KR102411906B1 - Method for consulting on exploration activities based on target departments and curriculum achievement standards - Google Patents

Abstract

The present invention is a technology developed through the 3rd Seoul Innovation Challenge (IC190048) "Artificial Intelligence-based Personalized Personalized University Admission Consulting Technology Development" by the Seoul Industry Promotion Agency. The inquiry activity consulting method based on the target department and subject achievement standards is started. The consulting method according to an embodiment of the present invention includes the steps of receiving user information on academic status and desired university, determining an academic achievement level and customized subject achievement standards based on the received user information, and determined academic achievement It may include the step of determining the inquiry task and inquiry material to be provided based on the level and the customized subject achievement standard, and the step of displaying the determined inquiry task and inquiry material.

Description

Consulting method for inquiry activity based on target department and subject achievement standards

The present invention relates to an inquiry activity consulting method based on target department and subject achievement standards. Specifically, based on artificial intelligence, it is possible to provide inquiry tasks and inquiry materials to perform inquiry activities according to a student's academic level and customized subject achievement criteria. It's about consulting methods.

With the advent of the 4th industrial revolution, changes in the work environment are accelerating. Accordingly, the paradigm of domestic education is also changing centered on career education. Representative examples are the free semester system that encourages students to explore various experiences and career paths in middle school, and the school credit system that supports career design and growth by expanding students' subject options in high school.

Contrary to the ideal policy goal, teachers do not know all the career paths suitable for all subjects, and there is a shortage of diverse and creative career education content to be used in career exploration activities, and teachers also have difficulties in creating educational content. In this regard, Edutech technology that utilizes next-generation information and communication technologies such as artificial intelligence and big data can effectively help students in career exploration and design, and provide content that connects career education and subject education, but Edu Tech data and technical research are also lacking.

Accordingly, voices from the field are increasing to provide high-quality, diverse curriculum-related career education contents based on next-generation edutech technology.
The present invention is to satisfy these needs, and is a technology developed through the 3rd Seoul Innovation Challenge (IC190048) "Artificial Intelligence-based Personalized Personalized University Admission Consulting Technology Development" by the Seoul Industry Promotion Agency.

Registered Patent Publication No. 10-1439464 (2014.09.02) Registered Patent Publication No. 10-1581921 (2015.12.24) Registered Patent Publication No. 10-2136388 (2020.07.15) Laid-open Patent Publication No. 10-2014-0078089 (2014.06.25)

The present invention is to solve the above-mentioned problems, it is possible to build a database about research topics, etc. based on a large number of literature contents, and provide appropriate research topics and research materials according to the student's academic level or customized academic achievement standards. It aims to provide a consulting method for inquiry activities based on target departments and subject achievement standards.

In order to achieve the above object, the inquiry activity consulting method based on the target department and subject achievement standard according to an embodiment of the present invention for achieving the above object includes the steps of receiving user information about the academic situation and desired university, based on the received user information Determining the academic achievement level and the customized subject achievement standard with the method, determining the inquiry task and inquiry material to be provided based on the determined academic achievement level and the customized subject achievement criterion, and displaying the determined inquiry task and inquiry material may include steps.

In addition, the user information includes grades for each subject, school enrolled, and desired university, and the determining step generates a grade index by giving different weights to grades in major subjects and other subjects with respect to grades for each subject, and , The step of generating a school index based on the type of school attended, the college admission result, and determining the academic achievement level based on the generated grade index and the school index and the acceptance cut-off data of the desired university may include

In addition, the user information includes the desired department, the currently enrolled subject, and the keyword of the field of interest, and the determining step includes matching the content of the currently enrolled subject with the desired department and the keyword of the field of interest to conduct an inquiry. It may include determining a subject and a unit and determining a plurality of achievement standards according to the determined subject and unit as a customized subject achievement standard.

And further comprising the step of building a database for the inquiry task and inquiry data, the determining step, extracting the inquiry task of the difficulty corresponding to the determined academic achievement level from the built database, the extracted A step of inferring an inquiry task suitable for the customized subject achievement standard among inquiry tasks using the database, and an inquiry task and inquiry data to provide the inquiry data containing the content of the inferred inquiry task and the inferred inquiry task It may include the step of determining

In addition, the step of inferring an inquiry task suitable for the customized subject achievement standard using the database includes determining a unit, department, and keyword corresponding to the customized subject achievement criterion, the determined unit among the extracted inquiry tasks, Filtering the inquiry task having a department and a keyword, and determining an inquiry task having the highest fitness level among the filtered inquiry tasks as the inquiry task to be provided.

And the step of building the database includes the steps of collecting document content, extracting a subject and a description of the subject, composing a document graph and the document content representing a relationship between the collected document content and the related document content extracting feature information of the collected document content based on a knowledge graph indicating a relationship between knowledge; inputting the extracted feature information into a learned database creation artificial intelligence model to describe the extracted subject and the subject Inferring related units, related departments, application cases, topic key keywords and topic difficulty for It may include storing the topic difficulty level together in a database.

According to various embodiments of the present invention as described above, various career education contents beyond the teacher's subject and major area can be collected using an artificial intelligence model, and career education is provided by providing an appropriate inquiry related to the student's chosen career path. A synergistic effect can be obtained between education and subject matter education.

1 is a block diagram for explaining an inquiry activity consulting system according to an embodiment of the present invention;
2 is a schematic block diagram for explaining the configuration of a consulting providing apparatus according to an embodiment of the present invention;
3 is a diagram for explaining deriving an academic achievement level;
4 is a view for explaining deriving a customized subject achievement standard;
5 is a diagram for explaining determining the inquiry task and inquiry material to be provided to the user;
6 is a schematic block diagram for explaining the configuration of a database building apparatus according to an embodiment of the present invention;
7 is a diagram showing the relationship of each module constituting the processor function in the database building apparatus;
8 is a view for explaining the operation of the collection unit according to an embodiment of the present invention;
9 is a view for explaining the operation of the knowledge unit according to an embodiment of the present invention;
10 is a diagram for explaining the operation of a learning unit, an inference unit, and an inspection unit according to an embodiment of the present invention;
11 is a flowchart for explaining a research activity consulting method based on a target department and subject achievement standards according to an embodiment of the present invention;
12 is a flowchart for explaining a method for generating an artificial intelligence-based academic inquiry topic database according to an embodiment of the present invention;
13 to 15 are exemplary diagrams for explaining a method of generating an artificial intelligence-based academic inquiry topic database according to an embodiment of the present invention.

