Method and device for identifying weak knowledge points of user, computer equipment and storage medium

Abstract

The invention provides a weak knowledge point identification method and device of a user, equipment and a storage medium, and the method comprises the steps: sending an inquiry instruction, and determining a to-be-answered question of the user according to the reply of the user to the inquiry instruction; a thought obtaining instruction is sent, a problem solving thought of the user for the to-be-solved problem is obtained according to the reply of the user for the thought obtaining instruction in combination with a target language model, and the problem solving thought comprises a problem solving answer; obtaining a standard question solving thought of the question to be answered, filling the standard question solving thought and the question solving thought into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the question solving thought of the user with the standard question solving thought through the target language model; and when the question solving thought of the user is incorrect, identifying weak knowledge points of the user in the question to be solved through the target language model.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, computer equipment and storage medium for identifying weak knowledge points of a user. Background Art

[0002] Virtual reality and augmented reality technologies are already widely used in education, including exercise devices used to develop children's learning abilities. These technologies create interactive and immersive teaching environments that allow students to practice and apply their learning skills in realistic and relevant situations.

[0003] However, the main way of teaching with current virtual reality and augmented reality technologies is to instill learning knowledge points and guidance into children, making it difficult to discover the difficulties and dilemmas in children's thinking. The specific reasons include: due to children's limited descriptive ability, it is difficult to form clear and complete problem-solving ideas, and the lack of effective comparison methods makes it impossible to identify children's weak knowledge points.

[0004] Therefore, in view of the shortcomings of the prior art solutions, the present invention provides a method for identifying weak knowledge points of a user. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device, equipment and storage medium for identifying weak knowledge points of a user in response to the above technical problems.

[0006] On the one hand, a method for identifying weak knowledge points of a user is provided, the method comprising: sending a query instruction, and determining the user's unanswered questions according to the user's reply to the query instruction; sending a solution acquisition instruction, and obtaining the user's solution ideas for the unanswered questions according to the user's reply to the solution acquisition instruction in combination with a target language model, wherein the solution ideas include solution answers; obtaining standard solution ideas for the unanswered questions, filling the standard solution ideas and the solution ideas into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the user's solution ideas with the standard solution ideas through the target language model; when the user's solution ideas are incorrect, identifying the user's weak knowledge points in the unanswered questions through the target language model.

[0007] Optionally, the sending of the inquiry instruction determines the unanswered questions of the user according to the user's reply to the inquiry instruction, including: sending inquiry questions to the user and receiving personal information of the user's answer; and matching the unanswered questions that match the user's personal information according to the user's personal information.

[0008] Optionally, the method further includes: obtaining reference teaching materials; extracting the title, solution ideas and answers of each question from the reference teaching materials to generate a data set; and fine-tuning the large language model based on the data set to obtain a target language model.

[0009] Optionally, the method further includes: creating a question description instruction, a guidance instruction and an answer confirmation instruction; and generating the idea description instruction according to the question description instruction, the guidance instruction and the answer confirmation instruction.

[0010] Optionally, the sending of the idea acquisition instruction is combined with the target language model to obtain the user's idea of ​​solving the problem to be answered according to the user's reply to the idea acquisition instruction, wherein the idea of ​​solving the problem includes the answer to the problem, including: sending the question description instruction to the user, wherein the question description instruction is used to instruct the user to describe the problem to be answered; receiving the question description content returned by the user, comparing the question description content with the problem to be answered, and determining whether the question description content is consistent with the problem to be answered; if not, resending the question description instruction to the user; if According to the problem to be solved, a corresponding guidance instruction is generated, wherein the guidance instruction includes a plurality of sub-guidance instructions, and each sub-guidance instruction is used to instruct the user to describe sub-problem-solving steps and sub-problem-solving answers; each sub-guidance instruction is sent to the user in turn, and the sub-problem-solving steps and sub-problem-solving answers returned by the user are received; when it is detected that the sub-problem-solving answer returned by the user is the answer to the problem to be solved, an answer confirmation instruction is sent to the user, and the question answer returned by the user is received; and according to the sub-problem-solving steps, the sub-problem-solving answers and the question answer, the user's solution idea for the problem to be solved is generated.

[0011] Optionally, generating corresponding guidance instructions according to the question to be answered, sending each sub-guidance instruction to the user in turn, and receiving sub-problem-solving steps and sub-problem-solving answers returned by the user includes: generating a corresponding first sub-guidance instruction according to the question to be answered, wherein the first sub-guidance instruction is used to instruct the user to describe a first sub-problem-solving step and a first problem-solving result calculated according to the first sub-problem-solving step; sending the first sub-guidance instruction to the user, and receiving the first sub-problem-solving step and the first problem-solving result returned by the user; parsing the first sub-problem-solving step returned by the user, and generating a second sub-guidance instruction according to the first sub-problem-solving step, wherein the second sub-guidance instruction is for guiding the second sub-problem-solving step of the question to be answered based on the first sub-problem-solving step; and receiving the second sub-problem-solving step and the second sub-problem-solving result returned by the user.

