System and method for generating personalized math problems for each student

Abstract

The present invention relates to a system and method for generating customized math problems for each student, and more specifically, to a system and method for generating customized math problems based on the student's individual learning level and learning data to enable self-directed learning at a level, thereby encouraging the learners' motivation and stimulating their interest, while simultaneously improving learning efficiency and learning productivity.

Description

Technology Field

[0001] The present invention relates to a system and method for generating customized math problems for each student, and more specifically, to a system and method for generating customized math problems based on the student's individual learning level and learning data to enable self-directed learning at a level, thereby encouraging the learners' motivation and stimulating their interest, while simultaneously improving learning efficiency and learning productivity. Background Technology

[0003] Generally, among the various subjects students learn in school, mathematics is the one they find most difficult to study.

[0004] Mathematics fundamentally deals with many abstract concepts—such as numbers, symbols, and formulas—that are not directly related to concrete objects or situations in the real world. Since it is a discipline where basic concepts build upon each other to advance to advanced concepts, if students move on to the next stage without sufficiently understanding the basic concepts, subsequent learning becomes increasingly difficult.

[0005] Furthermore, recent math question types are increasingly requiring problem-solving skills rather than simply relying on formulas or calculations; students may face significant difficulties in solving these problems if they lack experience in analyzing problems, finding solutions, and applying those methods.

[0006] When students repeatedly struggle to solve math problems, their confidence drops and they gradually lose interest in learning. Due to this problem, the number of students who have given up on math, known as "math dropouts," has been increasing recently.

[0007] In order to solve these problems, research and attempts have recently been made regarding customized learning tailored to the individual learning level of students or learning that helps with gradual understanding by solving problems while gradually increasing the difficulty level. As an example, Korean Registered Patent Publication No. 10-2636079 discloses an apparatus and method for providing customized math problems to learners.

[0008] The aforementioned prior art is characterized by the technical feature of setting a target level of concentration considering the learner's academic ability and learning environment, and providing customized math problems to the learner based on that level of concentration; however, there is a disadvantage in that the difficulty of the problems is determined and provided based on the level of concentration without accurately assessing the learner's level of understanding, which may lead to reduced learning efficiency.

[0009] In other words, since mathematics is a subject where fundamental concepts and principles are important, problem-solving must be performed with concentration while having a good understanding of these concepts or principles. However, if the difficulty of problems is determined based on the target level of concentration without accurately assessing whether the learner has a good understanding of the concepts or principles, there is a problem in that the accuracy of providing problems appropriate to the learner's level may be compromised.

[0010] Furthermore, although the aforementioned prior art includes generating customized math problems to be provided to learners by determining the problem structure and computational difficulty based on target concentration, it has the disadvantage of being insufficient for achieving optimal learning efficiency because it lacks a specific method for generating math problems by selecting a problem structure from a problem bank and determining the computational difficulty, as well as making it difficult for learners to engage in repetitive learning on their weak areas. Prior art literature

[0012] 1. Korean Patent Publication No. 10-2636079 (Registered Feb. 06, 2024) The problem to be solved

[0013] The present invention was devised to solve the problems of the aforementioned prior art technologies. The objective of the present invention is to provide a customized math problem generation system and method for each student that can simultaneously improve learning efficiency and learning productivity by combining OCR (Optical Character Recognition) technology, AI-based problem generation technology, and file format conversion technology to enable learners to intensively and repeatedly study the types of problems they are weak in. means of solving the problem

[0015] The present invention for achieving the above objectives is,

[0016] A mathematics problem generation system comprising a main server for generating mathematics problems to be provided to students, and student terminals and teacher terminals connected to the main server via a network, wherein the main server may include: a problem extraction unit that extracts a problem sheet original text containing only the mathematics problem portions excluding the student's solution portions from a problem sheet solved by a student using OCR technology; a problem generation unit that generates mathematics problems of the same or similar type as the problem extracted from the problem extraction unit using AI functions; and a problem sheet creation unit that selectively combines the mathematics problems extracted from the problem extraction unit and the mathematics problems generated from the problem generation unit to create a problem sheet to be provided to students through the student terminals.

[0017] At this time, the main server may further include a database for storing math problems to be provided to students; and a file format conversion unit for converting the file format of the original problem paper extracted by the problem extraction unit.

[0018] Additionally, the problem extraction unit may include: a font identification module that identifies a unique font used for question generation in a question paper through training a CNN-based font classification model and extracts a problem body that is distinguishable from the student's handwriting; a non-character extraction module that identifies and extracts shapes, graphs, and formulas included in the question paper; and a text reproduction module that generates a mathematical problem that includes the problem body extracted from the font identification module and optionally includes shapes, graphs, and formulas extracted from the non-character extraction module.

