KR102285665B1 - A method, system and apparatus for providing education curriculum - Google Patents

Abstract

According to an embodiment, the present invention relates to a method for providing a learning curriculum by analyzing academic achievement, performed by a control system, which includes the steps of: obtaining an identifier for specifying the identity of a first student as a learning target and a learning progress index for specifying the learning progress of the first student; obtaining an image obtained by photographing a problem sheet on which problem solving and scoring are completed during the learning of a first unit, by using an imaging unit of an electronic device; generating an input vector by encoding the image, and applying the same to an input layer of a pre-trained convolutional neural network; obtaining a first output value from an output layer of the convolutional neural network; inputting the first output value and an auxiliary output value stored in a database within the control system into a pre-trained neural network; obtaining a second output value from an output layer of the neural network; analyzing the type and difficulty level of a first problem according to the degree of similarity; determining the type and difficulty level of a second problem included in a second unit to be learned after the first unit; and providing a learning curriculum according to the type and difficulty level of the second problem. According to an embodiment, based on artificial intelligence, it is possible to determine the difficulty level of a curriculum in a subsequent learning field in consideration of an academic level of each student on each learning content.

Description

A METHOD, SYSTEM AND APPARATUS FOR PROVIDING EDUCATION CURRICULUM

The present invention relates to a method, system and apparatus for providing a personalized learning curriculum based on artificial intelligence, and more particularly, to a method, system and apparatus for providing a learning curriculum of appropriate difficulty based on an analysis of the academic achievement of a learning target; and It's about the device.

In the academic field where learning takes place, in most cases, a uniform learning curriculum is applied to a large number of students. However, in such a uniform, many-to-one type of academic field, it is impossible to consider each student's learning ability, weakness, or academic performance. For example, in the case of students with low learning ability compared to the average of many students in the academic field, it is difficult to follow a uniform learning curriculum, so they may not understand or feel frustrated. In this case, even though a more in-depth curriculum should be provided, there is a risk that the students may not be able to fully demonstrate their academic capabilities because a uniform learning curriculum is provided.

In addition, even for individual students, there may be areas of strength or areas of weakness depending on each subject and learning content. In other words, when planning an individual student's learning curriculum, instead of relying solely on the student's overall academic achievement, learning progress is provided such as providing an intensive curriculum in areas of strength or basic curriculum in areas of weakness to strengthen the basic skills Accordingly, the difficulty of the curriculum needs to be adjusted differently.

Korean Patent Registration No. 10-1044976 (June 29, 2011) Korean Patent Publication No. 10-2011-0082445 (2011.07.19.) Korean Patent Registration No. 10-2097777 (2020.04.06.) Korean Patent No. 10-2118980 (2020.06.04.)

The present invention has been made in order to solve the above problems, and an object thereof is to provide a curriculum optimized for each student in consideration of the academic level and efficiency of each student. More specifically, an object of the present invention is to analyze the student's academic performance results so that academic achievement can be easily determined, based on this, by determining the difficulty of the curriculum for the field of study to be conducted later, academic achievement It is about providing an efficient and appropriate curriculum in terms of aspects.

Objects of the present invention are not limited to the objects mentioned above, and other objects not mentioned will be clearly understood from the description below.

According to an embodiment, as a method of providing a learning curriculum by analyzing academic achievement, performed by a control system, an identifier for specifying the identity of a first student as a learning target and learning progress for specifying the learning progress of the first student obtaining an index; The learning progress of the first student is confirmed as the first unit through the learning progress index, and after learning of the first unit, problem solving and scoring are completed during the learning of the first unit using an imaging unit of an electronic device. acquiring a photographed image; generating an input vector by encoding the image, and applying it to an input layer of a pre-trained convolutional neural network of an embedded computer in the control system; obtaining, as a result of the operation of the convolutional neural network, a first output value from an output layer of the convolutional neural network, the first output value estimating the type and correct answer of the first problem extracted from the image; inputting the first output value and the auxiliary output value stored in a database in the control system into a pre-trained neural network, wherein the neural network evaluates the first output value based on the auxiliary output value; obtaining, from an output layer of the neural network, a second output value, wherein the second output value indicates the type of the first corrected problem and whether or not it is a correct answer; By comparing the first problem with the problems of the first unit stored in advance in the database, the degree of similarity is determined through keywords extracted from each question of the problem, and the type and difficulty of the first problem are analyzed according to the similarity. to do; determining the type and difficulty of a second problem included in a second unit to be learned after the first unit based on whether the first problem is correct, and the type and difficulty of the first problem; and providing a learning curriculum according to the type and difficulty of the second problem when it is confirmed that the learning progress of the first student has completed the learning of the first unit.

The obtaining of the first output value may include: encoding and inputting color information for each pixel of the image in a data sheet format, wherein the color information includes RGB color information, brightness information, and chroma information; separating the first question from the image based on the input, and separating and extracting the questions of the first question and the handwriting marked on the first question; recognizing a keyword from the items of the first problem, and classifying the type of the first problem through the recognized keyword; checking the color, brightness, and saturation of the handwriting graded on the first question, and determining whether the answer to the first question is correct according to preset first color values, first brightness values, and first saturation values; and outputting an output value including the type of the first problem and the correct answer to a total of 50 output layer nodes, wherein the convolutional neural network learns through a first learning signal according to a user input. wherein the upper 25 output layer nodes among the 50 output layer nodes indicate the type of the first problem, and the lower 25 output layer nodes among the 50 output layer nodes are nodes corresponding to the upper 25 output layer nodes - the lower 25 For the output layer nodes, the 26th output layer node corresponds to the 1st output layer node, the 27th output layer node corresponds to the 2nd output layer node, and so on, the 25+nth output layer node corresponds to the nth output layer node. It may indicate whether or not, and output a number '0' to a node indicating that the top 25 output layer nodes are not of the type of the first problem.

The neural network takes the first output value that is an output value of the convolutional neural network and the auxiliary output value stored in the database as inputs, evaluates the first output value based on the auxiliary output value, and a plurality of problems stored in the database Detects an abnormal signal based on the accumulated data of evaluation on the type and correct answer of Learning may be performed through the second learning signal.

The method for providing the learning curriculum includes: receiving, from a plurality of lecturer terminals, a notification that lecturer accounts logged into each of the lecturer terminals are set to a lecture standby mode; generating a lecture waiting list based on the lecturer accounts set to the lecture standby mode; receiving a lecture request for the second unit from a first student terminal logged in through the first student account; transmitting the lecture request to lecturer terminals in which each of the lecturer accounts included in the lecture waiting list are logged in; receiving whether or not to accept the lecture request from lecturer terminals in which each of the lecturer accounts included in the lecture waiting list are logged in; generating a lecture request acceptance list based on the lecture waiting list and whether the lecture has been accepted; transmitting the lecture request acceptance list to the first student terminal; receiving lecturer selection information from the first student terminal; transmitting the lecturer selection information to a first lecturer terminal in which a first lecturer account corresponding to the lecturer selection information is logged in; receiving from the first lecturer terminal whether to accept the lecturer matching; As a result of checking whether to accept the instructor matching, if the instructor matching is accepted, the first instructor account and the first student account are matched to create a chat space between the first instructor account and the first student account to do; receiving a notification from the first lecturer terminal that the lecture for the second unit has started; transmitting a request for approval to start the lecture of the second unit to the first student terminal; receiving approval to start the lecture of the second unit from the first student terminal; setting a time when the lecture start approval is received as a lecture start time; receiving a notification from the first student terminal that the lecture of the second unit has ended; transmitting a request for approval to end the lecture of the second unit to the first lecturer terminal; receiving approval to end the lecture of the second unit from the first lecturer terminal; setting a time when the lecture end approval is received as a lecture end time; calculating a cost to be paid by the first student account and a cost to be paid to the first lecturer account based on the lecture start time and the lecture end time; confirming a correspondence relationship between each preset staged outdoor environmental risk stage and an incentive per unit time; obtaining, based on the lecture request, whether an environment in which the first lecturer performs a lecture is outdoors; setting an expected incentive per unit time based on an outdoor environment forecast value at a lecture location and a lecture time according to the lecture request; and calculating an incentive of the first lecturer in response to an outdoor environment risk step including an outdoor environment value while the first lecturer is conducting a lecture.

The method for providing the learning curriculum includes: acquiring skin information of lecturers using lecturer accounts at predetermined intervals through a lecturer authentication means that is a camera capable of detecting the lecturer's skin; diagnosing whether each lecturer has a skin disease based on the lecturer's skin information; and if the first lecturer account is diagnosed as having a skin disease, and it is confirmed that the outdoor environment forecast value at the lecture place and lecture time according to the lecture request is included in the predetermined outdoor environment risk stage, It may further include the step of setting the acceptance not to be selected.

According to one embodiment, based on artificial intelligence, it is possible to determine the difficulty of the curriculum in the subsequent learning field in consideration of the academic level of each individual learning content. Therefore, learning can be carried out by reflecting the difficulty of the predetermined curriculum, minimizing trial and error or time wasted in academic performance, and more efficient and optimized learning than when a uniform curriculum is provided it is possible to do

On the other hand, effects according to the embodiments are not limited to those mentioned above, and other effects not mentioned will be clearly understood by those of ordinary skill in the art from the following description.

1 is a flowchart illustrating a method of providing a learning curriculum according to an embodiment.
2 is a schematic diagram for explaining a convolutional neural network used to provide a learning curriculum in an embodiment.
3 is a schematic diagram illustrating a neural network according to an embodiment.
4 is a schematic diagram for explaining a blockchain network in an embodiment.
5 is a diagram schematically showing the configuration of a system for providing a lecture service according to an embodiment.
6 is a view for explaining a lecturer matching process according to an embodiment.
7 is a view for explaining a course of a lecture according to an embodiment.
8 is a flowchart illustrating an incentive providing method according to an embodiment.
9 is a flowchart illustrating a method of setting whether a lecturer can provide a lecture service based on a health condition of the lecturer according to an exemplary embodiment.
10 is a diagram for explaining learning of an artificial neural network according to an embodiment.
11 is an exemplary diagram of a configuration of a control device according to an embodiment.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, since various changes may be made to the embodiments, the scope of the patent application is not limited or limited by these embodiments. It should be understood that all modifications, equivalents and substitutes for the embodiments are included in the scope of the rights.