[0012] Optionally, the method further includes: obtaining the weak knowledge points of the user in a plurality of questions to be answered; and summarizing the weak knowledge points of the user through the target language model.

[0013] On the other hand, a device for identifying weak knowledge points of a user is provided, and the device includes: a first processing module, which is used to send an inquiry instruction, and determine the user's unanswered questions according to the user's reply to the inquiry instruction; a second processing module, which sends an idea acquisition instruction, and obtains the user's solution idea for the unanswered questions according to the user's reply to the idea acquisition instruction in combination with a target language model, wherein the solution idea includes the answer to the question; a comparison module, which is used to obtain the standard solution idea for the unanswered questions, fill the standard solution idea and the solution idea into a preset comparison instruction template, generate a comparison instruction, send the comparison instruction to the target language model, and compare the user's solution idea with the standard solution idea through the target language model; an identification module, which is used to identify the user's weak knowledge points in the unanswered questions through the target language model when the user's solution idea is incorrect.

[0014] On the other hand, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program: sending a query instruction, and determining the user's unanswered question based on the user's reply to the query instruction; sending a thought acquisition instruction, and obtaining the user's solution idea for the unanswered question based on the user's reply to the thought acquisition instruction in combination with a target language model, wherein the solution idea includes an answer to the question; obtaining a standard solution idea for the unanswered question, filling the standard solution idea and the solution idea into a preset comparison instruction template, generating a comparison instruction, and sending the comparison instruction to the target language model, and comparing the user's solution idea with the standard solution idea through the target language model; when the user's solution idea is incorrect, identifying the user's weak knowledge points in the unanswered question through the target language model.

[0015] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: sending a query instruction, and determining the user's unanswered question based on the user's reply to the query instruction; sending a solution instruction, and obtaining the user's solution idea for the unanswered question based on the user's reply to the solution instruction in combination with a target language model, wherein the solution idea includes an answer to the question; obtaining a standard solution idea for the unanswered question, filling the standard solution idea and the solution idea into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the user's solution idea with the standard solution idea through the target language model; when the user's solution idea is incorrect, identifying the user's weak knowledge points in the unanswered question through the target language model.

[0016] The above-mentioned method, device, equipment and storage medium for identifying weak knowledge points of users, the method includes: sending an inquiry instruction, determining the user's unanswered questions according to the user's reply to the inquiry instruction; sending an idea acquisition instruction, combining with the target language model, and obtaining the user's solution ideas for the unanswered questions according to the user's reply to the idea acquisition instruction, wherein the solution ideas include the answer to the question; obtaining the standard solution ideas for the unanswered questions, filling the standard solution ideas and the solution ideas into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the user's solution ideas and the standard solution ideas through the target language model; when the user's solution ideas are incorrect, identifying the user's weak knowledge points in the unanswered questions through the target language model; in this way, by improving the existing large language model and combining it with fine instruction design, it is possible to help children generate clear and complete solution ideas, solve the problem that children are not good at expressing themselves, thereby realizing accurate identification of children's weak knowledge points in the learning process according to the solution ideas, and improving children's learning ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flowchart of a method for identifying weak knowledge points of a user in one embodiment;

[0018] Figure 2 A schematic diagram of a framework of a method for identifying weak knowledge points of a user in one embodiment;

[0019] Figure 3 is a structural block diagram of a device for identifying weak knowledge points of a user in one embodiment;

[0020] Figure 4 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0022] It should be understood that in the description of the present invention, unless the context clearly requires otherwise, words such as “include”, “including” and the like throughout the specification should be interpreted as including rather than being exclusive or exhaustive; that is, as including but not limited to “including”.

[0023] It should also be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0024] It should be noted that the terms "S1", "S2", etc. are only used for the purpose of describing the steps, and do not specifically refer to the order or sequence, nor are they used to limit the present invention. They are only for the convenience of describing the method of the present invention, and cannot be understood as indicating the order of the steps. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0025] In one embodiment, Figure 1 As shown, a method for identifying weak knowledge points of a user is provided, comprising the following steps:

[0026] S101: Sending a query instruction, and determining the user's unanswered questions according to the user's reply to the query instruction.

[0027] Here, according to the query instruction, you can directly ask the user about the questions to be answered, or you can ask the user about the scope of what he wants to learn, and match the corresponding questions to be answered according to the learning scope. You can also ask the user for information and automatically match the corresponding questions to be answered according to the user information.

[0028] Among them, the user's response can be obtained through text interaction or voice interaction with the user.

[0029] S102: Sending a thought acquisition instruction, combining with a target language model, and obtaining the user's thought for solving the problem to be solved according to the user's reply to the thought acquisition instruction, wherein the thought for solving the problem includes an answer to the problem.

[0030] Here, the idea acquisition instruction is used to guide the user to generate problem-solving ideas, and the target language model can be obtained by improving the existing language model.

[0031] Specifically, the idea acquisition instructions are input into the target language model, and the target language model generates multiple rounds of questions, interacts with the user by voice or text, and guides the user to answer the problem-solving steps step by step through multiple rounds of dialogue interaction until the answer to the problem is obtained. The target language model summarizes the problem-solving steps and obtains the user's solution ideas for the question.

