KR102400582B1 - Method and device for personalized learning amount recommendation based on big data and artificial intelligence - Google Patents

Abstract

Disclosed is a device for recommending a personalized learning amount based on big data and artificial intelligence (AI). The device for recommending the personalized learning amount based on big data and AI comprises: a learning data collection unit collecting a large number of learning data from student terminals corresponding to a large number of users; an academic achievement ability prediction unit predicting the type of academic achievement ability corresponding to each user of each of the student terminals by using the large number of learning data; a learning amount table generation unit pre-processing the learning data in accordance with the predicted type of academic achievement ability, and generating a learning amount table; and a recommended learning amount determination unit determining the recommended learning amount corresponding to a target user by referring to the learning amount table, and providing the determined recommended learning amount to the student terminal of the target user. The present invention aims to provide a device for recommending a personalized learning amount based on big data and AI, which is able to maximize learning effects.

Description

Recommendation method and device for personalized learning amount based on big data and artificial intelligence

The present invention relates to a learning amount recommendation technology, and more particularly, to a method and apparatus for recommending a personalized learning amount based on big data and artificial intelligence.

Parents' interest and enthusiasm for education in Korea is very high, and for this reason, many people are interested in students' school classes and university entrance exams through them.

Although students are more likely to achieve good learning outcomes by increasing the amount of learning as much as possible, it is physically impossible and physically counterproductive to increase the amount of learning blindly, so it is difficult to determine the optimal amount of learning within a given time becomes a very important issue.

In general, since the current learning result affects the next learning, the amount of next learning is individually determined by increasing or decreasing the next learning according to the current learning result, so the academic achievement results are often different.

In order to solve this problem, the prior art (Patent Application No. 10-2018-0135206) is based on a predetermined number of recently performed learning result data, the more recent the time, the higher the weight is assigned in advance. is applied to estimate the expected learning amount.

However, in the case of the prior art described above, there is a problem in that the prediction result can vary greatly depending on the pre-specified weight, and since the criterion for selecting the weight is determined very differently depending on the implementation method, it is difficult to expect a constant prediction result. there was

On the other hand, recently, research on artificial intelligence that imitates human intelligence to perform complex tasks is being actively conducted. ), and in the field of machine learning, deep learning implemented by mimicking human neurons is known as a slightly more advanced concept.

A deep neural network (DNN) based on deep learning is an artificial neural network (ANN) that learns various nonlinear relationships including multiple hidden layers and provides prediction results based on the learning results, A lot of interest and research is being done on this.

The deep neural network includes a deep trust neural network (DBN) based on unsupervised learning, a deep autoencoder, etc., depending on the algorithm, and is synthesized for processing two-dimensional data such as images. Convolutional neural networks (CNNs), recurrent neural networks (RNNs) useful for processing time series data, and the like are known.

An object of the present invention for solving the above problems is to provide a personalized learning amount recommendation method and apparatus based on big data and artificial intelligence.

One aspect of the present invention for achieving the above object provides a personalized learning amount recommendation device based on big data and artificial intelligence.

The big data and artificial intelligence-based personalized learning amount recommendation apparatus includes: a learning data collection unit for collecting a plurality of learning data from student terminals corresponding to a plurality of users; an academic achievement predicting unit for predicting an academic achievement type corresponding to the user of each of the student terminals by using the plurality of learning data; a learning amount table generation unit for generating a learning amount table by pre-processing the learning data according to the predicted academic achievement type; and a recommended learning amount determining unit for determining a recommended learning amount corresponding to a target user by referring to the learning amount table, and providing the determined recommended learning amount to a student terminal of the target user.

The academic achievement predicting unit determines a cluster corresponding to each of the users by clustering the users using the learning data according to the current curriculum, and corresponds to the academic achievement type assigned to the determined cluster. a clustering engine that predicts the academic achievement type of users belonging to the cluster; And supervised-learning using training data consisting of training input values corresponding to the learning data of each of the users and training output values labeled with an academic achievement type through the clustering engine, based on deep learning and an artificial neural network engine for predicting the academic achievement type corresponding to the target user.

The artificial neural network engine may include: an input layer configured to receive a learning feature vector obtained by transforming learning data according to a current curriculum of each of the users, and configured with the same number of neurons as the number of component values of the learning feature vector; a hidden layer that transmits an output vector calculated using the output values received from the input layer to the output layer; and an output layer that determines a probability corresponding to the output vector by applying an activation function to the output vector, and outputs an output vector having the highest determined probability.

The artificial neural network engine, when receiving the training input value, uses an output vector obtained as an output of the hidden layer and a training output vector indicating an academic achievement type according to the training output value to obtain a loss function according to the following equation Calculation and supervised learning so that the result value of the calculated loss function is minimized,

In the above equation, Ym is the m-th component (m is a natural number greater than or equal to 1) of the target output vector Ym, and Y'm is the m-th component of the output vector Y' obtained as an output of the hidden layer.

The artificial neural network engine further includes a self-attention layer that performs a technique according to a self-attention mechanism between the input layer and the hidden layer.

The self-attention concentration layer determines the importance of each of the intermediate operation vectors obtained as an output of the input layer, and selectively selects some of the intermediate operation vectors by excluding the intermediate operation vectors having a relatively low determined importance. transmitted to the hidden layer.

