of evaluation to the client 1, recruitment introducer 2, and aggregator 3, as described below. Examples of the predetermined criteria include skill match, position match, profit margin, partner company reliability, and training quota. The skill match degree refers to the degree of compatibility between the business skills required of personnel engaged in the work and the skills possessed by the personnel. The position match degree refers to the degree of suitability of the position, title, or position of the personnel assigned to a task in that task. Profit margin refers to the ratio of the profit generated by personnel assigned to a task to the cost of recruiting and retaining those personnel. The partner company reliability refers to the reliability that the client 1 has of the partner company 2b, which is the recruitment agent 2. A developmental position is a position for recruiting personnel who are unable to utilize their abilities in the position of their choice, with an emphasis on internal factors (suitability to the company culture, serious motivation, honesty, etc.). In assignment results 21, the human resource task allocation system 10 presents the results of assignment matching 20 to the recruitment introducer 2. In assignment proposal 22 , the human resource task allocation system 10 presents the results of assignment matching 20 to the requester 1 . In assignment candidate presentation 23 , the human resource task allocation system 10 presents the results of assignment matching 20 to the aggregator 3 . The aggregator 3 determines the assignment of actual tasks and personnel from the presented assignment candidate information.

[0028] (Business information) Documents related to tasks to be assigned include, for example, an overview of the tasks submitted by the task manager and requirements for personnel for the tasks. These documents describe the number of personnel required and the capabilities required of the personnel (skills, skill levels, past experience, etc.). These documents are written in natural language and are in an undefined format, making them unsuitable for computer-readable data conversion. Therefore, the human resource task assignment system 10 uses a large-scale language model to read and acquire the content of task information written in natural language in undefined formatted documents. The advantage of using a large-scale language model is that, unlike when trying to achieve a goal using a programming language, which would require complex code, it is possible to achieve the goal by issuing instructions to the large-scale language model using prompts written in natural language.Furthermore, a large-scale language model can handle unstructured information, so even if information of various formats is input into a large-scale language model, its content can be retrieved.

[0029] Next, the business information list 30 will be described with reference to FIG. FIG. 2 is a diagram showing an example of the business information list 30 according to this embodiment, and FIG. 2 is an excerpt showing a part of the business information list 30. As described above, the task information is provided by the client 1 to the human resource task allocation system 10 . The following describes a case where the client 1 provides the task information list 30 as task information to the human resource task allocation system 10. The business information list 30 displays a list of multiple business information items.

[0030] The business information list 30 includes summary information 31, detailed information 32, and personnel information 35. Summary information 31 is created for each of the multiple business information included in the business information list 30, serves as a heading for each business information, and includes a column for item number 31a, a column for status 31b, a column for customer / end user 31c, and a column for name 31d. The item number column 31a lists serial numbers assigned to each of the multiple pieces of business information included in the business information list 30. The item number is a unique number assigned to identify each piece of business information. The status column 31b describes the current status of each of the multiple pieces of business information included in the business information list 30. Types of status include, but are not limited to, proposed, preliminary, closed, time change, early start, withdrawn, and lost order. Preliminary means that part or an outline of the business information has been communicated in advance. In the customer / end user column 31c, the customer or client of each of the multiple business items included in the business information list 30 is entered. In the name column 31d, the name of each of the multiple pieces of business information included in the business information list 30 is written.

[0031] Detailed information 32 is created for each of the multiple pieces of business information included in the business information list 30, and refers to the details of each piece of business information, and includes a contract column 32a, a proposed amount column 32b, a work location column 32c, a start date column 32d, a period column 32e, a status column 33, and a technical information column 34. The contract column 32a describes the contract type of the personnel who will be hired as a staff member for each task of the multiple task information included in the task information list 30. The types of contract include, but are not limited to, a contract for commissioning, a dispatch contract, and a labor-hour contract. The proposed amount column 32b lists the budget allocated to each task of the multiple pieces of task information included in the task information list 30. The work location column 32c lists the work location of the personnel for each task of the multiple pieces of task information included in the task information list 30. The start date column 32d lists the start date for each task of the multiple pieces of task information included in the task information list 30. In the period column 32e, a period is entered for each task of the multiple pieces of task information included in the task information list 30. The status column 33 describes the status of each task of the multiple pieces of task information included in the task information list 30 . The technical information column 34 lists the business skills required of personnel for each task of the multiple pieces of business information included in the business information list 30. As described above, business skills include, but are not limited to, required skills and preferred skills. For example, preferred skills that are less important than required skills and preferred skills but are appreciated if personnel possess them may also be included in the business skills.

[0032] Personnel information 35 is summary information on the personnel for a task, and includes columns such as column 35a indicating whether personnel for the task need to be assigned, column 35b indicating the number of personnel at PL (Project Leader) level and their source of supply, column 35c indicating the number of personnel at AT (Advanced Technology) / senior programmer level and their source of supply, column 35d indicating the number of personnel at SE (System Engineer) / mid-level programmer level and their source of supply, and column 35e indicating the number of personnel at PG (Programmer) / junior programmer level and their source of supply. The column 35a for whether or not to assign (allocate) describes whether or not to assign personnel for each task of multiple task information included in the task information list 30. Types of whether or not to assign (allocate) include necessary, supervision request, not required, satisfied, and under consideration. The column 35b for the number of PL (Project Leader) level personnel and their supply sources lists the number of PL level personnel required and their supply sources for each task of multiple pieces of task information included in the task information list 30. The supply source refers to the origin and attributes of the personnel. Types of supply sources include BP (business partners), departments within the same company, and sole proprietors / freelancers. Column 35c for the number of AT (Advanced Technology) / senior programmer personnel and their supply sources lists the number of AT-level personnel required and their supply sources for each of the multiple business information tasks included in the business information list 30. Column 35d for the number of SE (System Engineer) / mid-level programmer personnel and their supply sources lists the number of SE-level personnel required and their supply sources for each of the multiple business information tasks included in the business information list 30. Column 35e of the number of PG (Programmer) / junior programmer personnel and their supply sources lists the number of PG-level personnel required and their supply sources for each of the multiple business information tasks included in the business information list 30.

[0033] Of the multiple pieces of business information included in the business information list 30, the business with item number 1652 will be used as an example for the following explanation. Business information for No. 1652 Status: Proposed Client / end user: Aaaa Corporation Name: EC system development Contract: Labor-hour contract Proposed amount: Around 1 million yen Work location:Telework Start date: Same day Period: Medium to long term Status: Inquired by client in December 2021, recruitment is ongoing Technical information: Required Skills ·Web API interface design experience (2+ years) ·Experience in logical and physical design of RDB (2+ years) ·Experience in requirements definition (2+ years) Preferred Skills · Knowledge of order and shipping processes -Ability to work while maintaining smooth communication with team members -Ability to research and resolve unclear points on one's own Need for assignment: Required Personnel information: One SE / intermediate, supplied by BP (business partner)

[0034] (Hardware configuration of the human resource task assignment system 10) The hardware configuration of the human resource task assignment system 10 will be described with reference to FIG. FIG. 3 is a block diagram illustrating an example of the hardware configuration of the human resource task assignment system 10. As shown in FIG. The human resource task allocation system 10 includes a communication unit 10a, a ROM 10b, a RAM 10c, a storage unit 10d, a calculation unit 10e, and an input / output interface 10f. Furthermore, the human resource task allocation system 10 includes external devices such as an input device 10g and an output device 10h that input and output data via an input / output interface 10f.

[0035] The communication unit 10a has a function of performing two-way communication with other information processing devices. The other information processing devices may be the information processing devices used by the client 1, the information processing devices used by the recruitment agent 2, other recruitment task allocation systems 10, and other communication devices. When the communication unit 10a performs two-way communication with other information processing devices, the communication unit 10a may perform the two-way communication via the information communication network 25, or may perform the two-way communication by directly connecting to the other information processing device. The communication unit 10a may use wired communication or wireless communication for communication with other information processing devices.

[0036] The ROM 10b can be used as a recording device, and stores a BIOS (Basic Input Output System) required for controlling the operation of each functional unit of the human resource task allocation system 10, various data used by the BIOS, and the like. BIOS is a program that manages the basic input / output functions of the human resource task assignment system 10. It is the first program to run when the power to the human resource task assignment system 10 is turned on, controlling hardware such as the communication unit 10a, ROM 10b, RAM 10c, memory unit 10d, calculation unit 10e, and input / output interface 10f, and preparing to start up the OS (Operating System).

