Homework correction method and system based on AI

Abstract

The invention discloses an AI-based homework correction method and system. The method comprises the following steps: acquiring a plurality of jobs under each cluster, and acquiring the job content of each job according to the job ID of each job in the plurality of jobs; and packaging an AI job correction parameter based on each cluster, and performing batch correction on the job content of each job by using an AI large model based on the AI job correction parameter. The technical problems of high homework correction cost and low accuracy in the prior art are solved.

Description

Technical Field

[0001] The present invention relates to the field of AI (artificial intelligence), and in particular to an AI-based homework grading method and system. Background Art

[0002] With the popularity of programming education, more and more schools and educational institutions are using programming languages ​​such as Python as one of their teaching contents. However, the problem that follows is how to efficiently and accurately grade a large number of programming assignments. Traditional Python assignment grading methods mostly rely on tools such as regular expressions. Although most of the assignment content can be graded, about 30% of the assignments cannot be accurately graded using regular expressions due to complex assignment logic and diverse programming styles. This results in a low accuracy rate for grading assignments, and requires a lot of manpower to handle some assignments that cannot be automatically graded, increasing the work cost and time consumption of educational institutions.

[0003] In addition, manual grading is not only time-consuming and labor-intensive, but also difficult to maintain consistent grading standards. Therefore, the existing technology urgently needs a new solution that can automatically handle more complex programming homework grading scenarios, improve grading accuracy, and reduce reliance on human resources.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide an AI-based homework grading method and system to at least solve the technical problems of high cost and low accuracy of homework grading in the prior art.

[0006] According to one aspect of an embodiment of the present invention, there is provided an AI-based homework grading method, comprising: obtaining multiple homeworks under each cluster, and obtaining the homework content of each homework according to the job ID of each homework in the multiple homeworks; encapsulating AI homework correction parameters based on each cluster, and batch-grading the homework content of each homework using an AI big model based on the AI ​​homework correction parameters.

[0007] According to another aspect of an embodiment of the present invention, there is also provided an AI-based homework grading system, including: a CRM system, configured to obtain multiple homeworks under each cluster, and obtain the homework content of each homework according to the job ID of each homework in the multiple homeworks; an AI big model, configured to encapsulate AI homework correction parameters based on each cluster, and use the AI ​​big model to batch correct the homework content of each homework based on the AI ​​homework correction parameters.

[0008] In the embodiment of the present invention, multiple jobs under each cluster are obtained, and the job content of each job is obtained according to the job ID of each job in the multiple jobs; AI job grading parameters are encapsulated based on each cluster, and the AI ​​big model is used to batch grade the job content of each job based on the AI ​​job grading parameters. The above solution solves the technical problems of high cost and low accuracy of grading jobs in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0010] Figure 1 is a flow chart of an AI-based homework grading method according to an embodiment of the present invention;

[0011] Figure 2 is a flow chart of another AI-based homework grading method according to an embodiment of the present invention;

[0012] Figure 3 is an architecture diagram of an AI-based homework grading system according to an embodiment of the present invention;

[0013] Figure 4 is a flow chart of a method for correcting a large AI model according to an embodiment of the present invention;

[0014] Figure 5 A schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0016] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] According to an embodiment of the present invention, a method embodiment of an AI-based homework grading method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0018] Figure 1 is an AI-based homework grading method according to an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0019] Step S102: obtaining multiple jobs under each cluster, and obtaining the job content of each job in the multiple jobs according to the job ID of each job.

[0020] Step S104, encapsulating AI homework grading parameters based on each cluster, and batch grading the homework content of each homework using the AI ​​big model based on the AI ​​homework grading parameters.

[0021] First, AI homework correction parameters are encapsulated based on each cluster.

[0022] Next, based on the AI ​​homework correction parameters, the AI ​​big model is used to batch correct the homework content of each homework.

[0023] Based on the AI ​​job correction parameters and the job content of each job, the architecture AI job correction interface is called in batches; the dialogue interface of the AI ​​big model is called through the architecture AI job correction interface, and the AI ​​big model is used to correct the job content of each job based on the AI ​​job correction parameters. For example, the dialogue interface of the AI ​​big model is called, and the job content of each job and the AI ​​job correction parameters are input to the AI ​​big model through the dialogue interface; the AI ​​big model parses and corrects each job based on the AI ​​job correction parameters, and returns the job correction result.