[0019] In addition, the problem generation unit may include: a problem analysis module that analyzes a problem extracted from the problem extraction unit to define a problem type and analyzes problem conditions and questions; a numerical transformation module that transforms numerical values ​​included in the problem conditions and questions analyzed by the problem analysis module; and a problem generation module that generates similar problems by applying numerical values ​​transformed by the numerical transformation module to the problem extracted from the problem extraction unit.

[0020] Meanwhile, the method for generating customized math problems for each student according to the present invention is,

[0021] The present invention relates to a method for generating customized math problems for each student, comprising a main server for generating math problems to be provided to students, and student terminals and teacher terminals connected to the main server via a network, wherein the method may include: a text extraction step for extracting and reproducing a text of a problem sheet containing only the math problem portions excluding the student's solution portions from a problem sheet provided by a student terminal that has been solved; a similar problem generation step for generating math problems of the same or similar type as the math problems extracted in the text extraction step; and a problem sheet generation and provision step for generating a problem sheet to be provided to students by selectively combining the math problems extracted in the text extraction step and the math problems generated in the similar problem generation step, and transmitting the result to the student terminal.

[0022] At this time, the original text extraction step may include: a font identification step that identifies the font used to create the question in the question paper through training a CNN-based font classification model and extracts the question body that is distinguishable from the student's handwriting; a non-character extraction step that identifies and extracts shapes, graphs, and formulas included in the question paper; and a problem reproduction step that generates a mathematical problem that includes the question body extracted from the font identification module and optionally includes shapes, graphs, and formulas extracted from the non-character extraction module.

[0023] Additionally, the above-mentioned similar problem generation step may include: a problem analysis step that analyzes the problem type, problem conditions, and question content of a mathematical problem reproduced in the original text extraction step through an AI function; a numerical transformation step that transforms numerical information included in the problem conditions and question content of the mathematical problem analyzed by the problem analysis step using an AI function; and a problem generation step that generates a similar problem by applying the numerical value transformed by the numerical transformation step to the mathematical problem reproduced from the original text extraction step.

[0024] In addition, the above-mentioned similar problem generation step may further include a difficulty setting step that sets the problem difficulty to level 9 by individually considering the number of conditions, the number of calculation steps, the number of necessary concepts, and the adjacency of multiple concepts included in the similar problem generated in the problem generation step. Effects of the invention

[0026] According to the present invention, since differentiated test papers can be generated according to the student's learning level, it has the excellent effect of providing optimal learning materials to students studying independently as well as to educational institutions such as academies and schools.

[0027] In addition, according to the present invention, by generating and providing customized math problems to enable self-directed learning at a level, it is possible to inspire learners' motivation and stimulate their interest, while simultaneously improving learning efficiency and learning productivity. Brief explanation of the drawing

[0029] FIG. 1 is a schematic diagram showing the configuration of a student-specific customized math problem generation system according to the present invention. FIG. 2 is a conceptual drawing of the present invention shown in FIG. 1. Figures 3(a) and 3(b) are diagrams exemplarily showing the original mathematical problem extracted by the problem extraction unit and the twin problem generated by the problem generation unit among the present invention shown in Figure 2. FIG. 4 is a flowchart sequentially illustrating a method for generating customized math problems for each student according to the present invention. Specific details for implementing the invention

[0030] The embodiments of the present disclosure are illustrative for the purpose of explaining the technical concept of the present disclosure. The scope of rights according to the present disclosure is not limited to the embodiments presented below or the specific description thereof.

[0031] All technical and scientific terms used in this disclosure, unless otherwise defined, have the meaning generally understood by those skilled in the art to which this disclosure pertains. All terms used in this disclosure are selected for the purpose of further clarifying this disclosure and are not selected to limit the scope of the rights under this disclosure.

[0032] Expressions such as “comprising,” “comprising,” “having,” etc. used in this disclosure should be understood as open-ended terms implying the possibility of including other embodiments, unless otherwise stated in the phrase or sentence containing such expressions.

[0033] Unless otherwise stated, singular expressions described in this disclosure may include a plural meaning, and this applies likewise to singular expressions described in the claims.

[0035] Hereinafter, preferred embodiments of the student-specific customized math problem generation system and method according to the present invention will be described in detail with reference to the attached drawings.

[0036] FIG. 1 is a schematic diagram showing the configuration of a student-specific customized math problem generation system according to the present invention, FIG. 2 is a conceptual diagram showing the present invention shown in FIG. 1, FIG. 3 (a) and (b) are diagrams showing an example of a math problem original text extracted by a problem extraction unit and a twin problem generated by a problem generation unit among the present invention shown in FIG. 2, FIG. 4 is a flowchart showing a student-specific customized math problem generation method according to the present invention in sequence.

