Question generation system, question generation method, and question generation program

Abstract

To generate a question using artificial intelligence, with a suitable difficulty in accordance with a learner's degree of understanding.SOLUTION: A question generation system includes a processor. The processor, in an input step, displays an input screen that receives an input of an answer to a first question which is generated based on a content of one or a plurality of teaching materials; and in an instruction step, executes an instruction to an artificial intelligence to generate a second question to be displayed next to the first question based on the input answer. The artificial intelligence has a learned model in which the first question and the answer are inputted, and the second question is output. The processor, in a display step, displays the second question output by the artificial intelligence based on the instruction.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a problem generation system, a problem generation method, and a problem generation program.

Background Art

[0002] A technique is known in which a computer displays problems to a learner and allows the learner to input answers.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] Patent Document 1 discloses an information processing apparatus that easily generates selected problems from an arbitrary document file. Patent Document 2 discloses a learning system that evaluates the proficiency of a learner based on the correctness of the problems answered by the learner and determines the problems to be presented next.

[0005] An object of the present invention is to provide a problem generation system that generates problems with an appropriate level of difficulty according to the understanding level of a learner.

Means for Solving the Problems

[0006] To achieve the above object, a problem generation system according to one aspect of the present invention is a problem generation system including a processor. In the input step, the processor displays an input screen for receiving an input of an answer to a first problem generated based on the content of one or more teaching materials. In the instruction step, based on the input answer, the processor gives an instruction to an artificial intelligence to generate a second problem to be displayed next to the first problem. The artificial intelligence has a learned model that takes the first problem and the answer as inputs and outputs the second problem. In the display step, the processor displays the second problem output by the artificial intelligence based on the instruction.

[0007] The learned model may be a model that machine-learns a plurality of learning elements included in one or more of the teaching materials and a learning order recommended for the plurality of learning elements.

[0008] In the instruction step, when the answer to the first problem is incorrect, the processor may give an instruction to the artificial intelligence to generate, as the second problem, a problem related to a learning element that comes earlier in the learning order among the learning elements necessary to solve the first problem.

[0009] The first problem is composed of a plurality of questions. In the input screen, the processor receives an answer input for each of the plurality of questions. In the instruction step, the processor may give an instruction to the artificial intelligence to generate the second problem based on the plurality of answers.

[0010] The learned model is a model that machine-learns a plurality of learning elements included in one or more of the teaching materials and a learning order recommended for the plurality of learning elements. In the instruction step, the processor gives an instruction to the artificial intelligence to extract incorrect answer questions in which the answer is incorrect among the plurality of questions, causes the artificial intelligence to extract common learning elements necessary to solve the incorrect answer questions, and gives an instruction to the artificial intelligence to generate a problem related to a learning element that comes earlier in the learning order among the extracted learning elements.

[0011] To achieve the above object, a problem generation method according to another aspect of the present invention includes an input reception step of causing a computer to display an input screen for receiving an input of an answer to a first problem generated based on the content of teaching materials, and based on the input answer, an instruction step of giving an instruction to an artificial intelligence having a learning model that takes the first problem and the answer as inputs and outputs the second problem to generate an instruction for generating a second problem to be displayed next to the first problem, and a display step of displaying the second problem output by the artificial intelligence based on the instruction.

[0012] To achieve the above object, a problem generation program according to still another aspect of the present invention causes a computer to execute an input reception step of displaying an input screen for receiving an input of an answer to a first problem generated based on the content of teaching materials, an instruction step of giving an instruction to an artificial intelligence having a learning model that takes the first problem and the answer as inputs and outputs the second problem to generate an instruction for generating a second problem to be displayed next to the first problem based on the input answer, and a display step of displaying the second problem output by the artificial intelligence based on the instruction. Note that the computer program can be stored in and provided on various data-readable recording media, or provided so as to be downloadable via a network such as the Internet.

Advantages of the Invention

[0013] According to the present invention, problems of appropriate difficulty levels corresponding to the learner's level of understanding can be generated.

Brief Description of the Drawings

[0014]

Figure 1

Figure 2

Figure 3

Embodiments for Carrying Out the Invention

[0015] Hereinafter, embodiments of the question generation system according to the present invention will be described with reference to the drawings.