After the learning amount table is created, the learning data collection unit acquires target learning data according to the current curriculum of the target user from the student terminal of the target user, and the academic achievement predicting unit obtains the target learning data Predict the academic achievement type corresponding to the target user based on the learning amount, and the recommended learning amount determiner is, in the learning amount table, a plurality of learning amounts according to the next curriculum corresponding to the academic achievement type predicted for the target user The recommended learning amount for the next training course of the target user is determined by using the statistical data for each.

The clustering engine performs clustering using a K-means algorithm.

The learning amount table generation unit, after generating the learning amount table, updates the learning amount table by using the learning data for the next educational course of the target user according to the recommended learning amount.

In the case of using the personalized learning amount recommendation method and apparatus based on big data and artificial intelligence according to the present invention as described above, the student's academic achievement ability is predicted using a pre-supervised artificial neural network, and according to the predicted learning achievement ability By determining the optimal amount of recommended learning based on big data and providing it to students, it is possible to guide the most efficient amount of learning to students and maximize the learning effect.

1 is a conceptual diagram illustrating an environment in which a personalized learning amount recommendation method based on big data and artificial intelligence is performed according to an embodiment.
FIG. 2 is a block diagram illustrating functional components of the personalized learning amount recommendation apparatus based on big data and artificial intelligence according to FIG. 1 .
FIG. 3 is a diagram exemplarily showing the configuration of a learning amount table according to FIG. 1 .
4 is a diagram illustrating a result of calculating a weighted score with reference to the learning amount table according to FIG. 2 .
FIG. 5 is a diagram for explaining the operation of the clustering engine according to FIG. 2 .
FIG. 6 is a view for explaining the operation of the artificial neural network engine according to FIG. 2 .
7 is a diagram specifically illustrating the structure of the artificial neural network engine according to FIG. 6 .
8 is a diagram for explaining a method of determining a K vector in a self-attention layer according to an embodiment.
9 is a diagram exemplarily showing a hardware configuration of a personalized learning amount recommendation device based on big data and artificial intelligence according to an embodiment.
10 is a graph showing statistical analysis results that can confirm the effect of the big data and artificial intelligence-based personalized learning amount recommendation method and apparatus according to an embodiment.

Since the present invention can have various changes and can have various embodiments, specific embodiments are illustrated in the drawings and described in detail in the detailed description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention. In describing each figure, like reference numerals have been used for like elements.

Terms such as first, second, A, and B may be used to describe various elements, but the elements should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, a first component may be referred to as a second component, and similarly, a second component may also be referred to as a first component. and/or includes a combination of a plurality of related listed items or any of a plurality of related listed items.

When an element is referred to as being “connected” or “connected” to another element, it is understood that it may be directly connected or connected to the other element, but other elements may exist in between. it should be On the other hand, when it is said that a certain element is "directly connected" or "directly connected" to another element, it should be understood that the other element does not exist in the middle.

The terms used in the present application are only used to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present application, 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 defined otherwise, all terms used herein, including technical and scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention 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

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a conceptual diagram illustrating an environment in which a personalized learning amount recommendation method based on big data and artificial intelligence is performed according to an embodiment.

Referring to FIG. 1 , the personalized learning amount recommendation method based on big data and artificial intelligence may be referred to as a learning amount recommendation server 100, which is also referred to as a personalized learning amount recommendation device based on big data and artificial intelligence, to be described below. ) can be done by

The learning amount recommendation server 100 may collect learning data in conjunction with a plurality of student terminals 200a to 200n.

Each of the student terminals 200a to 200n may have a pre-distributed learning application installed and driven, and as the learning application is executed, each of the student terminals 200a to 200n selects a plurality of learning problems according to the curriculum. It may be provided to the user of the student terminal 200, and the user's problem solving result may be stored.

The user of the student terminal 200 may access the learning amount recommendation server 100 by using the student terminal 200 in advance, and register user information in the learning amount recommendation server 100 through a user authentication procedure. The user information may include his or her age, grade, school, and the like.

In addition, each of the student terminals 200a to 200n provides learning data representing a problem solving result according to the learning problems provided to the user of the student terminal 200 in real time or at a predetermined period or at predetermined times, a learning amount recommendation server (100) can be provided.

In detail, the learning data may mean solving results for learning problems provided for at least one lesson according to an educational process (eg, an addition process) on a weekly or daily basis. The curriculum may correspond to a large classification of learning tasks provided according to a curriculum, and a lesson may correspond to a small classification of learning tasks provided within a specific curriculum. For example, the education course may have a predetermined order as an addition basic course, an advanced addition course, a basic subtraction course, an advanced subtraction course, and the like. Specifically, a training course to be learned after the basic addition course may be an advanced addition course, and a training course to be learned after the deep addition course may be a basic subtraction course.

For example, a lesson on the basics of addition may include addition to single digit numbers, addition to two digits, etc., while lessons for an advanced addition course include addition to three digits and addition to all three or more numbers. addition may be included.

The amount of learning recommendation server 100 collects a plurality of past learning data from a plurality of student terminals 200a to 200n, and based on the collected past learning data, the user of each of the student terminals 200a to 200n is a machine. By using the learning engine 10 to cluster (clustering), it is possible to determine the academic achievement type corresponding to each of the users.

Here, the machine learning engine 10 may include a clustering engine 102a operating based on a Gaussian mixture model (GMM). The clustering engine 102a may correspond to a kind of source code in which the Gaussian mixture model (GMM) is implemented in software or an operation program of the source code.

Since each person has different innate talents, memory, and mathematical arithmetic abilities, there are those who can easily solve the problems and those who can't solve them easily even if each user is given the same problem. In addition, even among users with the same academic achievement level, there is a tendency that certain people are good at math, but weak in Korean, while other specific people are relatively poor at math.