[0032] S103: Obtain the standard solution idea of ​​the question to be answered, fill the standard solution idea and the solution idea into a preset comparison instruction template, generate a comparison instruction, send the comparison instruction to the target language model, and compare the user's solution idea with the standard solution idea through the target language model.

[0033] Here, you can obtain the questions to be answered, search the resource library, retrieve the reference materials that have the questions to be answered, and get the standard solution ideas from the reference materials; you can also use the target language model to analyze the questions to be answered and get the standard solution ideas.

[0034] Specifically, a comparison instruction template can be preset, including an instruction description module, which is blank. The user's problem-solving ideas and standard problem-solving ideas are filled into the corresponding positions in the instruction description module of the comparison instruction template to generate a complete comparison instruction. The comparison instruction is input into the target language model to obtain the recognition result.

[0035] Further, exemplary, the comparison instruction may be as follows: You are an excellent children's mathematics teacher, good at comparing the differences between children's problem-solving ideas and correct answer ideas. Summarize the questions, and please answer strictly in Chinese.

[0036] User's solution:

[0037] [Question Analysis]: Calculate the addition of two-digit numbers, 28+17.

[0038] [Solution steps]: First calculate the tens digit 2+1, the result is 3; then calculate the units digit 8+7, the result is 15; finally, combine the tens digit and the units digit, the result is 35.

[0039] [Calculate item by item]: 2+1=3, 8+7=15.

[0040]

Final result

[0041] Standard problem-solving ideas:

[0042] [Question Analysis]: Calculate the addition of two-digit numbers, 28+17.

[0043] [Solution steps]: First calculate the sum of the two tens-digit numbers; then calculate the sum of the two units-digit numbers; finally add them together. Note that if the sum of the units-digit numbers exceeds 10, there will be a carry, so add an extra 1 to the tens-digit number.

[0044] [Calculate item by item]: 2+1=3, 8+7=15.

[0045]

Final result

[0046] Please compare the two solutions above and see if your solution is correct. If not, please highlight the error.

[0047] S104: When the user's solution idea is incorrect, the target language model is used to identify the user's weak knowledge points in the question to be answered.

[0048] Specifically, through the target language model, the difference between the user's problem-solving ideas and the standard problem-solving ideas is obtained, the steps of the user's errors are identified, the knowledge points contained in the errors are determined, and the user's weak knowledge points are obtained. Exemplarily, the recognition result may be: By comparing the user's problem-solving ideas with the standard problem-solving ideas, we find that there are some errors in the user's problem-solving ideas. The user's problem-solving ideas: 1. Calculate the tens digit 2+1 as 3, which is correct. 2. Calculate the units digit 8+7 as 15, which is correct. 3. Combine the units digit and the tens digit, and the calculation result is 35, which is wrong. The error is that the user did not consider the problem that the sum of the units digits exceeds 10, that is, 15 exceeds 10 and needs to be carried. The standard problem-solving ideas clearly point out that if the sum of the units digits exceeds 10, it is necessary to add another 1 to the tens digit. Therefore, although the user's problem-solving ideas are partially correct, there are still errors. It is necessary to guide users to pay attention to the problem that the sum of the units digits exceeds 10, and consider the problem of carry when solving addition problems.

[0049] It should be noted that the present invention provides a comprehensive and effective method to help users improve their learning performance by using NLP (Natural Language Processing) technology and machine learning algorithms. By generating problem-solving ideas, identifying weak knowledge points, and summarizing weak knowledge points, it helps to reduce learning costs and promote the formation of children's correct learning thinking.

[0050] In some specific implementations, the sending of the inquiry instruction and determining the user's unanswered question according to the user's reply to the inquiry instruction include:

[0051] Sending inquiry questions to users and receiving personal information answered by the users;

[0052] According to the personal information of the user, matching questions to be answered that match the personal information of the user.

[0053] Here, the user's personal information may include the user's name, age, gender, grade, subjects that the user wants to practice, etc.

[0054] Furthermore, multiple exercises of the user may be recorded based on the user's personal information, so as to facilitate summarizing the user's weak knowledge points in multiple exercises.

[0055] In some embodiments, the method further comprises:

[0056] Obtain reference materials;

[0057] Extract the title, solution ideas and answers of each question from the reference textbook to generate a data set;

[0058] The large language model is fine-tuned based on the data set to obtain a target language model.

[0059] Here, the large language model may be llama3.1, and correspondingly, the target language model may be llama-child.

[0060] Specifically, we collect the homework materials used by existing teachers, which include questions and standard answers, extract the questions and answers with problem-solving ideas, and form a triple: <question, problem-solving idea, answer>. Based on this data, we fine-tune the existing large model to form a large model specifically for children's education (i.e., the target language model).

[0061] In some embodiments, the method further comprises:

[0062] Create question description instructions, guidance instructions, and answer confirmation instructions;

[0063] The idea description instruction is generated according to the question description instruction, the guidance instruction and the answer confirmation instruction.