[0024] After returning the job correction result, associate the job correction result with its corresponding job ID to generate correction feedback data with the job ID; store the correction feedback data in a preset job correction database, and update the correction status of the corresponding job based on the job ID and the architecture message ID, wherein the correction status includes initiated, failed to initiate, or correction completed.

[0025] Based on the correction results in the correction feedback data, a student feedback report is generated, wherein the student feedback report includes error point markings, correct answer suggestions and homework scores; and the student feedback report is pushed to the corresponding student account.

[0026] Finally, all assignments are summarized and analyzed based on the correction feedback data and the student feedback report; based on the summary analysis, the correction accuracy is calculated and the AI ​​assignment correction parameters are optimized.

[0027] Figure 2 Another AI-based homework grading method according to an embodiment of the present invention is applied to an AI-based homework grading system, such as Figure 3 As shown, the system includes a CRM system 32 , a tool system 34 , an architecture system 36 and an AI big model 38 .

[0028] The CRM system in the present invention is a system for processing homework data and grading process management. It obtains Python homework submitted by students from the homework library, manages batch processing of homework data, coordinates the interaction of various systems, and is responsible for saving the grading results to the database after the homework is grading. This system is the core scheduler of the grading process, and ensures the orderly progress of grading tasks through interface interaction with the tool system, architecture system, and Baidu AI large model system.

[0029] The tool system is an intermediate system used to process the interaction between assignment requests and assignment data. Its main function is to obtain requests from the CRM system, extract and return the specific content of the assignment. Specifically, the tool system receives requests from the CRM system, and based on the passed assignment ID, extracts the content data of each assignment by accessing the student assignment database or other storage system. The tool system needs to ensure the accuracy and completeness of the assignment data to avoid data loss or erroneous transmission. After successfully obtaining the assignment content, the assignment data is packaged and sent back to the CRM system for subsequent grading process. In order to improve processing efficiency, the tool system usually performs batch processing of assignment data, such as processing 100 assignments at a time, to reduce system load.

[0030] The architecture system is used to connect the tool system and the AI ​​big model. It is responsible for scheduling the core logic of AI job correction and managing the interface calls with the AI ​​model. Specifically, after the CRM system encapsulates the AI ​​job correction parameters, it will pass these parameters to the architecture system. The architecture system is responsible for parsing the correction request and calling the AI ​​big model for job correction according to conditions such as job type and clustering. In addition, by optimizing the concurrent processing of correction requests, the architecture system can call the interface of the AI ​​big model in batches to improve the overall efficiency of job correction. In addition, the architecture system can also provide the ability of request queuing and task priority scheduling to ensure that the system responds efficiently to large-scale job correction needs. After receiving the correction results from the AI ​​big model, the architecture system packages the results into a specific format and returns them to the CRM system. At the same time, it will return the message ID list and job ID related to the job for subsequent data update and storage. Through the centralized management of the architecture system, the AI ​​call process of job correction is more standardized, which greatly improves the scalability and flexibility of the system and supports diversified needs in different scenarios.

[0031] The AI ​​big model is the core technology for realizing intelligent homework grading in the present invention. The AI ​​big model provides natural language processing and homework content analysis capabilities. The model can process a large number of Python jobs in batches, especially solving the complex homework content that cannot be processed by regular expressions in traditional homework grading. The AI ​​big model understands and parses the code content and comments in Python jobs, and identifies the correctness, efficiency, and standardization of the code logic. Especially in the homework grading involving complex algorithms and data structures, the AI ​​big model can perform deep semantic analysis to determine whether the functions and running results of the code meet the requirements. In addition, based on the homework content and correction parameters, the AI ​​big model can not only identify errors, but also give specific modification suggestions and feedback. Through natural language generation technology, the model can generate detailed and targeted correction feedback to help students better understand errors and improve homework. In order to meet the needs of large-scale homework grading, the AI ​​big model supports parallel processing and can process multiple job requests at one time, greatly improving the processing power and response speed of the system. The AI ​​big model accepts correction requests through the interface of the architecture system, parses the homework content, and returns the results to the architecture system after completing the correction task. The output of the AI ​​big model includes specific information such as correction results, scores, error prompts, and modification suggestions. Through the powerful processing capabilities of the AI ​​big model, the accuracy and efficiency of homework grading have been significantly improved, solving the problems of high cost, low efficiency and insufficient accuracy in the traditional homework grading process.