[0038] The present invention relates to a system and method for generating customized math problems for each student based on the student's individual learning level and learning data, which enables self-directed learning tailored to the level, thereby encouraging the learners' motivation and stimulating their interest, while simultaneously improving learning efficiency and learning productivity. The system for generating customized math problems for each student (10) according to the present invention may include, as shown in FIG. 1, a database (110), a main server (100), a student terminal (200), and a teacher terminal (300).

[0039] First, the above database (110) serves to integrally store and manage information regarding mathematical problems and related content provided in the present invention, and the information stored in the database (110) can be provided from the main server (100), student terminal (200), and teacher terminal (300) to be described later.

[0040] That is, the information that can be stored in the above database (110) may include math problems of various types and difficulty levels, font information required for extracting the original text of math problems, and information input from a student terminal (200) or a teacher terminal (300), and each of the information may be stored separately in a separate DB.

[0041] To explain in more detail, the above database (110) may include a problem DB (111), a font DB (112), a form DB (113), a student DB (114), and a teacher DB (115), as shown in FIG. 2. First, the problem DB (111) is configured to store math problems that may be included in a problem sheet to be transmitted to a student terminal (200) by the main server (100) to be described later. Various existing types of math problems, as well as new problems generated by the problem generation unit (140) of the main server (100) to be described later, can be stored separately by unit, type, and difficulty level.

[0042] Next, the font DB (112) is configured to store various font types to be learned in the problem extraction unit (120) of the main server (100) to be described later, and the form DB (113) is configured to store forms that can be used to create test papers in the test paper creation unit (150) of the main server (100) to be described later, and a more detailed explanation thereof will be provided later.

[0043] Next, the above student DB (114) is configured to store information of students using the student terminal (200) described later. Student information entered into the main server (100) through the student terminal (200) can be transmitted to the database (110) via a network and stored in the student DB (114).

[0044] At this time, student information that can be stored in the above student DB (114) may include name, gender, age, school, email, learning records, grade data, etc.

[0045] Next, the above teacher DB (115) is configured to store information of teachers using the teacher terminal (300) described later. Likewise, teacher information entered into the main server (100) through the teacher terminal (300), namely, teacher information including age, gender, contact information, email, major career, etc., can be transmitted to the database (110) via a network and stored in the teacher DB (115).

[0046] As described above, the database (110) may be connected to the main server (100) via a network, but as shown in FIG. 2, the database (110) may also be configured to be included in the main server (100).

[0047] Next, the main server (100) is connected to the database (110) via a network, builds and updates the database (110), and uses information stored in the database (110) to generate customized math problems to be provided to students through the student terminal (200). A more detailed description of the configuration will be provided later.

[0048] Next, the student terminal (200) and the teacher terminal (300) may be a desktop computer, laptop computer, notebook, smartphone, tablet PC, mobile phone, etc., which are capable of communication and are carried by the student and the teacher, respectively.

[0049] That is, the student terminal (200) and the teacher terminal (300) are equipped with communication functions and are connected to the main server (100) via a network. The student terminal (200) receives the problem sheet and problem solution provided by the main server (100) and can simultaneously upload the problem sheet with completed solution to the main server (100) in the form of an image file or a vector file.

[0050] In addition, the teacher terminal (300) can be used to set the difficulty level of math problems generated by the main server (100) or to correct errors, and a more detailed explanation thereof will be provided later.

[0051] That is, the student-specific customized math problem generation system (10) according to the present invention generates a problem sheet at the main server (100) using math problems stored in the database (110) and transmits it to the student terminal (200). When the students transmit the problem sheet they have solved to the main server (100) using the student terminal (200), the main server (100) checks the problem sheet solution results, extracts only the math problem part excluding the solution part of the original problem, for example, a math problem that the student did not get the correct answer to, and generates new problems of the same or similar type and transmits them again to the student terminal (200) in the form of a problem sheet.

[0052] As a result, students can focus on their weak areas, enabling self-directed learning tailored to their level; furthermore, by repeatedly solving various problems of similar types, they can simultaneously improve learning efficiency and productivity.

[0053] To explain in more detail, the main server (100) may include a problem extraction unit (120), a file format conversion unit (130), a problem generation unit (140), a problem paper creation unit (150), and a grading unit (160), as shown in FIG. 2. First, the problem extraction unit (120) may be configured to extract the original text of the problem, i.e., the original text of the problem paper, excluding the part solved by the student from the problem paper (hereinafter collectively referred to as 'problem paper'), such as a test paper solved by the student.

[0054] That is, the problem extraction unit (120) can extract only the problem text by distinguishing the problem text printed on the problem paper from the handwriting written by students during the problem-solving process, and to do this, it can perform learning to identify the font printed on the problem paper.

[0055] To explain in more detail, the problem extraction unit (120) may include a font identification module (122), a non-character extraction module (124), an original text reproduction module (126), and a first editing module (128). The font identification module (122) can identify unique fonts used in the problem sheet by learning fonts through a large amount of training data stored in the font DB (112) of the database (110) using a known CNN (Convolutional Neural Network) based font classification model, and thereby can effectively distinguish between the problem text and the student's handwriting.