[0037] The RAM 10c is used to configure the main memory accessed by the calculation unit 10e, and is also used to temporarily store various data acquired or generated by the human resource task assignment system 10 before storing it in the storage unit 10d.

[0038] The storage unit 10d is realized by an HDD (Hard Disk Drive), an SSD (Solid State Drive), online storage, etc., and stores the OS, the information processing program described below, other application software, various data used by these programs, etc. The storage unit 10d also stores various data acquired or generated by the human resource task assignment system 10. The storage unit 10d may be a database, and it is particularly desirable to use a key-value database. A key-value database stores information as a set of keys and values. The key functions as a unique identifier.

[0039] The calculation unit 10e includes a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphic Processing Unit), etc., and is realized by a logic circuit or a dedicated circuit formed by an integrated circuit (IC (Integrated circuit) chip, LSI (Large-Scale Integration)), etc. The input / output interface 10f is an interface for transmitting and receiving data to and from external devices such as the input device 10g and the output device 10h. The input / output interface 10f may use different standards depending on the data being handled, and may be compatible with multiple standards, such as HDMI (registered trademark), USB 2.0, USB 3.0, RS-232C, IEEE 1394, SCSI, and SASI.

[0040] The input device 10g includes a keyboard, a mouse, and the like, and receives input for operating the human resource task assignment system 10 by a user. The output device 10h includes a monitor, a printer, and the like, and displays data generated by the human resource task assignment system 10 to a user or the like.

[0041] (Functional configuration of the human resource task assignment system 10) Next, an example of the functional configuration of the human resource task assignment system 10 will be described with reference to FIG. FIG. 4 is a block diagram for explaining an example of the functional configuration of the human resource task assignment system 10. As shown in FIG. The human resource task assignment system 10 loads a human resource task assignment program (described later) stored in the storage unit 10d into a main memory configured with a RAM 10c, etc. The calculation unit 10e accesses the main memory into which the human resource task assignment program has been loaded and executes the human resource task assignment program. By executing the human resources task assignment program, the human resources task assignment system 10 has functional units such as a task skill information acquisition unit 40, a possessed skill information acquisition unit 41, a table information acquisition unit 42, a candidate selection unit 43, a level information acquisition unit 44, a weight information acquisition unit 45, a total score calculation unit 46, a ranking unit 47, and an output unit 48 in the calculation unit 10e.

[0042] (Regarding the business skill information acquisition unit 40) The business skill information acquisition unit 40 acquires business skill information relating to business skills required of personnel engaged in the business from non-routine business information relating to the business provided by a client 1 who desires the introduction of personnel engaged in the business. The job skill information acquiring unit 40 may acquire the job skill information from the job information using a large-scale language model (LLM) and store the information in the storage unit 10d as information in key-value format. The business skill information acquisition unit 40 may take charge of the assignment request reception 16 shown in FIG.

[0043] As described above, the business information is an unstructured document, and is written in the email sent from the client 1, in an attachment to the email, or in both the email and the attachment to the email. An attached file is a file created by various applications, a document written in a markup language such as HTML, an electronic document file such as PDF, etc. that is attached to an email and sent by the requester 1 along with the email. The business skill information acquisition unit 40 analyzes the email using a large-scale language model, and acquires business skill information from the business information contained in the email (the email body and attached files). Specifically, the business skill information acquisition unit 40 uses a large-scale language model to analyze the contents of both the body of the email and the attachment, and acquires business skill information for each business regarding the business skills required of personnel engaged in the business from the business information contained in the body of the email and the attachment.

[0044] In this embodiment, the business skill information is described in the technical information column 34 of the business information list 30, which is an attached file, separated by business (see FIG. 2). Therefore, the business skill information acquisition unit 40 uses a large-scale language model to analyze the technical information column 34 of the business information list 30 and acquires business skill information for each business. In addition, if no file is attached to the email, the business skill information acquisition unit 40 uses a large-scale language model to analyze only the content of the email body and acquire business skill information regarding the business skills required of personnel engaged in the work. The business skill information acquisition unit 40 may acquire information other than business skill information from the business information, such as business summary information 31, detailed information 32, and personnel information 35 (see Figure 2), and store the contents in the memory unit 10d as key-value format data organized into pairs of keys and values.In this case, the business skill information acquisition unit 40 may analyze the business information using a large-scale language model. In addition, among the business information, detailed information excluding business overview information and technical information may be included in the email body, and technical information and personnel information may be included in the attachment, and the business skill information acquisition unit 40 may integrate the information acquired from the email body and the information acquired from the attachment to acquire business skill information and information other than business skill information.

[0045] (Regarding the possessed skill information acquisition unit 41) The possessed skill information acquisition unit 41 acquires possessed skill information on skills possessed by each human resource from non-standard human resource information on the human resources introduced by the recruitment introducer 2. The possessed skill information acquiring unit 41 may acquire possessed skill information from human resource information using a large-scale language model, and store the acquired skill information in the storage unit 10d as information in key-value format. The possessed skill information acquisition unit 41 may be responsible for the personnel proposal reception 15 shown in FIG.

[0046] As described above, the personnel information is an unstructured document, and is written in an email sent from the personnel introducer 2, in an email attachment, or in both the email and the email attachment. The attached file refers to a resume and curriculum vitae, such as a file created by various applications, a document written in a markup language such as HTML, or an electronic document file such as PDF (registered trademark), which is attached to an email and sent together with the email from the recruitment agent 2. In this embodiment, the curriculum vitae may also be referred to as engineer information (see FIG. 6) and engineer curriculum vitae (see FIG. 7). The possessed skill information acquisition unit 41 analyzes the email using a large-scale language model, and acquires possessed skill information from the human resource information described in the email (the email body and attached files). Specifically, the possessed skill information acquisition unit 41 uses a large-scale language model to analyze the contents of both the body of the email and the attached file (such as a resume and a curriculum vitae), and acquires possessed skill information for each employee regarding the skills possessed by the employee from the personnel information contained in the body of the email and the attached file (such as a resume and a curriculum vitae).

[0047] Next, a first reference example, a second reference example, and a third reference example of prompts to be given to the large-scale language model when acquiring possessed skill information will be described below. A large-scale language model is a natural language processing model built using massive amounts of text data and advanced deep learning technology. A prompt is a sentence or word used to give instructions, questions, etc. to a large-scale language model. The "first example prompt" may be used when a portion of a candidate's resume is provided to the human resource job assignment system 10. The "second example prompt" may be used when the talent assignment system 10 is provided with a skills summary and work history excerpted from the candidate's resume. The “third reference example of the prompt” may be used when a part of the email body is provided to the human resource task assignment system 10.

[0048] "First example of a prompt" You are an AI assistant in talent acquisition, tasked with screening resumes. You are provided with a portion of a candidate's resume. Your job is to extract the following details from this information: To extract the details covered: Name, Gender, Age, Nationality, Place of Residence, Qualifications, Skills, Work Experience, and Japanese Language Level, please follow the rules below: · Make sure to extract summary information into the Skill Summary field. Extract only the work history details that cover a specific period into the work history details field. If there is no information associated with a field whose content type is a string, return an empty string for that field. If there is no information associated with a field whose content type is numeric, return 0 for that field.

[0049] "Second example of prompt" You are an AI talent acquisition assistant tasked with screening resumes. Below you are given the skills summary and work history excerpts from a candidate's resume. Your job is to extract information about specific target skills from the given candidate information. To extract information, follow these rules: -Do not extract any information about skills other than the target skill. · Extract only clearly stated facts. Do not generate statements that include information not explicitly stated. If you do not have any applicable information, please answer "None."

[0050] "Third example of prompt" You are an AI assistant for talent acquisition, helping with resume screening. You are given a portion of an email body containing a list of candidates. Your task is to extract the start and end lines of each candidate's personal information. To do this, follow these rules: Extract the first and last lines of the personnel personal information paragraph verbatim.

[0051] Next, the operation of the large-scale language model used by the business skill information acquisition unit 40 and the possessed skill information acquisition unit 41 will be described below. The human resource task allocation system 10 takes as input the text of emails and the contents of email attachments sent by the client 1 and the recruitment agent 2, and can accurately analyze the task information or human resource information using a large-scale language model to generate data in a specified key-value format. The business skill information acquisition unit 40 can analyze and acquire business skill information and other information with high accuracy from the body of the email sent from the client 1 and the contents of the email attachments by using a large-scale language model. The possessed skill information acquisition unit 41 can analyze and acquire possessed skill information and other information with high accuracy by using a large-scale language model from the body of the email sent from the recruitment agent 2 and the contents of the email attachment.