[0032] like Figure 2 As shown, the AI-based homework grading method includes the following steps:

[0033] Step 1: The CRM system obtains job data.

[0034] In the entire process of homework grading, the CRM system assumes the function of homework data management. Every day from midnight to 23:59:59, the CRM system automatically obtains the homework of each cluster from the homework library. Clustering is a classification mechanism based on the homework submitted by students, which can be divided by homework type, difficulty, subject or other specific dimensions. In this embodiment, the division is mainly based on the homework type.

[0035] Step 2: Process in batches.

[0036] In order to improve the efficiency of review and correction, the CRM system processes the jobs in batches of 100. Each batch of jobs is identified by a corresponding job ID. The core of batch management is to reduce the pressure of large-scale data transmission and ensure that the system can maintain stability and efficiency during the processing process.

[0037] Step 3: The CRM system requests the tool system to obtain the job content.

[0038] Once the CRM system determines the job ID to be reviewed, it will send a request to the tool system through the API interface to obtain the specific job content.

[0039] Step 4, return to the job content.

[0040] The tool system queries the homework library by the homework ID, finds the corresponding homework content, and returns it to the CRM system. The homework content usually includes Python code and comments submitted by students.

[0041] In order to ensure the security and integrity of data transmission, in some embodiments, an encrypted transmission protocol based on SSL / TLS can be used. At the same time, the system can also adopt a cache mechanism. For the acquired job content, the tool system will cache it in memory for a short period of time to avoid repeated requests and further improve the response speed.

[0042] Step 5: Encapsulate the AI ​​job correction parameters and call the AI ​​correction interface of the architecture system in batches.

[0043] After obtaining the homework content, the CRM system generates the corresponding AI homework correction parameters according to the characteristics of each cluster. The encapsulation of homework correction parameters is the core of the entire process, because these parameters determine the strategies and rules used by the AI ​​large model when grading. Specific correction parameters may include the following aspects: 1) Homework type: for example, basic Python syntax exercises or complex algorithm design. 2) Grading criteria: Different types of homework may have different scoring criteria, such as code correctness, readability, completeness of comments, etc. 3) Error feedback: The system can configure some common programming error types in advance for reference by the AI ​​correction model. 4) Other auxiliary information: such as students' historical homework grades, homework submission time, etc., AI can make more personalized corrections and feedback based on this. By encapsulating these correction parameters, the system can ensure that the AI ​​model can flexibly adjust the correction strategy according to the actual situation of the homework, thereby improving the accuracy and efficiency of the correction.

[0044] After the CRM system completes the encapsulation of the correction parameters, it will call the AI ​​job correction interface of the architecture system in batches. The architecture system assumes the function of the middle layer, which is responsible for decoupling the upper-level business logic from the underlying AI model. Here, the architecture system is not just a simple API transfer station, but also used for the following:

[0045] 1) Request verification and processing: The architecture system will first verify the request data sent by the CRM system to ensure the integrity and correctness of the parameters. If an abnormality is found, the architecture system will immediately return an error prompt and record a log for subsequent investigation.

[0046] 2) Batch processing and scheduling: Since the correction tasks are performed in batches, the architecture system will reasonably arrange the processing order of the batches according to the current system load to avoid performance bottlenecks caused by processing too many jobs at one time.

[0047] 3) Call the AI ​​big model: When everything is ready, the architecture system will send the correction parameters and job content to the dialogue interface of the AI ​​big model through the encapsulated API interface.

[0048] Step 6: Call the conversational interface of the AI ​​big model for correction.

[0049] The AI ​​big model has the ability to process natural language and understand code. In the present invention, the AI ​​big model is mainly responsible for analyzing and correcting Python jobs. Specifically, the correction process of the AI ​​big model can be as follows: Figure 4 The following steps are shown:

[0050] Step S402: code analysis.

[0051] The AI ​​big model parses the submitted Python code. The main goal of code parsing is to understand the basic structure of the code, check the syntax, and identify potential logical errors. It mainly includes the following steps:

[0052] 1) Syntax analysis.

