Exercise generation method and device based on large model, terminal and storage medium

Abstract

The invention provides an exercise generation method and device based on a large model, a terminal and a storage medium. The method comprises the following steps: extracting subtitle information in a received education recorded broadcast video; inputting the subtitle analysis cue word and the subtitle information into a large model, wherein the large model analyzes the subtitle information based on the subtitle analysis cue word to obtain knowledge point summary information; inputting the exercise generation cue words and the knowledge point summary information into the large model, wherein the large model automatically generates matched exercises related to each target knowledge point in the knowledge point summary information based on the exercise generation cue words and analyzes answers of the matched exercises. Through a large model technology and subtitle information analysis, the matched exercises corresponding to the knowledge points in the video and answer analysis of the matched exercises are automatically generated for the education recorded broadcast video, personalized exercise resources can be provided for learners, the learners can conveniently and directly use the generated matched exercises to practice after class, the learners do not need to search exercise resources by themselves, and the learning efficiency is improved. Therefore, a lot of time is saved and the learning efficiency of the learner is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, terminal and storage medium for generating exercises based on a large model. Background Art

[0002] With the rapid development of Internet technology, online education has flourished in recent years, which has directly led to a sharp increase in the number of recorded lectures. Watching these recorded lectures has become an indispensable way for learners to acquire knowledge.

[0003] However, these recorded videos lack supporting or diverse after-class exercise resources. If learners look for relevant exercises on their own, it will not only be time-consuming and laborious, but the search results will be unsatisfactory, making it difficult to effectively consolidate what they have learned in class. Summary of the invention

[0004] The technical problem to be solved by the present invention is that, in view of the above-mentioned defects of the prior art, a method, device, terminal and storage medium for generating exercises based on a large model are provided, aiming to solve the problem of lack of supporting or diversified after-class exercise resources in existing recorded videos.

[0005] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0006] Receiving an educational recorded video and extracting subtitle information from the educational recorded video;

[0007] Inputting the subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information;

[0008] The exercise generation prompt words and the knowledge point summary information are input into the large model, so as to utilize the large model to automatically generate matching exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

[0009] In one implementation, the extracting of subtitle information from the educational recorded video includes:

[0010] Parsing the educational recorded video to determine whether there is corresponding subtitle information in the educational recorded video;

[0011] When the subtitle information exists in the educational recorded video, extracting the subtitle information in the educational recorded video;

[0012] Or, when the subtitle information does not exist in the educational recorded video, the audio information in the educational recorded video is extracted, and the audio information is transcribed to obtain the subtitle information in the educational recorded video.

[0013] In one implementation, the extracting audio information from the educational recording includes:

[0014] The audio information in the educational recording is extracted using the FFmpeg tool.

[0015] In one implementation, transcribing the audio information to obtain the subtitle information in the educational recorded video includes:

[0016] The audio information is transcribed using an ASR tool to obtain the subtitle information in the educational recorded video.

[0017] In one implementation, the step of automatically generating matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words using the macro model includes:

[0018] The large model is used to generate prompt words based on the exercises, and a preset number of supporting exercises and answer analysis related to each target knowledge point in the knowledge point summary information are automatically generated.

[0019] In one implementation, the knowledge point summary information includes each target knowledge point and its introduction and the explanation time period corresponding to each target knowledge point.

[0020] In one implementation, after inputting the exercise generation prompt words and the knowledge point summary information into the macro model, and using the macro model to automatically generate matching exercises related to each target knowledge point in the knowledge point summary information and their answer analysis based on the exercise generation prompt words, the method further includes:

[0021] The identification number corresponding to the educational video recording, the knowledge point summary information and the matching exercises related to each target knowledge point and the answer analysis thereof are stored in a pre-constructed exercise database;

[0022] Among them, the supporting exercises related to each target knowledge point in the exercise database and the answer analysis are associated with the corresponding knowledge point explanation segment in the educational recorded video based on the explanation time period corresponding to each target knowledge point in the educational recorded video.