Hereinafter, various embodiments of the present document will be described with reference to the accompanying drawings. However, this is not intended to limit the technology described in this document to specific embodiments, and it should be understood that various modifications, equivalents, and/or alternatives of the embodiments of this document are included. In connection with the description of the drawings, like reference numerals may be used for like elements.

In this document, expressions such as "has", "may have", "includes" or "may include" indicate the existence of a corresponding characteristic (eg, a numerical value, function, operation, or component such as a part). , does not exclude the presence of additional features.

In this document, expressions such as “A or B”, “at least one of A and/or B” or “one or more of A and/or B” may include all possible combinations of the items listed together. For example, "A or B", "at least one of A and B" or "at least one of A or B" means (1) includes at least one A, (2) includes at least one B, or ( 3) It may refer to all cases including both at least one A and at least one B. Expressions such as "first", "second", "first", or "second" used in this document may modify various elements, regardless of order and/or importance, and convert one element to another. It is used only to distinguish it from an element, and does not limit the corresponding elements.

The expression "configured to (or configured to)" as used in this document, depending on the context, for example, "suitable for", "having the capacity to" "," "designed to", "adapted to", "made to" or "capable of" can be used interchangeably. The term “configured (or configured to)” may not necessarily mean only “specifically designed to” in hardware. Instead, in some contexts the expression “a device configured to” may mean that the device is “capable of” with other devices or parts. For example, "a processor configured (or configured to perform) A, B, and C" means a dedicated processor (eg, an embedded processor) or memory device for performing the corresponding operations by executing one or more software programs stored in the memory device. It may refer to a generic-purpose processor (eg, a CPU or an application processor) capable of performing the execution.

1 is a block diagram illustrating elements constituting a research activity consulting system 1000 based on target departments and subject achievement standards according to an embodiment of the present invention. Referring to FIG. 1 , the inquiry activity consulting system 1000 may include a consulting providing apparatus 100 , a database generating apparatus 200 , and a user terminal apparatus 300 . The division of each device is for convenience of explanation, and it is not necessarily limited to three devices that the inquiry activity consulting system 1000 must operate. For example, it goes without saying that the functions of the consulting providing apparatus 100 and the database generating apparatus 200 may be performed in one electronic device.

The consulting providing apparatus 100 may determine the academic achievement level and the customized subject achievement standard based on the user's information. In addition, the consulting providing apparatus 100 may analyze the determined academic achievement level and the customized subject achievement standard to determine an inquiry task and inquiry data to be provided to the user.

The database generating device 200 collects literature content to recommend research tasks and research materials to be provided to the user, and infers topics, units, departments, keywords, difficulty, etc. related to the collected literature content using an artificial intelligence model. So you can build a database. The database constructed in this way may be used to provide personalized inquiry activity consulting by matching user information in the consulting providing apparatus 100 .

The user terminal device 300 is a device that enables a user to input information on a learning status and the like online and to receive customized consulting. The user terminal device 300 may be implemented in various devices such as a smart phone, a tablet, a notebook computer, and a PC.

Hereinafter, the inquiry activity consulting method will be mainly described with the consulting providing apparatus 100 and the subject database generation method will be mainly described with the database generating apparatus 200 .

2 is a schematic block diagram for explaining the configuration of a consulting providing apparatus 100 according to an embodiment of the present invention. Referring to FIG. 2 , the consulting providing apparatus 100 may include a communication unit 110 , a memory 120 , and a processor 130 . Of course, various components such as an input unit (not shown), an output unit (not shown), and a display unit (not shown) may be additionally included in addition to the illustrated components.

The communication unit 110 may communicate with an external device such as the user terminal device 300 and an external database such as the subject database 245 generated by the database generating device 200 using a wired/wireless network. For example, the communication unit 110 may receive, from the user terminal device 300 , user information about an academic situation input by the user and a university to which he/she wants to go. The communication unit 110 may use various methods such as Bluetooth, Zigbee communication, WiFi, infrared (InfraRed, IR) communication, and near field communication (NFC) as a wireless communication method. In addition, the communication unit 110 may use a high-definition multimedia interface (HDMI), low voltage differential signaling (LVDS), a local area network (LAN), etc. as a wired communication method.

The memory 120 may store various modules, software, and data for driving the consulting providing apparatus 100 . For example, the memory 120 includes data on subject achievement standards according to subjects and units, software for generating grade indicators and school indicators, data acquired from the communication unit 110 (user information, etc.), external devices and Control signal data, such as a communication protocol of , may be stored.

The memory 120 may be implemented in the form of a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or the like. For example, the memory 120 is a ROM for storing a program and/or an application for performing the operation of the consulting providing apparatus 100 , and a RAM for temporarily storing data according to the operation of the consulting providing apparatus 100 . can be provided. In addition, the memory 120 may further include an Electrically Erasable and Programmable ROM (EEPROM) for storing various types of reference data.

The processor 130 may control the above-described configurations of the consulting providing apparatus 100 . For example, the processor 130 may control the communication unit 110 to transmit the determined inquiry task and inquiry data to the user terminal device 300 . The processor 130 may be manufactured in the form of one or a plurality of hardware chips and mounted on the consulting providing apparatus 100 . For example, the processor 130 may be manufactured as a general-purpose processor (eg, CPU or Application Processor) or may be manufactured in the form of a dedicated hardware chip for artificial intelligence.

The processor 130 may control the communication unit 110 to receive user information transmitted from the user terminal device 300 . The received user information may include information about the user's academic situation and desired university. Specifically, the user information may include at least one of grades for each subject (grades of major subjects and other subjects), school enrolled, desired university, desired department, currently enrolled subject, and keywords of interest. In another embodiment, the processor 130 may receive user information directly through an input unit (not shown) instead of receiving the user information from the user terminal device 300 .

Subsequently, the processor 130 may determine an academic achievement level, a customized subject achievement standard, a customized department series, and the like, based on the received user information. The academic achievement level may be expressed as an academic level.

Referring to FIG. 3 , first of all, the processor 130 may determine an academic achievement level by using at least one of information about grades for each subject, a school enrolled in, and information about a desired university among user information. The processor 130 may separate grades for each subject into grades in major subjects and grades in other subjects. In addition, the processor 130 may generate a grade index by giving different weights to grades in major subjects and grades in other subjects. The processor 130 may refer to information on a desired university, a desired department, and currently enrolled subjects when separating or determining different weights from main subject grades and other subject grades.