Specific structural or functional descriptions of the embodiments are disclosed for purposes of illustration only, and may be changed and implemented in various forms. Accordingly, the embodiments are not limited to a specific disclosure form, and the scope of the present specification includes changes, equivalents, or substitutes included in the technical spirit.

Although terms such as first or second may be used to describe various components, these terms should be interpreted only for the purpose of distinguishing one component from another. For example, a first component may be termed a second component, and similarly, a second component may also be termed a first component.

When a component is referred to as being “connected to” another component, it may be directly connected or connected to the other component, but it should be understood that another component may exist in between.

The terms used in the examples are used for description purposes only, and should not be construed as limiting. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present specification, terms such as “comprise” or “have” are intended to designate that a feature, number, step, operation, component, part, or combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiment belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

In addition, in the description with reference to the accompanying drawings, the same components are given the same reference numerals regardless of the reference numerals, and the overlapping description thereof will be omitted. In the description of the embodiment, if it is determined that a detailed description of a related known technology may unnecessarily obscure the gist of the embodiment, the detailed description thereof will be omitted.

The embodiments may be implemented in various types of products, such as personal computers, laptop computers, tablet computers, smart phones, televisions, smart home appliances, intelligent cars, kiosks, wearable devices, and the like.

1 is a flowchart illustrating a method of providing a learning curriculum according to an embodiment.

According to an embodiment, a control system for providing a learning curriculum may include a control device, a convolutional neural network, a neural network, a block chain network, a plurality of terminals, and the like. The control device can control the overall operation of the control system, and the operation of the control system will be mainly described with respect to the operation of the control device.

The control device may acquire an identifier for specifying the identity of a student who is to be provided with a learning curriculum and a learning progress index for specifying the learning progress ( S101 ). For example, when the learning target is the first student, the control device may acquire an identifier specifying the identity of the first student and a learning progress index specifying the learning progress of the first student.

According to an embodiment, the identifier may be obtained by, for example, inputting an ID and a password into an electronic device used by a student to perform learning. In another example, the identifier may be obtained by collecting and analyzing biometric identifiers such as iris information, facial information, and voice information of a student by an imaging unit included in the electronic device. The learning progress index may include an index specifying the learning progress.

The electronic device includes a processor, a memory, a user interface, and a communication interface, and may be connected to other electronic devices through a network. The communication interface may transmit/receive data to and from other electronic devices within a predetermined distance through a wired or wireless network or wired serial communication. The network enables wired and wireless communication between the electronic device and various entities according to an embodiment. The electronic device may communicate with various entities over a network, and the network may use standard communication technologies and/or protocols. In this case, the network may include, but is not limited to, the Internet, a local area network (LAN), a wireless local area network (Wireless LAN), a wide area network (WAN), a personal area network (PAN), and the like. It is obvious to a person skilled in the art that it may be a different type of network capable of transmitting and receiving data.

The electronic device may be any electronic device including a communication function, for example, a smartphone, a tablet personal computer (PC), a mobile phone, a video phone, and an E-book reader. ), Desktop PC (Desktop Personal Computer), Laptop PC (Laptop Personal Computer), Netbook Computer (Netbook Computer), PDA (Personal Digital Assistant), PMP (Portable Multimedia Player), MP3 Player, Camera (Camera), Wearable Device ( Wearable Device) may include at least one of. The wearable device is at least one of a Head-Mounted-Device (HMD) such as electronic glasses, an electronic garment, an electronic bracelet, an electronic necklace, an electronic accessory, an electronic tattoo, a smart car, or a smart watch. may include.

The imaging unit included in the electronic device is a full HD (High Definition) quality camera device for capturing an image, and may be a photographing device based on 30 to 60 frames. The imaging unit can automatically adjust exposure, shutter speed, and sensitivity, and manual adjustment is possible according to the user's convenience.

After the student learns, the control device may acquire an image of the question paper in which problem solving and scoring have been performed during learning using the imaging unit of the electronic device (S102). By acquiring the image obtained by photographing the question paper as described above, it is possible to collect not only the problem already data data in the electronic device, but also the learning result on the problem printed out. The control device acquiring the image may be accomplished by transmitting the image captured by the imaging unit to the control device by the communication unit of the electronic device. When recognizing that the student's learning is finished, the control device may prompt the user of the electronic device to take a picture by displaying a notification requesting photographing of a problem sheet for which problem solving and scoring have been completed during learning on the display of the electronic device. The notification may be displayed at a predetermined time every predetermined period. Accordingly, it becomes possible to periodically collect and systematize the student's learning results.

The control device checks the learning progress of the first student as the first unit through the learning progress index of the first student, and after learning the first unit, problem solving and scoring during the learning of the first unit using the imaging unit of the electronic device An image of the completed question paper may be acquired.

For example, if the first unit corresponds to pages 1 and 2 of the question paper A, the control device may acquire images obtained by photographing pages 1 and 2, and as a result of image analysis, recognize the number in which the page is written in the image Thus, it is possible to check which unit the image was acquired during the learning with the recognized page number.

The control device may encode the acquired image to generate an input vector, and apply it to an input layer of a pre-trained convolutional neural network of an embedded computer in the control system (S103). The convolutional neural network is configured to learn through the image generated from the problem paper on which the learning has been completed. As a more preferred embodiment, the control device may generate an input vector by encoding pixels of the obtained image with color information. The color information may include, but is not limited to, RGB color information, brightness information, and saturation information. The control device may convert color information into a numerical value, and may encode an image in the form of a data sheet including this value.

In an embodiment, the convolutional neural network is composed of a feature extraction neural network and a classification neural network, and the feature extraction neural network stacks an input signal with a convolutional layer and a pooling layer sequentially. The convolution layer contains convolution operations, convolution filters, and activation functions. The calculation of the convolution filter is adjusted according to the matrix size of the target input, but a 9X9 matrix is generally used. The activation function generally uses, but is not limited to, a ReLU function, a sigmoid function, and a tanh function. The pooling layer is a layer that reduces the size of the input matrix, and uses a method of extracting representative values by tying pixels in a specific area. In general, the average value or the maximum value is often used for calculation of the pooling layer, but is not limited thereto. The operation is performed using a square matrix, usually a 9x9 matrix. The convolutional layer and the pooling layer are repeated alternately until the corresponding input becomes small enough while maintaining the difference.

According to one embodiment, the classification neural network has a hidden layer and an output layer. In the convolutional neural network of the present invention, there are generally three or more hidden layers, and 100 nodes of each hidden layer are specified, but more may be specified in some cases. The activation function of the hidden layer uses a ReLU function, a sigmoid function, a tanh function, and the like, but is not limited thereto. The number of output layer nodes of the convolutional neural network may be 50 in total.

The control device may apply the generated input vector to the input layer of the academic achievement judgment model (S104). According to an embodiment, 50 output layer nodes of the convolutional neural network may include upper 25 output layer nodes and lower 25 output layer nodes. The top 25 output layer nodes among 50 output layer nodes of the convolutional neural network can indicate the type of each problem included in the problem paper. The lower 25 output layer nodes may indicate whether the problem solving result corresponding to the upper 25 output layer nodes is correct. The details of the convolutional neural network will be described later in detail with reference to FIG. 2 .

According to an embodiment, the control device may obtain a first output value from an output layer of the convolutional neural network as a result of an operation of the convolutional neural network (S104). The first output value may estimate the type of each question extracted from the image and whether the answer is correct.

That is, the control device may obtain a first output value for estimating the type of the first problem, which is one or more problems among each problem extracted from the image, and whether the answer is correct. Here, the first problem is a problem included in the first unit.

The control device may input the first output value and the auxiliary output value stored in the database in the control system to the pre-learned neural network ( S105 ). The neural network may evaluate the type of each problem estimated from the image and the correct answer based on the auxiliary output value. Details of the neural network will be described later in detail with reference to FIG. 3 .

The control device may obtain the second output value from the output layer of the neural network (S106). The second output value may indicate the type of each corrected question and whether the answer is correct by correcting the estimate of the first output value.

That is, the control device can check the type of the first question and whether the answer is correct through the first output value, and check the evaluation result of the type of the first question and whether the answer is correct based on the auxiliary output value through the second output value, As a result, if it is determined that the type of the first question requires correction, the type of the first question corrected and whether the correct answer is correct may be checked through the second output value.

According to an embodiment, the problem type may be classified according to the learning ability required to solve the problem, and the learning ability may include mental arithmetic, problem comprehension, reasoning ability, memorization ability, spatial perception ability, etc. can The difficulty level may be set according to the type of the problem. For example, the first problem may be set to intermediate difficulty as the problem type for judging mental arithmetic in unit 1, and the second problem may be set to advanced difficulty as the problem type for judging problem comprehension in unit 2 there is.

According to one embodiment, mental arithmetic means the ability to calculate with the head without going through a calculator or handwriting, problem comprehension means the ability to interpret a problem, and reasoning ability determines a solution according to the interpretation of the problem Memorization refers to the ability to indicate the level of knowledge that can only be solved by prior knowledge among formulas and explanations, and spatial perception refers to the ability mainly used to solve graph objects and figure objects. may be doing

According to one embodiment, the type of problem is preset for each problem, a QR code indicating the type of problem for each problem may be printed on a part of the problem, and the control device is By recognizing the QR code, you can identify the type of problem for each problem.

According to an embodiment, the type of the first problem may be estimated through the first output value and then corrected through the second output value. For example, the first problem type is estimated as the problem type for determining the first learning ability through the first output value, but may be modified to the problem type for determining the second learning ability through the second output value.

According to an embodiment, the control device compares the first problem with the problems of the first unit stored in advance in the database, determines the similarity through keywords extracted from each question of the problem, and determines the type of the first problem according to the similarity And it is possible to analyze the difficulty (S107). A keyword may mean a word or formula that can highlight the characteristics of a problem.

