Information processing system, information processing method, and program

Abstract

To provide an information processing system, an information processing method, and a program capable of further accurately evaluating applicants.SOLUTION: An information processing system includes: an application document acquisition unit which acquires an applicant's application document for applying employment of a business operator in which at least a part is created by a natural sentence; an evaluation item accepting unit which accepts input of a plurality of evaluation items regarding the applicant; a score acquisition unit which acquires a score for each of the plurality of evaluation items by inputting the application document and the plurality of evaluation items to an AI chat bot which has been learnt to answer a question when a natural sentence is input; and a derivation unit which derives unofficial employment rate being a value obtained by evaluating the applicants on the basis of the score for each of the plurality of evaluation items.SELECTED DRAWING: Figure 1

Description

[Technical field]

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

[0002] In popular companies, the number of applicants tends to be very large compared to the number of people expected to be hired. When the number of applicants becomes very large, the workload of the hiring manager increases, and it may become difficult to make appropriate hiring decisions. For this reason, attempts have been made to automatically narrow down the applicants by computer processing job data, etc. For example, Patent Document 1 discloses an invention of a hiring support device that calculates a score on the quality of job data, outputs scores for two or more evaluation elements including elements related to the expertise of the personnel required by the hiring manager, and calculates these elements based on whether the job data contains terms that represent the job content of the personnel required by the hiring manager. By automatically narrowing down the applicants, it becomes possible to give priority to interviews with applicants who are more likely to be hired (e.g., acceptance rate) or to have the business entity scout them. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7242944 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional techniques sometimes fail to provide sufficient accuracy in evaluating applicants.

[0005] The present invention has been made in consideration of the above circumstances, and one of its objects is to provide an information processing system, an information processing method, and a program that can evaluate applicants with greater accuracy. [Means for solving the problem]

[0006] One aspect of the present invention is an information processing system comprising an application document acquisition unit that acquires application documents of applicants applying for employment at a business, at least a portion of which are written in natural language; an evaluation item receiving unit that receives input of multiple evaluation items related to the applicant; a score acquisition unit that acquires a score for each of the multiple evaluation items by inputting the application documents and the multiple evaluation items into an AI chatbot that has been trained to answer questions when natural language is input; and a derivation unit that derives an acceptance rate, which is a value used to evaluate the applicant, based on the scores for each of the multiple evaluation items. Effect of the Invention

[0007] According to one aspect of the present invention, it is possible to provide an information processing system, an information processing method, and a program that can evaluate applicants with greater accuracy. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a configuration diagram of an information processing system 1. [Diagram 2] 1 is an example of an interface screen for inputting evaluation items, which the evaluation item receiving unit 20 causes to be displayed on a display device. [Diagram 3] FIG. 2 is a diagram conceptually illustrating a portion of an evaluation model 66. [Figure 4] FIG. 4 is a diagram showing an example of the flow of processing executed by the information processing system 1 in a learning stage. [Diagram 5] FIG. 2 is a diagram showing an example of the flow of processing executed by the information processing system 1 in the inference stage. [Figure 6] FIG. 1 is a configuration diagram of an information processing system 1A. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, with reference to the drawings, an embodiment of an information processing system, an information processing method, and a program of the present invention will be described. The information processing system is realized by one or more processors. The information processing system is a system that derives an employment acceptance rate, which is a value that evaluates an applicant who applies for employment at a business, based on the application documents of the applicant, at least a part of which is written in natural language. Note that in the following description, no particular distinction is made between components in the learning stage and components in the inference stage, but they may be different software components.

[0010] FIG. 1 is a configuration diagram of an information processing system 1. The information processing system 1 includes, for example, an application document acquisition unit 10, an evaluation item reception unit 20, a score acquisition unit 30, a derivation unit 40, and a storage unit 60. The components other than the storage unit 60 are realized by, for example, a hardware processor such as a CPU executing a program (software). Some or all of these components may be realized by hardware (including circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit), or may be realized by cooperation between software and hardware. The program may be stored in advance in a storage device such as an HDD (Hard Disk Drive) or a flash memory (a storage device having a non-transient storage medium), or may be stored in a removable storage medium such as a DVD or a CD-ROM (a non-transient storage medium), and may be installed in the storage device by mounting the storage medium in a drive device.

[0011] The storage unit 60 is a RAM (Random Access Memory), a HDD, a flash memory, etc. The storage unit 60 stores data and models such as past employment data 62, application document data 64, and an evaluation model 66.

[0012] An input device 80 and an output device 90 may be connected to the information processing system 1. The input device 80 is, for example, a keyboard, a mouse, a touch panel, a touch pad, etc. The output device 90 is, for example, a display device, a speaker, etc. The input device 80 and the output device 90 may be computer devices such as a personal computer, a tablet terminal, a smartphone, etc., instead of computer peripheral devices.