[0056] For example, in learning the above typeface, font thickness, alignment, and pattern may be considered, and whether the line thickness and size of the characters are consistent, whether the text is linearly aligned and has consistent spacing, and whether the character shape and pattern are standard may be considered.

[0057] At this time, the problem text extracted by the font identification module (122) can be selectively combined with a non-text portion including shapes, graphs, and formulas extracted by the non-text extraction module (124) described later, and used for creating a problem sheet in the problem sheet creation unit (150).

[0058] In addition, the student's solution portion distinguished by the above-mentioned font identification module (122) can be separately stored in the student DB (114) of the database (110) to establish a digitized learning environment.

[0059] And, the font identified by the font identification module (122) can be used for generating problems and writing problem sheets in the problem generation unit (140) and problem sheet writing unit (150) to be described later.

[0060] Next, the non-character extraction module (124) is configured to identify and extract non-character parts, such as shapes, graphs, and formulas included in the problem sheet. Likewise, the non-character extraction module (124) can determine whether the non-character part is included in the original text of the math problem by considering the thickness, alignment, pattern, etc. of the non-character part in the problem sheet.

[0061] At this time, the non-character extraction module (124) can extract the non-character part and convert it into a vector format.

[0062] In other words, when non-textual parts such as shapes or graphs are saved as images with extensions like PNG or JPG, they are stored in pixel units, which can cause the image to become distorted or unclear depending on the resolution. Therefore, by converting them into vector formats with extensions such as SVG or PDF, shapes, graphs, and function formula information can be preserved exactly as they are in the original text.

[0063] In addition, the above non-character extraction module (124) can be configured to maintain data consistency by converting special symbols and formulas included in the test paper into LateX format.

[0064] Next, the original text reproduction module (126) plays the role of reproducing a completed math problem by combining the problem body extracted through the font identification module (122) and the non-character extraction module (124) with the non-character part. In the case of a math problem where no non-character part exists, the math problem is generated using only the problem body extracted through the font identification module (122), and in the case of a math problem that includes a non-character part, the math problem is generated including the problem body extracted through the font identification module (122) and the shapes, graphs, and formulas extracted by the non-character extraction module (124).

[0065] Next, the first editing module (128) may be configured to edit and correct any incorrectly generated text or non-text parts in the mathematical problem reproduced by the original text reproduction module (126).

[0066] That is, when the reproduction of a math problem is completed by the original text reproduction module (126), teachers can verify it using a teacher terminal (300). If there is a part of the reproduced math problem that does not match the original, teachers can connect to the main server (100) using the teacher terminal (300) and edit the math problem by the first editing module (128).

[0067] Next, the file format conversion unit (130) plays the role of converting the file format of the math problem extracted and reproduced by the problem extraction unit (120) into a digital format file that can be written and edited based on text, such as PDF, HWP, or web text. Through such file format conversion, the problem sheet creation unit (150), which will be described later, can create the problem sheet more easily, and at the same time, the created problem sheet can be easily transmitted and accessed using a student terminal (200) or a teacher terminal (300).

[0068] Next, the problem generation unit (140) uses AI functions to generate math problems of the same or similar types as the problems extracted from the problem extraction unit (120), thereby enabling students to focus on learning and solving math problems of the types they are weak in.

[0069] That is, as will be described later, the problem extraction unit (120) can selectively extract and reproduce the incorrect problems among the problem sheets solved by the students, and the problem generation unit (140) can generate problems of the same or similar type as the reproduced incorrect problems (hereinafter referred to as 'twin problems').

[0070] To explain in more detail, the problem generation unit (140) may include a problem analysis module (141), a numerical transformation module (143), and a problem generation module (144). First, the problem analysis module (141) is configured to analyze a mathematical problem extracted from the problem extraction unit (120), and an AI function may be utilized for the problem analysis.

[0071] That is, when a problem extracted and reproduced through the problem extraction unit (120) is input into the problem analysis module (141), the input problem can be analyzed through an AI function, and this problem analysis may include problem type, problem conditions, question content, etc.

[0072] At this time, the problem type is intended to determine which type of problem the input problem belongs to among various mathematical problem types such as operation, equation, inequality, function, geometry, probability, statistics, sequence, combination, permutation, vector, matrix, calculus, etc., and the problem analysis module (141) can specifically define the problem type by recognizing the problem body transformed into a text-based format by the file format conversion unit (130).

[0073] For example, if the mathematical problem input to the problem analysis module (141) is as in (a) of FIG. 3, the problem type can be defined as a geometric sequence problem.

[0074] In addition, the above problem conditions are analyzed as to what conditions are given for solving the problem. Taking the problem shown in Fig. 3 (a) as an example, b4= 6 and b6= 3b5-24 may correspond to the problem conditions.