In addition, the processor 130 may generate a school attendance indicator based on the type of school attended and the result of entering a university. For example, the types of schools attended may include general high schools, special purpose high schools, autonomous private high schools, autonomous public high schools, and the like. In addition, the processor 130 may adjust the weight used to determine the academic achievement level by using the pass cut-line data of the desired university.

In this way, the processor 130 may generate a grade index and a school index based on the user information. In addition, the processor 130 may determine the academic achievement level based on the generated grade index, the school index, and the admission cut-off data of the desired university.

Referring to FIG. 4 , the processor 130 may determine a customized subject achievement standard by using at least one of a desired department, a subject currently taken, a subject taken in the past, and a keyword of interest among user information. Specifically, the processor 130 may determine the subject and unit in which the inquiry activity is to be conducted by matching the content of the currently enrolled subject, the desired department, and the keyword of the field of interest. When the desired department is not input, the processor 130 may determine the desired department information from the currently enrolled subjects and the keywords of the field of interest. Also, the processor 130 may derive a customized department series by generalizing desired department information. In addition, the memory 120 may store a subject achievement standard database in which subject achievement standards matching subjects and units are stored. Using the subject achievement standard database, the processor 130 may determine a plurality of achievement standards according to the determined subject and unit as the customized subject achievement standard.

Referring to FIG. 5 , the processor 130 may determine an inquiry task and inquiry data to be provided to the user based on the determined academic achievement level and the customized subject achievement standard. Information on the inquiry task and inquiry material may be included in the subject database 245 . The content of constructing the subject database 245 including the content of the research task and the research material will be described in detail in the description of the database generating device 200 .

The processor 130 may determine a difficulty level corresponding to the determined academic achievement level. In addition, the processor 130 may extract a research topic having the determined difficulty from the topic database 245 . Alternatively, the processor 130 may control the communication unit 110 to transmit a request signal to provide a research topic having a determined difficulty to the database building apparatus 200 having the topic database 245 . Through this, the processor 130 may first screen the research topic based on the determined academic achievement level.

Subsequently, the processor 130 may determine an inquiry task and inquiry data suitable for the determined customized subject achievement standard among the first extracted inquiry topics. The inquiry material may be a plurality of curated contents that need to be looked at in order to solve the inquiry task. In this case, the processor 130 may infer the inquiry task and inquiry data by using the artificial intelligence model based on the constructed subject database 245 . For example, the processor 130 may determine a unit, a department, and a keyword corresponding to the customized subject achievement standard. In addition, the processor 130 may filter research tasks matching the determined unit, department, and keyword from among the research tasks extracted based on the degree of difficulty. For example, by utilizing the fact that the determined unit may be matched with the input course subject, and the subject matches the input desired subject, the processor 130 recommends a subject and a subject that the user may be interested in rather than a simple mapping. can be mapped. The processor 130 may calculate the matching rate and determine the inquiry task with the highest matching rate as the inquiry task to be provided to the user. In addition, the processor 130 may determine the data containing the contents of the research task determined as the research task to be provided as the research data to be provided. In addition, the processor 130 may control the display unit (not shown) to display the determined inquiry task and inquiry data, or control the communication unit 110 to transmit the information to the user terminal device 300 .

In determining the inquiry task and inquiry data, the processor 130 needs not to provide the same inquiry task and inquiry data to a plurality of users. By using the additional information on the interests and interests of the individual users (students), the processor 130 may provide different inquiry tasks for each individual user. Even when the inquiry task is the same, the processor 130 may provide different inquiry data, which are documents recommended to solve the inquiry task.

To this end, the processor 130 may limit the issuance (search, inquiry, etc.) of duplicate inquiry tasks or inquiry materials by setting duplicate issuance exclusion criteria. For example, the processor 130 may not provide the same inquiry task or inquiry data to other users attending a school in a special city, metropolitan city, or province to which the attending school belongs based on address information of the attending school. In addition, the processor 130 may not provide the same inquiry task or inquiry data to users whose desired university or customized departments overlap. In addition, even if it does not correspond to the above-mentioned duplicate issuance exclusion criteria, the processor 130 counts the number of issuances to the inquiry task and inquiry data so that it is not provided to the user without modification every time it is increased by one.

6 is a block diagram for explaining the configuration of the database building apparatus 200 according to an embodiment of the present invention. Referring to FIG. 6 , the database building apparatus 200 may include a communication unit 210 , a memory 220 , and a processor 230 . Of course, various components constituting the electronic device may be included in addition to the illustrated components.

The communication unit 210 may communicate with an external server having document contents, the consulting providing apparatus 100, and the like using a wired/wireless network. For example, the communication unit 210 may receive information such as a research topic, a difficulty level, and a customized subject achievement standard from the consulting providing apparatus 100 . The communication unit 210 may use various methods such as Bluetooth, Zigbee communication, WiFi, infrared (InfraRed, IR) communication, and near field communication (NFC) as a wireless communication method. In addition, the communication unit 210 may use a High Definition Multimedia Interface (HDMI), a Low Voltage Differential Signaling (LVDS), a Local Area Network (LAN), or the like as a wired communication method.

The memory 220 may store various modules, software, data, artificial intelligence models, and the like for driving the database building apparatus 200 . For example, the memory 220 may store a literature graph 241 indicating a relationship between documents, a knowledge graph 243 indicating a relationship between knowledge, and a topic database 245 including information related to a learning topic. The memory 220 may be implemented in the form of a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or the like. For example, the memory 220 is a ROM for storing a program and/or an application for performing the operation of the database building apparatus 200 , and a RAM for temporarily storing data according to the operation of the database building apparatus 200 . can be provided. In addition, the memory 220 may further include an Electrically Erasable and Programmable ROM (EEPROM) for storing various types of reference data.

The processor 230 may control the above-described configurations of the database building apparatus 200 . For example, the processor 230 may control the communication unit 210 to receive various document contents from a document database of an external server. The processor 230 may be manufactured in the form of one or a plurality of hardware chips and mounted on the database building apparatus 200 . For example, the processor 230 may be manufactured as a general-purpose processor (eg, CPU or Application Processor) or may be manufactured in the form of a dedicated hardware chip for artificial intelligence.

Hereinafter, in order to explain the operation of the processor 230 , the modules of the collection unit 231 , the knowledge unit 233 , the learning unit 235 , the inference unit 237 , and the inspection unit 239 will be described. Each module may be configured as a separate chip, or a plurality of module operations may be performed on one chip. FIG. 7 is a diagram illustrating the relationship between each module constituting the function of the processor 230 .