[0023] The present invention also discloses a large model-based exercise generation device, wherein the device comprises:

[0024] Video receiving module, used for receiving educational recorded videos;

[0025] A subtitle extraction module, used to extract subtitle information from the educational recorded video;

[0026] A subtitle parsing module, used for inputting subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information;

[0027] The exercise generation module is used to input the exercise generation prompt words and the knowledge point summary information into the large model, so as to use the large model to automatically generate supporting exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

[0028] The present invention also discloses a terminal, which includes: a memory, a processor, and a large model-based exercise generation program stored in the memory and executable on the processor. When the large model-based exercise generation program is executed by the processor, the steps of the large model-based exercise generation method as described above are implemented.

[0029] The present invention also discloses a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program can be executed to implement the steps of the large model-based exercise generation method as described above.

[0030] The present invention provides a method, device, terminal and storage medium for generating exercises based on a big model. The method for generating exercises based on a big model includes: receiving an educational recorded video and extracting subtitle information in the educational recorded video; inputting subtitle parsing prompt words and the subtitle information into the big model so that the big model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information; inputting exercise generation prompt words and the knowledge point summary information into the big model so that the big model can automatically generate matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words. It can be seen from this that the present invention automatically generates supporting exercises and answer analyses corresponding to each knowledge point in the educational recorded video through big model technology and subtitle information analysis, that is, extracts subtitle information in the educational recorded video, and then uses the big model to parse the subtitle information based on subtitle parsing prompts, and organizes and summarizes the knowledge point summary information, and finally uses the big model to generate supporting exercises and answer analyses related to each target knowledge point in the knowledge point summary information based on exercise generation prompts, thereby providing personalized exercise resources for different learners, making it convenient for learners to directly use the generated supporting exercises for practice after class without having to find exercise resources by themselves, thereby saving a lot of time and improving learners' learning efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1It is a flow chart of a preferred embodiment of the method for generating exercises based on a large model in the present invention;

[0032] Figure 2 It is a schematic diagram of a specific large model for generating exercises based on prompt words disclosed by the present invention;

[0033] Figure 3 It is a flow chart of a specific method for generating exercises based on a large model disclosed in the present invention;

[0034] Figure 4 It is a flow chart of a specific method for generating exercises based on a large model disclosed in the present invention;

[0035] Figure 5 It is a flow chart of a specific method for generating exercises based on a large model disclosed in the present invention;

[0036] Figure 6 It is a schematic diagram of subtitles for an image-based course in first-grade mathematics disclosed by the present invention;

[0037] Figure 7 It is a functional principle block diagram of a preferred embodiment of the exercise generation device based on a large model in the present invention;

[0038] Figure 8 It is a functional principle block diagram of a preferred embodiment of the terminal in the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. 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.

[0040] See also Figure 1 , Figure 1 Flowchart of the method for generating exercises based on a large model in the present invention. Figure 1 As shown, the method for generating exercises based on a large model according to an embodiment of the present invention includes:

[0041] Step S11, receiving an educational recorded video, and extracting subtitle information from the educational recorded video.

[0042] In this embodiment, subtitle information of the educational recorded video input by the user is extracted, wherein the subtitle information includes text content related to the educational recorded video and the corresponding timeline.

[0043] Step S12: inputting the subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information.

[0044] In this embodiment, after extracting the subtitle information from the educational video, the subtitle information is parsed by the big model, and the knowledge points in the educational video are systematically summarized and organized, and the specific time period for explaining each target knowledge point is accurately marked, that is, the obtained subtitle parsing prompt words and subtitle information are input into the big model, and the big model parses the subtitle information based on the subtitle parsing prompt words to obtain the corresponding knowledge point summary information. Among them, the knowledge point summary information includes each target knowledge point and its introduction and the corresponding explanation time period of each target knowledge point.