[0052] Furthermore, the business skill information acquisition unit 40 and the possessed skill information acquisition unit 41 can use a large-scale language model to combine the contents of the body of the e-mail sent from the client 1 and the recruitment agent 2 with the contents of the e-mail attachments, and can handle cases where a single e-mail contains multiple pieces of business information and personnel information, and can also handle cases where an e-mail has multiple attachments. The data input to the large-scale language model used by the business skill information acquisition unit 40 and the possessed skill information acquisition unit 41 is characterized by the body of an unstructured email and an attachment file to the unstructured email with an indefinite format. The data output from the business skill information acquisition unit 40, which uses a large-scale language model, includes information on the skills required of personnel (essential skills, preferred skills, etc.), the number of personnel required, sources of human resources, and summary and detailed information about the business. The data output from the skill information acquisition unit 41, which uses a large-scale language model, includes personal information of the person being introduced (name, gender, date of birth, nationality, place of residence, qualifications, language level, unit price, etc.), a summary of information regarding the skills possessed by the person, and detailed information about the person's work history.

[0053] Next, the operation and function of the possessed skill information acquisition unit 41 will be explained in order below. (1) Using a large-scale language model, divide the content of the email sent by Recruiter 2 into chunks Cn, where each chunk contains information about only one candidate, where n is a positive integer and is the number of chunks. (2) Using a large-scale language model, personal information Pk and skill and experience information Sk for each candidate are extracted from chunk Cn. Skill and experience information refers to information about the skills possessed by the candidate and information about their work history. k is a positive integer and is the candidate's identification number. (3) If chunk Cn contains a URL for a resume or work history (called a soft attachment), the URL is accessed, the resume or work history file is downloaded, and personal information Qk and skill and experience information Tk are extracted using a large-scale language model. k is a positive integer that is the candidate's identification number. (4) When an attachment such as a resume or work history is directly attached to an email (this is called a hard attachment), the file is downloaded, and the name of the file is compared with the name of the person being introduced using a large-scale language model to determine which person's resume or work history the file belongs to, and personal information Qk and skill and experience information Tk are extracted. In addition, if the downloaded file contains table information represented in table format, the possessed skill information acquisition unit 41 extracts personal information Qk and skill experience information Tk from the table information using the table information acquisition unit 42 described below. (5) For each field of personal information and skill experience information (such as name and gender), the personal information Pk and skill experience information Sk extracted from the email body are integrated with the personal information Qk and skill experience information Tk extracted from the attachment. During this integration, if there is overlapping information among the personal information Pk, personal information Qk, skill experience information Sk, and skill experience information Tk, the information is extracted as a single piece of information. Overlapping information refers to information in the same field (for example, a key in a key-value format). In addition, if there is overlapping field information between the email body and the email attachment, only the consistent information will be retained. For example, in a key-value format, information with the same key but different values ​​will not be retained as it is considered unreliable information. Also, if only age information is available, the year of birth is calculated, and the place of residence information is obtained as precisely as possible (e.g., Shinjuku-ku, Tokyo, rather than Tokyo). (6) The possessed skill information acquisition unit 41 outputs the result in a predetermined format (key value, CSV, etc.).

[0054] (Regarding the table information acquisition unit 42) The table information acquisition unit 42 acquires the contents of table information expressed in a table format using a large-scale language model and an image recognition model. When the possessed skill information includes table information, the possessed skill information acquiring unit 41 acquires the contents of the table information using the table information acquiring unit . An image recognition model is a machine-learned learning model or manual algorithm for image recognition that extracts features, such as specific objects and their positions, from given image data and performs tasks such as identification, prediction, or analysis based on these features. A manual algorithm for image recognition is an algorithm that identifies, predicts, or analyzes given image data based on features such as specific objects and their positions in the image, in accordance with rules defined by humans. The image recognition model used in this embodiment recognizes the specifications of a table in table information given as image data. The table specifications include the number of cells, rows, and columns the table contains, and the orientation of the characters contained in the cells (vertical or horizontal writing).

[0055] Next, the operations and functions of the table information acquisition unit 42 and the possessed skill information acquisition unit 41 will be explained below in order. By using the table information acquisition unit 42, the possessed skill information acquisition unit 41 can flexibly read the contents of table information expressed in a table format included in a resume, curriculum vitae, or the like. The goal of using the table information acquisition unit 42 is to appropriately read personal information and possessed skill information from data presented in a table format, which is often found in resumes and curriculum vitae. The advantage of using the table information acquisition unit 42 is that it can extract the contents of table information expressed in table format in resumes and curriculum vitae, and can extract the contents of table information even if symbols are used.

[0056] The input to the table information acquisition unit 42 is a resume or curriculum vitae attached to an email. These are document data containing non-standard table information. The data format is an XLSX file, an XLS file, a DOCX file, a DOC file, a PDF (registered trademark) file, or the like. The output from the table information acquisition unit 42 is information about the personnel read from the table information contained in the attached file, and includes personal information, a skill summary, and details of work history. The personal information includes, for example, name, gender, date of birth, nationality, place of residence, qualifications, language level, and unit price. A skill summary may be, for example, a computer language, firmware, database, operating system, etc. that can be used.

[0057] The possessed skill information acquisition unit 41 and the table information acquisition unit 42 function in the following procedure. (1) The possessed skill information acquisition unit 41 reads the attached file of the e-mail and divides the contents into chunks. If the attached file is an XLSX file or an XLS file, the possessed skill information acquisition unit 41 recognizes all tables and uses the table information acquisition unit 42 to extract the contents of each table as a chunk. If the attached file is a DOCX file or a DOC file, the possessed skill information acquisition unit 41 acquires paragraphs (text) and tables as chunks, and extracts the contents of the paragraph (text) chunks using the possessed skill information acquisition unit 41, and extracts the contents of the table chunks using the table information acquisition unit 42. If the attached file is a PDF file, the possessed skill information acquisition unit 41 recognizes the layout of the PDF file using a PDF recognition AI (for example, Azure (registered trademark) Document Intelligence) and divides the paragraphs (text) or tables on each page into chunks, extracts the contents of the paragraph (text) chunks using the possessed skill information acquisition unit 41, and extracts the contents of the table chunks using the table information acquisition unit 42. (2) If the chunk is a paragraph (text), the possessed skill information acquiring unit 41 extracts the content from the paragraph (text) using a large-scale language model. (3) If the chunk is a table, the possessed skill information acquisition unit 41 uses the table information acquisition unit 42 to organize the table. Specifically, it merges cells within the same frame, deletes unnecessary columns and rows, and converts the vertical and horizontal orientation depending on the situation before reading. Specifically, if the text in the cell is written vertically, it is read vertically from top to bottom, and if the text in the cell is written horizontally, it is read horizontally from left to right. Furthermore, it recognizes the table type (personal information / skill summary / work history) using a large-scale language model and takes the following measures according to each type. (3.1) In the case of a skills summary table with a checklist, a large-scale language model is used to identify checkmarks and extract only checked items. In the case of a skills summary table with no checklist, a large-scale language model is used to extract skill information and qualification information. (3.2) In the case of the work history table, information is extracted for each row using a large-scale language model. (3.3) For tables that are not skill summaries or work histories, a large-scale language model is used to extract information from the entire table.

[0058] 5 to 7, the e-mail body 38 and its attached files sent from the recruitment introducer 2 will be described, and an example of the operation of the possessed skill information acquisition unit 41 and the table information acquisition unit 42 will be described. FIG. 5 is a diagram showing an example of the email body 38 sent by the recruitment agent 2, FIG. 6 is a diagram showing an excerpt of engineer information 71, which is a first example of a resume of a person introduced by the recruitment agent 2, and FIG. 7 is a diagram showing an excerpt of engineer resume 75, which is a second example of a resume of a person introduced by the recruitment agent 2. In this embodiment, the recruitment agent 2 sends personnel information to the personnel task assignment system 10 in the email body (see Figure 5) and its attached files, which are engineer information 71 (see Figure 6) and engineer resume 75 (see Figure 7), but this is not limited to this, and the personnel information may be sent to the personnel task assignment system 10 only in the email body, or only in the email attachment.