[0053] After receiving the Python code submitted by the job, the AI ​​model automatically performs syntax analysis, similar to the role of a compiler. The AI ​​model recognizes Python's key syntax structures, including variable definitions, function declarations, loops, conditional judgments, etc. It will parse the code line by line to ensure that the code complies with Python's syntax specifications, such as whether the brackets match and whether the indentation is correct.

[0054] The syntax analysis of the AI ​​large model of this application is different from the traditional compiler syntax check in that it can not only identify conventional syntax errors, but also understand the code according to the context. For example, the model can identify cross-function or cross-file reference errors, which is usually one of the common problems in student assignments. Specifically, the first eigenvector of the context code and the second eigenvector of the current code to be detected are extracted respectively, and the first eigenvector and the second eigenvector are respectively subjected to the order-up processing according to the extended rank criterion, and the first eigenvector and the second eigenvector are subjected to the feature coupling and fusion processing using a multi-channel convolution kernel to obtain the coupled main embedded feature matrix and the auxiliary embedded feature matrix. The above matrix is ​​expanded along the direction of the extended dimension to form multiple sub-matrices, and each sub-matrix is ​​subjected to singular value decomposition to extract the core subspace representation and the factor matrix. The principal component difference between the main feature matrix and the auxiliary feature matrix is ​​calculated based on the core subspace representation and the factor matrix to extract the core features, thereby effectively identifying logical errors and reference problems.

[0055] 2) Syntax error prompts.

[0056] When a syntax error is found, the AI ​​model immediately gives detailed prompts. For example, if the model detects that the variable naming does not conform to the specification, or a function is missing a parameter, the AI ​​model points out the line of code where the error is located and provides corresponding modification suggestions. These prompts are particularly important for students or teachers when correcting homework, helping to quickly locate errors and make corrections.

[0057] The AI ​​big model provides specific grammar repair suggestions instead of just prompting errors. For example, if the indentation format in the code does not meet Python requirements, the model will automatically adjust and generate a code version with recommended modifications.

[0058] Step S404: code logic check.

[0059] After completing the syntax check, the AI ​​model further enters the code logic check phase, which includes the following steps:

[0060] 1) Control flow analysis.

[0061] The AI ​​big model determines whether the execution path of the code is reasonable by analyzing the control flow structures in the code, such as loops, conditional branches, etc. For example, in Python code, common control flow structures include if, for, while, etc. The AI ​​big model verifies whether these structures are executed in a logical order. If the model detects unnecessary nested loops, meaningless judgment conditions, or dead loop problems, it points out the specific code segment and gives suggestions for improvement.

[0062] The AI ​​big model performs cross-function analysis. When a Python job contains multiple function calls, the AI ​​big model tracks the calling order and return value of each function to ensure that the logical relationship between functions is correct. For example, if the return value of a function is not used appropriately, the model will point it out and suggest appropriate processing of the return value.

[0063] 2) Exception handling analysis.

[0064] Exception handling is often involved in Python homework, such as the try-except structure. In the code logic check, the AI ​​big model pays special attention to the use of exception handling. It not only detects whether the code has an exception handling structure, but also evaluates the completeness and effectiveness of the exception handling. For example, if the model detects that a try-except block is too broad and ignores key exception types, it will make modification suggestions to guide students on how to accurately capture exceptions and provide appropriate handling mechanisms.

[0065] 3) Variable usage check.

[0066] The definition and use of variables is another important part of code logic checking. The AI ​​big model analyzes the declaration, use, scope, and life cycle of variables to ensure the rationality of variable use. If unused variables, duplicate defined variables, or variable types do not match, the AI ​​big model will give corresponding error prompts. For example, if the model detects that an uninitialized variable is directly used for calculations, it will immediately mark the problem and recommend initializing the variable.

[0067] Step S406: code style check.

[0068] After checking syntax and logic, the AI ​​model further checks the style of the Python code. This is mainly based on Python programming best practices and industry standards, such as the PEP 8 specification. Specifically, the following steps are included:

[0069] 1) Naming conventions

[0070] The AI ​​big model automatically checks the naming of variables, functions, and classes in the code, determines whether the naming conforms to the standards of the Python community, and points out problems with unclear or inconsistent naming. For example, variable names should use the "snake_case" style, while class names should use the "CamelCase" style. If a variable name is too short, unclear, or conflicts with a reserved word, it is recommended to use a more descriptive name.