In one embodiment of the present invention, considering that each person has different academic achievement abilities, cluster users to configure a plurality of clusters so as to recommend the most appropriate amount of learning according to academic achievement ability, The academic achievement type corresponding to each of the configured clusters is determined. Here, the academic achievement type is a kind of identification symbol that is individually assigned to each cluster to which the user belongs. For example, a symbol determined without superiority, such as A, B, C, etc. may be used, or various types of unique identification symbols may be used. . That is, the academic achievement type is used as a meaning representing the academic achievement ability of users belonging to the corresponding cluster.

When the academic achievement type corresponding to each of the users is determined through clustering, the learning amount recommendation server 100 classifies the learning data for users according to the academic achievement capability type and the amount of learning performed in the past, and pre-processing the amount of learning Table 20 can be created.

The amount of learning table 20 generated here is used as big data for later determining the amount of recommended learning for a specific user, collects learning data for as many users as possible, and determines the academic achievement type corresponding to each of the users. By determining, it may be desirable to configure the learning amount table 20 based on large data as much as possible.

When the learning amount table 20 is generated through clustering, the learning amount recommendation server 100 may operate to provide a recommended learning amount for a specific person.

First, the learning amount recommendation server 100 may receive target learning data from the student terminal 200 of the target user among users, and using the machine learning engine 10 based on the received target learning data, the target user It is possible to predict the type of academic achievement ability corresponding to

In this case, the aforementioned clustering engine 102a may be used as the machine learning engine 10 , but is not limited thereto.

For example, an artificial neural network engine 102b based on a deep neural network may be used as the machine learning engine 10 . The artificial neural network engine 102b uses training data composed of a set of a training input value and a training output value to receive the learning data of a specific user, an output value indicating an academic achievement type corresponding to the specific user It may be pre-supervised learning to output .

Accordingly, the learning amount recommendation server 100 predicts the academic achievement type corresponding to the cluster to which the target user belongs based on the target learning data using the clustering engine 102a, or uses the artificial neural network engine 102b to predict the target It may be configured to predict the academic achievement type corresponding to the target user based on the learning data.

The learning amount recommendation server 100 may analyze the learning amount table 20 based on the predicted academic achievement type and determine the recommended learning amount for the target user according to the predicted academic achievement type.

For example, the recommended learning amount (or learning amount) may mean the next training course (subtraction process on the drawing) to be learned after the current training course of the target user according to the target learning data, the number of lessons per week, the number of lessons per day, etc. .

The learning amount recommendation server 100 may transmit the determined recommended learning amount to the student terminal 200 of the target user.

Since the target user receives the recommended amount of learning expected to achieve the highest performance in his/her academic achievement type with the student terminal 200, the possibility of achieving optimal academic achievement results by learning according to the recommended amount of learning This can be high.

FIG. 2 is a block diagram illustrating functional components of the personalized learning amount recommendation apparatus based on big data and artificial intelligence according to FIG. 1 . FIG. 3 is a diagram exemplarily showing the configuration of a learning amount table according to FIG. 1 . 4 is a diagram illustrating a result of calculating a weighted score with reference to the learning amount table according to FIG. 2 .

Referring to FIG. 2 , the learning amount recommendation server 100 collects learning data from each of the student terminals 200 corresponding to a plurality of users. The academic achievement predicting unit 102 for predicting the academic achievement type corresponding to the users of each of the student terminals 200 by using the plurality of learning data, pre-processing the learning data according to the predicted academic achievement type ( pre-processing) to determine the recommended learning amount corresponding to the target user by referring to the learning amount table generation unit 103 for generating the learning amount table, and the learning amount table, and providing the determined recommended learning amount to the target user's student terminal 200 It may include a recommended learning amount determiner 104 .

The learning data collection unit 101 may collect learning data in conjunction with a plurality of student terminals 200 . To this end, learning data may be acquired by periodically communicating with the student terminals 200 through a wired or wireless network. For example, when learning according to the educational process is completed in each of the student terminals 200 , each of the student terminals 200 may transmit learning data according to the completion of learning to the learning data collection unit 101 .

The academic achievement predicting unit 102 may predict the academic achievement type corresponding to each of the users by using the learning data, and provide it to the learning amount table generating unit 103 .

Specifically, the academic achievement predicting unit 102 is a clustering engine that determines a cluster corresponding to each of the users by clustering users corresponding to each of the learning data, and determines the academic achievement type assigned to the determined cluster. (102a) may be included.

For example, the clustering engine 102a may perform clustering according to a Gaussian Mixture Model (GMM) or clustering according to a K-means algorithm.

Here, the K-means algorithm may be a machine learning-based clustering technique belonging to unsupervised learning, unlike the Gaussian mixture model. Specifically, the K-means algorithm sets a vector value corresponding to an initial center point for each of the K clusters, and then allocates users to a cluster having an initial center point close to the learning feature vector representing each of the users.

Next, when cluster assignment to all users is finished, the center point of each cluster is reset to the median or average value of the learning feature vectors of users belonging to the cluster, and cluster assignment of all users is re-executed based on the reset center point. do.

Users can be assigned to clusters by repeating the above-described center point resetting and cluster reallocation until there is no change in the center point.

The learning amount table generating unit 103 may generate the learning amount table 20 by preprocessing the learning data according to the academic achievement type determined by the clustering engine 102a.