[0059] The email body 38, engineer information 71, and engineer resume 75 provided by the recruitment agent 2 to the recruitment task allocation system 10 are divided into chunks by the possessed skill information acquisition unit 41. There are two types of chunks: paragraph (text) chunks and table chunks. A paragraph (text) chunk is a paragraph that consists of text only. A table chunk is a paragraph consisting of only a table. The email body 38 and its attachments are divided into chunks so that a single chunk does not contain a mixture of paragraphs (text) and tables, although a table containing paragraphs (text) may be divided into a single chunk. The possessed skill information acquisition unit 41 extracts the contents of paragraph (text) chunks using a large-scale language model, and extracts the contents of table chunks using a table information acquisition unit 42. The table information acquisition unit 42 extracts the contents of table chunks using a large-scale language model.

[0060] As shown in FIG. 5, email body 38 includes personnel summary information 38a called Mr. KM's engineer summary, personnel summary information 38b called Mr. AR's engineer summary, personnel detail information 38c called Mr. KM's details, and personnel detail information 38e called Mr. AR's details. The talent summary information 38a, 38b refers to information that summarizes the personal information and skill information of the talent to be introduced, and specifically includes name, age, gender, length of employment in the industry, core skills, etc. Note that core skills refer to the main skills possessed by the talent. The detailed personnel information 38c, 38e refers to detailed information about the personal information and skill information of the personnel, and specifically includes the nearest station, operation date (operation start date), unit price, estimated range, detailed skill information 38d, 38f, and an explanatory text about the personnel.

[0061] The possessed skill information acquisition unit 41 divides the email body 38 into chunks containing only information about one talent. Specifically, the possessed skill information acquisition unit 41 divides the email body 38 into a first chunk C1 containing talent summary information 38a for Mr. KM, a second chunk C2 containing talent summary information 38b for Mr. AR, a third chunk C3 containing talent detailed information 38c for Mr. KM, and a fourth chunk C4 containing talent detailed information 38e for Mr. AR. The possessed skill information acquisition unit 41 analyzes the first chunk C1 and the third chunk C3 to extract personal information P1 and skill experience information S1 of Mr. KM (k=1), and analyzes the second chunk C2 and the fourth chunk C4 to extract personal information P2 and skill experience information S2 of Mr. AR (k=2). Note that P1, S1, P2, and S2 are expressed in key-value format and are output in the format "key:value".

[0062] P1's details are: "Name: KM", "Age: 34", "Gender: Female", "Years in industry: 11 years and 8 months", "Nearest station: Takao Station", "Starting date: From December 1st", "Unit price: 830,000 yen", and "Settlement range: 140-180 hours". S1 is "Computer languages ​​used: Java (registered trademark), JavaScript (registered trademark), AWS (registered trademark) Lambda, PHP, C#", "Firmware: Spring (TERASOLUNA), SpringBoot, React (registered trademark), jQuery / Ajax", "Database: Oracle (registered trademark), PostgreSQL (registered trademark), MySQL (registered trademark)", "OS: Linux (registered trademark), Windows (registered trademark)", "Other: Eclipse (registered trademark), Git (registered trademark), Subversion, HeartCore (registered trademark), AWS (Lambda / S3 / SES, etc.)"

[0063] P2's details are: "Name: AR", "Age: 24", "Gender: Male", "Years in the industry: 1 year and 8 months", "Nearest station: Nakata Station", "Starting date: From December 1st", "Unit price: 620,000 yen", and "Settlement range: 140-180 hours". S2's computer languages ​​used are Java, PHP, React, SQL, firmware is Spring (TERASOLUNA), Laravel, React, jQuery, database is PostgreSQL, MySQL, operating system is Windows, Linux (CentOS), other is Eclipse, Git, Subversion, JUnit (registered trademark), and qualification is Oracle Java SE11 Silver.

[0064] 6 shows engineer information 71 in table format, which is an excerpt of Mr. KM's engineer information. The engineer information 71 functions as Mr. KM's work history. The engineer information 71 includes two chunks: a first chunk 72 and a second chunk 73. A chunk refers to a group of data. The possessed skill information acquisition unit 41 divides the engineer information 71 into a first chunk 72 and a second chunk 73 . The first chunk 72 is a table in which personal information is recorded, and the second chunk 73 is a table in which work history is recorded. The table information acquisition unit 42 analyzes the first chunk 72 and the second chunk 73 to extract personal information Q1 and skill experience information T1 of Mr. KM (k=1). The personal information Q1 and skill experience information T1 are stored in the memory unit 10d as key-value format data. Hereinafter, the personal information Q1 and skill experience information T2 will be described in the format of "key:value".

[0065] The first chunk 72 comprises a first box 72a, a second box 72b, and a third box 72c, with the name written in the first box 72a, the age in the second box 72b, and the highest level of education written in the third box. Personal information Q1 is "Name: KM", "Age: 34 years old", "Highest level of education: Department of Modern Law, Faculty of Modern Law, DDDD University", and "Year of graduation: 2012".

[0066] The second chunk 73 includes a first frame 73a, a second frame 73b, and a third frame 73c. The first box 73a of the second chunk 73 suggests the structure of the second box 73b and the third box 73c that follow below. That is, the contents of the second box 73b and the third box 73c correspond to the contents (period, system name, model / OS, language, and details) of the first box 73a at the corresponding positions.

[0067] Specifically, the contents of the second frame 73b of the second chunk 73 are as follows: Period: April 2012 to May 2012 System name: New employee training Model / OS: Windows / Android (registered trademark) Language: C, Java detail: ·Software Development Basics Course 4 / 1~4 / 3 Huffman coding implementation in C 4 / 4~4 / 20 Program design / coding / testing Android Programming Course 4 / 23~4 / 27 Learned the basics of Android programming by creating a simple application: a calculator app / application using the accelerometer. ·Creating a clock application with voice reading function from 5 / 1~ Specification creation / program design / coding / testing I basically researched and implemented the usage of the APIs I used (TextToSpeech, Preferences, etc.) myself.

[0068] Similarly, the contents of the third frame 73c of the second chunk 73 are as follows: Period: June 2012 to July 2012 System name: Android app development Model / OS: Android Language: Java detail: Recipe Collection (released on Google Play) An app for saving and managing cooking recipes -Add new features to the current version, investigate and fix bugs 〇ZIP! Camera (released on Google Play) Program-linked camera app for broadcasting stations -Incorporating SNS posting functions (mixi (registered trademark), Facebook (registered trademark), Twitter (registered trademark)) 〇The Little Pig's Revenge (released on au (registered trademark) Smart Pass) Game Apps - Bug investigation and correction 〇iMemorization (released on Google Play) Vocabulary app -Added advertising function to the current version - Creating a phone book exchange app using QR Code (registered trademark) Planning and design Skeleton Creation

[0069] Skill and experience information T1 includes "OS: Windows, Android", "Computer languages ​​used: C language, Java", "April 1, 2012 - April 3, 2012: Basic software development course April 1 - April 3", "April 4 - April 20: Implementation of Huffman coding using C language (program design / coding / testing)", "April 23 - April 27: Android programming course (learned the basics of Android programming through creating simple applications. Calculator application / application using accelerometer)", "May 1 - May 31: Creation of a clock application with a text-to-speech function (specifications creation / program design / coding / testing. I basically researched and implemented how to use the APIs (TextToSpeech, Preferences, etc.) myself)", "June 2012 - July 2012: Recipe collection (GooglePlay ( Registered trademark)) / An app for saving and managing cooking recipes (adding new features to the current version, investigating and fixing bugs)", "June 2012 - July 2012: ZIP! Camera (released on Google Play) / A camera app for broadcasting stations linked to programs (incorporating a function for posting to social media (mixi, Facebook, Twitter))", "June 2012 - July 2012: The Little Pig Strikes Back (released on au Smart Pass) / A game app (investigating and fixing bugs)", "June 2012 - July 2012: iMemorization (released on Google Play) / A vocabulary book app (adding an advertising function to the current version)", "June 2012 - July 2012: Creation / planning, design and skeleton creation of a phone book exchange app using QR codes".

[0070] The engineer resume 75 shown in Figure 7 is shown in table format and is an excerpt from Mr. AR's engineer resume. The engineer resume 75 functions as Mr. AR's work resume. The engineer resume 75 includes two chunks (a first chunk 76 and a second chunk 77). The possessed skill information acquisition unit 41 divides the engineer resume 75 into a first chunk 76 and a second chunk 77 . The first chunk 76 is a table in which personal information is recorded, and the second chunk 77 is a table in which work history is recorded. The first chunk 76 is a table in which personal information is recorded, and the second chunk 77 is a table in which work history is recorded. The table information acquisition unit 42 analyzes the first chunk 76 and the second chunk 77 to extract personal information Q2 and skill experience information T2 of Mr. AR (k=2). The personal information Q2 and skill experience information T2 are stored in the memory unit 10d as key-value format data. Hereinafter, the personal information Q2 and skill experience information T2 will be described in the format of "key:value".