[0045] It should be noted that the subtitle parsing prompt words are pre-customized and generated based on the subtitle parsing prompt word template, wherein the subtitle parsing prompt word template of the large model is as follows:

[0046] Subtitle parsing prompt word template for the large model:

[0047] Help parse the following subtitle information, organize relevant knowledge points, and the output format is as follows:

[0048] "Knowledge point", "Knowledge point introduction", "Corresponding video playback start time", "Corresponding video playback end time";

[0049] Subtitle information: %s.

[0050] For example, based on the subtitle information provided, the big model parses the following knowledge points and related information:

[0051] "Knowledge point", "Knowledge point introduction", "Corresponding video playback start time", "Corresponding video playback end time";

[0052] 1. "Recognition of three-dimensional figures", "Introduced and recognized three-dimensional figures in the picture, such as cuboids, cubes, cylinders, etc.", "00:00:02,840", "00:00:07,280";

[0053] 2. "Recognition of rectangular prism", "Specially point out that the rectangular prism is one of the three-dimensional figures learned last semester.", "00:00:17,460", "00:00:24,260";

[0054] 3. "Identification of the cube", "Point out that the cube is one of the three-dimensional figures in the picture.", "00:01:10,200", "00:01:13,720";

[0055] 4. "Identification of cylinder", "Point out that the cylinder is one of the three-dimensional figures in the picture.", "00:01:24,220", "00:01:26,380";

[0056] 5. "Introduction to the triangular prism", "Introduced the new three-dimensional shape of the triangular prism.", "00:01:49,660", "00:01:52,100";

[0057] 6. "Characteristics of a rectangular block", "Ask what characteristics a rectangular block has", "00:02:14,490", "00:02:16,450".

[0058] Step S13, inputting the exercise generation prompt words and the knowledge point summary information into the macro model, so as to utilize the macro model to automatically generate matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

[0059] In this embodiment, after the subtitle information is parsed by the large model and the knowledge point summary information is obtained, the exercise generation prompt words and the knowledge point summary information are input into the large model, so as to use the large model to automatically generate matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words. It should be pointed out that the exercise generation prompt words are customized and generated based on each target knowledge point in the knowledge point summary information in advance, that is, these customized prompt words are input into the large model to automatically generate exercises closely related to each target knowledge point and their detailed answer analysis. Specifically, the large model is used to generate prompt words based on the exercise to automatically generate a preset number of matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information, so that learners can easily consolidate the knowledge they have learned through exercise questions after class, without having to find exercise resources by themselves, thereby saving a lot of time and improving the learning efficiency of learners. In addition, learners can generate exercises at any time according to their needs through the large language model during the process of playing the educational recording video for learning, which is used for comparative learning of knowledge points that are not understood at the moment.

[0060] For example, the prompt word template for generating exercises based on subject knowledge is as follows:

[0061] Please help generate after-class exercises based on the knowledge point information. The exercise type is xxx, and xx questions are generated for each type:

[0062] Knowledge point information: %s;

[0063] Output format: "Question", "Answer", "Analysis".

[0064] For example, see Figure 2As shown, when the big model receives an exercise generation prompt about "characteristics of a cuboid", the big model generates corresponding exercises based on the prompt. A set of exercise generation templates is preset in the big model based on subject knowledge. The template covers exercises of different types and difficulties, and can provide personalized exercise resources for different learners, thereby meeting the diverse learning needs of different learners.

[0065] In this embodiment, after the exercises closely related to the knowledge points and their detailed answer analysis are generated by the large model, the identification number corresponding to the educational video recording, the knowledge point summary information and the matching exercises related to each target knowledge point and their answer analysis can be stored in the pre-built exercise database. For example, the video ID, video knowledge point, knowledge point segment start time, knowledge point segment end time, exercises, answers, analysis and other information are stored in the pre-built exercise database.