[0064] Here, the question description instruction is used to guide the user to describe the question to be answered, the guidance instruction is used to guide the user to step by step describe the solution steps and the calculation results obtained according to each step, and the answer confirmation instruction is used to guide the user to answer the final solution result.

[0065] The idea description instructions also include prompt words, which may specifically include: role, language, skill, etc.

[0066] Specifically, the instruction for obtaining ideas can be: You are an excellent children's mathematics teacher, who is good at communicating with children based on the questions, collecting information, and finally generating their problem-solving ideas. The content of the problem-solving ideas includes information such as [problem analysis], [problem-solving steps], [item-by-item calculation], and [final result]. If there is no relevant information, it is marked as "none". The whole communication adopts a step-by-step approach; the first step is to ask the child about the current question, which is the [problem analysis] part. The second step is to understand his [problem-solving steps] from the child. According to the question and your mathematical logic knowledge, further communicate with the child and collect his description. This part must be detailed and step by step. Ask the child one by one, and try to avoid asking multiple questions at once. The third step is to ask the child to calculate the results one by one according to his problem-solving steps. The fourth step is to ask the child to give his final answer. Note that the communication process is conducted in Chinese throughout. Now you start to communicate with a child named Qiqi from the second grade class, who is female and 4 years old.

[0067] In some specific implementations, the sending of the idea acquisition instruction is combined with the target language model to obtain the user's solution idea for the question to be answered according to the user's reply to the idea acquisition instruction, wherein the solution idea includes the answer to the question, including:

[0068] Sending the question description instruction to the user, wherein the question description instruction is used to instruct the user to describe the question to be answered;

[0069] Receiving the question description content returned by the user, comparing the question description content with the question to be answered, and determining whether the question description content is consistent with the question to be answered;

[0070] If they are inconsistent, resending the topic description instruction to the user;

[0071] If they are consistent, generating a corresponding guidance instruction according to the to-be-answered question, wherein the guidance instruction includes a plurality of sub-guidance instructions, each of which is used to instruct the user to describe sub-question-solving steps and sub-question-solving answers;

[0072] Sending each sub-guidance instruction to the user in sequence, and receiving the sub-problem-solving steps and sub-problem-solving answers returned by the user;

[0073] When detecting that the sub-question answer returned by the user is the answer to the question to be answered, sending an answer confirmation instruction to the user, and receiving the question answer returned by the user;

[0074] The user's solution idea for the question to be answered is generated according to the sub-question-solving steps, the sub-question-solving answers and the question answer.

[0075] Here, the answer to the question to be solved is the final answer obtained by the user according to his / her own solution steps, which may be a correct answer or an incorrect answer.

[0076] Specifically, the process of guiding users to generate problem-solving ideas may include: guiding users to describe the question, guiding users to describe their problem-solving steps and the process answers obtained based on each step step by step until the user gets the answer to the question, guiding users to confirm the answer to the question, and generating problem-solving ideas based on the problem-solving steps, process answers and question answers.

[0077] Exemplarily, the following dialogue interaction process may be included:

[0078] Artificial Intelligence: I am your math teacher! I am very happy to meet you, Qiqi, a four-year-old girl from the second grade, for a math exchange. Now let's do a math problem, I need you to tell me what this problem is? Please describe this problem, Qiqi.

[0079] User: Calculate the addition of two-digit numbers, such as 28+17=?

[0080] AI: OK! Qiqi, the question you described is about adding two-digit numbers. Next, I want to know how you solved the problem. Please tell me how you calculated 28+17? What method would you use to calculate? Please tell me step by step.

[0081] User: I will add up the tens digits first, which is 2+1, and the result is 3.

[0082] Artificial Intelligence: Qiqi, you said that you would add up the tens first, 2+1, and the result is 3! What would you do with the ones next? How would you calculate 8+7? Please continue to describe your steps to solve the problem.

[0083] User: 8+7 is more than 10, so I used addition within 20 and counted with my fingers and toes, and the result was 15.

[0084] Artificial Intelligence: Haha, Qiqi, you use addition within 20, count with your fingers and toes, calculate 8+7, and the result is 15! Now, you have calculated the result of the unit digit is 15. Next, you need to combine the result of the tens digit 3 and the result of the unit digit 15 to calculate the result of the whole number. Please continue to describe your solution steps.

[0085] User: Put the ones digit and tens digit of the result together, which is 3 and 15, which together is 35.

[0086] Artificial Intelligence: Qiqi, you said you put the ones and tens digits of the result together, 3 and 15, and the total is 35. Now you have calculated the whole number, which is 35. You have completed the task of calculating 28+17! The last question, please tell me, what is your answer? What do you think 28+17 equals?

[0087] Users:35.

[0088] AI: Qiqi, you said your answer is 35! OK, let me check your answer, and I can see that 28+17 equals 45, so your answer is wrong.

[0089] OK! Now, we can summarize your problem-solving ideas:

[0090] [Question Analysis]: Calculate the addition of two-digit numbers, 28+17.