[0016] ●Question Generation System (1)● FIG. 1 shows the configuration of a server device 10 according to an embodiment of the present invention. The server device 10 is an example of a question generation system. The server device 10 is a device that generates questions for a learner, presents them to the learner, and generates the next question according to the answer from the learner. The questions are for confirming or evaluating the understanding of the field to be learned. The questions are composed of one or more inquiries. Each inquiry may be a so-called selection type question including a plurality of options for the learner to make a selection input, or a descriptive question for answering by input such as a sentence.

[0017] As shown in FIG. 1, the server device 10 is configured to be communicable with a learner terminal 30 used by a learner via a network NW. The server device 10 may be composed of a hardware device, or some or all of its functions may be realized by a cloud computer. Also, each configuration of the server device 10 may be realized by an API (Application Programming Interface). In this embodiment, the mutual communication between the server device 10 and the learner terminal 30 is wireless, but some or all of the connections may be wired. Furthermore, the server device 10 may be composed of a plurality of hardware configurations. In this case, the plurality of hardware configurations may be connected by wire or wirelessly, and information may be transmitted and received between them.

[0018] ●Learner Terminal 30 The learner terminal 30 is a terminal operated by the learner, such as a smartphone, a tablet terminal, or a personal computer. The learner terminal 30 mainly comprises functional blocks including a display unit 31, an operation reception unit 32, and a communication processing unit 33, which are constituted by a CPU (Central Processing Unit), a computer program executed by the CPU, a RAM (Random Access Memory) and a ROM (Read Only Memory) that store the computer program and predetermined data.

[0019] The display unit 31 is realized by a display or the like for outputting data. On the display unit 31, questions for the learner received from the server device 10 are displayed.

[0020] The operation reception unit 32 is realized by a touch panel, a keyboard, a mouse, a microphone, or the like for inputting data. The operation reception unit 32 can input the field of the questions to be generated. Also, the operation reception unit 32 can input answers to the questions.

[0021] The communication processing unit 33 is a processing unit that can execute data transmission and reception processing according to a predetermined protocol with the server device 10 via a network NW such as the Internet, and is realized by an app or a web browser or the like.

[0022] ● Server device 10 Here, the configuration of the server device 10 will be described with reference to FIG. 1. The server device 10 is an information processing device that executes question generation processing while exchanging data with the learner terminal 30 via the network NW. The server device 10 stores a question database DB1 and a teaching material information database DB2. The question database DB1 stores phrases used in the question generation processing. The teaching material information database DB2 stores the content of teaching materials registered in advance by an administrator or the like. The server device 10 has a function of artificial intelligence (AI: Artificial Intelligence), and generates and outputs questions using the information stored in the question database DB1 and the teaching material information database DB2.

[0023] As shown in FIG. 1, the server device 10 mainly includes functional blocks such as a display control unit 11, a storage control unit 12, an AI control unit 13, an artificial intelligence unit 14, and a communication processing unit 20, which are composed of a CPU (Central Processing Unit, an example of the processor in the claims), a computer program executed by the CPU, a RAM (Random Access Memory), a ROM (Read Only Memory), etc. that store the computer program and predetermined data.

[0024] The display control unit 11 executes processing for causing the learner terminal 30 to display a system screen related to the problem generation system. The display control unit 11 performs processing such as generating and transmitting an HTML (Hyper Text Markup Language) file, for example, and causes the learner terminal 30 to display a web page showing the system screen. Note that the display control unit 11 may perform processing such as generating and transmitting display data for an application for using the problem generation system.

[0025] The display control unit 11 causes the learner terminal 30 to display a registration screen for question information related to the question field to be presented, and a question screen for displaying the question. The registration screen for question information may be displayed so that a plurality of question field candidates can be selected, or free description may be accepted.

[0026] The storage control unit 12 is a functional unit that controls an appropriate storage device provided in the server device 10 and performs writing and reading of data. The storage control unit 12 stores, for example, the question field registered on the registration screen in the storage device. Further, the storage control unit 12 may store the history of questions presented to the learner, the answer history, and the history of correct and incorrect answers in association with the identification information of the learner in the storage device.