8 is a view for explaining the operation of the collection unit 231 according to an embodiment of the present invention. The collection unit 231 may collect study-related research topics. For example, the collection unit 231 may collect and pre-process the document content. The collection unit 231 may access a document database through the communication unit 210 to collect various document contents such as thesis, article, reference book, and magazine.

The collection unit 231 may pre-process the document content in the form of text data that can be used in subsequent modules as an input. In addition, the collection unit 231 may extract and store the main text from the pre-processed text data. Also, the collection unit 231 may extract and store the relationship between the document contents by using information such as a list of cited documents or hyperlinks described in the document contents. The relationship between the stored document contents may then be used to generate a document graph in the knowledge unit 233 .

The collection unit 231 may determine a preliminary category of the document content by using the extracted main text. Specifically, the collection unit 231 may input the extracted main text to a pre-trained language model (Language Model, LM) to classify preliminary categories of document content. This can be expressed as a function as in Equation 1 below.

는 추론하고자 하는 문헌 콘텐츠의 예비 카테고리이다. 그리고 LM은 사전에 학습된 언어 모델을 말한다. 언어 모델은 대량의 텍스트를 학습하여 자연언어처리를 할 수 있는 모델이라면 어떤 것이든 사용 가능하다. 예를 들어, BERT, KoBERT, KoGPT-2 등 다양한 언어 모델이 이용될 수 있다. f는 사전 학습된 언어 모델에서 출력된 문헌 콘텐츠의 텍스트 임베딩 데이터를 받아 문헌 콘텐츠의 카테고리를 분류하는 미분 가능한 연속 함수이다.here

is a preliminary category of document content to be inferred. And LM refers to a pre-trained language model. As the language model, any model that can learn a large amount of text and perform natural language processing can be used. For example, various language models such as BERT, KoBERT, and KoGPT-2 may be used. f is a differentiable continuous function that receives text embedding data of document contents output from the pre-trained language model and classifies categories of document contents.

The collection unit 231 cross-entropy loss f according to whether the classified preliminary category matches the actual category to be determined in a later module (eg, the category of the literature content extracted based on the literature graph 241) can be learned through In addition, the collection unit 231 may adjust the pre-trained language model LM based on the loss generated in the process of learning f.

)와 설명(

)으로 나누어 추출할 수 있다. 그리고 수집부(231)는 추출된 주제와 설명을 주제 데이터베이스(245)에 저장할 수 있다.Subsequently, the collection unit 231 may extract a subject and a description of the subject from the document content that has been classified for each preliminary category. That is, the collection unit 231 receives the document content (text) called DOCUMENT as an input, and the subject (

) and description (

) can be divided by In addition, the collection unit 231 may store the extracted subject and description in the subject database 245 .

9 is a view for explaining the operation of the knowledge unit 233 according to an embodiment of the present invention. The knowledge unit 233 may refine the category and difficulty level of the document content by deriving a relationship between the plurality of document contents. Also, the knowledge conversion unit 233 may learn and infer knowledge extracted from document content by deriving a relationship between a plurality of knowledges. The category, document difficulty, and document keyword of the document content extracted from the knowledge unit 233 are selected from the learning unit 235 or the inference unit 237 at least one of related units, related departments, application cases, subject core keywords, and subject difficulty. It can be used as auxiliary information for inference.

Hereinafter, for convenience of explanation, content that is a target for extracting a category and difficulty in the knowledge unit 233 will be referred to as document content, and other content related to the document content will be referred to as related document content.

The knowledge forming unit 233 may extract the category and the document difficulty of the document content based on the document graph 241 indicating the relationship between the document content and the related document content. Specifically, the knowledge unit 233 may generate the literature graph 241 based on the citation relationship extracted from the collection unit 231 . In addition, the knowledge unit 233 may infer the feature information of the document content by inputting the structure of the generated document graph 241 and node feature information into a graph neural network. Subsequently, the knowledge forming unit 233 may extract the category of the document content and the difficulty of the document from the inferred feature information of the document content.

)에서 노드(node)는 각각의 문헌 콘텐츠이고, 엣지(edge)는 문헌 콘텐츠 간의 관계를 나타낸다. 그리고 노드가 가질 수 있는 피처는 문헌 콘텐츠의 카테고리와 문헌 난이도이다.The literature graph 241 may be generated based on the citation relationship between the documents extracted in the process of pre-processing the document content by the collection unit 231 . literature graph (241,

), a node is each document content, and an edge indicates a relationship between document contents. And the features that a node can have are the category of the document content and the document difficulty.

)는 인접행렬(adjacency matrix)

와 노드 셋(node set)

(with

)를 갖는다. 인접행렬은 그래프에서 어느 노드들이 엣지로 연결되었는지를 나타내는 정사각 행렬이다. 그리고 각각의 노드는 c dimension의 노드 피처(node features)를 갖고, 노드 속성은 수집부(231)에서 사용된 언어 모델(LM)을 통해 도출될 수 있다(

). 즉, u번째 문헌 콘텐츠(document)에 대한 노드 피처는

을 통해 추출될 수 있다. 이러한 경우에 지식화부(233)는 그래프 신경망을 통해 이웃 간의 관계 정보를 바탕으로 노드의 표현을 학습할 수 있다. 이를 수식으로 나타내면 아래의 수학식 2와 같다.literature graph (241,

) is an adjacency matrix

and node set

(with

) has The adjacency matrix is a square matrix indicating which nodes in the graph are connected by edges. And each node has c dimension node features, and the node properties may be derived through the language model (LM) used in the collection unit 231 (

). That is, the node feature for the u-th document is

can be extracted through In this case, the knowledge unit 233 may learn the representation of a node based on relationship information between neighbors through a graph neural network. If this is expressed as an equation, it is as Equation 2 below.

은 l번째 GNN layers를 거쳐 생성된 노드 임베딩(노드 속성) 정보이며, UPDATE 및 AGGREGATE 함수는 임의의 미분 가능한 함수이고, N(u)는 노드 u의 이웃 노드들의 집합이다. 그리고

은 초기 노드 피처인

로 초기화된다.From here

is node embedding (node property) information generated through the l-th GNN layers, UPDATE and AGGREGATE functions are arbitrary differentiable functions, and N(u) is a set of neighboring nodes of node u. and

is the initial node feature

is initialized to

를 바탕으로 아래의 수학식 3에 기재된 수식들과 같이 문헌 콘텐츠의 카테고리(

)와 문헌 난이도(

)를 추출할 수 있다.The knowledge unit 233 is the final feature information of the node u learned based on the relationship between documents.