[0091] [Solution steps]: First calculate the tens digit 2+1, the result is 3; then calculate the units digit 8+7, the result is 15; finally, combine the tens digit and the units digit, the result is 35.

[0092] [Calculate item by item]: 2+1=3, 8+7=15.

[0093]

Final result

[0094] Your solution has been generated! Please ask the teacher to help check it!

[0095] In this way, by using existing teaching materials, training a dedicated large language model, and combining it with sophisticated instruction design, it can help users with limited descriptive abilities generate clear and complete problem-solving ideas.

[0096] In some specific implementations, generating corresponding guidance instructions according to the to-be-answered question, sending each sub-guidance instruction to the user in sequence, and receiving sub-question-solving steps and sub-question-solving answers returned by the user include:

[0097] Generate a corresponding first sub-guidance instruction according to the to-be-answered question, wherein the first sub-guidance instruction is used to instruct the user to describe a first sub-question-solving step and a first question-solving result calculated according to the first sub-question-solving step;

[0098] Sending the first sub-guidance instruction to the user, and receiving the first sub-problem-solving steps and the first problem-solving result returned by the user;

[0099] Parsing the first sub-problem-solving step returned by the user, and generating a second sub-guidance instruction according to the first sub-problem-solving step, wherein the second sub-guidance instruction is to guide the second sub-problem-solving step of the problem to be solved based on the first sub-problem-solving step;

[0100] Receive the second sub-problem-solving steps and the second sub-problem-solving results returned by the user.

[0101] Here, the guidance instruction may include multiple sub-guidance instructions, including a first sub-guidance instruction, a second sub-guidance instruction, ... an Nth sub-guidance instruction. The second sub-guidance instruction is generated based on the first sub-problem-solving step, the third sub-guidance instruction is generated based on the second sub-problem-solving step, and so on.

[0102] Specifically, assuming the question is to calculate 28+17, the first sub-guidance instruction can be to guide the user to describe the first step of the calculation. After receiving the first step of calculating the tens digit returned by the user, the received content is parsed, and the second sub-guidance instruction is generated based on the parsed content to guide the user to describe the problem-solving idea of ​​calculating the units digit. After receiving the second step of calculating the units digit returned by the user, the received content is parsed, and the third sub-guidance instruction is generated based on the parsed content to guide the user to combine the units digit and the tens digit to calculate the result of the whole number.

[0103] In this way, different sub-guidance instructions can be generated according to different topics and different returned contents.

[0104] In some embodiments, the method further comprises:

[0105] Obtaining the user's weak knowledge points in multiple questions to be answered;

[0106] The user's weak knowledge points are summarized through the target language model.

[0107] Specifically, a summary instruction can be preset, and the summary instruction can be input into the target language model. The target language model obtains the user's historical weak knowledge points according to the summary instruction, parses and generates a comprehensive report, obtains the weak knowledge points where the user has made mistakes many times, and makes subsequent training suggestions for the user's weak knowledge points. For example, the summary instruction can be: Please summarize all errors and blind spots based on the solution ideas of all the questions above, generate a comprehensive report for the user, focus on the [blind spots in problem solving], and give subsequent [training suggestions]. Please reply strictly in Chinese.

[0108] In this way, multiple wrong questions are collected, and the target language model is called to analyze and classify them to form a summary of weak knowledge points, so as to accurately locate the personalized teaching direction for each user.

[0109] In some embodiments, the method further comprises:

[0110] Perform word segmentation on the received user answers;

[0111] Calculating the confidence of each segmentation in the user's answer, and identifying a target segmentation whose confidence is less than a preset confidence;

[0112] Obtain candidate participles corresponding to the target participle, wherein the candidate participles are similar words to the target participle;

[0113] The target word is replaced by the candidate word to obtain a target answer.

[0114] Specifically, multiple candidate participles corresponding to the target participle may be obtained, and the multiple candidate participles may be scored in multiple aspects, such as from the perspective of phonetic level features, character level features, semantic level features, etc., and the candidate participle with the highest comprehensive score may be used to replace the target participle.

[0115] In this way, the communication efficiency between the user and the target language model can be promoted, and the data processing efficiency of the target language model can be improved.

[0116] In one embodiment, Figure 2 Schematic diagram of the framework in the embodiment of the present invention, such as Figure 2 As shown, the framework of the present invention includes an interactive and immersive practice module, an AI large language model module and a summary and scoring module; wherein the interactive and immersive practice module includes: a computer question-setting module, a game interactive problem-solving module and a computer scoring module, the game interactive problem-solving module further specifically includes: intelligent question-answering, oral presentation of ideas, generation of a single module for problem-solving ideas, a blind spot identification module, and the summary and scoring module includes a computer scoring and a blind spot summary analysis table.