[0066] Among them, the supporting exercises and answer analyses related to each target knowledge point in the exercise database are associated with the corresponding knowledge point explanation segments in the educational recorded video based on the corresponding explanation time periods of each target knowledge point in the educational recorded video. In other words, by utilizing the timeline information of the subtitles, the exercises can be associated with the original knowledge point explanation segments, which is convenient for jumping, and the exercises are closely related to the answer analyses. When encountering wrong questions or having doubts about the analyses, you can quickly jump to the original video knowledge point explanation segments for review and study, which helps learners consolidate what they have learned and improve their problem-solving abilities, thereby greatly improving the efficiency and quality of learning, and creating a more efficient and personalized online learning environment for learners.

[0067] It can be seen that in the embodiment of the present invention, through the big model technology and subtitle information analysis, matching exercises and answer analysis corresponding to each knowledge point in the video are automatically generated for the educational recorded video, that is, the subtitle information in the educational recorded video is extracted, and then the subtitle information is parsed by the big model based on the subtitle parsing prompt words, and the knowledge point summary information is sorted out and summarized, and finally, the big model is used to generate matching exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words, thereby providing personalized exercise resources for different learners, making it convenient for learners to directly use the generated matching exercises for practice after class without having to find exercise resources by themselves, thereby saving a lot of time and improving learners' learning efficiency.

[0068] See also Figure 3 As shown, the embodiment of the present invention discloses a specific method for generating exercises based on a large model. Compared with the previous embodiment, this embodiment further illustrates and optimizes the technical solution.

[0069] Step S21, receiving an educational recorded video, and analyzing the educational recorded video to determine whether corresponding subtitle information exists in the educational recorded video.

[0070] In this embodiment, before extracting the subtitle information, the received educational recorded video may be parsed to determine whether there is corresponding subtitle information in the video.

[0071] Step S22: When the subtitle information exists in the educational recorded video, extract the subtitle information in the educational recorded video.

[0072] In this embodiment, when subtitle information exists in the educational recorded video, the subtitle information in the educational recorded video is directly extracted.

[0073] Step S23: inputting the subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information.

[0074] Step S24, inputting the exercise generation prompt words and the knowledge point summary information into the macro model, so as to utilize the macro model to automatically generate matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

[0075] For the specific contents of the above steps S23 to S24, reference may be made to the corresponding contents disclosed in the above embodiments, which will not be repeated here.

[0076] It can be seen that in the embodiment of the present invention, through the big model technology and subtitle information analysis, matching exercises and answer analysis corresponding to each knowledge point in the video are automatically generated for the educational recorded video, that is, the subtitle information in the educational recorded video is extracted, and then the subtitle information is parsed by the big model based on the subtitle parsing prompt words, and the knowledge point summary information is sorted out and summarized, and finally, the big model is used to generate matching exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words, thereby providing personalized exercise resources for different learners, making it convenient for learners to directly use the generated matching exercises for practice after class without having to find exercise resources by themselves, thereby saving a lot of time and improving learners' learning efficiency.

[0077] See also Figure 4 As shown, the embodiment of the present invention discloses a specific method for generating exercises based on a large model. Compared with the previous embodiment, this embodiment further illustrates and optimizes the technical solution.

[0078] Step S31: Receive the educational recorded video and parse the educational recorded video to determine whether there is corresponding subtitle information in the educational recorded video.

[0079] Step S32: When the subtitle information does not exist in the educational recorded video, the audio information in the educational recorded video is extracted, and the audio information is transcribed to obtain the subtitle information in the educational recorded video.

[0080] In this embodiment, when the educational recorded video is analyzed to determine that there is no subtitle information in the video, the audio information in the educational recorded video is directly extracted, and then the audio information is transcribed to obtain corresponding subtitle information. When extracting the audio information in the video, the FFmpeg tool can be used to extract the audio in the video, that is, the FFmpeg tool is used to extract the audio information in the educational recorded video, and then the ASR (Automatic Speech Recognition) tool is used to transcribe the audio information to obtain the subtitle information in the educational recorded video.

