JP7456589B1 - Learning support device, learning support system, learning support method, and program - Google Patents

Abstract

An object of the present invention is to provide a device that supports a user's learning by generating explanations for answers to questions.
A learning support device 200 receives a question 10 and a question answer 20 input by a user from a user terminal 100, transmits the question and answer to a generation AI 300, and further receives a question from the user terminal after sending the question and answer. A transmitting/receiving unit 210 that receives a question 30 input by a user and transmits the question to a generating AI, and a generating unit 220 that receives an answer 50 to the question from the generating AI and generates an explanation 40 for the user based on the answer. Be prepared. According to this, not only the questions but also the answers to the questions can be input into the generation AI, and then questions from the user can be input into the generation AI, and an explanation for the user can be generated based on the answers generated by the generation AI. , it is possible to answer the user's questions, that is, the questions that the user does not understand.
[Selection diagram] Figure 1

Description

The present invention relates to a learning support device, a learning support system, a learning support method, and a program that support a user's learning by generating explanations for answers to questions.

In learning, explanations of problems are important for deepening understanding. However, explanations are sometimes omitted due to space limitations, and sufficient explanations are often not provided. Even if you send a question to a generation AI and have it generate an explanation, it is difficult to obtain an explanation that will deepen the user's understanding. Here, Non-Patent Document 1 states that the March 2023 version of GPT-4 (registered trademark) showed 84% accuracy in identifying prime numbers and composite numbers, while the June version of the same year showed 84% accuracy in identifying prime numbers and composite numbers. The accuracy of identifying and composite numbers decreased to 51%, indicating that the behavior of large-scale language model services may change significantly in a relatively short period of time.
[Prior art documents]
[Non-patent literature]
[Non-patent Document 1] Chen et al. "How Is ChatGPT's Behavior Changing over Time?", arXiv:2307.09009 [cs. CL], July 18, 2023

In a first aspect of the present invention, there is provided a learning support device that supports a user's learning by generating explanations for answers to questions, the device receiving questions input by the user and answers to the questions from a user terminal, a transmitting/receiving unit that transmits the question and the answer to the generation AI, further receives a question input by the user from the user terminal after transmitting the question and the answer, and transmits the question to the generation AI; A learning support device is provided that includes a generation unit that receives an answer to the question from a generation AI and generates an explanation for the user based on the answer.

In a second aspect of the present invention, a system is provided that includes the problem explanation device of the first aspect, the user terminal, and the generation AI.

In a third aspect of the present invention, there is provided a learning support method for supporting a user's learning by generating explanations for answers to questions, the method comprising: receiving questions input by the user and answers to the questions from a user terminal; transmitting the question and the answer to the generation AI; and after transmitting the question and the answer, further receiving a question input by the user from the user terminal and transmitting the question to the generation AI; A learning support method is provided, comprising the steps of receiving an answer to the question from the generating AI, and generating an explanation for the user based on the answer.

In the fourth aspect of the present invention, in order to generate an explanation for the answer to the question, the computer receives the question input by the user from the user terminal and the answer to the question, and sends the question and the answer to the generation AI. means for transmitting the question and the answer; further, after transmitting the question and the answer, further receiving a question input by the user from the user terminal and transmitting the question to the generation AI; and receiving an answer to the question from the generation AI. and generating an explanation for the user based on the answer.

Note that the above summary of the invention does not list all the features of the invention. Furthermore, subcombinations of these features may also constitute inventions.

1 shows the configuration of a learning support system according to this embodiment. The operation flow of the learning support system is shown. An example of a user terminal screen at the start of operation is shown. An example of a user terminal screen when transmitting a question and an answer (state before transmitting the question) is shown. Here is an example of the problem. An example of an answer is shown below. The detailed flow of sending and receiving questions and answers is shown. 13 shows an example of a user terminal screen when a question and an answer are sent (the state after the question is sent but before the answer is sent). An example of a user terminal screen at the time of transmitting a question and an answer (state after transmitting the answer) is shown. An example of a user terminal screen when transmitting a question and an answer (in a state where the answer is not at hand) is shown. An example of a user terminal screen when sending a question is shown. Here is an example of a question. An example of an answer is shown below. The detailed flow of explanation generation in question answering is shown. An example of multiple key points is shown. An example of explanation is shown. An example of explanation of different generation methods is shown below. An example of a user terminal screen when receiving an explanation for a question is shown. An example of a user terminal screen when receiving an explanation for an additional question is shown. The operation flow of the learning support system when selecting main points is shown. The operation flow of the learning support system when grading answers is shown. An example of a user terminal screen when requesting grading is shown. An example of an answer is shown below. An example of a user terminal screen when receiving scoring results and explanations is shown. An example of the configuration of a computer is shown.

Hereinafter, the present invention will be described through embodiments of the invention, but the following embodiments do not limit the invention according to the claims. Furthermore, not all combinations of features described in the embodiments are essential to the solution of the invention.

FIG. 1 shows the configuration of a learning support system 1 according to this embodiment. The learning support system 1 is a system that generates highly accurate explanations in response to user questions and improves the user's learning effect, and includes a user terminal 100, a learning support device 200, a generative artificial intelligence (generative AI) 300, and An optical character recognition device (OCR device) 400 is provided. These are communicably connected to each other via a network 500 such as the Internet.

The user terminal 100 is a communication terminal owned or used by a user, and is, for example, a smartphone, a tablet terminal, or a personal computer. A plurality of user terminals 100 may be included in the learning support system 1, each having individual individual recognition information, and may transmit individual recognition information such as an ID, a tag, or a user name during communication. This allows a plurality of users to independently communicate with the learning support device 200 using the user terminals 100, respectively. In this embodiment, the configuration and functions of the learning support system 1 will be described with attention to one user terminal 100 used by one user.