[0117] In this way, by using an AI big language model and three modules, a structured solution is provided for identifying children's blind spots and providing suggestions for improvement. First, the "Generate Problem Solving Ideas Single Module" helps children actively describe their solution process and interact with the AI ​​big language model through voice or text. This helps children learn from experience and improve their confidence. Secondly, the "blind spot identification module" identifies children's mistakes by comparing their solutions with standard answers. Third, the "blind spot summary analysis module" summarizes all the children's mistakes and blind spots and identifies the key points they did not understand. This provides parents with suggestions for improvement to help them better support their children's academic progress. Thus providing a comprehensive and effective method to help children improve their academic performance, by providing a structured solution for identifying children's blind spots and providing suggestions for improvement, the invention helps reduce technical failures and improve children's academic success.

[0118] It should be understood that although Figure 1 and Figure 2The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 and Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0119] In one embodiment, Figure 3 As shown, a device for identifying weak knowledge points of a user is provided, and the device includes: a first processing module 301, which is used to send an inquiry instruction, and determine the user's unanswered questions according to the user's reply to the inquiry instruction; a second processing module 302, which sends an idea acquisition instruction, and combines with a target language model to obtain the user's solution idea for the unanswered question according to the user's reply to the idea acquisition instruction, wherein the solution idea includes the answer to the question; a comparison module 303, which is used to obtain the standard solution idea for the unanswered question, fill the standard solution idea and the solution idea into a preset comparison instruction template, generate a comparison instruction, send the comparison instruction to the target language model, and compare the user's solution idea and the standard solution idea through the target language model; an identification module 304, which is used to identify the user's weak knowledge points in the unanswered question through the target language model when the user's solution idea is incorrect.

[0120] As a preferred implementation, in an embodiment of the present invention, the first processing module 301 is specifically used to: send inquiry questions to the user, receive personal information answered by the user; and match the unanswered questions that match the user's personal information based on the user's personal information.

[0121] As a preferred implementation manner, in an embodiment of the present invention, the device also includes an improvement module, which is specifically used to: obtain reference teaching materials; extract the title, solution ideas and answers of each question from the reference teaching materials to generate a data set; and fine-tune the large language model based on the data set to obtain a target language model.

[0122] As a preferred implementation manner, in an embodiment of the present invention, the device also includes a generation module, which is specifically used to: create question description instructions, guidance instructions and answer confirmation instructions; and generate the idea description instructions based on the question description instructions, guidance instructions and the answer confirmation instructions.

[0123] As a preferred implementation mode, in an embodiment of the present invention, the second processing module 302 is specifically used for: sending the question description instruction to the user, wherein the question description instruction is used to instruct the user to describe the question to be answered; receiving the question description content returned by the user, comparing the question description content with the question to be answered, and determining whether the question description content is consistent with the question to be answered; if not, resending the question description instruction to the user; if consistent, generating corresponding guidance instructions according to the question to be answered, wherein the guidance instructions include multiple sub-guidance instructions, each sub-guidance instruction is used to instruct the user to describe sub-problem-solving steps and sub-problem-solving answers; sending each sub-guidance instruction to the user in turn, and receiving the sub-problem-solving steps and sub-problem-solving answers returned by the user; when detecting that the sub-problem-solving answer returned by the user is the answer to the question to be answered, sending an answer confirmation instruction to the user, and receiving the question answer returned by the user; generating the user's solution ideas for the question to be answered according to the sub-problem-solving steps, the sub-problem-solving answers and the question answer.

[0124] As a preferred implementation manner, in an embodiment of the present invention, the second processing module 302 is specifically used to: generate a corresponding first sub-guidance instruction based on the question to be answered, wherein the first sub-guidance instruction is used to instruct the user to describe the first sub-problem-solving steps and the first problem-solving result calculated according to the first sub-problem-solving steps; send the first sub-guidance instruction to the user, and receive the first sub-problem-solving steps and the first problem-solving result returned by the user; parse the first sub-problem-solving steps returned by the user, and generate a second sub-guidance instruction based on the first sub-problem-solving steps, wherein the second sub-guidance instruction is to perform a second sub-problem-solving step guidance on the question to be answered based on the first sub-problem-solving steps; and receive the second sub-problem-solving steps and the second sub-problem-solving result returned by the user.

[0125] As a preferred implementation, in an embodiment of the present invention, the device also includes a summarizing module, which is specifically used to: obtain the weak knowledge points of the user in multiple questions to be answered; and summarize the weak knowledge points of the user through the target language model.

[0126] The specific definition of the user's weak knowledge point identification device can be found in the above definition of the user's weak knowledge point identification method, which will not be repeated here. The various modules in the above-mentioned user's weak knowledge point identification device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0127] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for identifying weak knowledge points of a user is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0128] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0129] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the following steps are implemented when the processor executes the computer program: S1: sending a query instruction, and determining the user's unanswered question based on the user's reply to the query instruction; S2: sending a thought acquisition instruction, and obtaining the user's solution idea for the unanswered question based on the user's reply to the thought acquisition instruction in combination with a target language model, wherein the solution idea includes an answer to the question; S3: obtaining a standard solution idea for the unanswered question, filling the standard solution idea and the solution idea into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the user's solution idea with the standard solution idea through the target language model; S4: when the user's solution idea is incorrect, identifying the user's weak knowledge points in the unanswered question through the target language model.