Based on the categories of literature content (

) and literature difficulty (

) can be extracted.

와

은 미분 가능한 연속 함수이다. 지식화부(233)는 이들 함수가 각각 추론한 카테고리 및 난이도가 실제 카테고리 및 난이도와 일치하는지 여부를 바탕으로 이들 함수를 cross-entropy loss를 통해 학습시킬 수 있다. 또한 지식화부(233)는 UPDATE 및 AGGREGATE를

와

의 학습 과정에서 발생한 loss를 바탕으로 학습시킬 수 있다.

Wow

is a differentiable continuous function. The knowledge unit 233 may learn these functions through cross-entropy loss based on whether the categories and difficulties inferred by these functions respectively match the actual categories and difficulties. In addition, the knowledge unit 233 performs UPDATE and AGGREGATE

Wow

It can be trained based on the loss that occurred in the learning process of

The difficulty level derived through Equation 3 in this way may be stored in the future subject database 245 in the form of a label. The difficulty level may be matched with an academic achievement level (academic level) derived from the consulting providing apparatus 100 . For example, the difficulty level and the academic achievement level may all be divided into three levels of high, medium, and low, and these may be matched.

The knowledge unit 233 may refine the classification of the document content by using the preliminary category of the document content determined by the collection unit 231 and the category of the document content extracted by the knowledge unit 233 . The preliminary category determined by the collection unit 231 is a primary classification category, the category extracted based on the relationship between documents in the knowledge unit 233 is a secondary classification category, and the knowledge unit 233 uses primary and secondary classification categories. Thus, the refined category may be called a tertiary classification category.

)는 다음의 수학식 4와 같이 결정될 수 있다.Refined categories (tertiary classification categories,

) can be determined as in Equation 4 below.

The knowledge conversion unit 233 may extract a document keyword of the document content by using the knowledge graph 243 indicating the relationship between knowledge constituting the document content. Specifically, the knowledge unit 233 may express knowledge constituting document content as entities and relationships. An entity can be divided into a subject (e_subject) and an object (e_object). That is, the knowledge conversion unit 233 may convert the relationship between knowledge of the document content into a triplet structure of a subject (e_subject), a relationship (r), and an object (e_object). In addition, the knowledge unit 233 may generate the knowledge graph 243 through a large amount of transformed triplet sets.

The knowledge unit 233 may use the generated knowledge graph 243 as a vector representation in a vector space. To this end, the knowledge unit 233 must express entities and relationships in a vector space. When the vector space expression of each of the subject (e_subject), the relationship (r), and the object (e_object) is defined as h, l, and t, the method for learning the vector expression in the knowledge unit 233 is as shown in Equation (5).

는 참 값인 트리플렛(true triplet)의 집합을 나타내고,

는 거짓 값인 트리플렛(false triplet)의 집합이다. 지식화부(233)는 참 값인 트리플렛과 거짓 값인 트리플렛의 차이를 최대화하도록 학습하여 엔티티와 관계를 벡터 공간에 표현할 수 있다. 또한, 위의 수식에서 s는 score function을 나타내며

와 같이 정의될 수 있다. 특히 지식화부(233)에서는 참 값인 트리플렛과 거짓 값인 트리플렛의 차이를 최대화할 때, 참 값인 트리플렛의 score를 최대화하는 방향으로 학습할 수 있다. 위의 수식에서

는 margin loss에서 필요한 하이퍼 파라미터(hyperparameter)이다.here

represents the set of true triplets,

is a set of false triplets. The knowledge unit 233 may learn to maximize the difference between a triplet that is a true value and a triplet that is a false value to express entities and relationships in a vector space. Also, in the above formula, s represents the score function,

can be defined as In particular, when maximizing the difference between a triplet that is a true value and a triplet that is a false value, the knowledge unit 233 may learn in a direction that maximizes the score of the triplet that is a true value. in the above formula

is a hyperparameter required for margin loss.

As such, the knowledge conversion unit 233 may generate the knowledge graph 243 by the above-described method, and may extract document keywords of the document content based on the generated knowledge graph 243 .

In summary, the knowledge conversion unit 233 may generate content related to the high school academic research topic by using the literature graph 241 indicating the relationship between documents and the knowledge graph 243 indicating the relationship between knowledge. The content generated by the knowledge unit 233 may then be used in the learning unit 235 or the inference unit 237 .

), 관련 학과(

), 적용 사례(

), 주제 핵심 키워드(

) 및 주제 난이도(

) 중 적어도 하나를 포함할 수 있다. 이를 수식으로 표현하면 아래의 수학식 6과 같다.10 is a view for explaining the operations of the learning unit 235, the reasoning unit 237, and the inspection unit 239 according to an embodiment of the present invention. The learning unit 235 uses the basic information of the research topic and explanation extracted based on the document content by the collection unit 231 , and auxiliary information such as the category of the document content inferred from the knowledge conversion unit 233 , the difficulty of the document, and the document keyword. This is to learn a model that recommends content Y related to the research topic. The learning unit 235 may configure a subject database by collecting content information recommended by the learned model. Content Y related to the research topic is the related section (

), related departments (

), application case (

), subject key keywords (

) and topic difficulty (

) may include at least one of. If this is expressed as an equation, it is shown in Equation 6 below.

Features for learning that can be used to train the artificial intelligence model for generating the topic database 245 include a topic, a description of the topic, a document from which the topic is extracted, an entity matching within the topic text (topic or description), or Relationships, categories, literature difficulty, etc.

topic

description of the topic

Literature from which the subject was extracted

Matching entities or relationships within the subject text

category

Literature difficulty

The learning unit 235 may receive the feature X related to the academic research topic as an input, and train a model for inferring the content Y related to the academic research topic. That is, the feature X described above may be defined as a set for inferring the content Y. The elements constituting these sets have permutation invariant properties. Having permutation invariant properties means that it can produce the same output regardless of the order of the input vector elements. The learning unit 235 may define the permutation invariance function as in Equation 7 below.

,

과 같은 itemized element이고, pool은 합산 연산자(summation operator),

는 연속 함수, n은 셋의 elements 갯수이다.In Equation 7, x is defined in feature X.