In the process of comparing the problems of the first unit with the first problem, the control device can extract duplicate keywords that match each other by comparing keywords extracted from each question of the problem, and the more the number of duplicate keywords, the higher the similarity. can be judged as

For example, if the second problem and the third problem are pre-stored in the database as the problem of the first unit, the control device compares the first problem with the second problem, and as a result of the comparison, it is confirmed that 8 keywords match Then, 8 duplicate keywords may be extracted, and the similarity between the first problem and the second problem may be determined to be 80% similar. The control device may compare the first problem with the third problem, and when it is confirmed that the four keywords match as a result of the comparison, the control device may extract four duplicate keywords and determine the similarity to be 40%.

In the process of analyzing the type and difficulty of the first problem according to the degree of similarity, the control device may set the type and difficulty of the problem with the highest similarity among the problems of the first unit as the type and difficulty of the first problem.

For example, as a result of determining the similarity between the problems of the first unit and the first problem, if it is determined that the similarity between the second problem and the first problem is the highest, the control device checks the type and difficulty of the second problem, It can be analyzed by setting the type and difficulty of the second problem to the type and difficulty of the first problem.

According to an embodiment, the control device groups the identifier specifying the student's identity, the learning progress index specifying the learning progress performed at the time, and the second output value obtained by the neural network, and transmits it to the database and the blockchain network can do. The control device may group the identifier, the learning progress index, and the second output value corresponding thereto and store it in the database, and store it in the database by accumulating in the block chain network. The blockchain network can determine the relative academic achievement level of a student based on the grouped and accumulated stored data. In addition, the blockchain network may analyze each type of a plurality of learning contents and classify them according to the degree of similarity. Each network storage device of the blockchain network can be grouped to create and verify blocks formed based on accumulated and stored information. In addition, the blockchain network may analyze each type of a plurality of learning problems and classify and store them according to similarity with other types. The type classification of the learning problems may be, for example, classifying a learning progress index corresponding to a learning problem having a high degree of similarity. The details of the blockchain network will be described later in detail with reference to FIG. 4 .

According to an embodiment, the control device may acquire the relative academic achievement level from the blockchain network. According to an embodiment, the relative academic achievement level may be acquired by databaseizing grouped and cumulatively stored information, and analyzing information having the same learning progress index among them. For example, the relative academic achievement level may be obtained by selecting one group from among the groups of the predetermined academic achievement level (eg, best, upper, upper middle, middle, lower middle, lower, lowest groups), but is not limited thereto. it is not

According to an embodiment, the control device extracts a type with the highest similarity to the type of learning to be performed next from among the previously performed learning contents from the database of the block chain network, and then refers to the relative academic achievement level corresponding to the extracted type Thus, it is possible to determine the difficulty of the curriculum of the next learning content. The difficulty level of the curriculum may be determined by selecting a difficulty level of one group from among groups of predetermined difficulty levels (eg, advanced, basic, and basic groups), but is not limited thereto. For example, if the next learning content to be performed is a 'probability' type of mathematics, the blockchain network may have determined that the 'number of cases' type among the already performed learning content has the highest similarity. The control device may determine the curriculum of the 'probability' type as the 'basic' difficulty level, for example, referring to that the student's relative academic achievement level corresponding to the 'number of cases' type is 'low'. However, in the above example, the type of learning content is exemplified as a partial unit of a mathematics subject, but the type of learning content is not limited to a subject or unit and may be further subdivided, and similarities between various types across a plurality of subjects may be considered. there is.

According to an embodiment, the control device may determine the problem type and difficulty of the next unit to be learned based on whether the problem is correct or not, the type and difficulty of the problem (S108).

Specifically, the control device determines that the learning progress of the first student is the first unit, and based on the correct answer to the first question included in the first unit, the type and difficulty of the first problem, learning next to the first unit You can determine the type and difficulty of the second problem to be included in the second unit to be completed.

For example, if the type of the first problem is a problem type for determining mental arithmetic, the difficulty of the first problem is intermediate, and the correct answer of the first problem is correct, the control device determines the type of the second problem with mental arithmetic can be determined as the problem type for judging , and the difficulty level of the second problem can be determined as advanced.

In addition, when the type of the first problem is a problem type for determining mental arithmetic, the difficulty of the first problem is intermediate, and the correct answer to the first question is incorrect, the control device determines the mental arithmetic power as the type of the second problem It can be determined as the type of problem to be solved, and the difficulty level of the second problem can be determined as low.

In addition to the above, through various embodiments, the type and difficulty of the second question may be determined based on whether the first question is correct, and the type and difficulty of the first question.

The control device may provide the curriculum of the determined difficulty when the student performs the learning of the next learning content (S109). The control device may store information on the difficulty of the determined curriculum of the next learning content in the database, and may provide the curriculum of the determined difficulty when receiving an input of starting the next learning from the student.

Specifically, when the correct answer to the first question included in the first unit is confirmed as the correct answer, the control device may determine that the learning progress of the first student indicates that the learning of the first unit is completed. Here, the first problem may be the last problem among the problems included in the first unit.

When the learning progress of the first student confirms that the learning of the first unit is completed, the control device may provide a learning curriculum according to the type and difficulty of the second problem included in the second unit.

2 is a schematic diagram for explaining a convolutional neural network used to provide a learning curriculum in an embodiment.

The convolutional neural network 200 according to an embodiment takes as an input an input vector generated by encoding an image photographed of a question paper that has completed problem solving and scoring during learning, and the type of each problem extracted from the image (230) and an estimate of whether the answer is correct 240 may be output. The encoding according to an embodiment may be performed by storing color information for each pixel of the image in the form of a digitized data sheet, and the color information may include RGB color, brightness information, and saturation information of one pixel. However, the present invention is not limited thereto.

According to an embodiment, the convolutional neural network 200 may include a feature extraction neural network 210 and a classification neural network 220 . The feature extraction neural network 210 may perform the task of separating each problem from the image taken from the question paper, and the classification neural network 220 classifies each problem by type therefrom, and performs the task of determining whether the answer is correct. can

According to an embodiment, the feature extraction neural network 210 stacks the input signal by sequentially stacking a convolutional layer and a pooling layer. The convolution layer contains convolution operations, convolution filters, and activation functions. The calculation of the convolution filter is adjusted according to the matrix size of the target input, but a 9X9 matrix is generally used. The activation function may generally use a ReLU function, a sigmoid function, a tanh function, and the like, but is not limited thereto. The pooling layer is a layer that reduces the size of the input matrix, and uses a method of extracting representative values by tying pixels in a specific area. In general, the average value or the maximum value is often used for calculation of the pooling layer, but is not limited thereto. The operation is performed using a square matrix, which is usually a 9x9 matrix. The convolutional layer and the pooling layer are repeated alternately until the corresponding input becomes small enough while maintaining the difference.

The classification neural network 220 according to an embodiment may classify each problem separated by the feature extraction neural network 210 by type, and estimate whether the answer is correct. When the classification neural network 220 determines whether each question is correct, it may determine whether the answer is correct by extracting the handwriting information scored on the question paper, but is not limited thereto.

According to an embodiment, the classification neural network 220 has a hidden layer and an output layer. The convolutional neural network 200 of the present invention generally includes five or more hidden layers, and 80 nodes of each hidden layer can be designated, but it is also possible to set more than that in some cases. The activation function of the hidden layer uses a ReLU function, a sigmoid function, a tanh function, and the like, but is not limited thereto. The number of output layer nodes of the convolutional neural network 200 may be 50 in total.

According to an embodiment, the control device may apply an input value to the convolutional neural network 200 to obtain a first output value from the convolutional neural network 200 .

The control device may extract the image as color information for each pixel, encode the color information, and input it to the convolutional neural network 200 .

Based on the input of the convolutional neural network 200 , the control device may separate the first question from the image, and extract the items of the first question and handwriting scored on the first question separately.

Specifically, on the question paper, a boundary line for separating the part where the problem is displayed for each problem is displayed, and the control device recognizes the boundary line surrounding the first problem in the image to divide the area of the first problem, We can isolate the first problem from the image by identifying the problem area. When the first question is separated, the control device may separate and extract the questions of the first question and the handwriting marked on the first question in the area of the first question.

The control device may distinguish the item from the handwriting through the color information in classifying the item and the handwriting of the first question. For example, if the question in the first question is printed on the question paper with the first color value, the first brightness value, and the first chroma value, the control device analyzes the image to find the blank area where the text is not displayed and the text. The displayed text area can be distinguished, and color information is checked for each pixel of the text area, and the checked color value, brightness value, and saturation value all match the first color value, the first brightness value, and the first saturation value. An area may be divided into an area in which an item is written, and an area in which any one of the first color value, the first brightness value, and the first chroma value does not match may be divided into the area in which the handwriting is written. For example, the control device may classify a text area marked in black as items and extract a text area marked in red as handwriting.

The control device may recognize a keyword from the items of the first question, and classify the type of the first question through the recognized keyword.

For example, the control device divides the text area displayed in black in the area of the first problem into items, identifies the characters included in the item area to recognize the keyword from the item, and determines the mental arithmetic power of the recognized keyword. If it is confirmed as a used keyword, the type of the first problem may be classified into a problem type for determining mental arithmetic.

The control device may check the color, brightness, and saturation of the handwriting scored on the first question, and determine whether or not the first question is correct according to whether the first color value, the first brightness value, and the first saturation value are preset. .

Specifically, the first color value, the first brightness value, and the first saturation value are preset, the control device checks color information for each pixel of the writing area, and the checked color value, the brightness value, and the saturation value are the first values. If it is confirmed that the color value, the first brightness value, and the first saturation value all match, the first question is determined as the correct answer, and any one of the first color value, the first brightness value, and the first saturation value does not match If confirmed, the first question may be determined as an incorrect answer.

For example, when the handwriting marked on the first question is blue, the control device may determine the first question as a correct answer, and when the handwriting marked on the first question is red, the control device may determine the first question as an incorrect answer.