[0013] The information processing system 1 communicates with the AI ​​chatbot server 100 via a network NW. Details of the AI ​​chatbot server 100 will be described later.

[0014] The application document acquisition unit 10 reads and acquires application documents of applicants applying for employment at a business, written in natural language, from application document data 64 stored in the memory unit 60. Note that the application documents may include information written in a format other than natural language, such as a Q&A format. The application document acquisition unit 10 acquires the application documents of each applicant in sequence from the memory unit 60, for example, in the inference stage (described later).

[0015] The evaluation item receiving unit 20 receives input of multiple evaluation items related to an applicant. The evaluation item receiving unit 20 may receive input of evaluation items via the input device 80, may receive input of evaluation items by reading out the evaluation items from a portable storage device attached to a drive device (not shown), or may receive input of evaluation items via a network such as an intranet. The evaluation item receiving unit 20 receives input of multiple evaluation items, for example, in a learning stage (described later) of the evaluation model 66.

[0016] FIG. 2 is an example of an interface screen for inputting evaluation items that the evaluation item receiving unit 20 displays on a display device. When the input device 80 or the output device 90 is a computer device, the evaluation item receiving unit 20 may have a web server function for displaying the relevant screen on the computer device. For example, words such as the degree of achievement of business goals, the degree of achievement of task goals, planning ability, and motivation to learn are input as evaluation items, but the evaluation items may be freely determined according to the needs of the business entity that employs personnel. The evaluation accuracy of the information processing system 1 is improved when a specific word such as "ability to execute a task while considering costs" is input, rather than a highly abstract word such as "ability to execute". In this case, the evaluation item receiving unit 20 may receive input of a weight (importance) corresponding to each evaluation item.

[0017] The score acquisition unit 30 inputs the application documents and multiple evaluation items to an AI chatbot that has been trained to answer questions when natural language is input, thereby acquiring a score for each of the multiple evaluation items. For example, the score acquisition unit 30 inputs (transmits) the application documents, multiple evaluation items, and the range of values ​​that the evaluation items can take to the AI ​​chatbot provided by the AI ​​chatbot server 100 in the form of questions in natural language. For example, the score acquisition unit 30 deletes information such as names and addresses that can identify individuals from the application documents, extracts the parts written in natural language, and inputs them to the AI ​​chatbot.

[0018] The questions that the score acquisition unit 30 inputs into the AI ​​chatbot are part or all of the application documents (resume) and sentences such as, "For the following items, please rate what you can tell from the resume on a 5-point scale from 1 (not good at all) to 5 (excellent): degree of achievement of work goals, degree of achievement of task goals, planning ability, motivation to learn, etc."

[0019] The AI ​​chatbot server 100 periodically obtains information from information sources 200 such as internet sites by crawling, and provides an AI chatbot that has been trained to create various answers based on common sense. The AI ​​chatbot is known by names such as ChatGPT() and Marvin.

[0020] The derivation unit 40 derives an employment offer rate, which is an evaluation value of an applicant, based on the scores for each of the multiple evaluation items. The derivation unit 40 derives the employment offer rate, for example, by inputting the scores for each of the multiple evaluation items into an evaluation model 66. The evaluation model 66 is a model generated by machine learning (supervised learning) using, for example, scores for each of the multiple evaluation items acquired from an AI chatbot for past applicants as input data (learning data) and labels indicating whether or not an employment offer was made to each applicant as objective variables (labels, teacher data). This information is stored in the storage unit 60 as past employment data 62.

[0021] The evaluation model 66 is generated in the form of, for example, a decision tree (regression tree). FIG. 3 is a diagram conceptually illustrating a part of the evaluation model 66. In this case, the learning of the evaluation model 66 is performed, for example, by LightGBM (Light Gradient Boosting Machine). LightGBM is a method of performing decision tree analysis multiple times using a method called "gradient boosting." Alternatively, other decision tree analysis methods such as "random forest" may be used. In addition, the evaluation model 66 may be a model generated by machine learning using a network such as a neural network, and the derivation unit 40 may derive the employment acceptance rate by a rule-based method such as multiplying the scores for each of a plurality of evaluation items by weights and aggregating the scores.

[0022] By deriving the employment acceptance rate in this manner, applicants can be evaluated more accurately. If application documents are designated as an item enumeration type or a score is simply calculated based on whether or not a keyword is hit, it is expected that the true attributes of the applicant may be hidden. In this embodiment, the results of analysis by the AI ​​chatbot are input into the evaluation model, so that it is expected that an evaluation will be made that is more multifaceted and closer to an evaluation made by a human. Furthermore, since the accuracy of the AI ​​chatbot is automatically improved daily, the burden of updating the system can be reduced. The employment acceptance rate is displayed by the output device 90 or stored in the storage unit 60 as data for future reference.