,

is an itemized element such as , pool is a summation operator,

is a continuous function, and n is the number of elements in the set.

The learning unit 235 may generate the set encoder by stacking permutation-equivalent layers. This is because permutation invariance does not change when the set encoder is configured in this way. The permutation-equivalent layer may be defined as in Equation 8 below. For example, the encoding function f can be expressed by stacking three permutation-equivalent layers as f1(f2(f3())).

와

는 학습 가능한 스칼라 변수(scalar variables)이고,

는 비선형 활성화(nonlinear activation) 함수이다.In Equation 8, pool is a summation operator,

Wow

is a learnable scalar variable,

is a nonlinear activation function.

The learner 235 may encode the feature X using this set encoder. Through the set encoding function, the learning unit 235 may compress and express feature information for content recommendation related to an academic research topic in a vector space. Based on the vector expression, the learning unit 235 may train a model for inferring Y, which is content related to the research topic.

Hereinafter, the set expression that is the result value from the set encoding function will be referred to as Z. The learning unit 235 may infer each y constituting Y from the set expression Z using a continuous function g such as a multi-layer perceptron (MLP). This is defined through Equation 9 below.

)이라면,

이 가질 수 있는 모든 엘리먼트는 고등학생이 배울 수 있는 전체 고등학교 교과 단원이다. 이를 수학식 10과 같이 표현할 수 있다.In Equation 9, y represents all elements that specific content may have in the research topic, and N represents the number of all elements. For example, if a particular content y is related to the

), if

Every element of this possible element is a whole high school unit that a high school student can learn. This can be expressed as Equation (10).

And the softmax function is applied element-wise as in Equation 11 below.

The learner 235 may define a function g for set decoding for each content to be inferred. For example, the learning unit 235 may define a decoding function g_lec for inferring a related unit and a decoding function g_maj for inferring a related department differently.

를 도출할 수 있다. 학습부(235)는 이러한 결론을 도출하기 위한 셋 인코딩과 셋 디코딩 함수를 아래의 수학식 12와 같이 cross-entropy loss로 학습시킬 수 있다. 수학식 12에서 y'은 셋 인코딩 함수 및 셋 디코딩 함수를 거쳐 올바르게 추론되어야 하는 참 값인 콘텐츠를 의미한다.In summary, the learning unit 235 may receive as an input the feature X related to the academic research topic, and train a model for inferring the content Y related to the academic research topic through the previously defined set encoding and set decoding functions. Through this process, the subject database creation artificial intelligence model that has been learned through this process synthesizes each inferred content and finally creates a set of subject-related units, departments, application cases, subject key keywords, and subject difficulty.

can be derived. The learning unit 235 may learn the set encoding and set decoding functions for deriving such a conclusion with cross-entropy loss as shown in Equation 12 below. In Equation 12, y' means content that is a true value to be correctly inferred through a set encoding function and a set decoding function.

The inference unit 237 may infer information to be entered into the subject database 245 by using the subject database creation artificial intelligence model that has completed learning in the learning unit 235 . The reasoning unit 237 may use basic information such as a subject and a description of the subject and auxiliary information such as a category of document content, document difficulty level, and document keyword as input data. The inference unit 237 may receive the subject and description from the collection unit 231 . In the previous description of the learning unit 235, it has been explained that these input information can also be expressed as a feature X for a subject. And related units, related departments, application cases, key keywords, and topic difficulty become output data. In the previous description of the learning unit 235, it has been explained that these output data can also be expressed as content Y.

중에서 가장 높은 값은 갖는 엘리먼트를 선택하여 관련 단원으로 추론할 수 있다.That is, the inference unit 237 may input the input data set X for the subject extracted from the document content into the AI model, and infer the content Y related to the subject. For example, the inference unit 237 may infer a related unit by using the feature X_abc for the subject abc extracted from the literature content as an input value. The reasoning unit 237 is a high school unit

It can be inferred as a related unit by selecting the element having the highest value among them.

Subsequently, the reasoning unit 237 may store the subject, the related unit inferred along with the explanation of the subject, the related department, the application case, the subject key keyword, the subject difficulty, and the like in the subject database 245 . The accuracy of the content inferred through the inspection unit 239 may be reviewed before being stored in the subject database. In this respect, although the inspection unit 239 is represented in FIG. 10, it cannot be said to be an essential component.

The subject database 245 constructed in this way may be used to provide inquiry tasks and inquiry materials in response to user input information in the consulting providing apparatus 100 . According to the input user information, the consulting providing apparatus 100 may determine an academic achievement level and a customized subject achievement standard. The consulting providing apparatus 100 may first collect topics having a topic difficulty corresponding to the academic achievement level determined in the topic database 245 . Among the topics collected primarily, the consulting providing apparatus 100 may determine at least one topic corresponding to a related unit, a topic key keyword, and the like, which are information that can be known according to a customized curriculum achievement standard. Also, the consulting providing apparatus 100 may add filtering to topics having related departments corresponding to desired department information derived from or input from user information. Specifically, the academic achievement level (academic level) derived from the consulting providing device 100 matches the difficulty level finally stored in the subject database 245 , and the derived customized subject achievement standard is finally stored in the subject database 245 . It can be matched with the relevant unit.

The consulting providing apparatus 100 may bring information such as a subject, an explanation, an application case, and a document from which the subject is derived in the subject database 245 for the determined subject. In addition, the consulting providing apparatus 100 may configure the content of the research task with a topic, explanation, application case, etc., and may configure the content of research data with literature information from which the topic is derived.

11 is a flowchart illustrating a method for consulting inquiry activities based on subject achievement standards according to an embodiment of the present invention. Referring to FIG. 11 , the consulting providing apparatus 100 may receive user information about an academic situation and a desired university ( S1110 ). In this case, the received user information may include at least one of grades by subject, school enrolled, university desired, department desired to advance, subject currently enrolled, subject taken in the past, and keyword of interest.

In addition, the consulting providing apparatus 100 may determine an academic achievement level and a customized subject achievement standard based on the received user information (S1120). First of all, the consulting providing apparatus 100 may classify grades for each subject among the received user information into grades in major subjects and grades in other subjects. In addition, the consulting providing apparatus 100 may generate a grade index by giving different weights to grades in major subjects and grades in other subjects. Also, the consulting providing apparatus 100 may generate a school attendance index based on the type of school attended among the received user information and results of college admission. The consulting providing apparatus 100 may determine the academic achievement level based on the generated grade index, the school index, and the acceptance cut-off data of the desired university.