As for the output of the convolutional neural network 200 according to an embodiment, the top 25 output layer nodes among 50 output layer nodes may indicate the type 230 of each problem, and the lower 25 output layer nodes are the top 25 output layer nodes, respectively. It is possible to indicate whether the correct answer 240 corresponds to . The method of mapping the top 25 output layer nodes and the bottom 25 output layer nodes is a method of matching the top nth output layer node and the bottom nth output layer node, and the nth output layer node in total corresponds to the 25+nth output layer node in total , where n is an integer of any one of 1 to 25. For example, a first output layer node may correspond to a 26th output layer node, a second output layer node may correspond to a 27th output layer node, a 25th output layer node may correspond to a 50th output layer node, and so on. The type 230 of each problem may be output as code information corresponding to the type, but is not limited thereto. An output layer node having no output value among 50 output layer nodes of the convolutional neural network 200 may output the number '0' as its output value. Nodes including this number '0' among the top 25 output layer nodes are considered to have no corresponding problem type and may be excluded from calculations in future neural network calculations. If there are 25 types of classified problems, the remaining problems may be processed automatically after all output values generated in advance are processed.

According to an embodiment, the convolutional neural network 200 receives the first learning signal generated by the user's input when the user discovers that the classification of the problem type according to the convolutional neural network 200 or whether the correct answer is incorrectly estimated. can learn through The first learning signal according to an embodiment is created based on the error between the correct answer and the output value, and in some cases, SGD using delta, a batch method, or a method following a backpropagation algorithm may be used. According to the first learning signal, the convolutional neural network 200 performs learning by modifying the existing weights, and in some cases, momentum may be used. A cost function may be used to calculate the error, and a cross entropy function may be used as the cost function.

According to an embodiment, the control device may output an output value including the type of the first problem and whether the answer is correct to a total of 50 output layer nodes. At this time, the top 25 output layer nodes out of a total of 50 output layer nodes indicate the type of the first problem, and the lower 25 output layer nodes out of a total of 50 output layer nodes are nodes corresponding to the top 25 output layer nodes. can direct

For example, the first output layer node may indicate the type of mental arithmetic, the 26th output layer node may indicate whether the correct answer to the mental arithmetic power determination problem, the second output layer node may indicate problem comprehension, and the 27th output layer node can indicate whether or not the answer to the problem comprehension judgment problem is correct.

The control device may output a number '0' to a node indicating that the top 25 output layer nodes are not of the type of the first problem. For example, if the type of the first problem is a problem type that determines mental arithmetic, and the first output layer node indicates the mental arithmetic power type, the number '0' is assigned to the remaining nodes except for the first output layer node among the top 25 output layer nodes. ' can be printed.

3 is a schematic diagram illustrating a neural network according to an embodiment.

The neural network 300 according to an embodiment is a first output value including an estimate of the type 230 and correct answer 240 of each problem obtained from the output layer of the convolutional neural network, and is stored in advance in the database 330. The auxiliary estimated value 310 may be input. The auxiliary estimate value 310 stored in the database 330 may be accumulated data of evaluations of a plurality of types of questions and whether or not the answer is correct. More specifically, the input of the neural network 300 is the output of the convolutional neural network, which is a matrix of 1 row and 50 columns including the type 230 and correct answer 240 of each problem, and a matrix of 1 row and 500 columns of the auxiliary estimate 310. The matrix may consist of a total of 1 row and 550 columns in which the matrix is combined, but is not limited thereto. As an example, a total of 550 input layer nodes as inputs to the neural network 300 may be formed, and an output layer node may include 5 nodes.

The neural network 300 according to an embodiment is the output of the convolutional neural network based on the information included in the auxiliary estimation value 310 stored in the database 330. The type 230 of each problem and the correct answer 240 can be evaluated for The neural network 300 may detect an abnormal signal based on accumulated data of evaluations of the types of a plurality of questions stored in the database 330 and whether correct answers are correct.

The neural network 300 evaluates the type and correct answer of the first problem identified through the first output value, compares the type and correct answer of the first problem with a plurality of types and correct answers stored in the database, and accumulates the evaluation data An abnormal signal may be detected based on the

When an abnormal signal is detected, the neural network 300 modifies the type of the first problem to generate a second output value indicating the type of the corrected first problem and whether the correct answer is correct, and when no abnormal signal is detected, the first A second output value indicating the type of the first uncorrected question and whether the answer is correct may be generated without modifying the type of the question.

According to an embodiment, the neural network 300 of the present invention includes three or more hidden layers, and 100 nodes of each hidden layer are designated, but more may be designated in some cases. The activation function of the hidden layer uses a ReLU function, a sigmoid function, a tanh function, and the like, but is not limited thereto. The output layer node may be 5 nodes, but is not limited thereto. Each output layer node may include a type 230 of each problem, a correct answer 240, and an abnormal signal. The output value of the output layer node changes every 0.02 seconds, and the corresponding output value for each problem type can be extracted.

According to an embodiment, the neural network 300 uses the output value of the output layer node to represent the output values for a total of 25 types of problems at 0.02-second intervals, that is, based on the output values collected for 0.5 seconds. The modified second output value 320 may be displayed. The neural network 300 may modify output values related to at least one of the type 230 and the correct answer 240 of each problem included in the existing first output value, and is not included in the first output value made from the acquired image. The output values may be preserved as they are.

According to an embodiment, when the user determines that the classification of the learning type according to the neural network 300 or the determination of the correct answer is incorrect, a second learning signal for learning of the neural network 300 may be generated. The neural network 300 may perform learning based on the generated second learning signal. The second learning signal according to an embodiment is created based on the error between the correct answer and the output value, and in some cases, SGD using delta, a batch method, or a method following a backpropagation algorithm may be used. According to the second learning signal, the neural network 300 performs learning by modifying the existing weights, and may use momentum in some cases. A cost function may be used to calculate the error, and a cross entropy function may be used as the cost function.

4 is a schematic diagram for explaining a blockchain network in an embodiment. The block chain network 400 groups (not shown) the identifier, the learning progress index, and the second output value obtained from the neural network and accumulates and stores the blocks 410 and each block 410 according to the update order in chronological order. It may be a private blockchain network including connecting chains 420 and a plurality of network storage devices 430 for storing each block chain.

According to an embodiment, the blocks 410 of the block chain network 400 may group the identifier, the learning progress index, and the second output value and accumulate and store them according to the update order, and a new block 410 is generated every predetermined period. can be The already generated content of the block 410 is stored in all network storage devices 430 and has a structure that cannot be changed unless a majority of the content is changed within a time period.

For example, assuming that each network storage device has one block in the blockchain network 400 having a total of three network storage devices 430a, 430b, 430c, the majority within a limited time, that is, two or more If the content of the block cannot be changed, each network storage device can re-change the block with a content different from the majority through verification to indicate the same value as the blocks of the majority. In this description, it has been described that there are three network storage devices 430, but the number of network storage devices 430 included in the block chain network 400 is not limited thereto, and can actually reach tens to hundreds of thousands of higher security. can have gender.

Chains 420 according to an embodiment may be configured with hash values. The chains 420 may connect the blocks 410 in chronological order, and at this time, each block 410 may be connected using a hash value. The block 410 may store the database and hash values of the database, the previous header, and the header of the current block. Here, the header of the current block functions as a previous header in the next block, so that each block 410 can be organically connected. In addition, the hash value is transformed into a completely different form when the contents of the block 410 are changed even a little, thereby effectively preventing an attempt to change the contents of the database. Since the header of the current block 410 becomes a hash value of the total consensus including the database, the hash value of the database, and the previous header, it is difficult to succeed in an attempt to compromise security by modifying the database and the hash value. Because, when the database and hash value are modified, the contents of the header are also changed, and accordingly, the contents of the previous header entering the next block 410 are also changed. The previous header of 410 may be replaced again. That is, all subsequent blocks 410 must be hacked. Therefore, the security of the block chain can be further increased with the chain 420 through the hash value.

According to an embodiment, the network storage devices include a first network storage device 430a including an institution for managing a plurality of students, and a second network storage device including an institution for obtaining statistics of academic achievement for a plurality of students ( 430b), a third network storage device 430c including a personal information protection organization, and an Internet connection network 440 connecting each of the network storage devices 430a, 430b, and 430c. However, the number and types of network storage devices are not limited to those listed above, and the number and types may be determined as necessary.

According to an embodiment, the first network storage device 430a may include an institution that manages a plurality of students. There may be one or more institutions managing a plurality of students, and the number of first network storage devices 430a may be determined according to the number of storage devices used by each institution. The first network storage device 430a may determine the relative academic achievement for each type already learned of the students to be managed through the block chain network 400 , and may manage the academic achievement results of newly input students.

According to an embodiment, the second network storage device 430b may include an institution that obtains statistics of academic achievement for a plurality of students. Preferably, the second network storage device 430b collects information on the academic achievement of a plurality of students according to the type of each learning field, and calculates the average and/or standard deviation thereof to calculate the relative academic achievement level of each student. can provide indicators for judging In addition, the second network storage device 430b may present a criterion for determining the similarity by type by analyzing the contents of each of the learning fields, and may provide an index regarding the degree of similarity between the learning fields according to the similarity for each type.

According to an embodiment, the third network storage device 430c may include a personal information protection organization. A privacy organization is any organization that can manage student identification identifiers or personal information regarding academic achievement, and there may be one or more privacy organizations, depending on the number of storage devices used Accordingly, the number of third network storage devices 430c may be determined. The method and apparatus for providing a learning curriculum can make big data of information about students' academic achievement through the blockchain network 400 and use enhanced security.

The Internet connection network 440 is a connection network including wired, wireless, optical cable technology, etc., and includes a local area network (LAN), a wireless local area network (LAN), a wide area network (WAN), a personal area network (PAN), and the like. It may include, but is not limited to.

5 is a diagram schematically showing the configuration of a system for providing a lecture service according to an embodiment.

Referring to FIG. 5 , the system according to an embodiment may include a control device 510 capable of communicating with each other through a communication network, a plurality of student terminals 520 , and a plurality of instructor terminals 530 .

First, the communication network may be configured regardless of its communication mode, such as wired and wireless, and may be implemented in various forms so that communication between a server and a server and communication between a server and a terminal are performed.