[0027] The AI control unit 13 is a functional unit that controls the input to the artificial intelligence unit 14. For example, the AI control unit 13 generates a prompt indicating that it outputs a question related to the question field as input information entered on the registration screen. Further, when the learner's answer to the previously output first question is input, the AI control unit 13 generates a prompt indicating that it outputs a second question different from the first question with the first question and the answer as input. In this case, the AI control unit 13 may use the information on the question field entered on the registration screen as input together with the learner's answer.

[0028] The artificial intelligence unit 14 is an AI equipped with a language model such as a transformer including BART (Bidirectional and Auto-regressive Transformer), BERT (Bidirectional Encoder Representations from Transformers), or GPT (Generative Pretrained Transformer, including GPT-1, GPT-2, GPT-3), etc., particularly a learning model such as a large language model (LLM). The learning model (also referred to as a machine learning model) refers to a learning model based on a machine learning algorithm. Specific algorithms for machine learning include the nearest neighbor method, the naive Bayes method, decision trees, and support vector machines, etc. Further, deep learning (deep learning) that uses a neural network to generate its own feature quantities and combined weight coefficients for learning is also included. The artificial intelligence unit 14 can appropriately apply the above algorithms.

[0029] The artificial intelligence unit 14 has a trained model that has learned teaching materials. The teaching materials contain content related to a predetermined learning field and are so-called textbooks and reference books. The teaching materials are composed of, for example, any one or a combination of text data, image data, video data, and audio data. The teaching materials may be provided by an administrator or the like, or may include information collected from the Internet or the like. In the trained model, information on a plurality of learning elements included in the question field is learned in association with the question field. The information on the learning elements is information obtained by subdividing and labeling the content of the question field, and a plurality of learning elements are included in one question field.

[0030] Also, the trained model has learned the recommended learning order of the learning elements. For example, in the question field of "Internet and Security", the learning elements include "IP address", "domain name", "definition of DNS (Domain Name System)", "structure of DNS server", "DNS client", etc. To understand DNS, it is necessary to understand the IP address and domain name. In other words, the learning elements that are recommended to be learned first are the prerequisite knowledge for the learning elements that are recommended to be learned later. Therefore, it is recommended to learn the above-mentioned learning elements in this order academically. When a plurality of learning elements are included in one teaching material, the trained model may learn the order described in the teaching material as the learning order. Also, the trained model may learn the dependency relationship between the learning elements from the content of the text included in the teaching material. The learning elements are learned continuously without being limited to a certain range of learning fields. In this description, it may be interpreted that the learning order and the difficulty level are related, that is, the learning elements that are recommended to be learned first may be interpreted as having a lower difficulty level than the learning elements that are recommended to be learned later. The trained model has obtained appropriate learning data in advance and can also perform additional learning as appropriate.

[0031] The artificial intelligence unit 14 takes the information on the designated question field as input and outputs questions regarding the question field (an example of the "first question" in the claims) based on the prompt generated by the AI control unit 13. The artificial intelligence unit 14 creates sentences using the phrases stored in the question database DB1 and the teaching material information database DB2. Further, the artificial intelligence unit 14 takes the answer input by the learner via the learner terminal 30 as input and outputs the second question to be displayed after the first question.

[0032] The artificial intelligence unit 14 determines whether the answer to the first question is correct or incorrect, and outputs questions with different learning elements according to the correctness. That is, the artificial intelligence unit 14 may output the second question with the correct / incorrect information of the answer input by the learner as input. When the answer to the first question is incorrect, the artificial intelligence unit 14 generates a second question regarding the learning element preceding in the learning order among the learning elements referred to in the generation of the first question. Note that the first question referring to the correctness in the generation of the second question is not limited to the question answered immediately before the display screen of the second question, and may be a question displayed over a plurality of past screens.

[0033] According to this configuration, it is possible to generate and present to the learner questions for checking the prerequisite knowledge of the learning elements in which the learner's understanding is insufficient. Therefore, the learner can be appropriately made to study what is necessary to expand understanding. Note that the learning elements referred to in the generation of questions are not limited to a certain range of learning fields, and are continuously connected by machine learning. Therefore, according to the configuration of the present invention, it is possible to continue asking questions beyond the range of the learning fields defined in the teaching materials and the framework of the generally divided units. Therefore, it is possible to accurately perform question generation regarding the necessary learning elements.