[0075] And, the question content is an analysis of what the answer required in the problem is, and similarly, taking the problem shown in Figure 3 (a) as an example, the value of b7 may correspond to the question content.

[0076] Next, the numerical transformation module (143) is configured to transform numerical values ​​included in problem conditions and question content analyzed by the problem analysis module (141), and likewise, can randomly transform numerical information included in problem conditions and question content by utilizing AI functions.

[0077] For example, in the problem shown in Fig. 3 (a), the numerical transformation module (143) can randomly transform the value of b4 = 6 included in the problem condition into c3 = 4, and transform b6 = 3b5-24 into a form such as c5 = 2c4+10, and at the same time change the value of b7 included in the question content to the value of c6, randomly generate options corresponding to incorrect answers based on the calculation result of c6.

[0078] At this time, in the numerical transformation module (143), when the coefficient included in the problem condition is an integer, the transformed coefficient can also be set to an integer to maintain a constant problem difficulty, and to adjust the difficulty, methods such as transforming the coefficient into a rational number can be used to form multiple twin problems for a single original problem with different difficulty levels.

[0079] Next, the problem generation module (144) is configured to generate twin problems, i.e., similar problems, by applying a numerical value modified by the numerical modification module (143) to a problem extracted from the problem extraction unit (120), and the generated twin problems can likewise be saved as files in digital formats such as PDF, HWP, and web text.

[0080] That is, the problem generation module (144) can generate a twin problem as shown in FIG. 3 (b) by applying numerical values ​​modified by the numerical transformation module (143) to the original mathematical problem text shown in FIG. 3 (a).

[0081] Meanwhile, the problem generation unit (140) may further include a difficulty setting module (145), wherein the difficulty setting module (145) is configured to automatically set the difficulty of similar problems generated through the problem generation module (144), and the number of conditions, the number of calculation steps, the number of necessary concepts, and the adjacency of multiple concepts may be used as criteria for setting the difficulty.

[0082] That is, the difficulty setting module (145) can be configured to automatically distinguish the difficulty of the generated twin problems into 9 levels by setting the difficulty from 1 to 3 based on individual criteria such as the number of conditions of the generated problem, the number of calculation steps, the number of necessary concepts, and the proximity of multiple concepts, assigning difficulty 1 (difficulty of number of conditions 1, difficulty of number of calculation steps 1, difficulty of number of necessary concepts and proximity of multiple concepts 1) to the problem with the lowest difficulty and assigning difficulty 9 (difficulty of number of conditions 3, difficulty of number of calculation steps 3, difficulty of number of necessary concepts and proximity of multiple concepts 3) to the problem with the highest difficulty, and can be configured so that the problem sheet creation unit (150) described later can select problems to be included in the problem sheet by considering the difficulty.

[0083] Additionally, the problem generation unit (140) may further include a second editing module (146), and the second editing module (146) may be configured to correct incorrectly generated text or non-character parts in a math problem generated by the problem generation module (144), similar to the first editing module (128) described above.

[0084] The second editing module (146) can also be used for manually setting the difficulty level of the generated twin problems or modifying the problem type. Teachers can check the problems generated by the problem generation unit (140) using the teacher terminal (300), and then specify or modify the type and difficulty level of the problems through the second editing module (146).

[0085] At this time, the same difficulty setting criteria as described above may be applied to the difficulty setting or modification by the teacher terminal (300), and twin problems created and with their types and difficulty levels set in this way may be separated by unit, type, and difficulty level and stored in the problem DB (111) of the database (110).

[0086] Additionally, the problem generation unit (140) may further include a type determination module (142), and the type determination module (142) may select and retrieve mathematical problems corresponding to a type similar to the problem type defined by the problem analysis module (141) from the problem DB (111) of the database (110).

[0087] That is, if the test paper is composed only of twin problems generated by applying the numerical values ​​modified by the numerical transformation module (143), students may become bored with solving the problems or the improvement of their learning ability may be limited. Therefore, when a problem type is defined by the problem analysis module (141), the problem type determination module (142) allows mathematical problems belonging to that problem type to be selected and retrieved from the problem DB (111) of the database (110), thereby making the mathematical problems to be included in the test paper to be written by the test paper writing unit (150) described later more diverse.

[0088] At this time, the mathematical problems selected by the type determination module (142) can also be changed by the numerical transformation module (143) to numerical values ​​included in the problem conditions and question content, and the problem generation module (144) can additionally generate twin problems using the changed numerical values.

[0089] Next, the above-mentioned problem sheet creation unit (150) serves to create problem sheets to be provided to students through the student terminal (200), and the problem sheet may include original mathematical problem text extracted and reproduced by the problem extraction unit (120) and twin problems generated by the problem generation unit (140).