The user terminal 100 includes a central processing unit (CPU), a communication device, and a storage device (all not shown). By executing a dedicated program by the CPU, a recording function for question explanations is realized. Note that the dedicated program is, for example, stored in a ROM (not shown) and read by the CPU, or stored in a storage medium such as an SD card and read by the CPU using a reading device (not shown), or read out by the CPU using a reading device (not shown), or It is stored in (multiple distributed servers or multiple subsystems, etc.), and is activated by reading it out from the cloud via a communication device by the CPU and deploying it in RAM. The communication device is a means for mutually communicating with the learning support device 200 and the OCR device 400 via the network 500, and can communicate using a protocol such as TCP/IP, for example. The communication device may employ either wired communication or wireless communication. The storage device is a storage device such as a NAND flash memory that stores a history of problems 10, answers 20, questions 30, and explanations 40, which will be explained later. The user terminal 100 can communicate with the learning support device 200 via the network 500 and provide the learning support device 200 with the questions 10, answers 20, and questions 30 necessary for generating explanations.

Note that the user terminal 100 may have a photographing function such as a camera. In such a case, the user can use the user terminal 100 to convert the problem 10, the answer 20, and the question 30 into image information using the photographing function and provide the image information to the learning support device 200. The user may include each of Question 10, Answer 20, and Question 30 in individual image information, specify the area of each image information, and provide the image information. 100, read it out, and provide it. Further, two or more of the problem 10, the answer 20, and the question 30 may be included in one image information, and respective areas may be specified and provided.

The learning support device 200 is a computer device that uses the generation AI 300 to generate highly accurate explanations in response to user questions. The learning support device 200 includes a central processing unit (CPU), a communication device, and a storage device (all not shown). The commentary generation function is realized by executing a dedicated program by the CPU. Note that the dedicated program may be stored, for example, in a ROM (not shown) and read by the CPU, or may be stored in a storage medium such as a CD-ROM and read by the CPU using a reading device (not shown), or It is stored in the cloud (multiple distributed servers or multiple subsystems, etc.), and is activated by the CPU reading it from the cloud via a communication device and deploying it in RAM. The communication device is a means for mutually communicating with the user terminal 100, the generated AI 300, and the OCR device 400 via the network 500, and can communicate using a protocol such as TCP/IP, for example. The communication device may employ either wired communication or wireless communication. The storage device is a storage device such as a hard disk drive (HDD) that stores various information such as problems 10, answers 20, questions 30, and explanations 40. In this embodiment, the storage device is provided within the learning support device 200. The learning support device 200 communicates with the generation AI 300 via the network 500 to provide the generation AI 300 with questions 10, answers 20, and questions 30 necessary for explanation generation, and communicates with the user terminal 100 via the network 500. The generated commentary 40 can be provided to the user terminal 100 through communication. Note that an external storage device (not shown) may be used as the storage device.

The generation AI 300 is an AI that learns patterns and relationships based on collected training data and generates content. In this embodiment, text generation AI is used as the generation AI 300, but there is no need to limit it to this. The generation AI 300 may use a commercial generation AI, and examples of commercial text generation AI include GPT-4 (registered trademark), Llama2, PaLM2, and the like. By using commercially available generation AI, a system can be easily constructed without the need for preparations such as collecting training data.

The OCR device 400 is a device that extracts characters from image information and converts them into text information. The OCR device 400 may use a device equipped with commercial OCR software, a commercial OCR device, or a commercial online OCR service, and the commercial OCR software or online OCR service may include, for example, SmartRead (registered trademark), Examples include smart OCR (registered trademark) and Seisho (registered trademark). In this embodiment, the OCR device 400 is connected via the network 500, but the OCR device 400 is not limited to this and may be incorporated into the learning support device 200.

FIG. 2 illustrates an operational flow of the learning support system 1 according to this embodiment. FIG. 3 shows an example of the screen display of the user terminal 100 at the start of the operation. Three options are displayed on the screen of the user terminal 100: "Take a picture of the question," "Ask a question," and "Score." When the user selects "take a picture of a question" on the user terminal 100, "request explanation" is executed and the flow is started.

In step S10, the user photographs the question 10 and the answer 20 with the user terminal 100, and instructs the user terminal 100 to transmit the question 10 and the answer 20. As shown in FIG. 4, a message 1001 "Take a picture of the question you don't understand or select an image of the question" is displayed on the screen of the user terminal 100 to prompt the user to transmit the question 10. In response, the user photographs the question 10 and the answer 20.

FIG. 5A shows an example of Problem 10. Problem 10 is a problem described in, for example, a textbook or problem collection, and the subject is not particularly limited. Here, the user may generate image information of question 10 using the camera of the user terminal 100, and may cut out a portion including question 10 from the captured image information. Note that the image information for question 10 that was generated in the past and saved in the user terminal 100 may be read and used.

In this example, problem 10 is a problem regarding mathematical permutations. The big question includes the question ``Create permutations using all the letters in something. In this case, how many ways are there such as the following?'' "(2) O, m, and e are adjacent to each other." (3) "o, m are not adjacent to each other." (4) "h, i, n, and g remain in this order." It includes five questions: ``(5) s is to the left of t and h is to the right of t.''

Note that the subject or field of question 10 is not limited, but if the question includes a diagram, the diagram may be transmitted to the learning support device 200 using image information and the question text may be transmitted to the learning support device 200 using text information. Further, even if the user sends problem 10 as image information to the learning support device 200, the learning support device 200 deletes the diagram information if the solution can be derived from the problem text alone without the diagram. It's okay. If the problem cannot be sent at once, such as because the problem is too long, the problem may be sent and received in multiple parts.