Specifically, the amount of learning table generation unit 103 is, based on the learning data according to the user's current educational process, the academic achievement type determined by the clustering engine 102a and the amount of learning performed in the next educational process after the current educational process. It is possible to preprocess the learning data by classifying according to , and calculating statistical data for the learning data corresponding to the classified academic achievement type and learning amount.

That is, the academic achievement type is predicted by the clustering engine 102a based on the learning data according to the user's current educational process.

Referring to FIG. 3 , the learning amount table 20 classifies learning data according to the current educational process according to the academic achievement type and learning amount, and is statistical data obtained through statistical calculation processing on the classified learning data. It may include an evaluation pass rate, an evaluation score average, and the like for the acquired learning data.

Here, the amount of learning may mean the number of learning per week and the number of learning lessons per day performed in the next educational course of users of the corresponding academic achievement type. As another example, the amount of learning may mean the number of learning days and learning hours per week performed in the following curriculum. Here, the amount of learning may be set in units of hours, days, or months according to the criteria for setting the index, or may be set in units of a mixture thereof.

Here, the evaluation pass rate may be a ratio of users whose solution result according to the amount of learning provided in the next curriculum exceeds a specific reference score for users determined by the corresponding academic achievement type.

The average evaluation score may be an average value of evaluation scores in a solution result according to the amount of learning provided in the next educational course for users determined by the corresponding academic achievement ability type.

Referring to FIG. 3 , among users whose academic achievement ability type in the current curriculum belongs to cluster A, the number of learning times per week according to the amount of learning provided for the next curriculum is 3 times, and the number of users with 3 learning lessons per day is 169 It can be seen that the evaluation pass rate for the amount of learning provided as the next curriculum for the 169 users is 92%, and the average evaluation score for the amount of learning is 94 points.

In this way, the learning amount table 20 predicts the academic achievement type corresponding to each of the users based on the learning data according to the current curriculum of a very large number of users using the clustering engine 102a, and predicts the predicted According to the type of academic achievement, it may be configured in a manner of calculating statistical data for learning data according to the following curriculum.

As described above, the amount of learning table 20 is classified according to the type of academic achievement predicted based on the learning data according to the current curriculum, so the amount of learning according to the next curriculum based on the learning data according to the current curriculum of a specific target user It is easy to recommend

When the learning amount table 20 is generated, the function of providing the recommended learning amount is activated.

In addition, the learning amount table generating unit 103 continues by updating the learning amount table 20 by using the learning data performed for the next educational course of the target user according to the recommended learning amount after generating the learning amount table 20 . It is preferably configured so that new data is added.

Specifically, the learning data collection unit 101 may acquire target learning data according to the current curriculum of the target user from the student terminal 200 of the target user. Here, the current curriculum may refer to a stage of a curriculum currently in progress by the target user.

The academic achievement predicting unit 102 may predict the academic achievement type corresponding to the target user based on target learning data according to the current curriculum.

In an embodiment, the academic achievement predicting unit 102 determines a cluster corresponding to the target user using the clustering engine 102a, and sets the academic achievement type assigned to the determined cluster to the academic achievement corresponding to the target user. It can be predicted by the ability type.

In another embodiment, the academic achievement predicting unit 102 may predict the academic achievement type corresponding to the target user by using the deep learning-based artificial neural network engine 102b. The artificial neural network engine 102b may be supervised in advance using training data.

In this case, the training input value constituting the training data may be a learning feature vector generated using the learning data collected from each of a plurality of users, and the training output value corresponds to each of the users using the clustering engine 102a. It may be the type of academic achievement given to the cluster.

That is, since the artificial neural network engine 102b receives the training output value labeled with the academic achievement type through the clustering engine 102a, the difficulty of collecting massive learning data for supervising the artificial neural network can be solved. can

In addition, since the clustering engine 102a is implemented using a machine learning-based Gaussian mixture model, even if the training output value labeled through the clustering engine 102a is used, it can be labeled statistically and meaningfully. The prediction accuracy for the academic achievement type of the engine 102b is implemented relatively accurately.

The recommended learning amount determining unit 104 calculates a weighted score according to a predetermined equation using statistical data for each learning amount corresponding to the type of academic achievement predicted for the target user in the learning amount table 20 , and calculates the calculated weight A recommended learning amount may be determined based on the score.

For example, as shown in FIG. 4 , the weighted score is calculated for statistical data according to the type of academic achievement and the amount of learning, and may be specifically calculated according to Equation 1 below.

Referring to Equation 1, N is the number of learners corresponding to the academic achievement type and learning amount, Nmin is the minimum number of learners individually defined in advance according to the learning ability type and learning amount, and PP is the academic achievement ability It is the evaluation pass rate of the learners corresponding to the type and learning amount, acSC is the average of the evaluation scores of the learners corresponding to the academic achievement type and learning amount, and k is a weight constant defined individually according to the learning achievement type and learning amount in advance. , and avgN may be an average value of the number of learners for each academic achievement type.

For example, when the academic achievement type predicted for the target user corresponds to cluster A, as shown in FIG. 4 , statistical data for each of a plurality of learning amounts corresponding to the cluster A of the academic achievement type is used. Thus, it is possible to calculate a weighted score according to Equation 1 for each of the learning amounts.

In this case, the recommended learning amount determiner 104 may determine, as the recommended learning amount, a learning amount corresponding to the highest calculated weighted score among weighted scores calculated for each of the learning amounts according to Equation 1 .