[0071] The first chunk 76 contains the name, gender, date of birth, employment status, address, nearest station, qualifications, years of experience, and selling points. Personal information Q2 is "Name: AR", "Gender: Male", "Date of birth: May 1999", "Age: 24", "Employment type: Full-time employee", "Address: Yokohama, Kanagawa Prefecture", "Nearest station: Nakata Station (Yokohama Municipal Subway Blue Line)", "Qualifications: Oracle Java SE11 Silver", "Years of industry experience: 1 year and 8 months", "Appealing points: He was the most outstanding of the new graduates in 2010, having completed his training early and passing the Java qualification exam on the first try with a 90% correct answer rate. He was then assigned to the company's lab team and participated in contracted projects. The engineer who is the director of the lab team also gives his approval, saying that he has a high level of technical understanding, can converse logically, and is capable of functioning as a PG (programmer) even in the field. He reports, communicates, and consults politely, and is a smooth communicator."

[0072] The second chunk 77 includes a vertical reading frame 78 , a first frame (top frame) 79 , a second frame 80 , a third frame 81 , a fourth frame 82 , and a fifth frame 83 . The vertical reading frame 78 contains the technical background information written in the second frame 80, the third frame 81, the fourth frame 82, and the fifth frame 83. The table information acquisition unit 42 changes the reading direction from horizontal to vertical when reading the contents of the vertical reading frame 78. When the table information acquisition unit 42 reads the contents of a frame other than the vertical reading frame 78, i.e., a horizontal reading frame, the reading direction is horizontal. The first box (top box) 79 contains the contents of the columns that follow below it: second box 80, third box 81, fourth box 82, and fifth box 83. The first column 79a contains the item number of the technical background, the second column 79b contains the development content (industry), the third column 79c contains the role, the fourth column 79d contains the scope of work, the fifth column 79e contains the server OS, the sixth column 79f contains the development language, the seventh column 79g contains the tools and DB, and the eighth column 79h contains the period.

[0073] The second box 80 contains the content of technical experience (1), the third box 81 contains the content of technical experience (2), the fourth box contains the content of technical experience (3), and the fifth box contains the content of technical experience (4). Of the technical experience (1) to technical experience (4), skill experience information T2 relating to the content of technical experience (1) in the second box 80 is described below. Skill and experience information T2 includes "Technical background: (1)", "Development content (industry): TERASOLUNA (registered trademark) training / Team size: 4 people / TERASOLUNA tutorial / Matcha CRM function addition (developed a function to add "type" to cases and manage masters) / Detailed specification creation (screen specifications, ER diagram, business logic details) / Manufacturing (created functions to add, edit, and delete types) / Testing (unit test specifications and implementation)", "Development content (industry): Various additional technical training / SQL (learned everything from basic CRUD to aggregate functions, table joins, and subqueries) / Version control (learned basic usage of Subversion) / Unit testing (learned practical unit testing training using JUnit)", "Role: PG", "Scope of work: 3.4.5", "Server OS: Windows 10", "Development language: Java 8, SQL", "Tools and DB: TERASOLUNA 1 series, (Spring Framework), (MyBatis), Tomcat (registered trademark) 7, PostgreSQL 9.4.26, pgAdmin 3, JUnit, Eclipse, Subversion, "Period: 2022.04 to 2022.06 (3 months)".

[0074] The skill information of Mr. KM and Mr. AR is described in the engineer summaries (personnel summary information) 38a, 38b and skill detail information (38d, 38f) in the email body 38 (see FIG. 5). The possessed skill information acquisition unit 41 uses a large-scale language model to analyze the email body 38 and the attached files, engineer information 71 and engineer resume 75, to acquire the possessed skill information of Mr. KM and Mr. AR. In addition, if no file is attached to the email, the possessed skill information acquisition unit 41 uses a large-scale language model to analyze only the content of the body of the email and acquire possessed skill information regarding the skills possessed by the personnel to be introduced (Mr. KM and Mr. AR in this embodiment). The possessed skill information acquisition unit 41 may acquire information other than possessed skill information from the human resource information, for example, personal information of the human resource to be introduced, and store the content in the memory unit 10d as key-value format data organized into pairs of keys and values.In this case, the possessed skill information acquisition unit 41 may analyze the human resource information using a large-scale language model. In addition, among the personnel information, there are cases where personnel summary information is described in the email body and detailed skill information is described in an attachment, and the possessed skill information acquisition unit 41 integrates the information acquired from the email body and the information acquired from the attachment to acquire possessed skill information and information other than possessed skill information.

[0075] (Regarding Candidate Selection Section 43) The candidate selection unit 43 selects candidates for personnel to engage in the work from among the introduced personnel based on a comparison between the possessed skill information and the work skill information. The candidate selection unit 43 selects, as candidates, personnel whose possessed skill information has a similarity to the business skill information that is equal to or greater than a threshold value, by performing a semantic search based on the business skill information and the possessed skill information.

[0076] The work skill information is information relating to the work skills required of personnel engaged in the work, and is information acquired by the work skill information acquisition unit 40 from the work information. The possessed skill information is information about the skills possessed by each human resource, and is information acquired by the possessed skill information acquisition unit 41 from the human resource information. The candidate selection unit 43 converts the business skill information into vector business skill information, which is vector information, and converts the possessed skill information into vector possessed skill information, which is vector information, using a large-scale language model. The vector information referred to here is text information converted into a numerical representation, and is used to capture the semantic similarity of the text information. The candidate selection unit 43 performs a semantic search based on the similarity between the vector business skill information and the vector possessed skill information to select, as candidates, personnel whose possessed skill information has a similarity to the business skill information that is equal to or greater than a threshold value. The similarity between two vectors is evaluated using, for example, cosine similarity, which is calculated from the inner product of two vectors, and the closer the value is to 1, the more similar the vectors are considered to be. By using semantic search, the candidate selection unit 43 can understand the intent of the business skill information and possessed skill information, and select as candidates personnel whose possessed skill information has a similarity to the business skill information that is above a threshold, while taking into consideration the context, ambiguous expressions, and synonyms.

[0077] On the other hand, semantic search cannot fully understand the intent of job skill information and possessed skill information, and there may be issues with accuracy. Also, it is difficult for semantic search to fully consider context, synonyms, and different expressions, and this may also be an issue. Therefore, the candidate selection unit 43 may select as candidates personnel whose possessed skill information has a similarity to the business skill information that is greater than or equal to a threshold value by performing a hybrid search of full-text search and semantic search based on the business skill information and possessed skill information. Full-text search involves searching for keywords in the entire text of two documents and scoring the similarity based on how many common keywords they contain. This has the advantage of clearly indicating the similarity between two documents. Hybrid search selects candidates based on both the degree of keyword consistency between business skill information and possessed skill information and the similarity of vector information obtained through semantic search, i.e., the degree of keyword consistency and the similarity of meaning.By performing full-text search and semantic search simultaneously, it is possible to perform searches that combine the clarity of full-text search with the flexibility of vector-based semantic search, which takes into account the intent, context, synonyms, and different expressions. Furthermore, the candidate selection unit 43 may select, as candidates, personnel whose possessed skill information has a similarity to the business skill information that is equal to or greater than a threshold value, by performing a full-text search based on the business skill information and the possessed skill information.

[0078] The purpose of the candidate selection unit 43 is to extract candidates who meet the requirements from the personnel information provided by the recruitment agency 2 based on the non-standard business information provided by the client 1, and to create a list of candidate personnel for the business. The candidate selection unit 43 can handle cases where the business information and personnel information are non-standard documents and use ambiguous expressions, because it extracts the contents of the business information and personnel information using a large-scale language model. The inputs to the candidate selection unit 43 are non-routine business information and non-routine personnel information. The non-routine job information includes the job title, summary, required skills (must have), preferred skills (better to have), work location, and unit price. The non-standard personnel information includes a personnel skill summary, work history details, and personal information such as name, gender, date of birth, nationality, place of residence, qualifications, language level, and unit price.

[0079] The output from the candidate selection unit 43 is a list of candidates ranked by scores indicating the degree of match with the job. The candidate selection unit 43 operates as follows. (1) The business skill information of the essential skills Sri (i=1, 2,...) created from the business information using a large-scale language model is further converted into vector business skill information using a large-scale language model. (2) The personnel skill information created from the human resources information using a large-scale language model is further converted into vector skill information using a large-scale language model. (3) A list of candidates D = {M1, ···, Mk, ···} is obtained using a hybrid search (full-text search + semantic search using Vector business skill information and Vector possessed skill information).

