Question bank generation method and system based on large model and storage medium

Abstract

The invention provides a question bank generation method and system based on a large model and a storage medium, and the method comprises the steps: obtaining question bank description information, and carrying out the data matching of the question bank description information and a data vector database, and obtaining question bank related data; obtaining a question bank generation requirement, and combining the question bank generation requirement with the question bank associated data to obtain text prompt information; and inputting the text prompt information into a large model to generate a question bank to obtain question bank generation data. According to the embodiment of the invention, the question bank associated data can be automatically determined by performing data matching on the question bank description information and the data vector database, the question bank generation processing operation of a large model is facilitated by combining the question bank generation requirements and the question bank associated data, and the question bank is generated by inputting the text prompt information into the large model, so that the efficiency of generating the question bank is improved. And the question bank generation data meeting the question bank generation requirements and the question bank description can be automatically generated based on the text prompt information.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular, to a question bank generation method, system and storage medium based on a large model. Background Art

[0002] In various professional fields (such as aviation, rail transit, airport, etc.), there is a need for business personnel training. After the lecturer finishes explaining the materials, it is necessary to generate a corresponding question bank for an exam to test the business personnel's mastery of relevant knowledge. Therefore, the problem of question bank generation has been paid more and more attention.

[0003] In the existing process of question bank generation, generally, questions are selected manually to generate the question bank, resulting in cumbersome manual operations and reducing the efficiency of question bank generation. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide a question bank generation method, system and storage medium based on a large model to solve the problem of low efficiency of question bank generation in the prior art.

[0005] The embodiments of the present invention are implemented as follows. A question bank generation method based on a large model, the method includes:

[0006] Obtain question bank description information, and perform material matching between the question bank description information and a material vector database to obtain question bank-related materials;

[0007] Obtain question bank generation requirements, and combine the question bank generation requirements with the question bank-related materials to obtain text prompt information;

[0008] Input the text prompt information into a large model for question bank generation to obtain question bank generation data.

[0009] Preferably, inputting the text prompt information into a large model for question bank generation to obtain question bank generation data includes:

[0010] Analyze the knowledge points of the text prompt information according to the large model to obtain question knowledge points, and determine a knowledge point question bank according to the question knowledge points;

[0011] Determine the question type, the number of questions corresponding to the question type and the question difficulty value according to the text prompt information, and determine the number of knowledge point questions of the question knowledge points according to the number of questions;

[0012] Extract questions from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data.

[0013] Preferably, determining a knowledge point question bank according to the question knowledge points includes:

[0014] Match the knowledge points of the question with the question bank query table to obtain a matching question bank, and obtain the error-prone questions of the knowledge points of the question;

[0015] Add the error-prone questions to the matching question bank to obtain the knowledge point question bank.

[0016] Preferably, before performing knowledge point analysis on the text prompt information according to the large model, it further includes:

[0017] Perform a random sorting of the difficulty values according to the number of questions of the knowledge points and the difficulty value of the questions to obtain a question difficulty sequence, and the sum of the difficulty values between the sequence elements in the question difficulty sequence is the difficulty value of the questions;

[0018] Extract questions from the knowledge point question bank according to the difficulty values of each sequence element to obtain the extracted questions, and perform a random sorting on the extracted questions to obtain the question bank generation data.

[0019] Preferably, before performing knowledge point analysis on the text prompt information according to the large model, it further includes:

[0020] Obtain a sample text, and input the sample text into the large model for word segmentation to obtain text word segmentation;

[0021] Construct a vocabulary according to the text word segmentation, and perform token encoding on the text word segmentation according to the vocabulary to obtain token identifiers;

[0022] Perform vector conversion on the token identifiers to obtain word vectors, and perform feature extraction on the word vectors to obtain sample features;

[0023] Decode the sample features to obtain predicted knowledge points, and calculate a loss according to the predicted knowledge points to obtain a model loss;

[0024] Update the parameters of the large model according to the model loss until the large model converges, and output the converged large model.

[0025] Preferably, after extracting questions from the knowledge point question bank according to the difficulty values of each sequence element to obtain the extracted questions, it further includes:

[0026] Perform vector conversion on the extracted questions to obtain question vectors, and calculate the similarity between different question vectors to obtain question similarities;

[0027] If any of the question similarities is greater than a preset similarity, optimize the different extracted questions corresponding to the question similarity.