Next, the consulting providing apparatus 100 may determine the subject and unit in which the inquiry activity is to be conducted by matching the content of the currently enrolled subject, the desired department to advance to, and the keyword of the field of interest among the received user information. Based on the subject achievement standard database, the consulting providing apparatus 100 may determine a plurality of achievement standards according to the determined subject and unit as the customized subject achievement standard.

When the academic achievement level and the customized subject achievement standard are determined, the consulting providing apparatus 100 may determine an inquiry task and inquiry data to be provided (S1130). The consulting providing apparatus 100 may use a database for inquiry tasks and inquiry materials, such as the constructed task database 245 . The consulting providing apparatus 100 may extract inquiry tasks of difficulty corresponding to the determined academic achievement level from the constructed database. And the consulting providing apparatus 100 may infer an inquiry task suitable for the customized subject achievement standard from among the extracted inquiry tasks. For example, the consulting providing apparatus 100 may determine a unit, department, and keyword corresponding to the customized subject achievement standard. In addition, the consulting providing apparatus 100 may filter research tasks having at least one of a determined unit, department, and keyword from among the research tasks extracted based on the degree of difficulty. The consulting providing apparatus 100 may determine an inquiry task with the highest fitness score among the filtered inquiry tasks as an inquiry task to be provided, and inquiry data to provide data containing the content of the inquiry task determined as an inquiry task to be provided can be decided with The consulting providing apparatus 100 may give the highest fitness score to an inquiry task having all corresponding units, departments, and keywords. In addition, the consulting providing apparatus 100 may calculate the suitability score in various ways, such as the number of matching items and the summation of matching scores to which a preset weight is given, for a research task that only partially corresponds to a unit, department, or keyword.

In addition, the consulting providing apparatus 100 may display the determined inquiry task and inquiry data (S1140). The consulting providing apparatus 100 may display research tasks and research materials by using a display unit (not shown) among the components, and may display information on the research tasks and research materials determined by an external device such as the user terminal device 300 . It can also be transmitted and displayed on the corresponding external device.

12 is a flowchart illustrating a method of generating an artificial intelligence-based academic inquiry topic database according to an embodiment of the present invention. Referring to FIG. 12 , the database generating apparatus 200 may collect and pre-process various document contents from an external server ( S1210 ). For example, the document content collected from the external server may be the document of FIG. 13 . The database generating apparatus 200 may convert the document content into an usable text form. In addition, the database generating apparatus 200 may extract and store the main text, citation relationship, title, date of creation, etc. of the document content by using the converted text data. 14 is an example of extracting the title, type, abstract, keyword, creation date, etc. of document content from the document of FIG. 13 .

Also, the database generating apparatus 200 may determine a preliminary category of the document content based on the extracted main text. For example, the database generating apparatus 200 may infer a preliminary category by inputting the extracted main text to the learned language model. The language model can be fine-tuned according to whether the preliminary category inferred in the future is a category of actual document content. For example, in the case of the document of FIG. 13 , the database generating apparatus 200 may determine the preliminary categories as 'Computer Science', 'Machine Learning', and 'Graph'.

The database generating apparatus 200 may extract a subject and a description of the subject from the document content (S1220). As described above, the database generating apparatus 200 may extract the subject of the document content and sentences describing the subject by using the converted text data. 15 shows an example in which the subject and description of document content are extracted. In the case of FIG. 15, since the document content is in the form of a thesis as given in FIG. 13, the language model recognizes the keyword or abstract part and extracts the subject and description. If it is not in the form of a thesis, the language model recognizes the entire sentence and generates a summary sentence to extract the topic or explanation.

Next, the database generating apparatus 200 may extract the category of the document content, the difficulty of the document, and the like, based on the document graph representing the relationship between the document content and the related document content ( S1230 ). As described in the description of the database generating apparatus 200 above, the literature graph expresses the relationship between documents, and may be generated based on the citation relationship extracted in the pre-processing of the document content. By inputting the structure of the document graph generated in this way and feature information of nodes into the graph neural network, the database generating apparatus 200 may infer feature information of the document content. In addition, the database generating apparatus 200 may extract the category of the document content, the difficulty level of the document, and the like from the inferred feature information of the document content.

In the example of FIG. 13 , the database generating device 200 may learn information such as “[This paper] is cited by [Layer normalization].”, “[This paper] cite [Diffusion-convolutional neural networks].” , "[This paper] might be cited by [Edge contraction pooling for graph neural networks].", "[This paper] might cite [LIBSVM: A libraty for support vector machines]." This learned and inferred information can go to the literature graph.

Next, the database generating apparatus 200 may extract a document keyword of the document content based on the knowledge graph indicating the relationship between the knowledge constituting the document content ( S1240 ). The database generating apparatus 200 may convert the knowledge constituting the document content into a triplet structure of a subject, a relation, and an object. The database generating apparatus 200 may express the knowledge graph in a vector space by using the converted multiple triplet structures. In generating the knowledge graph in this way, a method can be used to learn to maximize the difference between a triplet, which is a true value, and a triplet, which is a false value, to express in a vector space. In addition, a method of learning to maximize the value of a triplet, which is a true value, may be used and expressing it in a vector space. The database generating apparatus 200 may extract a document keyword of the document content based on the generated knowledge graph.

In the example of FIG. 13 , the database generating apparatus 200 first performs "[They, validate, GST for graph classification, reconstruction, and generation tasks]", "[existing GNN with our parametric pooling operation, can be, as powerful as the WL test]”. In addition, the database generating apparatus 200 may include "[They, validate, GST for graph classification, reconstruction, and generation tasks]", "[Proposed pooling operation, can be, as powerful as the WL test]" from the extracted knowledge. It can learn information and infer information like "[They, distinguish, graphs as powerful as the WL test]". This learned and inferred information can go to the knowledge graph.

Subsequently, the database generating device 200 inputs feature information such as the subject extracted in the previous process, a description of the subject, the category of the document content, the document difficulty level, and the document keyword, into the subject database generating artificial intelligence model on which the learning has been completed. , related departments, application cases, topic key keywords, and topic difficulty information can be inferred (S1250). In the example of FIG. 13 , the database generating device 200 sets the related unit as 'information, math 2', the related department as 'computer department, math department', the application case as 'graph classification, graph generation', and the key keyword ' With Graph representation learning and Graph pooling', the subject difficulty can be inferred as 'best'.