[0080] (Regarding the level information acquisition unit 44) The level information acquisition unit 44 acquires, for each candidate, level information indicating the level of the candidate's work skill from the possessed skill information of the selected candidates. The level information acquisition unit uses a large-scale language model to acquire level information indicating the level of the candidate's work skills for each candidate from the candidate's possessed skill information, and stores the information in key-value format that combines the work skill information and level information for each candidate in the memory unit 10d. If the work skills include required skills, which are skills that are essential for personnel to have, and satisfactory skills, which are skills that are different from the required skills and that personnel are desired to have, the level information acquisition unit 44 acquires, for each candidate, required level information Trk,i (k, i = 1, 2, ...) indicating the level of the candidate's required skills and satisfactory level information Tbk,j (k, j = 1, 2, ...) indicating the level of the satisfactory skills from the skill information of the selected candidate. Note that k is the candidate's identification number, i is the required skill's identification number, j is the preferred skill's identification number, and k, i, and j are positive integers. Similarly, for k, i, and j described below, k is the candidate's identification number, i is the required skill's identification number, and j is the preferred skill's identification number, and k, i, and j are positive integers.

[0081] The operation of the level information acquisition unit 44 is as follows. (1) For each required skill Sri and preferred skill Sbj obtained from the job information, the corresponding required level information Trk,i (k, i = 1, 2, ...) and preferred level information Tbk,j (k, j = 1, 2, ...) of candidate Mk (k = 1, 2, ...) are extracted from the personnel information using a large-scale language model. (2) The standard level of the required skill Sri (i = 1, 2, ...) is set as Lri (i = 1, 2, ...), and the required skill relative score Prk,i (k, i = 1, 2, ...) for the required skill of candidate Mk (k = 1, 2, ...) is calculated using a large-scale language model based on the required level information Trk,i (k, i = 1, 2, ...) extracted from the human resources information. In addition, the standard level of the acceptable skill Sbj (j = 1, 2, ...) is set as Lbj (j = 1, 2, ...), and a large-scale language model is used based on the acceptable level information Tbk,j (k, j = 1, 2, ...) extracted from the human resources information to determine the acceptable relative score Pbk,j (k, j = 1, 2, ...) for the acceptable skill of candidate Mk (k = 1, 2, ...).

[0082] (Regarding the weight information acquisition unit 45) The weight information acquisition unit 45 acquires weight information indicating the importance of the work skill based on the work information. When the business skills include required skills, which are skills that are required for personnel, and preferred skills, which are skills that are different from the required skills and that personnel are desired to possess, the weight information acquisition unit 45 acquires required skill weight Wri information and preferred skill weight Wbj information, which indicate the importance of the required skills Sri and the preferred skills Sbj, respectively, based on the business information.

[0083] The weight information acquisition unit 45 operates as follows. (1) Using a large-scale language model based on task information, the system outputs required skill weights Wri for required skills Sri and recommended skill weights Wbj for recommended skills Sbj. These weights can be further adjusted by the user. (2) Divide candidates into two classes: those who meet the required skills and those who do not.

[0084] (Regarding the total score calculation unit 46) The total score calculation unit 46 calculates a total score Zk indicating the suitability of each candidate for the job based on the weight information and level information for each job skill. The total score calculation unit 46 calculates the total score Zk for each candidate using the following formula (1). The suitability of each candidate for a job refers to the degree to which the candidate is suited to the job, taking into account the job skills required of the personnel engaged in the job, and is expressed by the total score Zk, which is the calculation result of formula (1). When the work skills include required skills, which are skills that are required for personnel, and acceptable skills, which are skills that are different from the required skills and that personnel are desired to possess, the total score calculation unit 46 calculates a total score Zk for each candidate based on the required level information, acceptable level information, required skill weight information, and acceptable skill weight information. Desirable skills are skills that are not as important as essential skills, but are welcomed and will result in a better evaluation if the employee possesses them. They are also called desirable skills, non-essential skills, and optional skills.

[0085]

number

[0086] Note that Zk is the total score of candidate k. Wr is the overall weight of essential skills, which indicates the overall weight of essential skills. Wri is the essential skill weight indicating the weight of the essential skill Sri. Prk,i is the essential skill relative score indicating the relative score of the essential skill Sri of candidate k. Wb is the overall weight of the preferred skills, which indicates the overall weight of the preferred skills. Wbj is a preferred skill weight indicating the weight of the preferred skill Sbj. Pbk,j is the relative score of the skill Sbj of candidate k. Here, k is the candidate's identification number, i is the required skill's identification number, and j is the preferred skill's identification number.

[0087] The operation of the total score calculation unit is as follows. For each class, calculate the total score Zk of candidate Mk using the required skill weights Wri for required skills Sri and the preferred skill weights Wbj for preferred skills Sbj (Wr is the overall weight of required skills, and Wb is the overall weight of preferred skills. The default values ​​are Wr=2, Wb=1. These weights can be further adjusted by the user).

[0088] (Regarding ranking section 47) The ranking unit 47 assigns a recommendation ranking to the candidate for the job based on the total score Zk. The ranking unit 47 creates a list of candidates in descending order of total score Zk.

[0089] (Regarding the output unit 48) The output unit 48 outputs a candidate list in which the candidates are arranged in association with the recommendation ranking of the candidates.

[0090] An example of the output screen of the output unit 48 will be described with reference to FIG. FIG. 8 is a diagram for explaining an example of the output screen 59 of the human resource task assignment system 10 according to this embodiment. The output screen 59 displays the matching results, and includes a main screen 60 and a side menu 61. The side menu 61 displays multiple buttons for switching the content displayed on the main screen 60. The side menu 61 includes a job display button 61a, a job posting display button 61b, a staff display button 61c, a staff detail display button 61d, and a matching result display button 61e.

[0091] The case display button 61 a is a button for displaying on the main screen 60 a list of tasks for which the requester 1 has made an assignment request 12 . The job posting display button 61b is a button for displaying on the main screen 60 detailed information about the job for which the requester 1 has made an assignment request 12. The personnel display button 61c is a button for displaying on the main screen 60 a list of personnel for whom personnel proposals 11 have been made by the recruitment introducer 2. The personnel details display button 61 d is a button for displaying on the main screen 60 detailed information about personnel for whom a personnel proposal 11 has been made by a recruitment introducer 2 . The matching result display button 61e is a button for displaying the matching result on the main screen 60.

[0092] 8 shows a state in which a matching result display button 61e has been pressed on the side menu 61. The main screen 60 includes a candidate ID display field 62, a job posting ID display field 63, a required skill display field 64, a required skill overall weight display field 65, a further skill display field 66, a further skill overall weight display field 67, a result display field 68, a remarks display field 69, and an execution date and time display field 70. The candidate ID display field 62 is a field for displaying the candidate's ID. The job posting ID display field 63 is a field that displays the ID of the job posting.

[0093] The required skill display field 64 is a field that displays the required skill relative score Prk,i of the candidate, an example of which is shown below (see FIG. 8). Web API IF Design: Experienced: Relative Score: 0.8 RDB logical design, physical design: 2 years or more: Relative score: 0.9 Requirements definition experience: 2+ years: Relative score: 1

[0094] The required skill overall weight display field 65 is a field for displaying the required skill overall weight Wri, and as an example, 2 is displayed (see FIG. 8).

[0095] The preferred skill display column 66 is a column that displays the candidate's preferred skill relative score Pbk,j, an example of which is shown below (see FIG. 8). Knowledge of order / shipping process: Knowledgeable: Relative score: 0.8 Communication skills: Relative score: 0.9 Self-solving ability: Experienced: Relative score: 1

[0096] The satisfactory skill overall weight display field 67 is a field for displaying the satisfactory skill overall weight Wb, and as an example, 1 is displayed (see FIG. 8).

[0097] The result display field 68 is a field for displaying the matching results, an example of which is shown below (see FIG. 8). The relative score for "Required Skills" for candidate ID: "01HSJRBP8D60301VTEBDVS122S" is [0.8,0.9,1], and the relative score for "Preferred Skills" is "0.8,0.9,1", resulting in a matching score of "0.9". A score of 0.9 indicates that the candidate is highly likely to be suited to the job.

[0098] The remarks display column 69 is a column for indicating remarks, and is blank in the example shown in FIG. The execution date and time display field 70 is a field that displays the execution date and time when the matching result was performed, and as an example, 08:55:06 on Monday, March 25, 2024 is displayed (see FIG. 8).