[0028] Preferably, matching the question bank description information with the data vector database to obtain question bank-related data, including:

[0029] Converting the question bank description information into a vector to obtain a question bank description vector, and using the question bank description vector as a query vector to perform a similarity search in the data vector database;

[0030] According to the similarity search results, determining the data with a vector similarity greater than the similarity threshold as the question bank-related data.

[0031] Another object of the embodiments of the present invention is to provide a question bank generation system based on a large model, the system includes:

[0032] A data matching module, configured to obtain question bank description information, and match the question bank description information with the data vector database to obtain question bank-related data;

[0033] An information combination module, configured to obtain question bank generation requirements, and combine the question bank generation requirements with the question bank-related data to obtain text prompt information;

[0034] A question bank generation module, configured to input the text prompt information into the large model for question bank generation to obtain question bank generation data.

[0035] Preferably, the question generation module is further configured to:

[0036] Analyze the knowledge points of the text prompt information according to the large model to obtain question knowledge points, and determine a knowledge point question bank according to the question knowledge points;

[0037] Determine the question type, the number of questions corresponding to the question type, and the question difficulty value according to the text prompt information, and determine the number of knowledge point questions of the question knowledge points according to the number of questions;

[0038] Extract questions from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data.

[0039] In the embodiments of the present invention, by matching the question bank description information with the data vector database, the question bank-related data can be automatically determined. By combining the question bank generation requirements with the question bank-related data, the question bank generation processing operation of the large model is facilitated. By inputting the text prompt information into the large model for question bank generation, question bank generation data that meets the question bank generation requirements and the question bank description can be automatically generated based on the text prompt information, without the need to select questions manually, improving the question bank generation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1It is a flowchart of a question bank generation method based on a large model provided by the first embodiment of the present invention;

[0041] Figure 2 It is a schematic structural diagram of a question bank generation system based on a large model provided by the second embodiment of the present invention;

[0042] Figure 3 It is a specific implementation schematic diagram of a question bank generation system based on a large model provided by the second embodiment of the present invention;

[0043] Figure 4 It is a schematic structural diagram of a terminal device provided by the third embodiment of the present invention. Specific embodiments

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0045] In order to illustrate the technical solutions described in the present invention, the following will be illustrated through specific embodiments.

[0046] Embodiment 1

[0047] Please refer to Figure 1 , which is a flowchart of a question bank generation method based on a large model provided by the first embodiment of the present invention. The question bank generation method based on a large model can be applied to any device or system. The question bank generation method based on a large model includes the steps:

[0048] Step S10, obtain question bank description information, and perform data matching between the question bank description information and a data vector database of materials to obtain question bank-related materials;

[0049] Among them, the question bank description information is a type description of the data generated by the user for the question bank to be generated. For example, the question bank description information can be set according to the user's needs as "generate a question bank for the emergency plan for heavy snow weather on rail transit", "generate a question bank for the emergency plan at the airport in heavy snow weather", etc. In this step, training material documents are obtained through crawler data in advance, and the training material documents are vectorized and stored to obtain a data vector database of materials. When the user wants to generate a question bank of a certain type, relevant training content is recalled from the data vector database of materials to obtain question bank-related materials.

[0050] Optionally, performing data matching between the question bank description information and a data vector database of materials to obtain question bank-related materials includes:

[0051] Perform vector transformation on the question bank description information to obtain a question bank description vector, and use the question bank description vector as a query vector to perform similarity search in the material vector database; wherein, the vector similarity is obtained by calculating the similarity between the question bank description vector and the material vectors in the material vector database.

[0052] According to the similarity search results, determine the materials with vector similarity greater than the similarity threshold as the materials associated with the question bank; wherein, the similarity threshold can be set according to requirements. When the vector similarity is greater than the similarity threshold, it is determined that the materials corresponding to the vector similarity have a high similarity to the required materials of the question bank description information.

[0053] Step S20: Obtain the question bank generation requirements, and combine the question bank generation requirements with the materials associated with the question bank to obtain text prompt information.

[0054] Among them, the question bank generation requirements include the user's requirements for the question type, the number of questions corresponding to the question type, and the question difficulty value. By combining the question bank generation requirements with the materials associated with the question bank to obtain the text prompt information prompt, it facilitates the question bank generation processing operation of the large model.