The control device 510 may be a server owned by a person providing a service using the control device 510; It may be a cloud server; It may also be a peer-to-peer (p2p) set of distributed nodes.

Each of the plurality of student terminals 520 and each of the plurality of lecturer terminals 530 may be implemented as a computing device having a communication function, for example, a mobile phone, a desktop PC, a laptop PC, a tablet PC, and a smart phone. and the like, but is not limited thereto, and may be implemented as various types of communication devices that can be connected to an external server.

The plurality of student terminals 520 are terminals used by students who want to receive a lecture, and include a first student terminal 521 used by a first student, a second student terminal 522 used by a second student, and the like. The plurality of lecturer terminals 530 are terminals used by lecturers who want to give a lecture, and the first lecturer terminal 531 used by the first lecturer and the second lecturer terminal 532 used by the second lecturer ) and the like.

Hereinafter, for convenience of explanation, the operations of the first student terminal 521 and the first instructor terminal 531 will be mainly described, but the operation of the first student terminal 521 in other student terminals such as the second student terminal 522 . Of course, the operation of the first lecturer terminal 531 may be performed instead of the second lecturer's terminal 532 and other lecturer terminals such as the terminal 532 .

The first student terminal 521 may be a smartphone, tablet, computer, or the like. An application developed or distributed by a person or organization providing a service using the control device 510 may be installed in the first student terminal 521 . The first student terminal 521 may be linked with a corresponding application to perform functions related to lecture matching and lecture progress.

The first lecturer terminal 531 may be a smartphone, tablet, computer, or the like. The first lecturer terminal 531 may include an authentication means for acquiring body information, such as a microphone or a camera. An application developed or distributed by a person or organization providing a service using the control device 510 may be installed in the first lecturer terminal 531 . The first lecturer terminal 531 may perform functions related to lecture matching and lecture progress by interworking with a corresponding application.

The first student terminal 521 may receive membership registration information input by the student. Here, a student means an individual who wants to receive lecture services. Membership information includes your email address; password; Mobile Phone Number; name; Payment card information may be included.

The control device 510 may receive membership registration information from the first student terminal 521 . The control device 510 may store the received member registration information to form a database. Through this, the control device 510 may create a first student account corresponding to the first student. Here, the student account means an electronic account (user) of an individual who wants to receive a lecture service. After the creation of the student account is completed, the control device 510 may notify the first student terminal 521 that the membership registration is complete.

After the first student's membership registration is completed, the first student terminal 521 may receive login information input by the first student. The login information may consist of the email and password entered by the first student when registering for membership; It may be an event of clicking or touching an auto-login button or widget, etc.; It may be body information acquired through an authentication means capable of acquiring body information, such as a microphone or a camera.

The control device 510 may receive login information from the first student terminal 521 . The control device 510 may inquire the first student account corresponding to the received login information in the member database. When the first student account corresponding to the received login information exists, the control device 510 allows the first student account to log in to the first student terminal 521 while logging in to the first student terminal 521 is performed. You can tell it's done.

Next, the first lecturer terminal 531 may receive membership registration information input by the lecturer. Herein, a lecturer means an individual or an organization that conducts lectures as a business or a side job. Membership information includes your email address; password; Mobile Phone Number; name; It may include deposit account information and the like.

The control device 510 may receive membership registration information from the first lecturer terminal 531 . The control device 510 may store the received member registration information to form a database. Through this, the control device 510 may create a first lecturer account corresponding to the first lecturer. Here, the lecturer account means an electronic account (user) of an individual or group who performs lectures as a business or a side job. After the lecturer account creation is completed, the control device 510 may notify the first lecturer terminal 531 that the membership registration is complete.

After the lecturer's membership registration is completed or during the membership registration process, the first lecturer terminal 531 may receive profile information of the first lecturer. Profile information includes the instructor's profile picture; date of birth; gender; Education; career, etc. In addition, the profile information may include a category in which the lecturer can teach. The lectureable category may include subjects, units, and the like. The control device 510 may set the account of the first lecturer to receive a matching lecture within the available lecture category input by the first lecturer.

The control device 510 may receive profile information from the first lecturer terminal 531 . The control device 510 may update the profile information of the first lecturer account based on the profile information. The control device 510 may change the available lecture category of the first lecturer account based on the updated profile information. The control device 510 may set the account of the first lecturer so that the lecture can be matched within the changed lecture available category. The control device 510 may notify the first lecturer terminal 531 that the profile information has been updated.

After the lecturer's membership registration is completed, the first lecturer terminal 531 may receive login information input by the first lecturer. The login information may consist of an email and password entered by the first lecturer when signing up for membership; It may be an event of clicking or touching an auto-login button or widget, etc.; It may be body information acquired through an authentication means capable of acquiring body information, such as a microphone or a camera.

The control device 510 may receive login information from the first lecturer terminal 531 . The control device 510 may inquire the account of the first lecturer corresponding to the received login information from the member database. If the first lecturer account corresponding to the received login information exists, the control device 510 allows the first lecturer account to log in to the first lecturer terminal 531 while logging in to the first lecturer terminal 531 is performed. You can tell it's done.

6 is a view for explaining a lecturer matching process according to an embodiment.

Referring to FIG. 6 , first, the first lecturer terminal 531 may set the first lecturer account logged into the first lecturer terminal 531 as a lecture standby mode ( S601 ). The first lecturer terminal 531 may include a lecturer authentication means. The instructor authentication means may be a camera capable of photographing the skin condition of the instructor.

Next, the control device 510 may receive a notification that the first lecturer account logged into the first lecturer terminal 531 is set to the lecture standby mode (S602). The control device 510 is an artificial neural network for diagnosing the health condition of the lecturer, including whether the lecturer is logged in and whether the lecturer is set to a lecture standby mode through the lecturer's skin information obtained from the camera of the first lecturer terminal 531 . You can create input. Specific details regarding the artificial neural network will be described later with reference to FIG. 10 .

Subsequently, the control device 510 may generate the lecture waiting list by including the account of the first lecturer logged into the first lecturer terminal 531 in the lecture waiting list ( S603 ). Lecture waiting lists include lecturers' available courses; lectureable units; And, there may be a plurality of each for each available region range. The control device 510 may include a subject in which the first instructor using the first instructor account logged into the first instructor terminal 531 can teach; lectureable units; and the account of the first lecturer logged in to the first lecturer terminal 531 based on the range of the lectureable area, etc. may be included in the corresponding lecture waiting list.

That is, the control device 510 receives a notification from the lecturer terminals 530 that the lecturer accounts logged into each of the lecturer terminals are set to the lecture standby mode, and based on the lecturer accounts set to the lecture standby mode, the lecture standby By including the lecturer account set to the mode in the lecture waiting list, it is possible to create a lecture waiting list.

Meanwhile, the first student terminal 521 may receive a lecture request of the first student account logged into the first student terminal 521 ( S604 ). For example, after completing learning of the first unit, the first student may input a lecture request for a second unit to be learned after the first unit. Information included in a course request includes course subjects and units; lecture date; lecture time; estimating lectures per unit of time; and the specific content of the lecture; It may include lecture venues, and the like. At this time, if the lecture is conducted outdoors, and the outdoor environmental risk stage at the lecture time and the lecture location is expected to be included in the predetermined outdoor environmental risk stage, the unit corresponding to the predetermined outdoor environmental risk stage is included in the estimate of the lecture per unit time. Hourly incentives may be added. A detailed description related thereto will be described later with reference to FIG. 8 .

Next, the control device 510 may receive a lecture request of the first student account logged into the first student terminal 521 from the first student terminal 521 ( S605 ). The control device 510 may include lecture subjects and words included in the lecture request; And based on the lecture location, etc., it is possible to select at least one or more waiting lists for lectures with a high degree of agreement with the lecture request. For example, the control device 510 may receive a request for a lecture on the second unit from the first student terminal 521 , and select a waiting list for a lecture in which the second unit is set as a lectureable unit.

Subsequently, the control device 510 may transmit a lecture request to the lecturer terminals 530 in which each of the lecturer accounts included in the lecture waiting list are logged in ( S606 ). At this time, if the lecture is conducted outdoors, and it is expected that the outdoor environment risk step at the time of the lecture and the lecture place is included in the predetermined outdoor environment risk step, the lecturer terminals 530 are the outdoor environment where the lecture requires attention. may be notified, and incentives per unit time corresponding to a predetermined outdoor environmental risk level may be provided. A detailed description related thereto will be described later with reference to FIG. 8 .

Thereafter, the first lecturer terminal 531 may receive an input of acceptance or rejection of the first lecturer account for the received lecture request ( S607 ). When the first lecturer using the first lecturer terminal 531 accepts the lecture request, the first lecturer terminal 531 may transmit the lecture request acceptance to the control device 510 , and the first lecturer terminal 531 . When the first lecturer who uses , rejects the lecture request, the first lecturer terminal 531 may transmit the lecture request rejection to the control device 510 . For example, the first lecturer using the first lecturer terminal 531 may have accepted the lecture request.

Next, the control device 510 may receive from the first lecturer terminal 531 whether the lecture request of the first lecturer account is accepted ( S608 ). If the lecture request acceptance is not three or more, the control device 510 may treat the lecture matching as incomplete, and may additionally ask whether the lecture request is accepted by sending the lecture request to the lecturer terminals of the other lecture waiting lists. there is.

That is, the control device 510 may receive whether to accept the lecture request from the lecturer terminals in which each of the lecturer accounts included in the lecture waiting list are logged in.

Subsequently, the control device 510 may generate a lecture request acceptance list based on the lecture waiting list and whether the lecture request has been accepted ( S609 ). That is, the control device 510 may generate a lecture request acceptance list by extracting lecturer accounts that have accepted the lecture request of the first student terminal 521 from among lecturer accounts included in the lecture waiting list. At this time, if the lecture request acceptance of the lecturer accounts included in one lecture waiting list is not three or more, the lecture matching may be processed as incomplete, and the first student terminal ( 521), lecturer accounts that have accepted the lecture request may be included in the lecture request acceptance list.