[0090] The above test paper can be created by including the original math problem text and twin problems in the test paper form stored in the form DB (113) of the database (110), and can be created using the same font as the original text identified by the font identification module (122) and saved as a digital file such as PDF, HWP, or web text.

[0091] Next, the main server (100) may further include a grading unit (160) and a communication unit (170). The grading unit (160) is configured to grade completed problem sheets transmitted from a student terminal (200) to verify whether each math problem is correct, and the communication unit (170) is configured to perform wired and wireless communication processes with a database (110), a student terminal (200), and a teacher terminal (300) when configured independently of the main server (100). Various known wired and wireless communication networks such as Ethernet, Internet, WIFI, and LTE may be utilized.

[0092] The scoring unit (160) and the communication unit (170) mentioned above are known configurations and are not claimed as rights in the present invention; therefore, a more detailed description thereof will be omitted.

[0094] Meanwhile, the method for generating customized math problems for each student according to the present invention relates to a method for generating customized math problems for each student using the aforementioned system (10), and as shown in FIG. 4, the configuration may largely include a problem provision and grading step (S10), a text extraction step (S20), a file format conversion step (S30), a similar problem generation step (S40), and a problem sheet generation and provision step (S50).

[0095] First, the problem provision and grading step (S10) is a process for verifying the student's learning level to generate customized problems for each student. The main server (100) transmits the problem sheet to the student terminal (200) through the communication unit (170), and the student can print out the transmitted problem sheet, solve it, and then photograph and transmit the problem sheet including the solution process and answers.

[0096] The grading unit (160) of the main server (100) can grade the test paper transmitted through the student terminal (200) and check whether each question is correct.

[0097] Next, the above original text extraction step (S20) is a process for extracting and reproducing the original text of a test paper that includes only the math problem parts excluding the student's solution part from the completed test paper provided from the student terminal (200), and the extraction and reproduction of the original text of the test paper may be carried out mainly on math problems identified as incorrect answers as a result of grading by the grading unit (160).

[0098] In addition, the math problems extracted and reproduced in the above original text extraction step (S20) may include not only problems identified as incorrect answers but also past exam questions by problem type or problems containing essential concepts that must be known.

[0099] To explain in more detail, the above original text extraction step (S20) may include a font identification step (S22), a non-character extraction step (S24), and a problem reproduction step (S26). First, the above font identification step (S22) is a process for identifying the font used to create the problem in the original text of the problem sheet. The font identification module (122) of the main server (100) can identify the unique font used to create the problem through learning a CNN-based font classification model, and accordingly, the main text of the problem can be extracted separately from the student's handwriting.

[0100] Next, the above-mentioned non-character extraction step (S24) is a process for identifying and extracting non-character parts included in a math problem. As described above, the non-character extraction module (124) of the main server (100) can determine whether the non-character parts are included in the original text of the math problem by considering the thickness, alignment, pattern, etc. of the non-character parts, such as shapes, graphs, and formulas included in the problem sheet, and extract them.

[0101] At this time, in the above non-character extraction step (S24), the extracted non-character parts can be converted into a vector format having an extension such as SVG or PDF so that shapes, graphs, and function formula information can be maintained as they are in the original text, and special symbols and formulas included in the question paper can be converted into a LateX format so that data consistency can be maintained.

[0102] Next, the problem reproduction step (S26) is a process for reproducing a completed math problem by combining the problem body and non-character parts extracted through the font identification step (S22) and the non-character extraction step (S24). In the original text reproduction module (126) of the main server (100), as described above, in the case of a math problem where no non-character parts exist, a math problem is generated using only the problem body, and in the case of a math problem where non-character parts are included, a math problem is generated including the problem body and non-character parts such as shapes, graphs, and formulas.

[0103] Meanwhile, the above original text extraction step (S20) may further include a first editing step (S28). The first editing step (S28) is a process for editing and correcting incorrectly generated text or non-character parts that differ from the original text in the math problem reproduced through the problem reproduction step (S26). Teachers can complete the math problem identical to the original text by editing the parts that differ from the original text through the first editing module (128) after connecting to the main server (100) using a teacher terminal (300).

[0104] Next, the file format conversion step (S30) is a process for converting the file format of the mathematical problem reproduced in the original text extraction step (S20) into a digital format file that can be created and edited based on text, such as PDF, HWP, or web text. Since such file format conversion technology is already known in various fields, a detailed explanation thereof will be omitted.

[0105] Next, the above-mentioned similar problem generation step (S40) is a process for generating mathematical problems of the same or similar type as the mathematical problem extracted and reproduced in the original text extraction step (S20). As described above, the problem generation unit (140) of the main server (100) can generate twin problems of the same or similar type as the reproduced original mathematical problem text using an AI function.