Note that in this embodiment, question 10 is generated as image information, but the problem is not limited to this and may be generated as text information or audio information. In such a case, the user terminal 100 may prompt the user to generate question 10 as text information or audio information by displaying it on the screen.

FIG. 5B shows an example of the answer 20. The answer 20 is the answer to the question 10, and is described in, for example, a textbook or an answer book or an explanation book attached to the problem book, and the subject is not limited. Here, the user may generate image information of the answer 20 using the camera of the user terminal 100, and may cut out a portion including the answer 20 from the photographed image information. Further, image information of the answer 20 generated in the past and saved in the user terminal 100 may be read out and used.

In this example, answer 20 is the answer to problem 10 regarding mathematical permutations. The answers to (1) and (2) are shown as the answer to Problem 10, and the answers are shown as follows: ``(1) There are two ways to enter s and g.Other letters can be arranged normally, so 2 x 7 !=10080 (street)" (2) If o, m, and e are considered as one, the whole can be considered as 7 characters. There are 7! ways to arrange them, and for each of them, o, m , there are 3! ways of permuting e, so 7!×3!=30240 (ways).''

The answer 20 only needs to include the part of the answer that the user wants to ask a question about, and when the user asks a question about a small question, it does not need to include the answers to all the small questions included in the large question. In the case of explanations with diagrams inserted, the diagram information is missing when transmitting anything other than image information, so the diagrams may be transmitted to the learning support device 200 using image information, and the questions may be transmitted to the learning support device 200 using text information. Further, even if the user transmits the answer 20 as image information to the learning support device 200, the learning support device 200 may delete the information in the diagram if it is not necessary.

Note that in this embodiment, the answer 20 is generated as image information, but the answer 20 is not limited to this and may be generated as text information or audio information. In such a case, the user terminal 100 may prompt the user to generate the answer 20 as text information or audio information by displaying it on the screen.

When the user operates the user terminal 100 (by tapping the screen, etc.), the generated questions 10 and answers 20 are transmitted from the user terminal 100 to the learning support device 200.

In step S100, the learning support device 200 receives the question 10 and the answer 20 from the user terminal 100. FIG. 6 shows a detailed flow of step S100.

In step S110, the transmitter/receiver 210 receives question 10 (see FIG. 5A) from the user terminal 100. As shown in FIG. 7, when the learning support device 200 receives the question 10 from the user terminal 100, the learning support device 200 sends a message 1003 prompting the user terminal 100 to send the answer 20, “Take a picture of the answer or select an image of the answer. Please send. Note that in FIG. 7 and other drawings showing the display screen of the user terminal 100, the message displayed on the left side of the screen of the user terminal 100 is a message sent from the learning support device 200 and received by the user terminal 100. , the message displayed on the right side is a message sent from the user terminal 100 to the learning support device 200. Further, the user terminal 100 displays "I don't have the answer" and "Photograph the answer" which the user can select.

In step S120, the transmitting/receiving unit 210 receives the answer 20 (see FIG. 5B) from the user terminal 100. As shown in FIG. 8A, when the learning support device 200 receives the answer 20 from the user terminal 100, the transmitted answer 20 is displayed on the screen.

Note that, as shown in FIG. 8B, if the user selects "I don't have the answer" when the answer 20 cannot be sent from the user terminal 100, the learning support device 200 sends a warning message regarding the decrease in the accuracy of the explanation. 1005 ``If there is no answer, the accuracy of the explanation may decrease'' is sent to the user terminal 100. If the transmitting/receiving unit 210 cannot obtain the answer 20 from the user terminal 100, it searches for an explanation of the same question or a question with a similar structure from the user's question history recorded by the learning support device 200 or the transmission history of all users. You may also obtain it by doing so.

In step S130, question 10 and answer 20, which are image information, are converted into text information. Here, the transmitting/receiving unit 210 transmits the image information of the question 10 and the answer 20 to the OCR device 400. The OCR device 400 converts the image information into text information and sends it back to the transmitter/receiver 210. If the question 10 and the answer 20 are text information, step S130 is omitted. According to this, the question 10 or the answer 20 input as image information can be input into the generation AI 300 as text information, and the accuracy of the answer 50 is increased compared to inputting the image information as is into the generation AI 300. It is possible to generate a highly accurate explanation 40 for the user based on the , and answer the user's question 30 , that is, the question 10 and the answer 20 .

In step S140, the transmitting/receiving unit 210 transmits the question 10 and the answer 20, which are text information, to the generation AI 300. Thereby, step S100 is completed, and the learning support device 200 enters a standby state. On the other hand, the generation AI 300 receives the question 10 and the answer 20.

After the learning support device 200 goes into standby mode, when the user selects "Ask a question" on the screen of the user terminal 100 (see FIG. 3), the process proceeds to step S20. In response, the learning support device 200 transmits a message 1101 to the user terminal 100 prompting the user to send a question 30. As a result, the message 1101 "What's your question today?" is displayed on the screen of the user terminal 100, as shown in FIG. 9.

In step S20, the user inputs the question 30 on the user terminal 100 and instructs the user terminal 100 to send the question 30. Here, the user creates a question 30 on the user terminal 100.

FIG. 10 shows an example of the question 30. Question 30 is created from a user's question regarding question 10 and answer 20. In this embodiment, the question 30 is generated as text information, but the question 30 is not limited to this and may be generated as image information or audio information.

In this example, question 30 is a question about problem 10 and answer 20 regarding mathematical permutations. Question 30 asks the meaning of (2) in answer 20 of question 10, "The way they are arranged is 7!" The question 30 may be any content that can specify which part of the question 10 or the answer 20 the user is asking about. For example, it may specify the number of the subquestion or the number of lines in the answer 20, or it may specify the number of the question to be asked. You may quote the text of the passage.

When the user operates the user terminal 100 (e.g., by tapping the screen), the created question 30 is sent from the user terminal 100 to the learning support device 200.