[0055] Step S30: Input the text prompt information into the large model for question bank generation to obtain question bank generation data.

[0056] Among them, the question bank generation data is saved in the business system, and business personnel can retrieve it through the exam question system for exam training. In this embodiment, by introducing a vector database and a large model to replace the work of manually writing question banks, the writing efficiency is improved. Once the training materials change, the corresponding question banks can also be updated quickly; at the same time, vectorized retrieval of the question bank theme is performed, and within the scope of the recalled relevant knowledge, the question banks automatically generated by the large model greatly increase the accuracy of the question banks.

[0057] Optionally, inputting the text prompt information into the large model for question bank generation to obtain question bank generation data includes:

[0058] Analyze the knowledge points of the text prompt information according to the large model to obtain question knowledge points, and determine the knowledge point question bank according to the question knowledge points; wherein, based on the large model to analyze the knowledge points of the text prompt information, the question knowledge points of the materials associated with the question bank in the text prompt information can be effectively extracted, and the corresponding knowledge point question bank can be effectively determined based on the question knowledge points.

[0059] Determine the question type, the corresponding number of questions and the question difficulty value according to the text prompt information, determine the number of knowledge point questions of the question knowledge points according to the number of questions, and extract questions from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data; among them, obtain the question type, the number of questions and the question difficulty value stored in the question bank generation requirements in the text prompt information, calculate the quotient of the number of questions and the number of knowledge points of the question knowledge points to obtain the number of knowledge point questions, and through the number of knowledge point questions and the question difficulty value, the operation of automatically extracting questions from the knowledge point question bank can be realized to obtain question bank generation data.

[0060] Further, determine the knowledge point question bank according to the question knowledge points, including:

[0061] Match the question knowledge points with the question bank query table to obtain the matching question bank, and obtain the error-prone questions of the question knowledge points; among them, the question bank query table stores the corresponding relationship between different question knowledge points and the corresponding matching question banks, match the question knowledge points with the wrong question database to obtain the error-prone questions, and the wrong question database stores the corresponding relationship between different question knowledge points and the corresponding error-prone questions;

[0062] Add the error-prone questions to the matching question bank to obtain the knowledge point question bank; among them, by adding the error-prone questions to the matching question bank, the question quality of the knowledge point question bank is effectively improved.

[0063] Even further, extract questions from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data, including:

[0064] Perform a random sorting of the difficulty values according to the number of knowledge point questions and the question difficulty value to obtain a question difficulty sequence; among them, the sum of the difficulty values between the sequence elements in the question difficulty sequence is the question difficulty value;

[0065] Extract questions from the knowledge point question bank according to the difficulty value of each sequence element to obtain the extracted questions, and perform a random sorting on the extracted questions to obtain the question bank generation data; among them, randomly extract questions with the corresponding difficulty value in the knowledge point question bank according to the difficulty value of each sequence element to obtain the extracted questions.

[0066] Preferably, before analyzing the knowledge points of the text prompt information according to the large model, it further includes:

[0067] Obtain a sample text, and input the sample text into the large model for word segmentation to obtain text word segmentation; among them, match the sample text with the word segmentation dictionary, and perform word segmentation on the sample text according to the matching result to obtain text word segmentation;

[0068] Construct a vocabulary based on the text tokenization, and perform token encoding on the text tokenization according to the vocabulary to obtain token identifiers; among them, constructing a vocabulary based on the text tokenization and the number of text tokenizations, and performing token encoding on the text tokenization through the vocabulary can effectively convert the text tokenization into corresponding token identifiers;

[0069] Perform vector conversion on the token identifiers to obtain word vectors, perform feature extraction on the word vectors to obtain sample features, perform decoding on the sample features to obtain predicted knowledge points, calculate the model loss according to the predicted knowledge points, update the parameters of the large model according to the model loss until the large model converges, output the converged large model, and based on the converged large model, can effectively analyze the knowledge points of the input text prompt information to obtain corresponding topic knowledge points.