[0130] In one embodiment, when the processor executes the computer program, the following steps are also implemented: sending a query question to the user and receiving the personal information of the user's answer; and matching the unanswered questions that match the user's personal information based on the user's personal information.

[0131] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining reference teaching materials; extracting the title, solution ideas and answers of each question from the reference teaching materials to generate a data set; and fine-tuning the large language model based on the data set to obtain a target language model.

[0132] In one embodiment, when the processor executes the computer program, the following steps are also implemented: creating a question description instruction, a guide instruction and an answer confirmation instruction; and generating the idea description instruction according to the question description instruction, the guide instruction and the answer confirmation instruction.

[0133] In one embodiment, when the processor executes the computer program, the following steps are also implemented: sending the question description instruction to the user, wherein the question description instruction is used to instruct the user to describe the question to be answered; receiving the question description content returned by the user, comparing the question description content with the question to be answered, and determining whether the question description content is consistent with the question to be answered; if not, resending the question description instruction to the user; if consistent, generating corresponding guidance instructions according to the question to be answered, wherein the guidance instructions include multiple sub-guidance instructions, each sub-guidance instruction is used to instruct the user to describe sub-problem-solving steps and sub-problem-solving answers; sending each sub-guidance instruction to the user in turn, and receiving the sub-problem-solving steps and sub-problem-solving answers returned by the user; when it is detected that the sub-problem-solving answer returned by the user is the answer to the question to be answered, sending an answer confirmation instruction to the user, and receiving the question answer returned by the user; generating the user's solution ideas for the question to be answered according to the sub-problem-solving steps, the sub-problem-solving answers and the question answer.

[0134] In one embodiment, when the processor executes the computer program, the following steps are also implemented: generating a corresponding first sub-guidance instruction according to the question to be answered, wherein the first sub-guidance instruction is used to instruct the user to describe the first sub-problem-solving steps and the first problem-solving result calculated according to the first sub-problem-solving steps; sending the first sub-guidance instruction to the user, and receiving the first sub-problem-solving steps and the first problem-solving result returned by the user; parsing the first sub-problem-solving steps returned by the user, and generating a second sub-guidance instruction according to the first sub-problem-solving steps, wherein the second sub-guidance instruction is to guide the second sub-problem-solving steps of the question to be answered based on the first sub-problem-solving steps; and receiving the second sub-problem-solving steps and the second sub-problem-solving results returned by the user.

[0135] In one embodiment, when the processor executes the computer program, the following steps are also implemented: obtaining the weak knowledge points of the user in multiple questions to be answered; and summarizing the weak knowledge points of the user through the target language model.

[0136] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: S1: sending a query instruction, and determining the user's unanswered question according to the user's reply to the query instruction; S2: sending a thought acquisition instruction, and combining a target language model, obtaining the user's solution idea for the unanswered question according to the user's reply to the thought acquisition instruction, wherein the solution idea includes the answer to the question;

[0137] S3: Obtain the standard solution ideas for the questions to be answered, fill the standard solution ideas and the solution ideas into a preset comparison instruction template, generate comparison instructions, send the comparison instructions to the target language model, and compare the user's solution ideas and the standard solution ideas through the target language model; S4: When the user's solution ideas are incorrect, identify the user's weak knowledge points in the questions to be answered through the target language model.

[0138] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: sending a query question to the user and receiving the personal information of the user's answer; and matching the unanswered questions that match the user's personal information based on the user's personal information.

[0139] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining reference teaching materials; extracting the title, solution ideas and answers of each question from the reference teaching materials to generate a data set; and fine-tuning the large language model based on the data set to obtain a target language model.

[0140] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: creating a question description instruction, a guide instruction and an answer confirmation instruction; and generating the idea description instruction according to the question description instruction, the guide instruction and the answer confirmation instruction.

[0141] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: sending the question description instruction to the user, wherein the question description instruction is used to instruct the user to describe the question to be answered; receiving the question description content returned by the user, comparing the question description content with the question to be answered, and determining whether the question description content is consistent with the question to be answered; if not, resending the question description instruction to the user; if consistent, generating corresponding guidance instructions according to the question to be answered, wherein the guidance instructions include multiple sub-guidance instructions, each sub-guidance instruction is used to instruct the user to describe sub-problem-solving steps and sub-problem-solving answers; sending each sub-guidance instruction to the user in turn, and receiving the sub-problem-solving steps and sub-problem-solving answers returned by the user; when it is detected that the sub-problem-solving answer returned by the user is the answer to the question to be answered, sending an answer confirmation instruction to the user, and receiving the question answer returned by the user; generating the user's solution ideas for the question to be answered according to the sub-problem-solving steps, the sub-problem-solving answers and the question answer.

[0142] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: generating a corresponding first sub-guidance instruction based on the question to be answered, wherein the first sub-guidance instruction is used to instruct the user to describe the first sub-problem-solving steps and the first problem-solving result calculated according to the first sub-problem-solving steps; sending the first sub-guidance instruction to the user, and receiving the first sub-problem-solving steps and the first problem-solving result returned by the user; parsing the first sub-problem-solving steps returned by the user, and generating a second sub-guidance instruction based on the first sub-problem-solving steps, wherein the second sub-guidance instruction is to guide the second sub-problem-solving steps of the question to be answered based on the first sub-problem-solving steps; and receiving the second sub-problem-solving steps and the second sub-problem-solving result returned by the user.