[0106] To explain in more detail, the above-mentioned similar problem generation step (S40) may include a problem analysis step (S41), a type selection step (S42), a numerical transformation step (S43), a problem generation step (S44), a difficulty setting step (S45), and a second editing step (S46). First, the above-mentioned problem analysis step (S41) is a process for analyzing a mathematical problem reproduced in the original text extraction step (S20). The problem analysis module (141) of the main server (100) can recognize the problem body transformed into a text-based format by the file format conversion step (S30) through an AI function to analyze the problem type, problem conditions, question content, etc.

[0107] Next, the type selection step (S42) is a process for selecting and retrieving math problems corresponding to a type similar to the problem type analyzed by the problem analysis step (S41) from the problem DB (111) of the database (110), and math problems of a similar problem type selected by the type determination module (142) of the main server (100) may be included in the problem sheet to be provided through the student terminal (200).

[0108] Next, the numerical transformation step (S43) is a process for transforming numerical information included in the problem conditions and question content of the original math problem text analyzed by the problem analysis step (S41) and numerical information included in a similar type of math problem selected through the type selection step (S42). The numerical transformation module (143) of the main server (100) can randomly transform the numerical information included in the math problem by utilizing AI functions.

[0109] In addition, in the numerical transformation step (S43), numerical information can be transformed according to specific rules using an AI algorithm, for example, the common difference in a sequence problem can be changed, or the size or ratio of a shape in a geometric figure problem can be transformed, and numerical values ​​included in the original problem or a similar type of math problem can be transformed into different numerical values ​​by predicting the change in y value according to the change in x value in a function problem through regression analysis.

[0110] Next, the problem generation step (S44) is configured to generate a twin problem, i.e., a similar problem, by applying a numerical value modified by the numerical transformation step (S43) to a mathematical problem original text reproduced from the original text extraction step (S20) or a mathematical problem of a similar type selected in the type selection step (S42). The twin problem generated by the problem generation module (144) of the main server (100) can be saved as a file in a digital format such as PDF, HWP, or web text.

[0111] Next, the difficulty setting step (S45) is a process for automatically setting the difficulty of twin problems generated in the problem generation step (S44). The difficulty setting module (145) of the main server (100) can set the difficulty of the twin problems to 9 levels from 1 to 9 by individually considering the number of conditions included in the twin problems, the number of calculation steps, the number of necessary concepts, and the adjacency of multiple concepts.

[0112] Next, the second editing step (S46) is a process for editing and correcting the twin problem generated in the problem generation step (S44) if there are incorrectly generated text or non-character parts. Teachers can check the generated twin problem using a teacher terminal (300) and then edit the incorrectly generated parts through the second editing module (146) of the main server (100).

[0113] Additionally, as described above, in the second editing step (S46), teachers can use the teacher terminal (300) to manually specify the difficulty level of the twin problem or correct an incorrectly set problem type or difficulty level.

[0114] Twin problems completed up to the second editing stage (S46) through the above process can be classified by unit, problem type, and difficulty level and stored in the problem DB (111) of the database (110).

[0115] Next, the above-mentioned problem sheet generation and provision step (S50) is a process for generating a problem sheet to be provided to students by selectively combining the math problem extracted in the original text extraction step (S20) and the math problem generated in the similar problem generation step (S40), i.e., twin problems, and transmitting it to the student terminal (200). In the problem sheet creation unit (150) of the main server (100), the math problem extracted in the original text extraction step (S20) and the twin problems generated in the similar problem generation step (S40) are selectively combined in the problem sheet form stored in the form DB (113) of the database (110) to generate a problem sheet to be provided to students, and then the generated problem sheet can be transmitted to the student terminal (200) through the communication unit (170).

[0116] At this time, the math problems included in the above test paper may include past exam questions by question type or questions containing core concepts by question type, and the problem DB (111) of the above database (110) may be configured so that math problems that must be included in the test paper by question type are managed separately, and that the separately managed math problems are included whenever a test paper for the corresponding question type is created.

[0117] As described above, when the problem sheet generated by the student terminal (200) is transmitted, the problem-solving and grading step (S10) starts again, and through the repetition of this process, students can intensively study the core concepts and weak areas for each problem type.

[0118] Accordingly, the system and method for generating customized math problems for each student according to the present invention as described above can generate differentiated problem sheets based on the student's learning level, thereby providing optimal learning materials to students who study independently as well as to educational institutions such as academies and schools. Furthermore, by generating and providing customized math problems to enable self-directed learning tailored to the level, it can inspire the learners' motivation and stimulate their interest, while simultaneously improving learning efficiency and productivity.

[0120] The systems and methods described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. The processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0121] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave in order to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0122] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and may be recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to a person with ordinary knowledge in the field of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory.