[0070] In this embodiment, after extracting questions from the knowledge point question bank according to the difficulty values of each sequence element, it further includes:

[0071] Perform vector conversion on the extracted questions to obtain question vectors, and calculate the similarity between different question vectors to obtain question similarity; among them, by calculating the similarity between different question vectors, it can effectively determine whether there is a phenomenon of similar questions among different extracted questions;

[0072] If any of the question similarities is greater than the preset similarity, optimize the questions corresponding to the question similarity; the preset similarity can be set according to requirements. If any question similarity is greater than the preset similarity, it is determined that there is a duplicate phenomenon between the different extracted questions corresponding to the question similarity. By retaining one extracted question, the effect of optimizing the extracted questions is achieved, thereby improving the accuracy of the extracted questions.

[0073] In this embodiment, by matching the question bank description information with the data vector database, the question bank associated data can be automatically determined. By combining the question bank generation requirements with the question bank associated data, the question bank generation processing operation of the large model is facilitated. By inputting the text prompt information into the large model for question bank generation, the question bank generation data that meets the question bank generation requirements and the question bank description can be automatically generated based on the text prompt information, without the need to select questions manually, improving the question bank generation efficiency.

[0074] Embodiment Two

[0075] Please refer to Figure 2 , which is a schematic structural diagram of a question bank generation system 100 based on a large model provided by the second embodiment of the present invention, including:

[0076] The data matching module 10 is used to obtain the question bank description information, and match the question bank description information with the data vector database to obtain the question bank associated data. The question bank description information is the type description of the data generated by the user for the to-be-generated question bank. For example, the question bank description information can be set according to the user's needs as "generate a question bank for the emergency plan for heavy snow weather on rail transit", "generate a question bank for the emergency plan at the airport in heavy snow weather", etc. In this step, training material documents are obtained through crawler data in advance, and the training material documents are vectorized and stored to obtain the data vector database. When the user wants to generate a question bank of a certain type, relevant training content is recalled from the data vector database to obtain the question bank associated data.

[0077] Optionally, the data matching module 10 is further used to: perform vector conversion on the question bank description information to obtain a question bank description vector, and perform similarity search on the data vector database with the question bank description vector as the query vector;

[0078] According to the similarity search results, the data with a vector similarity greater than the similarity threshold is determined as the question bank associated data.

[0079] The information combination module 11 is used to obtain the question bank generation requirements, and combine the question bank generation requirements with the question bank associated data to obtain text prompt information. The question bank generation requirements include the user's requirements for the question type, the number of questions corresponding to the question type, and the question difficulty value. By combining the question bank generation requirements with the question bank associated data to obtain the text prompt information prompt, it facilitates the question bank generation processing operation of the large model.

[0080] The question bank generation module 12 is used to input the text prompt information into the large model for question bank generation to obtain question bank generation data.

[0081] Optionally, the question bank generation module 12 is further used to: perform knowledge point analysis on the text prompt information by the large model to obtain question knowledge points, and determine a knowledge point question bank according to the question knowledge points;

[0082] Determine the question type, the number of questions corresponding to the question type, and the question difficulty value according to the text prompt information, and determine the number of knowledge point questions of the question knowledge points according to the number of questions;

[0083] Perform question extraction on the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data. Among them, obtain the question type, the number of questions, and the question difficulty value stored in the question bank generation requirements in the text prompt information, calculate the quotient value between the number of questions and the number of knowledge points of the question knowledge points to obtain the number of knowledge point questions. Through the number of knowledge point questions and the question difficulty value, the operation of question extraction from the knowledge point question bank can be automatically realized to obtain question bank generation data.

[0084] Further, the question bank generation module 12 is further configured to: match the question knowledge points with the question bank query table to obtain a matching question bank, and obtain the error-prone questions of the question knowledge points;

[0085] Add the error-prone questions to the matching question bank to obtain the knowledge point question bank.

[0086] Furthermore, the question bank generation module 12 is further configured to: perform a random sorting of difficulty values according to the number of knowledge point questions and the question difficulty value to obtain a question difficulty sequence, and the sum of the difficulty values between the sequence elements in the question difficulty sequence is the question difficulty value;

[0087] Extract questions from the knowledge point question bank according to the difficulty values of the sequence elements, obtain the extracted questions, and perform a random sorting on the extracted questions to obtain the question bank generation data.