[0081] Step S33: input the subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information.

[0082] Step S34, inputting the exercise generation prompt words and the knowledge point summary information into the macro model, so as to utilize the macro model to automatically generate matching exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

[0083] For the specific contents of the above-mentioned step S31 and steps S33 to S34, reference may be made to the corresponding contents disclosed in the aforementioned embodiments, which will not be repeated here.

[0084] It can be seen that in the embodiment of the present invention, through the big model technology and subtitle information analysis, matching exercises and answer analysis corresponding to each knowledge point in the video are automatically generated for the educational recorded video, that is, the subtitle information in the educational recorded video is extracted, and then the subtitle information is parsed by the big model based on the subtitle parsing prompt words, and the knowledge point summary information is sorted out and summarized, and finally, the big model is used to generate matching exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words, thereby providing personalized exercise resources for different learners, making it convenient for learners to directly use the generated matching exercises for practice after class without having to find exercise resources by themselves, thereby saving a lot of time and improving learners' learning efficiency.

[0085] For example, see Figure 5As shown, the video information of the input educational recorded video is parsed to determine whether there is a subtitle file in the video information of the educational recorded video. If there is a subtitle file in the video information of the educational recorded video, the subtitle file is extracted, and the large model is directly used to parse the video subtitle information in the subtitle file, and the knowledge point summary information is sorted out, wherein the subtitle file is a file in the SRT format; if there is no subtitle file in the video information of the educational recorded video, FFmpeg can be used to extract the audio file in the educational recorded video, and then the audio file is processed by ASR technology to extract the corresponding text to obtain the subtitle file, for example, AsrTool is used to process the audio file in the video to transcribe the audio therein into text to obtain the corresponding subtitle file, and the format in the subtitle file is set to SRT format, wherein the subtitle file in the SRT format has the start time and end time corresponding to the recorded text, for example, see Figure 6 The subtitle information of the graphic explanation course in the first grade mathematics shown in the figure is partial, and the subtitle information contains multiple lines, and the information of each line is: "start time", "end time", "duration", "text", after extracting the subtitle information from the educational recording video, the extracted subtitle information is systematically summarized and organized using the large model, and each knowledge point in the video is marked. At the same time, the specific time period of each knowledge point explanation is recorded so that subsequent learners can quickly locate the explanation segment of each knowledge point in the video. Finally, the large model automatically generates a specified number of supporting exercises and their answer analysis related to each target knowledge point in the knowledge point summary information according to the input exercise generation prompt words, and stores the video id, video knowledge point, knowledge point segment start time, knowledge point segment end time, exercises, answers, analysis and other information in the system library. It should be pointed out that in the technical solution of the application, the generative adversarial network (GAN) can also be used to replace the large model, so that the generated exercises can have higher error correction and verification capabilities.

[0086] In one embodiment, if Figure 7 As shown, based on the above-mentioned large model-based exercise generation method, the present invention also provides a large model-based exercise generation device, including:

[0087] The video receiving module 11 is used to receive educational recorded videos;

[0088] A subtitle extraction module 12, used to extract subtitle information from the educational recorded video;

[0089] A subtitle parsing module 13, used for inputting subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information;

[0090] The exercise generation module 14 is used to input the exercise generation prompt words and the knowledge point summary information into the large model, so as to use the large model to automatically generate supporting exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

[0091] Figure 8 A schematic diagram of the structure of a terminal provided in an embodiment of the present application. The terminal may include:

[0092] A memory 501 , a processor 502 , and a computer program stored in the memory 501 and executable on the processor 502 .

[0093] When the processor 502 executes the program, the large model-based exercise generation method provided in the above embodiment is implemented.

[0094] Furthermore, the terminal further includes:

[0095] The communication interface 503 is used for communication between the memory 501 and the processor 502 .

[0096] The memory 501 is used to store computer programs that can be executed on the processor 502 .