In addition, the database generating apparatus 200 may store the inferred information in the database together with the feature information (S1260). The automatically generated subject database is used together with the user information input from the consulting providing apparatus 100, so that it is possible to provide a customized inquiry topic and inquiry data suitable for the user.

According to various embodiments of the present invention as described above, various career education contents beyond the subject area and major area of the teacher may be collected using an artificial intelligence model. It can also provide students with appropriate inquiry tasks related to their chosen career path. In particular, it has the effect of providing different inquiry tasks according to each student's situation (eg, relevance to the student's desired major, relevance to subjects taken, relevance to unit, appropriateness of difficulty, etc.).

Meanwhile, as used herein, the term “unit” or “module” includes a unit composed of hardware, software, or firmware, and may be used interchangeably with terms such as, for example, logic, logic block, part or circuit. have. “Part” or “module” may be an integrally formed part or a minimum unit that performs one or more functions or a part thereof. For example, the module may be configured as an application specific integrated circuit (ASIC).

Various embodiments of the present invention may be implemented as software including instructions stored in a machine-readable storage medium (eg, a computer). The device is a device capable of calling a stored command from a storage medium and operating according to the called command, and may include an electronic device (eg, the consulting providing device 100 ) according to the disclosed embodiments. When the instruction is executed by the processor, the processor may directly or use other components under the control of the processor to perform a function corresponding to the instruction. Instructions may include code generated or executed by a compiler or interpreter. The device-readable storage medium may be provided in the form of a non-transitory storage medium. Here, 'non-transitory' means that the storage medium does not include a signal and is tangible, and does not distinguish that data is semi-permanently or temporarily stored in the storage medium.

According to an embodiment, the method according to various embodiments disclosed in this document may be provided as included in a computer program product. Computer program products may be traded between sellers and buyers as commodities. The computer program product may be distributed in the form of a machine-readable storage medium (eg, compact disc read only memory (CD-ROM)) or online through an application store (eg, Play Store™). In the case of online distribution, at least a portion of the computer program product may be temporarily stored or temporarily generated in a storage medium such as a memory of a server of a manufacturer, a server of an application store, or a relay server.

Each of the components (eg, a module or a program) according to various embodiments may be composed of a singular or a plurality of entities, and some sub-components of the aforementioned sub-components may be omitted, or other sub-components may be various It may be further included in the embodiment. Alternatively or additionally, some components (eg, a module or a program) may be integrated into a single entity, so that functions performed by each corresponding component prior to integration may be performed identically or similarly. According to various embodiments, operations performed by a module, program, or other component may be sequentially, parallelly, repetitively or heuristically executed, or at least some operations may be executed in a different order, omitted, or other operations may be added. can

1000: inquiry activity consulting system
100: consulting providing device
200: database creation device
300: user terminal device

Claims (9)

Step, by the database generating device, to build a database for the inquiry task and inquiry data;
Receiving user information about academic status and desired university through the communication unit;
determining, by the processor, an academic achievement level and a customized subject achievement standard based on the received user information;
determining, by the processor, inquiry tasks and inquiry materials to be provided based on the determined academic achievement level and customized subject achievement standards using the built-up database; and
Including, by the processor, providing the determined inquiry task and inquiry data;
The inquiry task is a subject having at least one common item with a unit, department, and keyword according to the customized subject achievement standard, and the inquiry material is a plurality of curated contents to solve the inquiry task,
Building the database includes:
collecting, by the database generating device, document content, and extracting a subject and a description of the subject;
The database generating device generates feature information of the collected document contents based on a document graph indicating a relation between the collected document contents and related document contents and a knowledge graph indicating a relation between knowledge constituting the document contents. extracting;
The database generating device inputs the extracted feature information into the learned database creation artificial intelligence model, and the extracted topic and related units for the description of the topic, related departments, application cases, topic key keywords and topics inferring the difficulty; and
Storing, by the database generating device, the extracted subject and the description of the subject and the inferred related unit, related department, application case, subject key keyword, and subject difficulty together in a database; and
The step of extracting feature information of the collected document content based on the knowledge graph comprises:
The database generating device converts the knowledge constituting the document content into a triplet structure of subject, relationship, and object, and learns to maximize the difference between a triplet that is a true value and a triplet that is a false value based on the set of the converted triplet structures to generate a knowledge graph by expressing it in a vector space, and extracting a literature keyword of the document content based on the generated knowledge graph. According to claim 1,
The user information includes grades by subject, school enrolled, and desired university,
The determining step is
generating, by the processor, a grade index by giving different weights to grades in a major subject and other subjects with respect to the grades for each subject, and generating an index of attending school based on the type of school attended and results of college admission; and
The processor, the step of determining the academic achievement level based on the generated grade index and the school index, and the acceptance cut-off data of the desired university to advance; inquiry activity consulting method comprising a. According to claim 1,
The user information includes the desired department, currently enrolled subjects, and keywords in the field of interest,
The determining step is
determining, by the processor, a subject and unit to conduct an inquiry activity by matching the content of the currently enrolled subject with the desired department and the field of interest keyword; and
Inquiry consulting method comprising a; determining, by the processor, a plurality of achievement standards according to the determined subject and unit as customized subject achievement standards. According to claim 1,
The determining step is
extracting, by the processor, an inquiry task of difficulty corresponding to the determined academic achievement level from the built-up database;
The processor determines a unit, department, and keyword corresponding to the customized subject achievement standard, filters the inquiry task having the determined unit, department, and keyword among the extracted inquiry tasks, and a fitness level among the filtered inquiry tasks determining the highest inquiry task as the inquiry task to be provided; and
The process of determining, by the processor, the inquiry data containing the contents of the determined inquiry task as the inquiry data to be provided; inquiry activity consulting method comprising a. delete delete According to claim 1,
The determining step is
The processor, based on the address information of the attending school included in the user information, the history of providing the determined inquiry task and inquiry data to other users attending schools in the special city, metropolitan city, and province to which the attending school belongs An inquiry activity consulting method, characterized in that it is determined whether there is, and if there is a history provided as a result of the judgment, the inquiry task and inquiry data are re-determined. According to claim 1,
The determining step is
The processor, based on the desired university and the desired department included in the user information, the determined inquiry task and inquiry data is a history provided to another user in which at least one of the user and the desired university and the desired department to advance to the same Inquiry consulting method, characterized in that it is determined whether there is, and if there is a history provided as a result of the determination, the inquiry task and inquiry data are re-determined. delete

Images