That is, in the case of FIG. 4, since the weighted score was calculated as the highest value when the number of learning times per week was 3 and the number of daily learning lessons was 3 as the amount of learning, the recommended learning amount in the next curriculum is the number of learning times per week. It is 3 times, and the number of learning lessons per day is determined by the amount of learning that is three.

FIG. 5 is a diagram for explaining the operation of the clustering engine according to FIG. 2 .

The clustering engine 102a may generate a learning feature vector representing the user's current curriculum by using the learning data collected for each user.

For example, referring to FIG. 5 , the clustering engine 102a uses the learning data according to the current curriculum collected for each user, the number of problem solving according to the current curriculum and lesson, the number of evaluation passes, and the number of times in the curriculum. It is possible to obtain preprocessing learning data including the average number of passes, the average score in the curriculum, and the test score in the lesson.

The clustering engine 102a may generate a learning feature vector corresponding to the user by using at least some of the indices included in the obtained pre-processing learning data.

For example, the clustering engine 102a uses each of the identifier of the lesson in the current curriculum, the number of problem solving in the preprocessing learning data, the number of evaluation passes, the average number of passes, the average score, and the test score in the lesson as component values. It is possible to create a learning feature vector that

One learning feature vector may be generated for each lesson belonging to the current curriculum.

The clustering engine 102a may classify each of the users into one of a plurality of clusters using at least one learning feature vector according to the current curriculum of each of the users. For example, in the case of a Gaussian mixture model (GMM), each of K (K is a natural number greater than 1) Gaussian distributions specified in advance corresponds to a cluster. It operates to classify users into one of K Gaussian distributions.

Specifically, in the case of the clustering engine 102a, the Gaussian distribution having the highest result value of the Gaussian distribution selection function (γ) according to Equation 2 below among K Gaussian distributions using Bayes' theorem is selected, and the selected Gaussian distribution is determined as the academic achievement ability type corresponding to the current curriculum of the user.

In Equation 2, (γ) is a selection function, x _n is a learning feature vector according to the nth lesson in the corresponding curriculum, and z _nk is k (k is 1 and K in a Gaussian mixture model when a learning feature vector is given) It is a binary variable with a value of 1 if the Gaussian distribution is selected, otherwise 0. Also, μ and Σ in Equation 2 are parameters according to the Gaussian mixture model, and are values determined in advance through a learning process for the Gaussian mixture model. For the learning process for determining the parameters according to the Gaussian mixture model, it can be understood by referring to Christopher Bishop's Pattern Recognition and Machine Learning, 2008.03.18, so a detailed description will be omitted.

FIG. 6 is a view for explaining the operation of the artificial neural network engine according to FIG. 2 . 7 is a diagram specifically illustrating the structure of the artificial neural network engine according to FIG. 6 . 8 is a diagram for explaining a method of determining a K vector in a self-attention layer according to an embodiment.

6, when the deep learning-based artificial neural network engine 102b sequentially receives learning feature vectors for each lesson according to the current curriculum as training input values constituting training data, the artificial neural network engine 102b Supervised learning is performed by comparing the academic achievement ability type obtained as the output value of , with the training output value, and tuning parameters constituting the artificial neural network engine 102b according to the comparison result.

Specifically, the deep learning-based artificial neural network engine 102b generates an output vector indicating the academic achievement type obtained as an output value of the artificial neural network engine 102b and a training output vector indicating the academic achievement ability type according to the training output value. It is calculated according to a pre-defined loss function by using it, and parameters constituting the artificial neural network engine 102b may be adjusted so that the result value of the calculated loss function is minimized.

In this case, the loss function may be a cross entropy function.

Referring to FIG. 7 , the artificial neural network engine 102b receives a learning feature vector (X), and the input layer 11 is composed of the same number of input nodes as the number of component values of the received learning feature vector (N). ), a hidden layer 12 that transmits an output vector Y′ calculated using the output values received from the input layer 11 to the output layer 13, and an output vector by applying an activation function to the output vector Y′ The output layer 13 may include an output layer 13 that determines a probability p corresponding to (Y′) and outputs an output vector Y′ having the highest determined probability p. In the present invention, each of the nodes constituting the artificial neural network engine 102b may be referred to as a neuron, which is an expression commonly used in the technical field to which the present invention belongs.

Specifically, when the artificial neural network engine 102b receives a learning feature vector (X) provided as a training input value, an output vector (Y′) obtained as an output of the hidden layer 12 and academic achievement ability provided as a training output value A loss function is calculated using the training output vector (Y) indicating the type, and supervised learning is performed so that the result of the calculated loss function is minimized.

For example, the loss function H(Y, Y`) may be a cross entropy function. The cross entropy (H(Y, Y')) between the output vector (Y') and the training output vector (Y) may be defined as in Equation 3 below.

In Equation 3, Ym may be the m-th component (m is a natural number greater than or equal to 1) of the training output vector (Y), and Y'm may be the m-th component of the output vector (Y').

The input layer 11 receives a learning feature vector (X), and applies one or more connection strength values corresponding to input nodes to each of the components of the received learning feature vector (X) to the hidden layer 12 . can transmit

For example, one or more connection strength values corresponding to each of the input nodes may be expressed as a first connection strength matrix W _N×M having a size of N×M. In this case, N may be the same number as the input nodes, and M is set sufficiently smaller than N by 1/10 times or less. The first connection strength matrix (W _N×M ) may be a parameter that is set to an arbitrary initial value and is continuously updated through supervised learning.

In summary, the input layer 11 may transmit an intermediate operation vector (X) obtained by matrix multiplication operation of the received learning feature vector (X) with the first connection strength matrix (W _N×M ) to the hidden layer 12 . .