In step S200, the transmitting/receiving unit 210 transmits and receives the question 30. First, the transmitter/receiver 210 receives the question 30 (see FIG. 10) from the user terminal 100. In this embodiment, the question 30 is text information, but if the question 30 is image information, the transmitting/receiving unit 210 converts the question 30 into text information as in step S130, and if the question 30 is audio information, it is embedded or online. Converts to text information using voice recognition software. Then, the transmitting/receiving unit 210 transmits the question 30, which is text information, to the generation AI 300. At this time, the transmitting/receiving unit 210 may instruct generation AI 300 to create an answer to question 30 by transmitting a message "Please answer question 30 based on question 10 and answer 20." On the other hand, the generation AI 300 receives the question 30 and further the message.

In step S30, after the generation AI 300 receives the question 30, the generation AI 300 generates an answer 50 to the question 30 based on the question 10 and the answer 20, and transmits it to the learning support device 200. FIG. 11 shows an example of the answers 50. Here, in response to question 30 (see Figure 10), ``I don't know the meaning of 7!'', we answered, ``'7!' is pronounced as 'factorial of 7.' This means everything from 7 to 1. It means multiplying by positive integers. In other words, 7 x 6 x 5 x 4 x 3 x 2 x 1 = 5040.'' is the answer. However, the answer 50 is 7! It does not include an explanation of what it means in this problem.

In step S300, the generation unit 220 generates the explanation 40. FIG. 12 shows a detailed flow of explanation generation step S300.

In step S310, the generation unit 220 receives the answer 50 generated based on the question 10, the answer 20, and the question 30 from the generation AI 300.

In step S320, the generation unit 220 instructs the generation AI 300 to summarize the plurality of key points 60. The plurality of key points 60 are knowledge required when solving the problem 10 or understanding the answer 20. For example, the generation unit 220 sends the generation AI 300 a message saying, "Please list all the knowledge necessary to solve this problem. Please also tell us the specific properties and theorems included in that knowledge." By inputting , the generation AI 300 can compile the plurality of main points 60 of the answer 20.

In step S<b>40 , the generation AI 300 summarizes the plurality of key points 60 and sends them back to the learning support device 200 .

FIG. 13 shows an example of a plurality of key points 60 in question 10 (see FIG. 5A) and answer 20 (see FIG. 5B). In this example, the multiple points 60 are "Permutations": A permutation represents a method or pattern of arranging a series of elements in a specific order. When calculating the number of permutations, start from n elements. We need a formula for selecting r elements." "Combinations: A combination represents a method or pattern for selecting a series of elements regardless of order. When calculating the number of combinations, n We need a formula to select r elements from elements.''Factorial: Factorial represents the product of consecutive natural numbers.The factorial of n is expressed as n! n!=n×(n-1)×(n-2)×…×2×1” and “Permutation formula: When calculating the number of permutations, the following formula is used: P( It consists of four items: n, r)=n!/(n-r)!.

In step S330, the generation unit 220 receives the plurality of key points 60 returned from the generation AI 300. Note that the summary of a plurality of key points 60 may be skipped by selection, and in such a case, steps S320 to S330 may be omitted.

In step S340, the generation unit 220 generates the explanation 40 by adding all or part of the plurality of key points 60 following the answer 50. As a result, not only question 10 but also the answer 20 of question 10 is input to the generation AI 300, and then the question 30 from the user is further input to the generation AI 300, and the generation AI 300 is given an answer to question 30 based on question 10 and answer 20. 50 and generate an explanation 40 for the user based on the answer 50, it is possible to answer the user's question 30, that is, the question 10 and the answer 20. Further, according to this, by including the main points 60 of the question 10 and the answer 20 in the explanation 40, the user can grasp the main points 60 of the question 10 and the answer 20 and identify the main points that are not understood. It can enhance the learning effect.

Note that when transmitting and receiving at least one of Question 10, Answer 20, or Question 30 as image information, the user's text input work can be reduced by inputting Question 10, Answer 20, and Question 30 as image information. Furthermore, by having the generation AI 300 generate a question 10 including a figure, an answer 20, or an answer 50 to the question 30, and generating an explanation 40 for the user based on the answer 50, a question 10 including a figure, Even with the answer 20 and the question 30, it is possible to answer the user's question 30, that is, the question in the question 10 and the answer 20.

FIG. 14A illustrates an explanation 40 generated by the generation unit 220 based on the problem 10, the answer 20, and the question 30. The explanation 40 corresponds to the explanation for the question 30 based on the question 10 and the answer 20. As described above, the explanation 40 includes the answer 50 to the question 30 based on the question 10 and the answer 20 generated by the generation AI 300, and the contents of some of the main points 60 in the question 10 and the answer 20 generated by the generation AI 300. The generation unit 220 combines and generates the following. Note that the explanation 40 may be stored in the storage device of the learning support device 200.

In this example, the explanation 40 is an explanation for the question 30 for the question 10 and the answer 20 regarding mathematical permutations. Explanation 40 says, ``'7!' is read as 'factorial of 7.' This means multiplying all positive integers from 7 to 1. In other words, 7 x 6 x 5 x 4 x 3 x 2 x 1 = 5040.'' The answer is 50, and ``In this problem, ``7!'' has 7 elements (letters in this case) and represents all possibilities of how they can be arranged. For example, in the case of ``abc'', there are 6 (3!) ways: ``abc,'' ``acb,'' ``bac,'' ``bca,'' ``cab,'' and ``cba.'' and ``Based on this. , Returning to problem (2), if we consider "o, m, and e" as one element, and consider each other letter as one element, there are a total of seven elements.How can we define these seven elements? It consists of two main points 60: ``All possibilities of arranging ``are ``7!''''.