[0034] In addition, when the answer to the first question is correct, the artificial intelligence unit 14 generates a question regarding the learning element that the learning element included in the first question should systematically have next. According to such a configuration, the learner can efficiently proceed with systematic learning. Also, through the content of the generated question and the correct / incorrect information of the answer, the learner can more accurately self-evaluate how well they understand the teaching material and can apply it in practice. As a result, it can also contribute to improving the confidence of the learner according to the present invention.

[0035] Alternatively, instead of the above-described configuration, the artificial intelligence unit 14 may generate a second question regarding the learning element referred to in the generation of the first question, so-called a similar question to the first question. The artificial intelligence unit 14 may generate a similar question when the first question is answered correctly, or may be configured to generate a similar question when the first question is answered incorrectly. Also, the artificial intelligence unit 14 may determine whether to present a similar question regarding the same learning element as the first question or a question regarding a learning element different from the first question. In this case, the artificial intelligence unit 14 may determine the learning element to be referred to in the generation of the second question using the answer history or correct / incorrect history of the learner as an input.

[0036] The question presented by the artificial intelligence unit 14 may be composed of a plurality of questions as shown in FIGS. 3(a) to 3(b), for example. In this case, on the input screen shown in FIGS. 3(a) to 3(b), answers input for each of the plurality of questions are received. Also, it may be configured such that one question is presented on one screen, and when answering a plurality of questions, one or more questions to be displayed next are generated according to the plurality of answers.

[0037] When there is an incorrect answer among the answers to a plurality of questions, the artificial intelligence unit 14 may extract learning elements necessary to correct the questions with incorrect answers. When there are a plurality of incorrect answers, the artificial intelligence unit 14 may extract common learning elements necessary to correct the questions with incorrect answers. According to the configuration of extracting necessary learning elements based on the correctness of a plurality of questions, learning elements lacking in the learner's understanding can be extracted more accurately. The artificial intelligence unit 14 generates a plurality of questions regarding the learning elements preceding in the learning order among the extracted learning elements. Note that when common learning elements cannot be extracted or according to appropriate settings, the artificial intelligence unit 14 may extract some of the learning elements of the questions with incorrect answers.

[0038] The communication processing unit 20 is a processing unit capable of performing data transmission and reception processing according to a predetermined protocol with the learner terminal 30 via a network NW such as the Internet. The communication processing unit 20 receives learner information and answers to questions from the learner terminal 30. Further, the communication processing unit 20 transmits the questions generated by the artificial intelligence unit 14 to the learner terminal 30.

[0039] As described above, according to the server device 10 according to the present invention, questions of an appropriate difficulty level corresponding to the learner's understanding level can be generated. Further, with this configuration, the learner can sufficiently obtain opportunities for outputting knowledge as well as simply inputting knowledge, and can improve practical knowledge application ability. Furthermore, the learner can avoid wasting time answering questions of inappropriate difficulty levels and can concentrate on answering questions necessary for ability improvement.

[0040] ●Processing flow in which questions are generated As shown in FIG. 2, first, the learner terminal 30 inputs the question field (step S101) and transmits it to the server device 10. The AI control unit 13 gives an instruction to generate a plurality of questions regarding the question field received in step S101 to the artificial intelligence unit 14 (step S102). Next, the server device 10 transmits the questions output from the artificial intelligence unit 14 to the learner terminal 30 (step S103). The learner terminal 30 displays these questions on the display unit 31 (see FIG. 3) and accepts input from the learner. When an answer is input to the learner terminal 30 (step S104), the learner terminal 30 transmits the answer to the server device 10.

[0041] The AI control unit 13 generates an instruction to determine whether the learner's answer obtained via the learner terminal 30 is correct or incorrect for the artificial intelligence unit 14 (step S105). If there is an incorrect answer among the answers to the plurality of questions (N in step S105), the AI control unit 13 transmits an extraction instruction to extract the learning elements necessary to correct the questions with incorrect answers to the artificial intelligence unit 14 (step S106). If there are a plurality of incorrect answers, the necessary learning elements common to correcting the questions with incorrect answers may be extracted. Next, the AI control unit 13 extracts the learning elements related to the learning elements preceding the extracted learning elements (the "second learning element" in FIG. 2) for the artificial intelligence unit 14 (step S107), and gives an instruction to generate a plurality of questions regarding the second learning element (step S108). Then the process proceeds to step S103, and the generated questions are transmitted to the learner terminal 30.