Thereafter, the control device 510 may transmit the lecture request acceptance list to the first student terminal 521 ( S610 ). The lecture request acceptance list may be configured to check the profiles of lecturers who have accepted the lecture request.

Next, the first student terminal 521 may receive an input of selecting a lecturer of the student account (S611). When the first student using the first student terminal 521 selects a specific lecturer from the lecture request acceptance list, the first student terminal 521 may transmit lecturer selection information to the control device 510 . For example, the first student who uses the first student terminal 521 may select the first lecturer who uses the first lecturer terminal 531 .

Subsequently, the control device 510 may receive lecturer selection information from the first student terminal 521 ( S612 ). Meanwhile, when the first student account does not select a lecturer, the control device 510 may process the lecture matching incomplete. In this case, the control device 510 terminates the lecture matching, or retransmits or newly transmits the lecture request of the first student terminal 521 to the first lecturer terminal 531 in which the lecturer accounts of one or more lecture waiting lists are logged in. there is.

Thereafter, the control device 510 may transmit the lecturer selection information to the first lecturer terminal 531 in which the first lecturer account corresponding to the lecturer selection information is logged in ( S613 ). That is, since the first student using the first student terminal 521 selected the first instructor using the first instructor terminal 531 , in this case, the control device 510 controls the first student using the first instructor selection information. The lecturer selection information is transmitted to the first lecturer terminal 531 in which the lecturer account is logged in.

Next, the first lecturer terminal 531 may receive an input of accepting or rejecting the lecturer matching for the received lecturer selection (S614). When the first lecturer using the first lecturer terminal 531 accepts the lecturer matching, the first lecturer terminal 531 may transmit the lecturer matching acceptance to the control device 510 ; When the lecturer using the first lecturer terminal 531 rejects the lecturer matching, the first lecturer terminal 531 may transmit the lecturer matching rejection to the control device 510 . For example, the lecturer using the first lecturer terminal 531 may select to accept the lecturer matching.

Subsequently, the control device 510 may receive whether or not to accept the instructor matching from the first instructor terminal 531 in which the first instructor account corresponding to the instructor selection information is logged in ( S615 ). The control device 510 may record that the first instructor account logged in to the first instructor terminal 531 and the first student account logged in the first student terminal 521 match each other, and specific details of lecture matching items can be recorded. In addition, the control device 510 may record the number of times of acceptance or rejection of matching of the first instructor account logged in to the first instructor terminal 531 , and the instructor account with a high number of matching acceptance receives benefits, and the number of matching rejections is high. Instructor account can apply a penalty.

Thereafter, the control device 510 may transmit whether the lecturer matching is accepted to the first student terminal 521 . Through this, the first student using the first student terminal 521 can check whether or not the first student is matched with the first lecturer selected by the student, and thereafter, a full-fledged discussion and actual lecture for the course of the lecture can be conducted.

7 is a view for explaining a course of a lecture according to an embodiment.

Referring to FIG. 7 , first, the control device 510 checks whether to accept the lecturer matching received from the first lecturer terminal 531 , and as a result of accepting the lecturer matching, the first lecturer selection information corresponding to the first lecturer matching information The lecturer account and the first student account may be matched, and a chat space may be created between the matched first lecturer account and the first student account ( S701 ).

The first student terminal 521 and the first instructor terminal 531 may access a chat space between the first instructor account and the first student account ( S702 ). Through this, the first lecturer who uses the first lecturer terminal 531 and the first student who uses the first student terminal 521 may specify a lecture schedule while chatting. The first student and the first lecturer can chat with the lecture date; lecture time; lecture duration; and lecture locations can be discussed or changed. The control device 510 may include a lecture date between the first student and the first instructor; lecture time; lecture duration; If the lecture location is changed, it can be recorded.

Next, the first lecturer terminal 531 may receive an input that the lecture has started (S703). The control device 510 may determine that the lecture start input does not match the lecture date, lecture time, and lecture location according to the lecture request; If the changed lecture date, lecture time, and lecture location according to the chat contents between the student and the lecturer do not match, the lecture start input may be treated as a false input.

Subsequently, the control device 510 may receive a notification that the lecture has started from the first lecturer terminal 531 to which the first lecturer account is logged ( S704 ). For example, when the first student requests a lecture on the second unit, a notification that the lecture on the second unit has started may be received.

Next, the control device 510 may transmit a lecture start approval request to the first student terminal 521 (S705).

Next, the first student terminal 521 may receive a lecture start approval input (S706).

Next, the control device 510 may receive a lecture start approval from the first student terminal 521 (S707).

Next, the control device 510 may record that the lecture has started, and set the time when the lecture start approval is received as the lecture start time ( S708 ). The control device 510 may start calculating the cost after the lecture is recorded as starting.

After the lecture between the student and the lecturer is made, the student should notify the control device 510 that the lecture is over. To this end, the first student terminal 521 may receive an input that the lecture has ended (S709).

Next, the control device 510 may receive a notification that the lecture has ended from the first student terminal 521 (S710). For example, when the first student requests a lecture on the second unit, a notification that the lecture on the second unit has ended may be received.

Next, the control device 510 may transmit a lecture termination approval request to the first lecturer terminal 531 to which the first lecturer account is logged ( S711 ).

Next, the first lecturer terminal 531 may receive a lecture termination approval input (S712).

Next, the control device 510 may receive a lecture termination approval from the first lecturer terminal 531 (S713).

Thereafter, the control device 510 may record that the lecture has ended, and set the time at which the lecture end approval is received as the lecture end time ( S714 ).

The control device 510 may calculate and determine the cost to be paid by the first student account and the cost to be paid to the first lecturer account based on the lecture start time and the lecture end time.

The control device 510 performs a lecture start recording based on the lecture start input received from the first lecturer by the first instructor terminal 531, and the first student terminal 521 receives the lecture end input from the first student. By performing the record of the end of the lecture based on the That is, when the start of the lecture is received from the first student terminal 521 instead of from the first lecturer terminal 531 , the student may intentionally input the lecture start late, and in this case, the control device 510 calculates the cost The lecturer is at a loss because it starts late. In addition, when the end of the lecture is received from the first lecturer terminal 531 instead of from the first student terminal 521 , the lecturer may intentionally input the end of the lecture early, and in this case, the control device 510 calculates the cost Students are at a loss because they finish late. Therefore, in order for both the student and the lecturer to not suffer unreasonable damage and to reasonably calculate the cost, the control device 510 records the start of the lecture based on the lecture start input received from the first lecturer terminal 531, and the first student terminal ( 521), the end of the lecture is recorded based on the input of the end of the lecture.

The lecture between the lecturer and the student according to an embodiment may be conducted indoors or outdoors. In particular, in the case of a lecture related to an exercise subject, the lecture may be frequently conducted outdoors. If the lecture is held outdoors, the health condition of the lecturer may deteriorate due to the outdoor environment requiring attention.

Specifically, if the lecturer frequently lectures outdoors on a day with a lot of ultraviolet rays, there is a risk of leading to skin diseases of the lecturer. Therefore, there may be a phenomenon in which lecturers avoid conducting lectures in an outdoor environment on a day when the outdoor environment is not good. In this case, there is a high possibility that a smooth lecture matching service will not be provided, which includes a person or an organization that provides a lecture matching service through the control device 510; student; and harm to all lecturers.

Therefore, when the lecturer performs a lecture in an outdoor environment that requires attention, a compensation system for the lecturer is required. In addition, when it is determined that the health of a lecturer has deteriorated due to an outdoor environment requiring attention, a method of restricting the lecturer from working in an outdoor environment requiring attention is required. In addition, since each person is affected by health conditions and outdoor environments differently, lecturers who are confident in their health or who are judged to be insignificant to the extent of being affected by the outdoor environment can actively teach even in outdoor environments that require attention. There is a need for a way to support this.

In order to solve or improve the above problems, the control device 510 may perform the methods described with reference to FIGS. 8 to 9 . Through this, the control device 510 may provide an incentive system for the lecturer when the lecturer performs a lecture in an outdoor environment requiring attention. In addition, when it is determined that the health of the lecturer is deteriorated due to the outdoor environment requiring the lecturer's attention, the control device 510 may protect the health of the lecturer by limiting the ability to work in the outdoor environment requiring the lecturer's attention.

8 is a flowchart illustrating an incentive providing method according to an embodiment.

Referring to FIG. 8 , first, the control device 510 may check a correspondence relationship between each preset staged outdoor environmental risk stage and an incentive per unit time ( S801 ).

To this end, the control device 510 may pre-stage the outdoor environment risk stage. Specifically, the outdoor environmental risk stage may be a WHO UV index or other values indicating the degree of outdoor ultraviolet rays divided by numerical values. The control device 510 sets the outdoor environmental risk level to “safe” if the WHO UV index is less than 8.0; If the WHO UV Index is greater than or equal to 8.0 and less than 11.0, the outdoor environmental risk level is classified as “dangerous”; If the WHO UV index is above 11.0, the outdoor environmental risk level can be pre-staged as “very dangerous”.

The control device 510 may set a corresponding relationship between each outdoor environmental risk level and an incentive per unit time. The control device 510 sets the incentive per unit time to 0% if the WHO UV index is less than 8.0 and the outdoor environmental risk level is “safe”; If the WHO UV Index is greater than or equal to 8.0 and less than 11.0 and the outdoor environmental risk level is “dangerous”, the incentive is 10% per unit of time; If the WHO UV index is 11.0 or higher and the outdoor environmental risk level is “very dangerous”, the incentive can be set at 20% per unit time.

In addition, the control device 510 may pre-stage the outdoor environmental risk stage based on various values indicating the degree of outdoor ultraviolet rays, and may set various correspondence relationships between the pre-staged outdoor environmental risk stage and the incentive per unit time. there is.

Next, the control device 510 may acquire whether an environment in which a lecture is performed is outdoors ( S802 ).

Specifically, the control device 510 may acquire, based on the lecture request received from the first student terminal 521 , whether the environment in which the first lecturer using the first lecturer account performs a future lecture is outdoors. there is. In other words, since the lecture request input through the first student terminal 521 includes a lecture place and a lecture category, the control device 510 receiving the lecture request provides information such as a lecture place and a lecture category included in the lecture request. Based on , it may be obtained whether the environment in which the first lecturer using the first lecturer account performs a future lecture is outdoors.