In the explanation 40, if any of the problem 10, the answer 20, and the question 30 are transmitted and received as image information, and a figure is included in the image information, the figure may be extracted and cited. Furthermore, if the content of the question 30 corresponds to multiple locations in the question 10 and the answer 20, and the content of the explanation differs depending on the location, the generation unit 220 may prepare and output separate explanations individually.

FIG. 14B illustrates an explanation 41 of different generation methods. The explanation 41 is an explanation of the question 30 for the question 10 and the answer 20 regarding mathematical permutations, and is more suitable for users who do not understand what part of the question 10 and the answer 20 than the explanation 40 does. The explanation 41 is generated in a form that lists an answer 50 (see FIG. 11) and a plurality of main points 60 (see FIG. 13). Such a generation method may be selected when the user sets to view all the main points on the user terminal 100. When displayed on the user terminal 100, the explanation 41 can be displayed in a form in which a plurality of main points 60 can be selected respectively, or by having the user reply with the number of the main points, so that the user can find out more details among the plurality of main points 60. You may also be able to select the key points for which you would like to know the explanation.

Note that the summary of key points 60 (steps S320 and S40) may be omitted. In such a case, the learning support device 200 generates the explanation 40 using only the answer 50 received from the generation AI 300 in step S310.

In step S400, the explanation 40 generated in step S300 is transmitted to the user terminal 100.

In step S50, the user terminal 100 receives the explanation 40 from the output unit 230 and displays it on the screen. FIG. 15 shows an example of the screen display of the user terminal 100 in step S50. In this example, the user sends the question 30 to the learning support device 200 on the user terminal 100 in step S20, saying, "I don't understand the meaning of 7!", and in step S40, the learning support device 200 sends the question " Of course, I'll explain it to you! "7!" means..." We have received explanation 40. Thereby, the explanation 40 can be output to the user terminal 100, and the explanation 40 can be confirmed on the user terminal 100. The user can also check the explanation 40 and input additional questions 70.

In step S60, the user inputs an additional question 70 on the user terminal 100 and instructs transmission. Thereby, the additional question 70 is transmitted from the user terminal 100 to the learning support device 200.

In step S500, the transmitter/receiver 210 transmits and receives the additional question 70. First, the transmitting/receiving unit 210 receives the additional question 70 from the user terminal 100. In this embodiment, the additional question 70 is text information, but if the additional question 70 is image information, the transmitting/receiving unit 210 converts the additional question 70 into text information as in step S130, and If the information is voice information, it is converted to text information using built-in or online voice recognition software. Then, the transmitting/receiving unit 210 transmits an additional question 70, which is text information, to the generation AI 300. According to this, the user's additional question 70 regarding the explanation 40 is further inputted to the generation AI 300, the answer 51 to the additional question 70 is generated, and the explanation 42 for the user is generated based on the answer 51. Even if 40 is not sufficient, the user's additional question 70, that is, the question in explanation 40, can be answered.

Step S70, subsequent step S600, step S80, step S700, and step S90 are the same processes as step S30, step S300, step S40, step S400, and step S50, respectively. Details of these steps are omitted.

FIG. 16 shows an example of the screen display of the user terminal 100 in step S90. In this example, the user sends a question 30 to the learning support device 200 on the user terminal 100 in step S20 saying, "I don't understand the meaning of 7!", and in step S50 the learning support device 200 sends the question " Of course, I'll explain it to you! "7!" means..." Explanation 40 is received, and then in step S60, "Why do you think of "o, m, e" as one element? ” is sent to the learning support device 200, and an explanation 42 starting with “For problem (2),...” is received from the learning support device 200.

Note that the flow from steps S60 to S90 may be repeated until the user's doubts are resolved.

FIG. 17 shows an operational flow of gist selection by the learning support system 1. In response to the explanation 41 displayed on the user terminal 100 in step S50 of the above-mentioned operation flow (see FIG. 2), the user displays a plurality of main points 60 listed in the explanation 41 (see FIG. 14B) on the user terminal 100. You may choose either one. Such processing is executed in step S61. As a result, a request for further explanation of the main points selected from among the plurality of main points 60 is transmitted from the user terminal 100 to the learning support device 200.

In step S501, the transmitter/receiver 210 and the generator 220 generate a further explanation of the selected key point. First, the transmitter/receiver 210 instructs the generation AI 300 to explain the selected key points. For example, after inputting a plurality of key points 60 to the generation AI, an explanation of the selected key point can be requested by inputting "Please explain the third key point in more detail."

In step S71, the generation AI 300 generates an explanation 65 of the selected key point 75, and returns the generated explanation 65 to the generation unit 220.

Then, the generation unit 220 receives the explanation 65 generated from the generation AI 300 and uses it as the explanation 43 of the selected main point. According to this, the main point 75 selected by the user among the main points 60 of the question 10 and the answer 20 is inputted to the generation AI 300 to generate an explanation 65 for the main point 75 selected by the user, and based on the explanation 65. By generating the explanation 43 for the user based on the information, the learning support device concentrates the explanation on the key points 75 selected by the user, thereby deepening the user's understanding.

In step S700 and step S90, the same processing as step S400 and step S50 is performed, respectively. Details of these steps are omitted.

FIG. 18 shows an operation flow of automatic grading by the learning support system 1. Here, the steps of step S10, step S100, steps S400 to S700, and step 90 are the same as those described above. This flow starts when the user selects "Score" on the screen of the user terminal 100 (see FIG. 3) at the start of the operation. At this time, a message 1201 "Please grade" is sent from the user terminal 100 to the learning support device 200. In response, the learning support device 200 sends a message to the user terminal 100 prompting the user to send the answer 80 after step S100 ends. As a result, as shown in FIG. 19, messages 1202 "Of course" and "Can you show me your answers?" are displayed on the screen of the user terminal 100.