[0097] The memory 501 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0098] If the memory 501, the processor 502 and the communication interface 503 are implemented independently, the communication interface 503, the memory 501 and the processor 502 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0099] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 503 are integrated on a chip, the memory 501, the processor 502 and the communication interface 503 can communicate with each other through an internal interface.

[0100] The processor 502 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0101] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned large model-based exercise generation method is implemented.

[0102] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in this application. The specification and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the claims.

[0103] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0104] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, which can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor or other system that can read instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with these instruction execution systems, apparatuses or devices.

[0105] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA, Programmable Gate Array), a field programmable gate array (FPGA, Field-Programmable Gate Array), etc.

[0106] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for generating exercises based on a large model, characterized in that: The method comprises: Receiving an educational recorded video and extracting subtitle information from the educational recorded video; Inputting the subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information; The exercise generation prompt words and the knowledge point summary information are input into the large model, so as to utilize the large model to automatically generate matching exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

2. The method for generating exercises based on a large model according to claim 1, characterized in that: The extracting of subtitle information from the educational recorded video includes: Parsing the educational recorded video to determine whether there is corresponding subtitle information in the educational recorded video; When the subtitle information exists in the educational recorded video, extracting the subtitle information in the educational recorded video; Or, when the subtitle information does not exist in the educational recorded video, the audio information in the educational recorded video is extracted, and the audio information is transcribed to obtain the subtitle information in the educational recorded video.

3. The method for generating exercises based on a large model according to claim 2, characterized in that: The extracting of audio information from the educational recording comprises: The audio information in the educational recording is extracted using the FFmpeg tool.

4. The method for generating exercises based on a large model according to claim 2, characterized in that: The step of transcribing the audio information to obtain the subtitle information in the educational video recording includes: The audio information is transcribed using an ASR tool to obtain the subtitle information in the educational recorded video.

5. The method for generating exercises based on a large model according to claim 1, characterized in that: The method of automatically generating matching exercises and answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words by using the large model includes: The large model is used to generate prompt words based on the exercises, and a preset number of supporting exercises and answer analysis related to each target knowledge point in the knowledge point summary information are automatically generated.

6. The method for generating exercises based on a large model according to any one of claims 1 to 5, characterized in that: The knowledge point summary information includes each target knowledge point and its introduction and the explanation time period corresponding to each target knowledge point.

7. The method for generating exercises based on a large model according to claim 6, characterized in that: After inputting the exercise generation prompt words and the knowledge point summary information into the macro model, and using the macro model to automatically generate matching exercises related to each target knowledge point in the knowledge point summary information and their answer analysis based on the exercise generation prompt words, the method further includes: The identification number corresponding to the educational video recording, the knowledge point summary information and the matching exercises related to each target knowledge point and the answer analysis thereof are stored in a pre-constructed exercise database; Among them, the supporting exercises related to each target knowledge point in the exercise database and the answer analysis are associated with the corresponding knowledge point explanation segment in the educational recorded video based on the explanation time period corresponding to each target knowledge point in the educational recorded video.

8. A large model-based exercise generation device, characterized in that: The device comprises: Video receiving module, used for receiving educational recorded videos; A subtitle extraction module, used to extract subtitle information from the educational recorded video; A subtitle parsing module, used for inputting subtitle parsing prompt words and the subtitle information into the large model, so that the large model parses the subtitle information based on the subtitle parsing prompt words to obtain corresponding knowledge point summary information; The exercise generation module is used to input the exercise generation prompt words and the knowledge point summary information into the large model, so as to use the large model to automatically generate supporting exercises and their answer analysis related to each target knowledge point in the knowledge point summary information based on the exercise generation prompt words.

9. A terminal, characterized in that: include: A memory, a processor, and a large model-based exercise generation program stored in the memory and executable on the processor, wherein the large model-based exercise generation program, when executed by the processor, implements the steps of the large model-based exercise generation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which can be executed to implement the steps of the large model-based exercise generation method as described in any one of claims 1 to 7.

Images