The hidden layer 12 applies one or more connection strengths corresponding to each of the hidden nodes to a feature vector (F) obtained from an intermediate operation vector (X) transmitted from the input layer 11 to an output vector (Y). `), and the generated output vector (Y`) may be transmitted to the output layer 13 .

For example, the hidden layer 12 may obtain the feature vector F, which is a diagonal matrix satisfying Equation 4 below, from the intermediate operation vector X.

In Equation 4, R is a matrix that reversibly satisfies the relation according to Equation 4 between the diagonal matrix feature vector F and the intermediate operation vector X, and corresponds to a coordinate transformation matrix. The operation according to Equation 4 may be referred to as one of the diagonalization operations, and since it can be easily understood by those skilled in the art, a detailed description thereof will be omitted.

In this case, one or more connection strength values corresponding to each of the hidden nodes may be expressed as a second connection strength matrix (U _M×Q ) having a size of M×Q. That is, the second connection strength matrix (UM×Q) increases the feature vector F mapped in M dimensions to Q dimensions again. Q is set to a value sufficiently larger than M by 10 times or more.

On the other hand, after the initial value of the second connection strength matrix (U _M×Q ) is set to an arbitrary value, the output generated by the matrix multiplication operation between the feature vector (F) and the second connection strength matrix (U _M×Q ) The vector (Y`) may be continuously updated to become the training output vector (Y), which is the training output value. That is, the second connection strength matrix (U _M×Q ) may also be a parameter that is updated as the training data is continuously supervised.

That is, the hidden layer 12 converts the intermediate operation vector (X) received from the input layer 11 into a feature vector, and applies the connection strength to the feature vector to generate an output vector (Y'), so it is like a CNN A function similar to the convolutional layer responsible for feature extraction of a (convolutional neural network) may be configured to be performed in the hidden layer 12 .

The output layer 13 determines a probability p corresponding to the output vector Y′ by applying an activation function to the output vector Y′ received from the hidden layer 12, and the determined probability p has the highest You can output an output vector (Y`). The activation function has the effect of converting values having various ranges into probabilities by expanding or reducing them to values between 0 and 1. For example, the activation function may be a ReLU function or a Softmax function, but is not limited thereto.

On the other hand, in an embodiment, between the input layer 11 and the hidden layer 12, a value corresponding to the power of 2 is a p multiple of 256 (p is a natural number equal to or greater than 2, for example 3). The configured self-attention layer 14 may be further included. The self-attention layer (14) is a layer that performs a technique according to the self-attention mechanism, and is one of the neural network structures that integrates information from the sequence transmitted from the input layer (11). it is kind

For self-focusing mechanisms, Vaswani , output probability to another component or to another sys- A. , Shazeer , N. , Parmar , N. , Uszkoreit , J. , Jones , L. , Gomez , tem , and the like . Examples of the output terminal may 45 A. N. , Kaiser , L. and Polosukhin , I. , 2017. Since it can be understood with reference to, the basic explanation has been simplified.

On the other hand, in an embodiment of the present invention, the self-attention layer 14 determines the importance of each of the intermediate operation vectors X provided as an output of the input layer 11, and excludes information having a relatively low importance. By operating to transfer to the hidden layer 12, it serves to increase the prediction result. In particular, when a learning feature vector (X) for each lesson according to one training course is sequentially input to the input layer 11, a lesson with a large influence on the next training course and a lesson with a small influence are distinguished. It is preferable to configure so as to increase the prediction success rate by passing to (12).

Specifically, the self-attention layer 14 is an intermediate operation received from the input layer 11 using vectors corresponding to each of three variables, namely a query, a key, and a value. After calculating the self-attention value (α) indicating the importance for each vector (X), among the intermediate operation vectors (X) received from the input layer 11, the intermediate value in which the calculated self-attention value is calculated the highest It may be configured to select an operation vector (X) and transmit it to the hidden layer (12).

In order to calculate the self-attention value (α), a q vector corresponding to a query variable, a K vector corresponding to a key variable, and a V vector corresponding to a value variable must first be determined. Here, the q vector is an object for which importance is to be determined, and in the present invention, it may be an intermediate operation vector (X) provided as an output of the input layer 11 . The K vector can serve to derive a reference value for determining importance by being calculated with the q vector. The V vector plays a role in deriving the self-attention value (α) of an appropriate size for the probability value obtained as a result of the operation between the K vector and the q vector. The q vector, the K vector, and the v vector are all vectors having the same dimension.

Referring to FIG. 8 , in one embodiment, the K vector is the total number of lessons according to the current curriculum, the ratio of the number of evaluation passes according to the number of problem solving in the lesson, and the ratio of the average score according to the average number of passes in the current curriculum An initial K vector having each as a component value may be constructed.

Next, the initial K vector may be padded using a preset reference value to have a dimension corresponding to the q vector, and the K vector may be determined as a result value obtained by performing a shift dot product operation with a mask vector. A mask vector is a vector that performs data smoothing on component values to be calculated. Since various types of smoothing vectors exist, a person skilled in the art can selectively use one of well-known smoothing vectors.

The shift dot product operation may mean sequentially performing a dot product operation with the padded initial K vector while the mask vector moves one component value. That is, whenever the dot product operation is performed, sequentially obtained result values become the component values of the K vector.

The V vector may be a vector having a fixed component value preset by an administrator.

Based on the determined q vector, K vector, and V vector, the self-attention value α according to the following equation can be calculated as follows.