In step S21, the user inputs the answer 80 on the user terminal 100 and instructs transmission. FIG. 20 shows an example of the answer sheet 80. Answer 80 is an answer created by the user for question 10. The answer sheet 80 is transmitted from the user terminal 100 to the learning support device 200.

In step S201, the transmitting/receiving unit 210 transmits and receives the answers 80. First, the transmitting/receiving unit 210 receives the answer 80 from the user terminal 100. Note that if the answer 80 is input as image information, the answer 80 may be converted into text information as in step S130. Then, the transmitting/receiving unit 210 transmits the answer 80 to the generation AI 300 and requests the answer 80 to be graded by transmitting a message "Please grade."

In step S31, the generation AI 300 receives the answer 80 from the learning support device 200, scores it, and transmits it to the learning support device 200.

In step S301, the answers are graded. First, the generation unit 220 receives the scoring result 90 generated by the generation AI 300 based on the question 10, the answer 20, and the answer 80 from the generation AI 300. Next, the generation unit 220 requests the generation AI 300 to extract a plurality of key points 60. In step S40, the generation AI 300 extracts a plurality of key points 60 and sends them back to the learning support device 200. Thereafter, the generation unit 220 receives the plurality of main points 60 returned from the generation AI 300. Then, the generation unit 220 generates the explanation 45 based on the scoring result 90 and the plurality of key points 60.

As shown in FIG. 21, an example of the screen display of the user terminal 100 in step S50 is shown. In this example, the answer that the user sent in step S21 is displayed on the user terminal 100 as a message 1203 from above, and a message 1204 "Your answer is wrong" is displayed as the scoring result 90 generated by the generated AI in step S30. ” and the message 1205 ``Let's make sure you fully understand the intent of the problem...'', which is part of the multiple points 60 generated by the generation AI in step 40, are displayed as the explanation 45. There is.

This completes the scoring of the answers. According to this, by inputting the user's answer 80 to the generation AI 300 and having the generation AI 300 grade the answer, the effort of grading can be reduced and the grading work can be performed independently of the grader. Further, the user's additional question 71 for the scoring result 90 of the generation AI 300 is input to the generation AI 300, the generation AI 300 generates an answer 52 to the additional question 71 based on the question 10 and the answer 20, and based on the answer 52. By generating the explanation 44 for the user, it is possible to answer the user's questions regarding the scoring result 90.

As explained above, according to the learning support system 1 according to the present embodiment, the learning support device 200 inputs not only the question 10 input from the user terminal 100 but also the answer 20 of the question to the generation AI 300, and then further By inputting the question 30 from Be able to answer questions, that is, doubts in problems and answers.

Further, the learning support device 200 according to the present embodiment is a device that supports the user's learning by generating explanations for the answers to the questions, and includes a question 10 and an answer 20 to the question 10 input by the user from the user terminal 100. a transmitting/receiving unit that receives the question 10 and the answer 20 to the generation AI 300, further receives the question 30 input by the user from the user terminal 100 after transmitting the question 10 and the answer 20, and transmits the question 30 to the generation AI 300. 210; and a generation unit 220 that receives an answer 50 to the question 30 from the generation AI 300 and generates an explanation 40 for the user based on the answer 50. According to this, not only question 10 but also the answer 20 of question 10 is input to the generation AI 300, and then a question 30 from the user is further input to the generation AI 300, and the generation AI 300 generates the question 30 based on the question 10 and the answer 20. By generating an answer 50 to the question 50 and generating an explanation 40 for the user based on the answer 50, it is possible to answer the user's question 30, that is, the question 10 and the answer 20 that the user does not understand.

Further, the learning support system 1 according to the present embodiment is provided as a system including a learning support device 200, a user terminal 100, and a generation AI 300. According to this, highly accurate explanations can be generated in response to 30 questions from the user, and the learning effect of the user can be improved.

Further, the learning support method according to the present embodiment is a learning support method that supports the user's learning by generating an explanation 40 for the answer 20 of the problem 10, and includes the problem 10 and the problem input by the user from the user terminal 100. 10 and transmitting the question 10 and the answer 20 to the generation AI 300, and after transmitting the question 10 and the answer 20, further receiving the question 30 input by the user from the user terminal 100 and generating the question 30. It includes the steps of transmitting the answer to the AI 300, and receiving the answer 50 to the question 30 from the generating AI 300, and generating the explanation 40 for the user based on the answer 50. According to this, not only question 10 but also the answer 20 of question 10 is input to the generation AI 300, and then a question 30 from the user is further input to the generation AI 300, and the generation AI 300 generates the question 30 based on the question 10 and the answer 20. By generating an answer 50 to the question 50 and generating an explanation 40 for the user based on the answer 50, it is possible to answer the user's question 30, that is, the question 10 and the answer 20 that the user does not understand.

In addition, the learning support program according to the present embodiment receives the question 10 and the answer 20 of the question 10 input by the user from the user terminal 100 into the computer, and generates an explanation about the answer to the question. 20 to the generation AI 300; means for further receiving the question 30 input by the user from the user terminal 100 after transmitting the question 10 and the answer 20; and means for transmitting the question 30 to the generation AI 300; A program is provided that executes means for receiving an answer 50 to a user and generating an explanation 40 for a user based on the answer 50. According to this, not only question 10 but also the answer 20 of question 10 is input to the generation AI 300, and then a question 30 from the user is further input to the generation AI 300, and the generation AI 300 generates the question 30 based on the question 10 and the answer 20. By generating an answer 50 to the question 50 and generating an explanation 40 for the user based on the answer 50, it is possible to answer the user's question 30, that is, the question 10 and the answer 20 that the user does not understand.