[0023] [Processing flow] (Learning stage) 4 is a diagram showing an example of the flow of processing executed by the information processing system 1 in the learning stage. The score acquisition unit 30 acquires multiple sets of data indicating past application documents and whether or not a job offer has been received (S1). Next, the evaluation item receiving unit 20 receives input of multiple evaluation items (S2).

[0024] Next, the score acquisition unit 30 inputs questions that associate multiple evaluation items with each of multiple past application documents into the AI ​​chatbot to acquire a score for each evaluation item (S3).Then, the derivation unit 40 trains an evaluation model 66 based on the scores for each evaluation item and the data on whether or not the applicants who obtained those scores received job offers (S4).

[0025] (Inference stage) FIG. 5 is a diagram showing an example of the flow of processing executed by the information processing system 1 in the inference stage. First, the application document acquisition unit 10 acquires one application document (S11). Next, the score acquisition unit 30 inputs the application document acquired in S11 and multiple evaluation items input in the learning stage (i.e., the evaluation items that are the premise of the evaluation model 66) into the AI ​​chatbot and acquires a score for each evaluation item (S12). Then, the derivation unit 40 inputs the score for each evaluation item into the evaluation model 66 to derive the employment acceptance rate (S13). After that, it is determined whether the processing has been completed for all applicants (S14), and if not, the processing returns to S11.

[0026] In FIG. 1, the AI ​​chatbot is provided by a device external to the information processing system, but the AI ​​chatbot may be an internal function of the information processing system. FIG. 6 is a configuration diagram of an information processing system 1A. In the configuration shown in FIG. 6, the information processing system 1A has an AI chatbot functional unit 50. The function of the AI ​​chatbot functional unit 50 is equivalent to the function of the AI ​​chatbot server 100 in FIG. 1.

[0027] The information processing system described above can be used as described in Fig. 5 and Fig. 6, and after verifying the accuracy in the inference stage, it is also possible to change the multiple evaluation items and perform re-learning to further improve the accuracy. Also, instead of having only one set of multiple evaluation items and evaluation models 66, multiple sets of multiple evaluation items and evaluation models 66, such as a set for the first test of adoption and a set for the second test, may be prepared in advance and used.

[0028] According to the embodiment described above, applicants can be evaluated with greater accuracy.

[0029] The above describes the form for carrying out the present invention using an embodiment, but the present invention is not limited to such an embodiment, and various modifications and substitutions can be made within the scope that does not deviate from the gist of the present invention. [Explanation of symbols]

[0030] 10 Application Document Acquisition Department 20 Evaluation Item Reception Section 30 Score Acquisition Section 40 Derivation part 50 AI Chatbot Function Department 60 Storage section 62 Past Recruitment Data 64 Application Document Data 66 Evaluation Model 100 AI Chatbot Servers 200 sources

Claims

1. an application document acquisition unit that acquires application documents of applicants applying for employment at a business, the application documents being at least partially written in natural language; an evaluation item receiving unit that receives input of a plurality of evaluation items related to an applicant; a score acquisition unit that acquires a score for each of the plurality of evaluation items by inputting the application documents and the plurality of evaluation items into an AI chatbot that has been trained to answer questions when natural language is input; a derivation unit that derives an employment acceptance rate, which is a value that evaluates the applicant, based on the scores for each of the plurality of evaluation items; An information processing system comprising:

2. The AI ​​chatbot is a function provided by a device external to the information processing system.

2. The information processing system according to claim 1.

3. The derivation unit derives the employment acceptance rate by inputting the scores for each of the plurality of evaluation items into an evaluation model.

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

4. The evaluation model is a model generated by machine learning using scores for each of a plurality of evaluation items as input data and past employment history as training data.

4. The information processing system according to claim 3.

5. An information processing method executed by an information processing system, comprising: A process of acquiring application documents of applicants for employment at a business, the application documents being at least partially written in natural language; A process for accepting input of a plurality of evaluation items related to an applicant; A process of inputting the application documents and the plurality of evaluation items into an AI chatbot that has been trained to answer questions when natural language is input, thereby obtaining a score for each of the plurality of evaluation items; A process of deriving an employment acceptance rate, which is a value obtained by evaluating the applicant, based on the scores for each of the plurality of evaluation items; An information processing method comprising:

6. Information processing system, A process of acquiring application documents of applicants for employment at a business, the application documents being at least partially written in natural language; A process for accepting input of a plurality of evaluation items related to an applicant; A process of inputting the application documents and the plurality of evaluation items into an AI chatbot that has been trained to answer questions when natural language is input, thereby obtaining a score for each of the plurality of evaluation items; A process of deriving an employment acceptance rate, which is a value obtained by evaluating the applicant, based on the scores for each of the plurality of evaluation items; A program for executing the above.

Images