[0088] Preferably, the question bank generation module 12 is further configured to: obtain a sample text, input the sample text into the large model for word segmentation to obtain text word segmentation;

[0089] Construct a vocabulary table according to the text word segmentation, and perform token encoding on the text word segmentation according to the vocabulary table to obtain token identifiers;

[0090] Perform vector conversion on the token identifiers to obtain word vectors, and perform feature extraction on the word vectors to obtain sample features;

[0091] Decode the sample features to obtain predicted knowledge points, and calculate a loss according to the predicted knowledge points to obtain a model loss;

[0092] Update the parameters of the large model according to the model loss until the large model converges, and output the converged large model.

[0093] In this embodiment, the question bank generation module 12 is further configured to: perform vector conversion on the extracted questions to obtain question vectors, and calculate the similarity between different question vectors to obtain question similarity;

[0094] If any of the question similarities is greater than a preset similarity, optimize the different extracted questions corresponding to the question similarity.

[0095] Please refer to Figure 3 , by introducing a vector database, vectorize and store the training material documents; when the user wants to generate a certain type of question bank, first recall relevant training content from the vector database, and then use the large model to generate a question bank from this content, specifically:

[0096] 1. Slice the training materials into paragraph data to obtain material paragraphs, and store the data of the material paragraphs, such as a lotus, into the vector database;

[0097] 2. Obtain the description information of the question bank type to be generated (for example: generate a question bank for the emergency plan for heavy snow weather in rail transit);

[0098] 3. Use the description information of the question bank type for vectorized query, and recall relevant knowledge content from the vector database;

[0099] 4. Assemble the relevant knowledge content and the requirements for generating the question bank (question types, questions, correct answers, etc.) into a prompt;

[0100] 5. Send the prompt to the large model for understanding, and automatically generate a question bank from the relevant knowledge content;

[0101] 6. Save the question bank to the business system, and business personnel can retrieve it through the exam question system for exam training.

[0102] By introducing a vector database and a large model to replace the work of manually writing question banks, the writing efficiency is improved. Once the training materials change, the corresponding question banks can also be updated quickly; at the same time, vectorized retrieval is performed on the question bank theme, and within the scope of the recalled relevant knowledge, the question banks automatically generated by the large model greatly increase the accuracy of the question banks.

[0103] In this embodiment, by matching the question bank description information with the material vector database, the question bank-related materials can be automatically determined. By combining the requirements for generating the question bank with the question bank-related materials, it facilitates the question bank generation processing operation of the large model. By inputting the text prompt information into the large model for question bank generation, the question bank generation data that meets the requirements for generating the question bank and the question bank description can be automatically generated based on the text prompt information, without the need to select questions manually, improving the question bank generation efficiency.

[0104] Embodiment 3

[0105] Figure 4 It is a structural block diagram of a terminal device 2 provided in the third embodiment of the present application. As Figure 4 shown, the terminal device 2 of this embodiment includes: a processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the processor 20, such as a program for the method of generating a question bank based on a large model. When the processor 20 executes the computer program 22, the steps in each of the above embodiments of the method for generating a question bank based on a large model are implemented.

[0106] Exemplarily, the computer program 22 may be divided into one or more modules, which are stored in the memory 21 and executed by the processor 20 to complete the present application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program 22 in the terminal device 2. The terminal device may include, but is not limited to, a processor 20 and a memory 21.

[0107] The so-called processor 20 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0108] The memory 21 may be an internal storage unit of the terminal device 2, such as the hard disk or memory of the terminal device 2. The memory 21 may also be an external storage device of the terminal device 2, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the terminal device 2. Further, the memory 21 may also include both the internal storage unit and the external storage device of the terminal device 2. The memory 21 is used to store the computer program and other programs and data required by the terminal device. The memory 21 may also be used to temporarily store data that has been output or is to be output.

[0109] In addition, in each embodiment of the present application, the various functional modules may be integrated in one processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit. The above integrated unit may be implemented in the form of hardware or in the form of a software functional unit.

[0110] When an integrated module is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium can be non-volatile or volatile. Based on such an understanding, to implement all or part of the processes in the above-mentioned embodiment methods of this application, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable storage medium can include: any entity or device that can carry computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable storage medium does not include electrical carrier signals and telecommunication signals.

[0111] The above-described embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of this application, and should all be included in the protection scope of this application.