[0099] (About the structure of the human resource assignment program) Next, with reference to FIG. 9, the structure of a human resource task assignment program 50, which will be described later, will be described. FIG. 9 is an example of a block diagram for explaining the structure of the human resource task assignment program 50. In FIG. The human resource assignment program 50 comprises a back end 51 and a front end 52 . The backend 51 is located behind the application and refers to the area of ​​the programs and databases of the human resource task allocation system 10 that is invisible to the user, while the frontend 52 refers to the part that the user directly interacts with, and specifically, is responsible for building interfaces such as web pages or applications.

[0100] (About Backend 51) The backend 51 includes an email analysis process 51a, an information analysis process 51b for a human resources information analysis service, a business information analysis process 51c, an information acquisition process 51d, a matching process 51e, and an information registration process 51f.

[0101] In the email analysis process 51a, the human resources task allocation system 10 analyzes the email from the client when Google (registered trademark) App Script (GAS) 55 receives a notification that it has received an email 54 from the client, who is the requester 1 or the recruitment introducer 2. GAS is a programming language for automating services provided by Google (registered trademark), and in this embodiment, Gmail (registered trademark) is used. The email analysis process 51 a may be performed by the task skill information acquisition unit 40 and the possessed skill information acquisition unit 41 of the human resource task assignment system 10 .

[0102] In the information analysis processing 51b of the human resources information analysis service, the human resources task allocation system 10 receives a notification of an update request sent from the human resources information analysis service when the information of the human resources information analysis service is updated 56, and performs processing to analyze the information of the human resources information analysis service. The human resources information analysis service is, for example, Talent Palette (registered trademark), and when the human resources information provided by the recruitment agent 2 is updated or new human resources information is received, a notification of an update request is sent to the human resources task allocation system 10. A human resources information analysis service is a service that centrally manages personal information and skills of personnel and uses them for evaluation, recruitment, training, etc. It centralizes information such as personal information, skills, and work experience of personnel and uses it as a human resources database. The information analysis process 51b of the human resource information analysis service may be performed by the possessed skill information acquisition unit 41 of the human resource task allocation system 10.

[0103] In the business information analysis process 51c, the human resource business assignment system 10 accesses the in-house business management sheets, etc. to periodically check whether the information in the in-house business management sheets, etc. has been updated 57, and if the information has been updated, acquires the updated business information, etc. The task information analysis process 51c may be performed by the task skill information acquisition unit 40 of the human resource task assignment system 10.

[0104] In the information acquisition process 51d, the human resource task allocation system 10 acquires information required for the matching process 51e. The information required for the matching process 51e refers to the required skill overall weight Wr, the required skill weight Wri, the required skill relative score Prk,i, the satisfactory skill overall weight Wb, the satisfactory skill weight Wbj, and the satisfactory skill relative score Pbk,j. The information acquisition process 51d may be performed by the total score calculation unit 46 of the human resource task assignment system 10.

[0105] In the matching process 51e, the human resource / task assignment system 10 matches candidates with jobs. The human resource task assignment system 10 calculates a total score Zk indicating the suitability of each candidate for the task by the total score calculation unit 46.

[0106] In the information registration process 51 f, the human resource task assignment system 10 registers the human resource information, task information, and assignment candidate information in the database 53 . The database 53 may be the storage unit 10d. The information registration process 51f may be performed by a business skill information acquisition unit 40 that acquires business information and business skill information, a possessed skill information acquisition unit 41 that acquires human resource information and possessed skill information, a candidate selection unit 43 that selects candidates for personnel engaged in the business, a level information acquisition unit 44 that acquires level information indicating the level of the candidates' business skills, and a total score calculation unit 46 that calculates a total score Zk for each candidate, etc.

[0107] (About Front End 52) The front end 52 includes a display processing unit 52a and an information operation processing unit 52b.

[0108] (Regarding display process 52a) In the display process 52a, the human resource task allocation system 10 uses the monitor of the output device 10h to present information about tasks, personnel, and matching results to the user. The display process 52a may be performed by the output unit 48 of the human resource task assignment system 10.

[0109] (Regarding information manipulation processing 52b) In the information manipulation process 52b, the human resource task assignment system 10 accepts information manipulation by the user. The information operation refers to operation information relating to the user's operation of the human resource task allocation program 50, and specifically refers to operations such as registering and updating information, and executing matching.

[0110] (Human resource task allocation method and human resource task allocation program 50) Referring to FIG. 10, a human resource task assignment program 50 according to an embodiment of the present invention will be described together with a human resource task assignment method. FIG. 10 is an example of a flowchart of the human resource task assignment program 50. The human resource task allocation method is executed by the computing unit 10e of the human resource task allocation system 10 based on a human resource task allocation program. The human resource task allocation program includes a task skill information acquisition step S40, a possessed skill information acquisition step S41, a table information acquisition step S42, a candidate selection step S43, a level information acquisition step S44, a weight information acquisition step S45, and a total score calculation step S46. The human resource task assignment program causes the calculation unit 10e of the human resource task assignment system 10 to realize a task skill information acquisition function, a possessed skill information acquisition function, a table information acquisition function, a candidate selection function, a level information acquisition function, a weight information acquisition function, and a total score calculation function. These functions are executed in the order shown in the flowchart of Figure 10, but the order can also be changed as appropriate. Note that each function overlaps with the description of the various functional units of the human resource task assignment system 10 described above, and therefore detailed description thereof will be omitted.

[0111] The business skill information acquisition function acquires business skill information regarding the business skills required of personnel engaged in the business from non-standard business information regarding the business provided by a requester who wishes to be introduced to personnel engaged in the business (step S40: business skill information acquisition step).

[0112] The possessed skill information acquisition function acquires possessed skill information about skills possessed by each human resource from non-standard human resource information about the human resources introduced by the recruitment introducer (step S41: possessed skill information acquisition step).

[0113] The table information acquisition function acquires the contents of table information expressed in a table format using a large-scale language model and an image recognition model (step S42: table information acquisition step).

[0114] The candidate selection function selects candidates for personnel to engage in the work from among the introduced personnel based on a comparison between the possessed skill information and the work skill information (step S43: candidate selection step).

[0115] The level information acquisition function acquires, for each selected candidate, level information indicating the level of the candidate's work skill from the possessed skill information of the candidate (step S44: level information acquisition step).

[0116] The weight information acquisition function acquires weight information indicating the importance of the job skill based on the job information (step S45: weight information acquisition step).

[0117] The total score calculation function calculates a total score Zk indicating the suitability of each candidate for the job based on the weight information and level information for each job skill (step S46: total score calculation step).

[0118] (Other embodiments of the human resource task assignment method and human resource task assignment program 50) Next, a human resource task assignment program 50 according to another embodiment of the present invention will be described together with a human resource task assignment method according to another embodiment with reference to Fig. 11. Fig. 11 is an example of a flowchart of the human resource task assignment program 50 according to another embodiment. The flowchart of the human resource task assignment program 50 according to another embodiment shown in FIG. 11 differs from the flowchart of the human resource task assignment program 50 shown in FIG. 10 in that a ranking step S47 and an output step S48 have been added. The human resource task assignment method according to the other embodiment is executed by the calculation unit 10e of the human resource task assignment system 10 based on the human resource task assignment program 50 according to the other embodiment shown in FIG. The human resource task assignment program 50 according to another embodiment shown in FIG. 11 includes a task skill information acquisition step S40, a possessed skill information acquisition step S41, a table information acquisition step S42, a candidate selection step S43, a level information acquisition step S44, a weight information acquisition step S45, a total score calculation step S46, a ranking step S47, and an output step S48.

[0119] 11, a human resource task assignment program 50 according to another embodiment performs functions such as a task skill information acquisition function, a possessed skill information acquisition function, a table information acquisition function, a candidate selection function, a level information acquisition function, a weight information acquisition function, a total score calculation function, a ranking function, and an output function in the calculation unit 10e of the human resource task assignment system 10. These functions are executed in the order shown in the flowchart of FIG. 11, but the order can also be changed as appropriate. The following describes the human resource task assignment method and human resource task assignment program 50 according to another embodiment shown in FIG. 11, focusing only on the differences from the human resource task assignment method and human resource task assignment program 50 shown in FIG. Furthermore, since each function overlaps with the description of the various functional units of the human resource task assignment system 10 described above, detailed description thereof will be omitted.

[0120] The ranking function assigns a recommendation ranking for the job to the candidate based on the total score Zk (step S47: ranking assignment step).