Referring to Equation 5 above, activate is an activation function that outputs a probability value with respect to an input value, and may be the aforementioned ReLU function or Softmax function. q is the q vector, K is the K vector, V is the V vector, and d is a constant that lowers the operation result between the q vector and the K vector to less than or equal to a value of a preset size, and may be a preset value. .

In an embodiment of the present invention, the hidden layer 12 has the same number of neurons and may be composed of a plurality of different and sequentially connected hidden layers. For example, the hidden layer 12 may include a first hidden layer, a second hidden layer, a third hidden layer, and a fourth hidden layer that are different and sequentially connected. In this case, the first to fourth hidden layers may be linear layers, and at least one of the first to fourth hidden layers may be a dense layer. Here, the designation of a linear layer or a dense layer can be simply set through a predefined function in Tensorflow, so a description will be omitted.

Also, drop out may be applied to at least one of the first to fourth hidden layers according to a probability p determined according to the number of hidden layers and the number of neurons in each of the hidden layers.

In an embodiment of the present invention, dropout refers to setting the weights of at least some neurons randomly selected according to probability p among neurons constituting a corresponding layer to 0 or a predefined constant, so that some of the neurons are It can mean disable or configure to interfere.

For example, if the first hidden layer consists of 768 neurons, and the probability p is 0.2, then for 20% of randomly selected neurons among 768 neurons (ie, 768/5 neurons) The weight can be set to 0 or a predefined constant.

For example, the probability p may be defined as in Equation 6 below.

In Equation 6, NRn may be the average number of neurons included in the first to fourth hidden layers, and HDn may be the total number of hidden layers and may be 4 when the first to fourth hidden layers are configured.

As such, when dropout is applied to at least one of the first to fourth hidden layers, it is possible to compensate or offset the tendency of overfitting that may occur as the number of hidden layers and the number of each neuron increases, so that a higher prediction effect can be obtained. It can help to bring

In an embodiment, each of the first hidden layer and the third hidden layer to the fourth hidden layer among the first to fourth hidden layers may be configured with the same number of neurons. For example, each may be composed of 768 neurons.

In this case, the number of neurons constituting the second hidden layer connected behind the first hidden layer may be at least 4 times greater than the number of neurons in each of the other hidden layers. When the number of neurons constituting the second hidden layer is set to be larger than the reference value or more than the number of neurons in each of the other hidden layers, there is an advantage in that the tendency of the artificial neural network engine 102b to overfit can be reduced.

9 is a diagram exemplarily illustrating a hardware configuration of an apparatus for recommending personalized learning amount based on big data and artificial intelligence according to an embodiment.

Referring to FIG. 9 , the learning amount recommendation server 100 stores at least one processor 110 and instructions for instructing the at least one processor 110 to perform at least one operation. It may include a memory (memory, 120).

Here, at least one operation is interpreted as including at least some of the operations or functions of the above-described learning amount recommendation server 100, and detailed descriptions are omitted to prevent overlapping descriptions.

Here, the at least one processor 110 may mean a central processing unit (CPU), a graphics processing unit (GPU), or a dedicated processor on which methods according to embodiments of the present invention are performed. can The memory 120 may be configured as at least one of a volatile storage medium and a non-volatile storage medium.

The learning amount recommendation server 100 may further include a storage device 160 for non-temporarily storing input data, intermediate processing data, temporary data, output data, etc. according to performing at least one operation.

For example, the memory 120 may be one of a read only memory (ROM) and a random access memory (RAM), and the storage device 160 is a flash-memory. , a hard disk drive (HDD), a solid state drive (SSD), or various memory cards (eg, micro SD card).

In addition, the learning amount recommendation server 100 may further include a transceiver 130 for performing communication through a wireless network. In addition, the learning amount recommendation server 100 may further include an input interface device 140 , an output interface device 150 , and the like. Each of the components included in the learning amount recommendation server 100 may be connected by a bus 170 to communicate with each other.

For example, the amount of learning recommendation server 100, a communicable desktop computer (desktop computer), a laptop computer (laptop computer), a notebook (notebook), a smart phone (smart phone), a tablet PC (tablet PC), a mobile phone (mobile) phone), smart watch, smart glass, e-book reader, PMP (portable multimedia player), portable game console, navigation device, digital camera, DMB (digital multimedia broadcasting) ) player, digital audio recorder, digital audio player, digital video recorder, digital video player, PDA (Personal Digital Assistant), and the like.

10 is a graph showing statistical analysis results that can confirm the effect of the big data and artificial intelligence-based personalized learning amount recommendation method and apparatus according to an embodiment.

Referring to FIG. 10 , the recommended learning amount provided by the big data and artificial intelligence-based personalized learning amount recommendation method or the big data and artificial intelligence-based personalized learning amount recommendation device according to an embodiment of the present invention is used in the next curriculum. A probability distribution (with AI) according to the evaluation result (X _i ) for the first user group that has been trained is shown.

In addition, referring to FIG. 10 , according to an embodiment of the present invention, learning in the next curriculum is not carried out with the recommended learning amount provided according to an embodiment of the present invention, but learning in the next curriculum is performed with the learning amount according to the user's own selection. 2 The probability distribution (General) according to the evaluation result (X _i ) for the user group is shown.

Here, the current curriculum and the next curriculum performed by the first user group and the second user group are the same, and the number of users belonging to each group is the same. In the drawing, multiplication, division, and fraction/decimal mean the following curriculum, respectively, and have the sequence of addition -> multiplication -> division -> fraction/decimal curriculum.