Further, according to the learning support system 1 according to the present embodiment, the learning support device 200 inputs not only the question 10 input from the user terminal 100 but also the answer 20 of the question to the generation AI 300, and then further questions 30 from the user. is input to the generation AI 300, the generation AI 300 generates an answer 50 to the question 30 based on the question 10 and the answer 20, and generates an explanation 40 for the user based on the answer 50. The correct answer rate, which varied between 84% and 84%, can be stabilized at 98%. According to this, it is possible to find a solution to the world of generative AI where hallucination is an issue.

Various embodiments of the invention may be described with reference to flowcharts and block diagrams, where the blocks represent (1) a stage in a process at which an operation is performed, or (2) a device responsible for performing the operation. may represent a section of Certain steps and sections may be implemented by special purpose circuitry, programmable circuitry provided with computer readable instructions stored on a computer readable medium, and/or a processor provided with computer readable instructions stored on a computer readable medium. It's fine. Dedicated circuitry may include digital and/or analog hardware circuits, and may include integrated circuits (ICs) and/or discrete circuits. Programmable circuits include logic AND, logic OR, logic Reconfigurable hardware circuits may include reconfigurable hardware circuits, including, for example.

A computer-readable medium may include any tangible device capable of storing instructions for execution by a suitable device, such that the computer-readable medium having instructions stored thereon is illustrated in a flowchart or block diagram. An article of manufacture will be provided that includes instructions that can be executed to create a means for performing the operations. Examples of computer readable media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, and the like. More specific examples of computer readable media include floppy disks, diskettes, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), Electrically Erasable Programmable Read Only Memory (EEPROM), Static Random Access Memory (SRAM), Compact Disk Read Only Memory (CD-ROM), Digital Versatile Disk (DVD), Blu-ray Disk, Memory may include sticks, integrated circuit cards, etc.

Computer-readable instructions may include assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state configuration data, or instructions such as Smalltalk®, JAVA®, C++, etc. any source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as may include.

Computer-readable instructions may be transmitted to a processor or programmable circuitry of a programmable data processing device, such as a general purpose computer, special purpose computer, or other computer, either locally or over a wide area network, such as a local area network (LAN), the Internet, etc. The computer readable instructions may be provided over a network (WAN) and executed to create a means for performing the operations specified in the flowchart or block diagram. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, and the like.

FIG. 22 illustrates an example computer 1200 in which aspects of the invention may be implemented, in whole or in part. A program installed on computer 1200 may cause computer 1200 to function as an operation or one or more sections of an apparatus according to an embodiment of the present invention, or to perform an operation or one or more sections of an apparatus according to an embodiment of the present invention. Sections and/or computer 1200 may be caused to perform a process or a step of a process according to an embodiment of the invention. Such programs may be executed by CPU 1212 to cause computer 1200 to perform certain operations associated with some or all of the blocks in the flowcharts and block diagrams described herein.

Computer 1200 according to this embodiment includes a CPU 1212, RAM 1214, graphics controller 1216, and display device 1218, which are interconnected by host controller 1210. The computer 1200 also includes input/output units such as a communication interface 1222, a hard disk drive 1224, a DVD-ROM drive 1226, and an IC card drive, which are connected to the host controller 1210 via an input/output controller 1220. There is. The computer also includes legacy input/output units, such as ROM 1230 and keyboard 1242, which are connected to input/output controller 1220 via input/output chip 1240.

The CPU 1212 operates according to programs stored in the ROM 1230 and RAM 1214, thereby controlling each unit. Graphics controller 1216 obtains image data generated by CPU 1212, such as in a frame buffer provided in RAM 1214 or itself, and causes the image data to be displayed on display device 1218.

Communication interface 1222 communicates with other electronic devices via a network. Hard disk drive 1224 stores programs and data used by CPU 1212 within computer 1200. DVD-ROM drive 1226 reads programs or data from DVD-ROM 1201 and provides the programs or data to hard disk drive 1224 via RAM 1214. The IC card drive reads programs and data from and/or writes programs and data to the IC card.

ROM 1230 stores therein programs that are dependent on the computer 1200 hardware, such as a boot program that is executed by the computer 1200 upon activation. Input/output chip 1240 may also connect various input/output units to input/output controller 1220 via parallel ports, serial ports, keyboard ports, mouse ports, and the like.

A program is provided by a computer readable medium such as a DVD-ROM 1201 or an IC card. The program is read from a computer readable medium, installed on hard disk drive 1224, RAM 1214, or ROM 1230, which are also examples of computer readable media, and executed by CPU 1212. The information processing described in these programs is read by the computer 1200 and provides coordination between the programs and the various types of hardware resources described above. An apparatus or method may be configured to implement the manipulation or processing of information according to the use of computer 1200.

For example, when communication is performed between the computer 1200 and an external device, the CPU 1212 executes a communication program loaded into the RAM 1214 and sends communication processing to the communication interface 1222 based on the processing written in the communication program. You can give orders. The communication interface 1222 reads transmission data stored in a transmission buffer processing area provided in a recording medium such as a RAM 1214, a hard disk drive 1224, a DVD-ROM 1201, or an IC card under the control of the CPU 1212, and transmits the read transmission data. Data is transmitted to the network, or received data received from the network is written to a reception buffer processing area provided on the recording medium.

Further, the CPU 1212 causes the RAM 1214 to read all or a necessary part of a file or database stored in an external recording medium such as a hard disk drive 1224, a DVD-ROM drive 1226 (DVD-ROM 1201), an IC card, etc. Various types of processing may be performed on data on RAM 1214. The CPU 1212 then writes back the processed data to the external recording medium.