[0121] The output function outputs a candidate list in which the candidates are arranged in association with the recommendation order of the candidates (step S48: output step).

[0122] (Effects of the human resource task assignment system 10 according to the embodiment) According to the human resource task assignment system 10 according to the embodiment described above, even non-standard human resource information and task information can be analyzed to acquire task skill information and possessed skill information. Furthermore, according to the human resource task assignment system 10 according to the embodiment described above, it is possible to calculate a total score Zk indicating the suitability for a task for each candidate. Furthermore, according to the human resource task assignment system 10 of the above-described embodiment, in addition to the required skills to be assigned to personnel, it is possible to calculate a total score Zk indicating the suitability of each candidate for the task by taking into consideration preferred skills, which are skills that are different from the required skills but that are desirable for personnel to possess. Furthermore, according to the human resource task allocation system 10 of the above-described embodiment, the client 1 and the recruitment agent 2 can freely select the format of the task information and the human resource information, thereby reducing the hassle of creating the task information and the human resource information. Furthermore, the human resource task allocation system 10 according to the embodiment described above can reduce the burden on the aggregator 3 in matching tasks with personnel, and can improve the accuracy of the matching. Furthermore, according to the human resource task assignment system 10 of the above-described embodiment, it is possible to accurately analyze business information or human resource information using a large-scale language model based on the text of emails and email attachments sent from the client 1 and the recruitment agent 2, and generate data in a predetermined key-value format. Furthermore, according to the human resource task assignment system 10 of the above-described embodiment, even if the table information is presented in table format in a resume or curriculum vitae attached to an e-mail, the contents of that information can be extracted, and even if symbols are used in the table information, the contents of that information can be extracted. Furthermore, according to the human resource task assignment system 10 of the above-described embodiment, by using semantic search, it is possible to understand the intent of the business skill information and possessed skill information, and then select as candidates human resources whose possessed skill information has a similarity to the business skill information that is above a threshold, while taking into consideration the context, ambiguous expressions, and synonyms.

[0123] The present invention is not limited to the human resource task assignment system 10, the human resource task assignment method, and the human resource task assignment program according to the above-described embodiments, and various other modifications or applications are possible without departing from the spirit of the present invention as set forth in the claims. Also, although the term "information" is used in the above-described embodiments, the term "information" can be replaced with "data," and the term "data" can be replaced with "information." Additionally, the term "non-standard" can be replaced with "undefined," meaning that there is no set format, shape, or style. [Explanation of symbols]

[0124] 1. Requester 1a Project 1b End Users 2. Recruitment Agents 2a Department 2b Partner companies 2c Self-employed / Freelance 3. Aggregators 3a Assignment Office 10. Human Resources Task Allocation System 10a Communications Department 10b ROM (Read Only Memory) 10c RAM (Random Access Memory) 10d storage section 10e Calculation unit 10f Input / Output Interface 10g input device 10h output device 11 Personnel Proposal 12 Assignment Request 13 Assignment candidates 15. Receipt of personnel proposals 16 Assignment request reception 17 Assignment DB 20 Assignment Matching 21 Assignment Results 22 Assignment Proposal 23 Assignment candidate presentation 25 Information and Communications Networks 30 Business Information List 31 General Information 38 Email Body 40 Business Skills Information Acquisition Department 41 Skill Information Acquisition Department 42 Table information acquisition section 43 Candidate Selection Department 44 Level information acquisition section 45 Weight information acquisition unit 46 Total score calculation section 47 Ranking Section 48 Output section 50 Human Resources Task Allocation Program 51 Backend 52 Front End 52a Display processing 52b Operation acceptance process 53 databases 54 Emails from clients 55 Google App Script 56 Human Resources Information Analysis Service Information Updated 57 Update information on internal business management sheets, etc. 59 Output Screen 60 Main Screen 61 Side Menu 79 1st slot (top slot) 80 2nd slot

Claims

1. a business skill information acquisition unit that acquires business skill information related to business skills required for personnel engaged in the business from non-routine business information related to the business provided by a client who desires to introduce personnel engaged in the business; a skill information acquisition unit that acquires, for each talent, skill information relating to skills possessed by the talent from non-standard talent information relating to talent introduced by a talent introducer; a candidate selection unit that selects candidates for personnel to engage in the work from among the introduced human resources based on a comparison between the possessed skill information and the work skill information; a level information acquisition unit that acquires, for each of the selected candidates, level information indicating the level of the business skill of the candidate from the possessed skill information of the candidate; a weight information acquisition unit that acquires weight information indicating the importance of the business skill based on the business information; a total score calculation unit that calculates a total score indicating suitability for the job for each of the candidates based on the weight information for each of the job skills and the level information; A human resources task allocation system comprising:

2. 2. The human resource task allocation system according to claim 1, further comprising a ranking unit that ranks the candidates for the task based on the total score.

3. 3. The human resource task allocation system according to claim 2, further comprising an output unit that outputs a candidate list in which the candidates are arranged in association with the recommendation ranking of the candidates.

4. The human resource task allocation system described in claim 1, characterized in that the business skill information acquisition unit uses a large-scale language model to acquire the business skill information from the non-standard business information and store it in the memory unit as key-value format information.

5. The human resources task allocation system described in claim 1, characterized in that the possessed skill information acquisition unit uses a large-scale language model to acquire the possessed skill information from the non-standard human resources information and store it in the memory unit as key-value format information.

6. The human resource task allocation system of claim 1, characterized in that the candidate selection unit selects as candidates the human resources whose possessed skill information has a similarity to the business skill information that is greater than or equal to a threshold value through a semantic search based on the business skill information and the possessed skill information.

7. The human resources task allocation system described in claim 1, characterized in that the level information acquisition unit uses a large-scale language model to acquire level information indicating the level of the candidate's work skills from the candidate's possessed skill information for each candidate, and stores the information in a memory unit as key-value format information combining the work skill information and the level information for each candidate.

8. The business skills include essential skills that are essential skills required of the personnel, and desirable skills that are different from the essential skills and are desirable for the personnel to possess, the level information acquisition unit acquires, for each of the selected candidates, essential level information indicating the level of the essential skill of the candidate and satisfactory level information indicating the level of the satisfactory skill from the possessed skill information of the candidate; the weight information acquisition unit acquires essential skill weight information and preferred skill weight information indicating the importance of the essential skill and the preferred skill, respectively, based on the job information; 2. The human resources task allocation system of claim 1, wherein the total score calculation unit calculates a total score for each candidate based on the required level information, the acceptable level information, the required skill weight information, and the acceptable skill weight information.

9. a table information acquisition unit that acquires the contents of table information expressed in a table format using a large-scale language model and an image recognition model; Furthermore, The human resource task allocation system according to claim 1, characterized in that the possessed skill information acquisition unit acquires the contents of the table information using the table information acquisition unit when the possessed skill information includes the table information.

10. The computer a task skill information acquisition step of acquiring task skill information relating to task skills required for personnel engaged in the task from non-routine task information relating to the task provided by a client who desires to introduce personnel engaged in the task; a possessed skill information acquisition step of acquiring possessed skill information regarding skills possessed by each of the human resources from non-standard human resource information regarding the human resources introduced by the recruitment introducer; a candidate selection step of selecting candidates for personnel to engage in the work from among the introduced human resources based on a comparison between the possessed skill information and the work skill information; a level information acquisition step of acquiring, for each of the selected candidates, level information indicating the level of the business skill of the candidate from the possessed skill information of the candidate; a weight information acquisition step of acquiring weight information indicating the importance of the business skill based on the business information; a total score calculation step of calculating a total score indicating suitability for the job for each of the candidates based on the weight information for each of the job skills and the level information; A method for allocating human resources to tasks, comprising:

11. On the computer, a business skill information acquisition function that acquires business skill information related to business skills required for personnel engaged in the business from non-routine business information related to the business provided by a client who desires to introduce personnel engaged in the business; A skill information acquisition function for acquiring skill information relating to skills possessed by each of the personnel from non-standard personnel information relating to the personnel introduced by the personnel introducer; a candidate selection function for selecting candidates for personnel to engage in the work from among the introduced personnel based on a comparison between the possessed skill information and the work skill information; a level information acquisition function for acquiring, for each of the selected candidates, level information indicating the level of the business skills of the candidates from the possessed skill information of the candidates; a weight information acquisition function for acquiring weight information indicating the importance of the business skill based on the business information; a total score calculation function for calculating a total score indicating suitability for the job for each of the candidates based on the weight information for each of the job skills and the level information; A human resources task allocation program characterized by realizing the above.

Images