[0143] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining the weak knowledge points of the user in multiple questions to be answered; and summarizing the weak knowledge points of the user through the target language model.

[0144] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0145] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0146] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for identifying weak knowledge points of a user, characterized in that: The method comprises: Sending a query instruction, and determining the user's unanswered question according to the user's reply to the query instruction; Sending a thought acquisition instruction, combining with the target language model, and obtaining the user's thought for solving the problem to be solved according to the user's reply to the thought acquisition instruction, wherein the thought for solving the problem includes the answer to the problem; Obtaining a standard solution idea for the question to be answered, filling the standard solution idea and the solution idea into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the user's solution idea with the standard solution idea through the target language model; When the user's solution idea is incorrect, the target language model is used to identify the user's weak knowledge points in the question to be answered.

2. The method for identifying weak knowledge points of a user according to claim 1, characterized in that: The sending of the inquiry instruction and determining the question to be answered by the user according to the user's reply to the inquiry instruction include: Sending inquiry questions to users and receiving personal information answered by the users; According to the personal information of the user, matching questions to be answered that match the personal information of the user.

3. The method for identifying weak knowledge points of a user according to claim 1, characterized in that: The method further comprises: Obtain reference materials; Extract the title, solution ideas and answers of each question from the reference textbook to generate a data set; The large language model is fine-tuned based on the data set to obtain a target language model.

4. The method for identifying weak knowledge points of a user according to claim 1, characterized in that: The method further comprises: Create question description instructions, guidance instructions, and answer confirmation instructions; The idea description instruction is generated according to the question description instruction, the guidance instruction and the answer confirmation instruction.

5. The method for identifying weak knowledge points of a user according to claim 4, characterized in that: The sending of the idea acquisition instruction, in combination with the target language model, obtains the user's solution idea for the question to be answered according to the user's reply to the idea acquisition instruction, wherein the solution idea includes the answer to the question, including: Sending the question description instruction to the user, wherein the question description instruction is used to instruct the user to describe the question to be answered; Receiving the question description content returned by the user, comparing the question description content with the question to be answered, and determining whether the question description content is consistent with the question to be answered; If they are inconsistent, resending the topic description instruction to the user; If they are consistent, generating a corresponding guidance instruction according to the to-be-answered question, wherein the guidance instruction includes a plurality of sub-guidance instructions, each of which is used to instruct the user to describe sub-question-solving steps and sub-question-solving answers; Sending each sub-guidance instruction to the user in sequence, and receiving the sub-problem-solving steps and sub-problem-solving answers returned by the user; When detecting that the sub-question answer returned by the user is the answer to the question to be answered, sending an answer confirmation instruction to the user, and receiving the question answer returned by the user; The user's solution idea for the question to be answered is generated according to the sub-question-solving steps, the sub-question-solving answers and the question answer.

6. The method for identifying weak knowledge points of a user according to claim 1, characterized in that: The method of generating a corresponding guidance instruction according to the to-be-answered question, sending each sub-guidance instruction to the user in sequence, and receiving the sub-question-solving steps and sub-question-solving answers returned by the user includes: Generate a corresponding first sub-guidance instruction according to the to-be-answered question, wherein the first sub-guidance instruction is used to instruct the user to describe a first sub-question-solving step and a first question-solving result calculated according to the first sub-question-solving step; Sending the first sub-guidance instruction to the user, and receiving the first sub-problem-solving steps and the first problem-solving result returned by the user; Parsing the first sub-problem-solving step returned by the user, and generating a second sub-guidance instruction according to the first sub-problem-solving step, wherein the second sub-guidance instruction is to guide the second sub-problem-solving step of the problem to be solved based on the first sub-problem-solving step; Receive the second sub-problem-solving steps and the second sub-problem-solving results returned by the user.

7. The method for identifying weak knowledge points of a user according to claim 1, characterized in that: The method further comprises: Obtaining the user's weak knowledge points in multiple questions to be answered; The user's weak knowledge points are summarized through the target language model.

8. A device for identifying weak knowledge points of a user, characterized in that: The device comprises: The first processing module is used to send a query instruction and determine the user's unanswered questions according to the user's reply to the query instruction; The second processing module sends a thought acquisition instruction, combines the target language model, and obtains the user's thought for solving the problem to be solved according to the user's reply to the thought acquisition instruction, wherein the thought includes the answer to the problem; A comparison module, used for obtaining the standard solution ideas of the to-be-answered questions, filling the standard solution ideas and the solution ideas into a preset comparison instruction template, generating a comparison instruction, sending the comparison instruction to the target language model, and comparing the user's solution ideas with the standard solution ideas through the target language model; The identification module is used to identify the user's weak knowledge points in the question to be answered through the target language model when the user's solution idea is incorrect.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images