At this time, as shown in FIG. 9 , in the following curriculum according to almost all academic achievement types except for the case where the academic achievement type is D type and the next educational process is a fraction/decimal, all according to an embodiment of the present invention It is confirmed that a higher evaluation result is obtained when the recommended learning amount is implemented.

Therefore, when learning with the recommended amount of learning provided according to an embodiment of the present invention, the learning effect can be further increased, and thus academic efficiency can be greatly increased. In addition, younger students who find it difficult to select or plan their own learning volume may have a higher academic achievement increase effect when provided with a recommended learning amount.

The methods according to the present invention may be implemented in the form of program instructions that can be executed by various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer-readable medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software.

Examples of computer-readable media may include hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions may 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 device described above may be configured to operate as at least one software module to perform the operations of the present invention, and vice versa.

In addition, the above-described method or apparatus may be implemented by combining all or part of its configuration or function, or may be implemented separately.

Although the above has been described with reference to the preferred embodiment of the present invention, those skilled in the art can variously modify and change the present invention within the scope without departing from the spirit and scope of the present invention as described in the claims below. You will understand that it can be done.

10: Machine Learning Engine
20: Learning amount table
100: learning amount recommendation server
101: learning data collection unit
102: academic achievement predicting unit
102a: clustering engine
102b: artificial neural network engine
103: learning amount table generation unit
104: recommended learning amount determining unit
200: student terminal

Claims (8)

As a personalized learning amount recommendation device based on big data and artificial intelligence,
a learning data collection unit for collecting a plurality of learning data from student terminals corresponding to a plurality of users;
an academic achievement predicting unit for predicting an academic achievement type corresponding to the user of each of the student terminals by using the plurality of learning data;
a learning amount table generation unit for generating a learning amount table by pre-processing the learning data according to the predicted academic achievement type; and
A recommended learning amount determining unit that determines a recommended learning amount corresponding to a target user by referring to the learning amount table, and provides the determined recommended learning amount to a student terminal of the target user;
The academic achievement predicting unit,
By clustering the users using the learning data according to the current curriculum, a cluster corresponding to each of the users is determined, and the academic achievement type assigned to the determined cluster is assigned to the academic achievement ability of the user belonging to the cluster. a clustering engine that predicts by type; and
Supervised-learning is performed using training data composed of training input values corresponding to the learning data of each of the users and training output values labeled with an academic achievement type through the clustering engine, and based on deep learning, the An artificial neural network engine that predicts the academic achievement type corresponding to the target user;
The artificial neural network engine,
an input layer configured to receive a learning feature vector obtained by transforming learning data according to the current curriculum of each of the users, and configured with the same number of neurons as the number of component values of the learning feature vector;
a hidden layer that transmits an output vector calculated using the output value received from the input layer to the output layer;
an output layer that determines a probability corresponding to the output vector by applying an activation function to the output vector, and outputs an output vector having the highest determined probability; and
a self-attention layer that performs a technique according to a self-attention mechanism between the input layer and the hidden layer;
The self-attention concentration layer calculates a self-attention concentration value (α) for each of the intermediate operation vectors obtained as an output of the input layer, and sets the calculated intermediate operation vector with the highest self- attention concentration value of the input layer. It is selected as an output value and delivered to the hidden layer,
The self-attention value (α) is calculated according to the following equation,

In the above equation, activate is an activation function that outputs a probability value for an input value, and is a ReLU function or a Softmax function, q is a q vector representing each of the intermediate operation vectors provided as an output of the input layer, and V is a preset by an administrator. It is a V vector having a set fixed component value, and K is a K vector derived as a result of a shift dot product operation with a mask vector after padding on the initial K vector,
The mask vector is a vector predefined to perform data smoothing on the component values to be calculated, and the initial K vector is the total number of lessons according to the current curriculum and the number of problem solving in the lesson. A personalized learning amount recommendation device based on big data and artificial intelligence, which is a vector having, as a component value, each of the ratio of the number of passing evaluations and the ratio of the average score according to the average number of passes in the current curriculum. delete In claim 1,
The artificial neural network engine, when receiving the training input value, uses an output vector obtained as an output of the hidden layer and a training output vector indicating an academic achievement type according to the training output value to obtain a loss function according to the following equation supervised learning so that the calculation and the result value of the calculated loss function are minimized,

In the above equation, Ym is the mth component (m is a natural number greater than or equal to 1) of the target output vector (Ym), and Y`m is the mth component of the output vector (Y`) obtained as an output of the hidden layer, big data and A personalized learning amount recommendation device based on artificial intelligence. delete delete In claim 1,
After the learning amount table is created,
The learning data collection unit obtains target learning data according to the current curriculum of the target user from the student terminal of the target user,
The academic achievement predicting unit predicts the academic achievement type corresponding to the target user based on the target learning data,
The recommended learning amount determining unit,
In the learning amount table, the recommended learning amount for the next curriculum of the target user is determined by using statistical data for each of a plurality of learning amounts according to the next curriculum corresponding to the academic achievement type predicted for the target user A personalized learning amount recommendation device based on big data and artificial intelligence. In claim 6,
The clustering engine is
A personalized learning amount recommendation device based on big data and artificial intelligence that performs clustering using the K-means algorithm. In claim 6,
The learning amount table generation unit,
After creating the learning amount table,
A personalized learning amount recommendation device based on big data and artificial intelligence that updates the learning amount table by using the learning data for the next training course of the target user according to the recommended learning amount.

Images