Next, the control device 510 may set an expected incentive per unit time based on the predicted outdoor environment value at the place and time at which the lecture is held ( S803 ). To this end, the control device 510 may be linked with a weather information provider through an API to receive the UV forecast. The control device 510 may request the weather information provider to provide the UV concentration at the lecture location and the lecture time according to the lecture request through the API. The control device 510 may set an expected incentive per unit time based on the received predicted value of the UV concentration at the lecture location and the lecture time. The control device 510 may inform the first student terminal 521 and the first lecturer terminal 531 of the expected incentive per unit time.

For example, the control device 510 may be linked with a weather information provider through an API to receive a WHO UV index forecast. The control device 510 may request the weather information provider to provide the WHO UV index at Namsan Tower (the lecture place) of the upcoming Arbor Day (class time) according to a lecture request to the weather information provider through the API. The control device 510 may set the expected incentive per unit time to 10% based on the WHO UV index at Namsan Tower on the upcoming Arbor Day of 9.0. The control device 510 may notify the first student terminal 521 and the first instructor terminal 531 that the expected incentive per unit time is 10% when a lecture is held at Namsan Tower on the upcoming Arbor Day.

Through this, the student account can check the expected incentives per unit time of the lecture location and lecture time according to the lecture request, so that the final cost to be provided to the lecturer can be confirmed in advance, which is convenient. In addition, the lecturer account can check the lecture location according to the lecture request and the expected incentive per unit time of the lecture time. Acceptance of matching can be accepted. Through this, the control device 510 may provide an incentive system for a lecturer who lectures in an outdoor environment that requires attention while improving the usability of the multi-user lecture system.

Thereafter, the control device 510 may calculate and provide the incentive of the first lecturer in response to the outdoor environment risk stage including the outdoor environment value while the lecturer is performing the lecture ( S804 ).

Specifically, the control device 510 may request and receive real-time UV information from the weather information provider through the API at the start time of the lecture. Thereafter, the control device 510 may request and receive real-time UV information from the weather information provider through the API in a predetermined time unit, for example, one hour unit. The control device 510 may receive the real-time UV information for each predetermined time unit, for example, one hour by repeatedly performing the above request until the end time of the lecture.

The control device 510 may determine the outdoor environmental risk level for each predetermined time unit based on the received ultraviolet light information for each predetermined time unit; based on the outdoor environmental risk level for each predetermined time unit, it is possible to determine the instructor's incentive for each predetermined time unit; The instructor's incentive may be given based on the instructor's incentive for each predetermined time unit.

The control device 510 determines the cost to be paid by the first student account and the cost to be paid by the first instructor account based on the lecture start time and the lecture end time, Each cost can be determined by reflecting the incentives. The control device 510 may provide an incentive for the first lecturer when the first lecturer performs a lecture in an outdoor environment requiring attention.

For example, the control device 510 requests to provide a real-time WHO UV index near Namsan Tower, where the lecture is held, from the weather information provider through the API at 12:00, the start time of the lecture. You can receive the WHO UV index of 9.0 near Namsan Tower at 12:00. Thereafter, the control device 510 may request and receive the real-time WHO UV index from the weather information provider through the API every hour. The control device 510 may receive the real-time WHO UV index per hour by repeatedly performing the above request until 14:30, which is the end time of the lecture. Hourly real-time WHO UV index is 8.0 at 13:00; It can be 6.0 at 14:00.

The control device 510 may determine the outdoor environmental risk level for each hour based on the WHO UV index received for each hour. Specifically, the control device 510 sets the outdoor environmental risk level as “dangerous” because the WHO UV index from 12:00 to 13:00 is 9.0; The WHO UV Index from 13:00 to 14:00 is 8.0, so the outdoor environmental risk level is “dangerous”; Since the WHO UV index from 14:00 to 14:30 is 6.0, the outdoor environmental risk level can be set as “safe”.

Next, the control device 510 may determine the instructor's incentive for each hour based on the outdoor environment risk level for each hour. Specifically, the control device 510 sets the incentive to 10% because the outdoor environment risk stage from 12:00 to 13:00 is “risk”; Since the outdoor environmental risk level from 13:00 to 14:00 is “risk”, the incentive is 10%; Since the outdoor environmental risk level from 14:00 to 14:30 is “safe”, the incentive can be set at 0%.

Thereafter, the control device 510 may provide the instructor's incentive based on the instructor's incentive for each hour. Specifically, the control device 510 provides an incentive of 10% for lectures from 12:00 to 13:00; 10% incentive for lectures from 13:00 to 14:00; 0% incentive can be given for lectures from 14:00 to 14:30.

The control device 510 determines the cost to be paid by the first student account and the cost to be paid by the first instructor account based on 12:00, the start time of the lecture and 14:30, which is the end time of the lecture, the outdoor environment risk step Incentives for instructors awarded according to ―10% from 12:00 to 13:00; 10% from 13:00 to 14:00; From 14:00 to 14:30, 0%- can be reflected to determine each cost. Through this, the control device 510 may provide an incentive system for the lecturer when the lecturer performs a lecture in an outdoor environment requiring attention.

9 is a flowchart illustrating a method of setting whether a lecturer can provide a lecture service based on a health condition of the lecturer according to an exemplary embodiment.

First, the control device 510 may acquire the body information of the lecturer at predetermined intervals ( S901 ). The predetermined period may be, for example, one month. The control device 510 may receive and obtain body information of lecturers using lecturer accounts obtained through each lecturer authentication means included in the lecturer terminals 530 from the lecturer terminals 530 .

For example, the control device 510 may acquire skin information of lecturers every month. The skin information may be a face of a lecturer photographed with a camera. In order to obtain the skin information, the lecturer terminals 530 may have the lecturer using the lecturer account take a picture every month. Lecturer terminals 530 may transmit the lecturer's skin information photographed by the camera, which is an authentication means, to the control device 510, through which the control device 510 may obtain skin information of the lecturers using lecturer accounts. there is.

Next, the control device 510 may diagnose whether there is an abnormality in the lecturer's health state based on the lecturer's body information ( S902 ). Specifically, the control device 510 may diagnose whether the lecturer's skin has a disease based on the acquired lecturer's skin information. To this end, the learned artificial neural network of the control device 510 may analyze the skin change for each lecturer's cycle, determine whether the lecturer's skin has a disease caused by UV exposure, and output it, and the control device ( 510) may diagnose whether there is an abnormality in the health state of each lecturer based on the output of the artificial neural network. Specific details regarding the learning of the artificial neural network will be described later with reference to FIG. 10 .

Subsequently, the control device 510 may set the lecturer diagnosed as having an abnormal health condition not to provide the lecture service when the outdoor environment forecast value is included in the predetermined outdoor environment risk stage ( S903 ). In short, the control device 510 determines that the first lecturer account has a skin disease, is diagnosed as having an abnormal health condition, and is a forecast value for the degree of ultraviolet light in the lecture place and lecture time included in the lecture request. If it is confirmed that it is included in this predetermined outdoor environment risk step, it may be set so that it is not possible to select acceptance of the lecture request or acceptance of the lecturer matching.

For example, the control device 510 sets the WHO UV index forecast value to the outdoor environmental risk level “danger” for a lecturer diagnosed as having an abnormal health condition based on the output of the learned artificial neural network. Or, it can be set so that it cannot provide lecture services when it is included in “very risky”. The lecture location included in the lecture request is near Namsan Tower; Arbor Day is approaching lecture time; When the outdoor environmental forecast value is predicted as WHO UV index 9.0, the outdoor environmental forecast value is included in “danger”, so the control device 510 is the first lecturer used by the first lecturer diagnosed as having an abnormal health condition You can set your account so that it cannot choose to accept lecture requests or accept instructor matching.

Through this, when it is determined that the health of the lecturer is deteriorated due to the outdoor environment requiring the lecturer's attention, the control device 510 may protect the health of the lecturer by restricting the lecture in the outdoor environment requiring the lecturer's attention.

On the other hand, despite the fact that the artificial neural network of the control device 510 diagnoses that there is an abnormality in the health condition of the lecturer, the lecturer uses the control device 510 through the first lecturer terminal 531 to obtain a doctor's opinion that there is no abnormality in the lecturer's skin. You may submit a statement of opinion. When the control device 510 receives a doctor's note made of an image file or the like, the control device 510 may reset the lecturer account corresponding to the lecturer to accept the lecture request or to accept the lecturer matching.

Through this, the control device 510 may supplement the reasoning of the artificial neural network, which may be incomplete, and allow a lecturer with no health abnormalities to normally participate in lecture matching without being penalized and increase profits.

10 is a diagram for explaining learning of an artificial neural network according to an embodiment.

The artificial neural network can receive the lecturer's skin information and diagnose whether a disease has occurred on the lecturer's skin due to ultraviolet rays. The artificial neural network may be a component included in the learning apparatus, and may be learned through the learning apparatus or a separate learning apparatus. Hereinafter, a process in which an artificial neural network is trained will be described.

First, the learning device collects human skin information at an arbitrary point in time and human skin information at a different point in time to generate skin information of people at two different points in time as training data. can Among them, people who did not have a skin disease at any time, but who developed a skin disease at any time and other time points may be included. The skin information may be a face of a person photographed by a camera.

In the learning apparatus, the skin information of people who did not have a skin disease at any time point, but had a skin disease at any time point and other time points, was 1; By assigning a value of 0 to skin information of other people, a label corresponding to each training data may be generated. For example, a label of a person who shows a normal skin condition at an arbitrary time point, but is in a skin condition with spots or wrinkles due to ultraviolet rays or the like, may be generated at a different time point than at an arbitrary time point.

Referring to FIG. 10 , the learning apparatus may generate an input from training data ( S1001 ). The learning apparatus may use people's skin information as it is as an input to the artificial neural network at two different viewpoints, or may generate an input of the artificial neural network after a normal process of removing unnecessary information, such as noise, from the skin information.