[0042] In step S105, when the answer is correct, the AI control unit 13 gives an instruction to generate a plurality of questions regarding the learning elements that should be systematically next for the learning elements included in the first question to the artificial intelligence unit 14 (step S109). Then the process proceeds to step S103, and the generated questions are transmitted to the learner terminal 30. With such a configuration, the learner terminal 30 receives repeated questions from the server device 10 and displays them. That is, an infinite number of questions are continuously displayed on the learner terminal 30 in theory. Therefore, with this configuration, the learner can not only obtain input of knowledge but also fully obtain opportunities for output of knowledge, and can realize an improvement in practical knowledge application ability.

[0043] ● Display example FIG. 3 is an example of a question displayed on the learner terminal 30. As shown in FIG. 3(a), a plurality of questions are displayed on the learner terminal 30. Also, the learner inputs an answer on the screen. At this time, for example, if the answers to Q1 to Q3 are incorrect, the artificial intelligence unit 14 extracts learning elements common to Q1 to Q3. That is, the artificial intelligence unit 14 extracts that knowledge about DNS is required as prerequisite knowledge to answer Q1 to Q3 correctly. Then, the artificial intelligence unit 14 estimates that knowledge about domain and IP address is required as prerequisite knowledge to understand DNS. Then, the artificial intelligence unit 14 generates questions about domain and IP address, and the questions are displayed on the learner terminal 30 as shown in FIG. 3(b).

[0044] As described above, according to the question generation system according to the present invention, questions of an appropriate difficulty level according to the learner's understanding level can be generated.

Explanation of reference numerals

[0045] 10 Server device (question generation system) 11 Display control unit 12 Memory control unit 13 AI control unit 14 Artificial intelligence unit 20 Communication processing unit DB1 Question database DB2 Teaching material information database 30 Learner terminal

Claims

1. A question generation system, comprising a processor, the processor in an input step, displays an input screen for receiving an input of an answer to a first question generated based on the content of one or more teaching materials, in an instruction step, based on the input answer, gives an instruction to an artificial intelligence to generate a second question to be displayed next to the first question, the artificial intelligence having a learned model that takes the first question and the answer as inputs and outputs the second question, in a display step, displays the second question output by the artificial intelligence based on the instruction, a question generation system.

2. The learned model is a model that has machine-learned a plurality of learning elements included in one or more of the teaching materials and a learning order recommended for the plurality of learning elements, The question generation system according to Claim 1.

3. In the instruction step, when the answer to the first question is incorrect, an instruction is given to the artificial intelligence to generate, as the second question, a question regarding a learning element that comes before in the learning order among the learning elements necessary to solve the first question, The question generation system according to Claim 2.

4. The first question is composed of a plurality of questions, on the input screen, answers input for each of the plurality of questions are received, in the instruction step, an instruction to generate the second question based on the plurality of answers is given to the artificial intelligence, The question generation system according to Claim 1.

5. The learned model is a model that has machine-learned a plurality of learning elements included in one or more of the teaching materials and a learning order recommended for the plurality of learning elements, in the instruction step, an extraction instruction to extract incorrect answer questions in which the answer is incorrect among the plurality of questions is given to the artificial intelligence, the common learning elements necessary to solve the incorrect answer questions are extracted, and an instruction to generate a question regarding a learning element that comes before in the learning order among the extracted learning elements is given to the artificial intelligence, The question generation system according to Claim 4.

6. By a computer, an input reception step of displaying an input screen for receiving an input of an answer to a first question generated based on the content of a teaching material, An instruction step of giving an instruction to an artificial intelligence having a learning model that takes the first question and the answer as inputs and outputs the second question, for generating an instruction for the second question to be displayed next to the first question based on the input answer; A display step of displaying the second question output by the artificial intelligence based on the instruction; Executing, Question generation method.

7. Causing a computer to, An input reception step of displaying an input screen for receiving an input of an answer to a first question generated based on the content of teaching materials; An instruction step of giving an instruction to an artificial intelligence having a learning model that takes the first question and the answer as inputs and outputs the second question, for generating an instruction for the second question to be displayed next to the first question based on the input answer; A display step of displaying the second question output by the artificial intelligence based on the instruction; Causing to execute, Question generation program.

Images