[0123] Although the embodiments have been described above with reference to limited examples and drawings, those skilled in the art can make various modifications and variations from the description above. For example, suitable results can be achieved even if the described techniques are performed in a different order than described, and / or the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents. Explanation of the symbols

[0125] 10: Math Problem Generation System 100 : Main Server 110 : Database 111 : Problem DB 112 : Font DB 113 : Form DB 114 : Student DB 115 : Teacher DB 120 : Problem Extraction Unit 122 : Font Identification Module 124 : Non-character extraction module 126 : Original text reproduction module 128 : 1st Edit Module 130 : File format conversion section 140 : Problem Generation Section 141 : Problem Analysis Module 142 : Type determination module 143 : Numerical transformation module 144 : Problem Generation Module 145 : Difficulty setting module 146 : 2nd Edit Module 150 : Question Paper Preparation Department 160 : Scoring Department 170 : Communications Department 200 : Student terminal 300 : Teacher terminal S10: Problem-solving and grading stage S20 : ​​Original text extraction stage S22 : Font identification stage S24: Non-character extraction step S26 : Problem reproduction stage S28 : 1st editing stage S30: File format conversion step S40: Similar problem generation stage S41: Problem Analysis Phase S42: Type Selection Step S43: Numerical transformation stage S44: Problem Generation Phase S45: Difficulty setting stage S46 : 2nd editing stage S50: Question paper generation and provision stage

Claims

Claim 1 A mathematics problem generation system comprising a main server for generating mathematics problems to be provided to students, and student terminals and teacher terminals connected to the main server via a network, wherein the main server comprises: a problem extraction unit that extracts a problem sheet original text containing only the mathematics problem portions excluding the student's solution portions from a problem sheet solved by a student using OCR technology; a problem generation unit that generates mathematics problems of the same or similar type as the problem extracted from the problem extraction unit using AI functions; and a problem sheet creation unit that selectively combines the mathematics problems extracted from the problem extraction unit and the mathematics problems generated from the problem generation unit to create a problem sheet to be provided to students through the student terminals; a student-specific customized mathematics problem generation system. Claim 2 In claim 1, the main server further comprises: a database storing math problems to be provided to students; and a file format conversion unit converting the file format of the original problem paper extracted by the problem extraction unit; a student-specific customized math problem generation system. Claim 3 In claim 1, the problem extraction unit comprises: a font identification module that identifies a unique font used for question generation in a question paper through CNN-based font classification model learning and extracts a question body that is distinguishable from the student's handwriting; a non-character extraction module that identifies and extracts shapes, graphs, and formulas included in the question paper; and a text reproduction module that generates a math problem that includes the question body extracted from the font identification module and optionally includes shapes, graphs, and formulas extracted from the non-character extraction module, thereby creating a customized math problem generation system for each student. Claim 4 In claim 2, the problem generation unit comprises: a problem analysis module that analyzes a problem extracted from the problem extraction unit to define a problem type and analyzes problem conditions and questions; a numerical transformation module that transforms numerical values ​​included in the problem conditions and questions analyzed by the problem analysis module; and a problem generation module that generates similar problems by applying numerical values ​​transformed by the numerical transformation module to the problem extracted from the problem extraction unit; a student-specific customized math problem generation system. Claim 5 The method for generating customized math problems for each student includes a main server for generating math problems to be provided to students, and a student terminal and a teacher terminal connected to the main server via a network, comprising: a text extraction step for extracting and reproducing a text of a problem sheet containing only the math problem portion excluding the student's solution portion from a problem sheet provided from a student terminal that has completed the solution; a similar problem generation step for generating math problems of the same or similar type as the math problem extracted in the text extraction step; and a problem sheet generation and provision step for generating a problem sheet to be provided to students by selectively combining the math problem extracted in the text extraction step and the math problems generated in the similar problem generation step, and transmitting the result to the student terminal. Claim 6 In claim 5, the above-mentioned text extraction step comprises: a font identification step that identifies the font used to create the question in the question paper through training a CNN-based font classification model and extracts the question body that is distinguishable from the student's handwriting; a non-character extraction step that identifies and extracts shapes, graphs, and formulas included in the question paper; and a problem reproduction step that generates a math problem including the question body extracted from the font identification module and optionally including shapes, graphs, and formulas extracted from the non-character extraction module; a method for generating customized math problems for each student. Claim 7 In claim 5, the above-mentioned similar problem generation step comprises: a problem analysis step that analyzes the problem type, problem conditions, and question content of a math problem reproduced in the original text extraction step through an AI function; a numerical transformation step that transforms numerical information included in the problem conditions and question content of the math problem analyzed by the problem analysis step using an AI function; and a problem generation step that generates a similar problem by applying the numerical value transformed by the numerical transformation step to the math problem reproduced from the original text extraction step; a method for generating customized math problems for each student. Claim 8 In claim 7, the above-mentioned similar problem generation step further includes a difficulty setting step that sets the problem difficulty to 9 levels by individually considering the number of conditions, the number of calculation steps, the number of necessary concepts, and the adjacency of multiple concepts included in the similar problem generated in the problem generation step.

Images