Next, the learning apparatus may apply the input to the artificial neural network (S1002). The artificial neural network included in the control device 510 may be an artificial neural network that is learned according to supervised learning. Through this, the artificial neural network can easily learn people's skin change patterns at two different time points. The artificial neural network may be a convolutional neural network (CNN) or recurrent neural network (RNN) structure suitable for learning through supervised learning.

Then, the learning apparatus may obtain an output from the artificial neural network (S1003). The output of the artificial neural network may be an inference of whether a specific person has developed a skin disease at a time other than an arbitrary time point. Specifically, the artificial neural network extracts a pattern of skin information at an arbitrary point in time of a specific person and skin information at a different time point and at a different point in time, and based on the difference in the extracted pattern, the person at a different point in time It is possible to output the result of inferring whether or not a skin disease has occurred.

Thereafter, the learning apparatus may compare the output with the label (S1004). The process of comparing the output of the artificial neural network corresponding to inference and the label corresponding to the correct answer may be performed by calculating a loss function. As the loss function, a known mean squared error (MSE), cross entropy error (CEE), etc. may be used. However, the present invention is not limited thereto, and loss functions used in various artificial neural network models may be used if the deviation, error, or difference between the output of the artificial neural network and the label can be measured.

Next, the learning apparatus may optimize the artificial neural network based on the comparison value (S1005). By updating the weights of the nodes of the artificial neural network so that the learning device comparison value becomes smaller and smaller, the output of the artificial neural network corresponding to inference and the label corresponding to the correct answer can be gradually matched, and through this, the artificial neural network can It can be optimized to output inferences close to the correct answer. Specifically, the learning apparatus may optimize the artificial neural network by repeating the process of resetting the weight of the artificial neural network so that the loss function corresponding to the comparison value approaches the estimated value of the minimum value. A known backpropagation algorithm, stochastic gradient descent, or the like may be used for optimization of an artificial neural network. However, the present invention is not limited thereto, and a weight optimization algorithm used in various neural network models may be used.

The learning apparatus may learn the artificial neural network by repeating the above process. Through this, the learning device infers whether or not the person has a skin disease at a different point in time and at a different point in time based on the skin information at an arbitrary point in time and the skin information at a different point in time. An artificial neural network may be provided.

The control device 510 equipped with the learned artificial neural network collects the skin information of the lecturers every predetermined period, and the specific lecturer shows a normal skin condition in the first period, and spots and wrinkles caused by ultraviolet rays in the second period. If the skin condition that has occurred is shown, the lecturer can diagnose that a skin disease has occurred due to ultraviolet rays. Through this, lecturers whose health has deteriorated by conducting lectures in an outdoor environment requiring attention are restricted from performing lectures, thereby protecting the health of lecturers.

11 is an exemplary diagram of a configuration of a control device according to an embodiment.

The control device 510 according to an embodiment includes a processor 511 and a memory 512 . The processor 511 may include at least one of the devices described above with reference to FIGS. 1 to 10 , or perform at least one method described above with reference to FIGS. 1 to 10 . A person or group using the control device 510 may provide services related to some or all of the methods described above with reference to FIGS. 1 to 10 .

The memory 512 may store information related to the above-described methods or a program in which methods to be described below are implemented. Memory 512 may be volatile memory or non-volatile memory.

The processor 511 may execute a program and control the control device 510 . The code of the program executed by the processor 511 may be stored in the memory 512 . The control device 510 may be connected to an external device (eg, a personal computer or a network) through an input/output device (not shown), and may exchange data through wired/wireless communication.

The control device 510 may be used to train an artificial neural network or use the learned artificial neural network. Memory 512 may include a learning or learned artificial neural network. The processor 511 may learn or execute an artificial neural network algorithm stored in the memory 512 . The control device 510 for learning the artificial neural network and the control device 510 for using the learned artificial neural network may be the same or may be separate.

The embodiments described above may be implemented by a hardware component, a software component, and/or a combination of a hardware component and a software component. For example, the apparatus, methods and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate (FPGA). array), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

The software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or device, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computer systems, and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

As described above, although the embodiments have been described with reference to the limited drawings, those skilled in the art may apply various technical modifications and variations based on the above. For example, the described techniques are performed in a different order than the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (5)

A method of providing a learning curriculum by analyzing academic achievement, performed by a control system, the method comprising:
obtaining an identifier for specifying the identity of a first student as a learning target and a learning progress index for specifying the learning progress of the first student;
Through the learning progress index, the learning progress of the first student is confirmed as the first unit, and after learning the first unit, problem solving and scoring are completed during the learning of the first unit using an imaging unit of an electronic device. acquiring a photographed image;
generating an input vector by encoding the image, and applying it to an input layer of a pre-trained convolutional neural network of an embedded computer in the control system;
obtaining, as a result of the operation of the convolutional neural network, a first output value from an output layer of the convolutional neural network, the first output value estimating the type and correct answer of the first problem extracted from the image;
inputting the first output value and an auxiliary output value stored in a database in the control system into a pre-trained neural network, wherein the neural network evaluates the first output value based on the auxiliary output value;
obtaining, from an output layer of the neural network, a second output value, wherein the second output value indicates the type of the first corrected problem and whether or not it is a correct answer;
By comparing the first problem with the problems of the first unit stored in advance in the database, the degree of similarity is determined through keywords extracted from each question of the problem, and the type and difficulty of the first problem are analyzed according to the similarity. to do;
determining the type and difficulty of a second problem included in a second unit to be learned after the first unit based on whether the first problem is correct, and the type and difficulty of the first problem;
providing a learning curriculum according to the type and difficulty of the second problem when it is confirmed that the learning progress of the first student has completed the learning of the first unit;
Receiving a notification from a plurality of lecturer terminals that lecturer accounts logged into each of the lecturer terminals are set to a lecture standby mode;
generating a lecture waiting list based on the lecturer accounts set to the lecture standby mode;
receiving a lecture request for the second unit from a first student terminal logged in through the first student account;
transmitting the lecture request to lecturer terminals in which each of the lecturer accounts included in the lecture waiting list are logged in;
receiving whether or not to accept the lecture request from lecturer terminals in which each of the lecturer accounts included in the lecture waiting list are logged in;
generating a lecture request acceptance list based on the lecture waiting list and whether the lecture has been accepted;
transmitting the lecture request acceptance list to the first student terminal;
receiving lecturer selection information from the first student terminal;
transmitting the lecturer selection information to a first lecturer terminal in which a first lecturer account corresponding to the lecturer selection information is logged in;
receiving from the first lecturer terminal whether to accept the lecturer matching;
As a result of checking whether to accept the lecturer matching, if the lecturer matching is accepted, the first lecturer account and the first student account are matched to create a chat space between the first lecturer account and the first student account to do;
receiving a notification from the first lecturer terminal that the lecture for the second unit has started;
transmitting a request for approval to start the lecture of the second unit to the first student terminal;
receiving approval to start the lecture of the second unit from the first student terminal;
setting a time when the lecture start approval is received as a lecture start time;
receiving a notification from the first student terminal that the lecture of the second unit has ended;
transmitting a request for approval to end the lecture of the second unit to the first lecturer terminal;
receiving approval to end the lecture of the second unit from the first lecturer terminal;
setting a time when the lecture end approval is received as a lecture end time;
calculating a cost to be paid by the first student account and a cost to be paid to the first lecturer account based on the lecture start time and the lecture end time;
confirming a correspondence relationship between each preset staged outdoor environmental risk stage and an incentive per unit time;
obtaining, based on the lecture request, whether an environment in which the first lecturer performs a lecture is outdoors;
setting an expected incentive per unit time based on an outdoor environment forecast value at a lecture location and a lecture time according to the lecture request; and
Comprising the step of calculating the incentive of the first lecturer in response to the outdoor environment risk stage including the outdoor environment value during the lecture by the first lecturer,
How to deliver a learning curriculum. According to claim 1,
Obtaining the first output value comprises:
encoding and inputting color information for each pixel of the image in the form of a data sheet, wherein the color information includes RGB color information, brightness information, and chroma information;
separating the first question from the image based on the input, and separating and extracting the questions of the first question and the handwriting marked on the first question;
recognizing a keyword from the items of the first problem, and classifying the type of the first problem through the recognized keyword;
checking the color, brightness, and saturation of the handwriting scored on the first question, and determining whether the answer to the first question is correct according to preset first color values, first brightness values, and first saturation values; and
and outputting an output value including the type of the first problem and the correct answer to a total of 50 output layer nodes,
The convolutional neural network is characterized in that it learns through a first learning signal according to a user input,
The top 25 output layer nodes among the 50 output layer nodes indicate the type of the first problem,
Among the 50 output layer nodes, the lower 25 output layer nodes correspond to the upper 25 output layer nodes. In the lower 25 output layer nodes, the 26th output layer node corresponds to the first output layer node, and the 27th output layer node is the second node. In a manner corresponding to the output layer node, the 25+nth output layer node corresponds to the nth output layer node, respectively, indicating whether the answer is correct,
Characterized in that the number '0' is outputted to a node indicating that it is not the type of the first problem in the top 25 output layer nodes,
How to deliver a learning curriculum. According to claim 1,
The neural network is
Taking the first output value that is an output value of the convolutional neural network and the auxiliary output value stored in the database as inputs,
evaluating the first output value based on the auxiliary output value;
Detecting an abnormal signal based on the accumulated data of the evaluation of the types and correct answers of a plurality of questions stored in the database,
generating the second output value indicating the type of the first problem corrected and whether the answer is correct;
The neural network is characterized in that it learns through a second learning signal according to a user input,
How to deliver a learning curriculum. delete According to claim 1,
acquiring skin information of lecturers using lecturer accounts at predetermined intervals through a lecturer authentication means, which is a camera capable of detecting the lecturer's skin;
diagnosing whether each lecturer has a skin disease based on the lecturer's skin information; and
If the first lecturer account is diagnosed as having a skin disease, and it is confirmed that the outdoor environment forecast values at the lecture location and the lecture time according to the lecture request are included in the predetermined outdoor environment risk stage, accept the lecture request further comprising the step of not being able to select
How to deliver a learning curriculum.

Images