Various types of information, such as various types of programs, data, tables, and databases, may be stored on a recording medium and subjected to information processing. CPU 1212 performs various types of operations, information processing, conditional determination, conditional branching, unconditional branching, and information retrieval on data read from RAM 1214 as described elsewhere in this disclosure and specified by the program's instruction sequence. Various types of processing may be performed, including /substitutions, etc., and the results are written back to RAM 1214. Further, the CPU 1212 may search for information in a file in a recording medium, a database, or the like. For example, if a plurality of entries are stored in a recording medium, each having an attribute value of a first attribute associated with an attribute value of a second attribute, the CPU 1212 search the plurality of entries for an entry that matches the condition, read the attribute value of the second attribute stored in the entry, and thereby associate it with the first attribute that satisfies the predetermined condition. The attribute value of the second attribute may be acquired.

The programs or software modules described above may be stored on computer readable media on or near computer 1200. Additionally, a storage medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can be used as a computer-readable medium, thereby providing the program to the computer 1200 via the network. do.

Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the range described in the above embodiments. It will be apparent to those skilled in the art that various changes or improvements can be made to the embodiments described above. It is clear from the claims that such modifications or improvements may be included within the technical scope of the present invention.

The execution order of each process such as operation, procedure, step, and stage in the apparatus, system, program, and method shown in the claims, specification, and drawings is specifically defined as "before" or "before". It should be noted that they can be implemented in any order unless the output of the previous process is used in the subsequent process. Even if the claims, specifications, and operational flows in the drawings are explained using "first," "next," etc. for convenience, this does not mean that it is essential to carry out the operations in this order. It's not a thing.

1...Learning support system, 10...Problem, 20...Answer, 30...Question, 40, 41, 42, 43, 44, 45...Explanation, 50, 51, 52...Answer, 60, 61...Point, 65...Explanation, 70, 71...Additional question, 75...Selected main point, 80...Answer, 90...Result, 100...User terminal, 200...Learning support device, 210...Transmission/reception section, 220...Generation section, 230...Output section, 400 ...OCR device, 500...Network, 1001, 1002, 1003, 1004, 1005, 1101, 1102, 1103, 1104, 1105, 1201, 1202, 103, 1204, 1205...Message, 1200...Computer, 1210...Host controller, 1212 ...CPU, 1216...Graphic controller, 1218...Display device, 1220...I/O controller, 1222...Communication interface, 1224...Hard disk drive, 1226...DVD-ROM drive, 1240...I/O chip, 1242...Keyboard

Claims (13)

A learning support device that supports user learning by generating explanations for answers to questions,
In response to the user's request, a message prompting the user to submit the question and a message prompting the user to submit the answer described in the textbook or problem book for the question is sent to the user terminal, and the user inputs from the user terminal. receive the question and the answer to the question, send the question and the answer to the generation AI, further receive the question input by the user from the user terminal after sending the question and the answer, and send the question to the generation AI. a transmitting/receiving unit that transmits to the generation AI;
a generation unit that receives an answer to the question from the generation AI and generates an explanation for the user based on the answer;
A learning support device equipped with. The transmitting/receiving unit further receives an additional question input by the user from the user terminal, and transmits the additional question to the generation AI.
The learning support device according to claim 1, wherein the generation unit receives an answer to the additional question from the generation AI, and generates an explanation for the user based on the answer. The learning support device according to claim 1, further comprising an output unit that outputs the explanation to the user terminal. The transmitting/receiving unit receives an answer input by the user from the user terminal, and transmits the answer to the generation AI for scoring,
The learning support device according to claim 1, wherein the generation unit receives the scoring result of the answer from the generation AI. The transmitting/receiving unit further receives a question regarding the scoring result input by the user from the user terminal, and transmits the question to the generation AI,
The learning support device according to claim 4, wherein the generation unit receives an answer to the question from the generation AI, and generates an explanation for the user based on the answer. The learning support device according to claim 1, wherein the generated AI is a generated AI connected to the learning support device via a network. The learning support device according to claim 1, wherein the transmitting/receiving unit receives at least one of the problem, the answer, or the question as image information. The learning method according to claim 6, wherein the transmitting/receiving unit recognizes characters included in at least one of the problem, the answer, or the question received as image information, and transmits the character information to the generation AI as character information. Support equipment. The learning support device according to claim 1, wherein the generation unit further causes the generation AI to extract a plurality of main points from the answer, and adds the plurality of main points to the explanation. The generation unit further causes the generation AI to extract a plurality of key points from the answer, provides the plurality of key points to the user terminal, transmits the key points selected by the user to the generation AI, and causes the generation AI to extract a plurality of key points from the answer. The learning support device according to claim 1, wherein the learning support device receives an answer to the main point from the user and generates an explanation for the user based on the answer. The learning support device according to any one of claims 1 to 10;
The user terminal;
The generation AI;
A system equipped with A learning support method for supporting user learning by generating explanations for answers to questions, the method comprising:
transmitting to the user terminal, in response to a user's request, a message prompting the user to submit a question and a message prompting the user to submit an answer written in a textbook or problem book for the question;
receiving a question input by a user and an answer to the question from the user terminal, and transmitting the question and the answer to a generation AI;
After transmitting the question and the answer, further receiving a question input by the user from the user terminal and transmitting the question to the generation AI;
receiving an answer to the question from the generating AI and generating an explanation for the user based on the answer;
A learning support method that provides to the computer to generate explanations for the answers to the questions.
a step of transmitting to a user terminal, in response to a user's request, a message prompting the user to submit a question and a message prompting the user to submit an answer written in a textbook or problem book for the question;
means for receiving a question input by a user and an answer to the question from the user terminal, and transmitting the question and the answer to a generation AI;
After transmitting the question and the answer, further receiving a question input by the user from the user terminal, and transmitting the question to the generation AI;
means for receiving an answer to the question from the generating AI and generating an explanation for the user based on the answer;
A program to run.