Claims

1. A method for generating a question bank based on a large model, characterized in that: The method comprises: Acquire question bank description information, and perform data matching on the question bank description information and the data vector database to obtain question bank related data; Obtaining a question bank generation requirement, and combining the question bank generation requirement with the question bank associated data to obtain text prompt information; The text prompt information is input into the large model to generate a question bank, and question bank generation data is obtained.

2. The method for generating a question bank based on a large model as claimed in claim 1, characterized in that: The text prompt information is input into the large model to generate a question bank, and the question bank generation data is obtained, including: Performing knowledge point analysis on the text prompt information according to the large model to obtain question knowledge points, and determining a knowledge point question bank according to the question knowledge points; Determine the question type, the number of questions corresponding to the question type and the question difficulty value according to the text prompt information, and determine the number of knowledge point questions of the question knowledge point according to the number of questions; Questions are extracted from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data.

3. The method for generating a question bank based on a large model as claimed in claim 2, characterized in that: Determine a knowledge point question bank according to the knowledge points of the question, including: Match the question knowledge point with the question bank query table to obtain a matching question bank, and obtain easy-to-make mistakes questions of the question knowledge point; The easy-to-make mistakes questions are added to the matching question bank to obtain the knowledge point question bank.

4. The method for generating a question bank based on a large model as claimed in claim 2, characterized in that: Extracting questions from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data includes: Randomly sort the difficulty values ​​according to the number of knowledge point questions and the difficulty values ​​of the questions to obtain a question difficulty sequence, wherein the sum of the difficulty values ​​of the sequence elements in the question difficulty sequence is the question difficulty value; Extracting questions from the knowledge point question bank according to the difficulty value of each sequence element to obtain extracted questions, and randomly sorting the extracted questions to obtain the question bank generation data.

5. The method for generating a question bank based on a large model as claimed in claim 2, characterized in that: Before performing knowledge point analysis on the text prompt information according to the large model, the method further includes: Obtaining a sample text, and inputting the sample text into the large model for word segmentation to obtain text segmentation; Constructing a vocabulary according to the text segmentation, and performing word unit encoding on the text segmentation according to the vocabulary to obtain word unit identifiers; Performing vector conversion on the word element identifier to obtain a word vector, and performing feature extraction on the word vector to obtain a sample feature; Decoding the sample features to obtain predicted knowledge points, and performing loss calculation based on the predicted knowledge points to obtain model loss; The parameters of the large model are updated according to the model loss until the large model converges, and the converged large model is output.

6. The method for generating a question bank based on a large model as claimed in claim 4, characterized in that: Extracting questions from the knowledge point question bank according to the difficulty value of each sequence element, and obtaining the extracted questions, further comprising: Performing vector conversion on the extracted questions to obtain question vectors, and calculating similarities between different question vectors to obtain question similarity; If the similarity of any of the questions is greater than a preset similarity, then the questions corresponding to the different extracted questions of the question similarity are optimized.

7. The method for generating a question bank based on a large model as claimed in claim 1, characterized in that: The question bank description information is matched with the data vector database to obtain question bank related data, including: Performing vector conversion on the question bank description information to obtain a question bank description vector, and using the question bank description vector as a query vector to perform similarity search in the data vector database; According to the similarity search results, the data whose vector similarity is greater than the similarity threshold is determined as the question bank related data.

8. A question bank generation system based on a large model, characterized in that: The system comprises: A data matching module is used to obtain question bank description information, and to match the question bank description information with a data vector database to obtain question bank related data; An information combination module, used to obtain question bank generation requirements, and combine the question bank generation requirements with the question bank related information to obtain text prompt information; The question bank generation module is used to input the text prompt information into the large model to generate the question bank and obtain question bank generation data.

9. The large model-based question bank generation system according to claim 8, characterized in that: The topic generation module is also used for: Performing knowledge point analysis on the text prompt information according to the large model to obtain question knowledge points, and determining a knowledge point question bank according to the question knowledge points; Determine the question type, the number of questions corresponding to the question type and the question difficulty value according to the text prompt information, and determine the number of knowledge point questions of the question knowledge point according to the number of questions; Questions are extracted from the knowledge point question bank according to the number of knowledge point questions and the question difficulty value to obtain question bank generation data.

10. A computer-readable storage medium storing a computer program, 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