[0071] 2) Code indentation and formatting.

[0072] Python has strict requirements for code indentation, and indentation errors may cause the code to fail to execute. The AI ​​big model can not only detect indentation errors, but also check whether the format of the entire code file is consistent. For example, it detects blank lines between code blocks, spaces on both sides of operators, etc., which are in line with the recommendations of PEP 8. The AI ​​big model gives automatic formatting suggestions to help adjust the code to a standardized format.

[0073] 3) Redundant code and optimization.

[0074] During the code style check, the AI ​​big model is also used to identify redundant code snippets, such as repeated code logic, redundant conditional judgments, or unused functions. After discovering these redundant parts, the model automatically suggests simplifying them. For example, the AI ​​big model suggests merging two functions with similar functions, or using more efficient built-in functions to replace complex custom logic.

[0075] Step S408: code optimization suggestions.

[0076] After analyzing the structure, logic, and style of the code, the AI ​​big model generates optimization suggestions to help improve the execution efficiency and maintainability of the code. Specifically, the following steps are included:

[0077] 1) Algorithm optimization.

[0078] The AI ​​big model identifies whether there are performance bottlenecks in the code, especially in jobs involving large-scale data processing or complex algorithms. The model analyzes the time and space complexity of the code to determine whether there are more efficient algorithms that can be used instead. For example, if the model detects that an operation in a for loop can be implemented using Python's built-in functions, it recommends using a more efficient map or filter function to speed up execution.

[0079] 2) Library and framework recommendations.

[0080] The AI ​​big model automatically recommends the use of suitable Python libraries or frameworks based on the code content. For example, if a student writes a common data processing function by himself, the AI ​​big model recommends using mature libraries such as pandas or numpy to simplify code development and provide more powerful functions. At the same time, the model will also remind developers to avoid using outdated or no longer maintained third-party libraries to ensure the long-term maintainability of the code.

[0081] The grading process of this application provides all-round feedback to the submitter of the assignment from grammar, logic, style to optimization suggestions. The AI ​​big model solves complex logical errors that cannot be recognized in traditional grading tools through its powerful code parsing and natural language processing capabilities, especially those that regular expressions cannot handle. Through the application of the AI ​​big model, the present invention greatly improves the accuracy and efficiency of Python assignment grading, and greatly reduces the cost and time of manual grading.

[0082] Step 7: Return the correction result.

[0083] After the AI ​​model completes the grading, it will return the grading results to the architecture system through the interface. The grading results include not only the score and feedback of the student's assignment, but also the specific grading status of the assignment (for example, whether there are errors that prevent it from running). In addition to the above information, the grading results can also include: syntax error prompts and repair suggestions; code logic problems and optimization suggestions; code style check results; possible algorithm optimization and library recommendations.

[0084] The review results are sent back to the CRM system through the architecture system and are ultimately fed back to the submitter of the assignment. The submitter can modify and optimize the code based on this feedback.

[0085] Step 8, return the work-id and architecture message ID list.

[0086] The architecture system will further process the correction results, package the correction results, work_id, and architecture message id list, and send them to the CRM system.

[0087] Step 9: Data storage and status update.

[0088] After receiving the grading results, the CRM system will save them to the homework grading database (such as tbl_ai_work_review). The database records the grading status, score, feedback content, etc. of each homework, ensuring that the grading results can be tracked and reviewed later.

[0089] At the same time, the CRM system will update the AI ​​model’s review status based on the results returned by the architecture system. If the review is successful, the status will be set to “Initiated”; if the review fails, the status will be set to “Failed to initiate”. This ensures that the system can monitor the execution of the review task in real time, facilitating subsequent exception handling.

[0090] The present invention provides an efficient and intelligent Python homework grading method and system by combining the collaborative work of multiple systems and the powerful capabilities of the AI ​​big model. It not only solves 30% of complex homework problems that cannot be handled by traditional grading methods, but also greatly improves the accuracy and efficiency of grading. Through the present invention, educational institutions can reduce their dependence on human resources and improve the overall teaching quality and efficiency in the process of large-scale homework grading. At the same time, the decoupling design of the architecture system and the AI ​​big model also makes the system have good scalability, and it can be further applied to other programming languages ​​and teaching scenarios in the future.

[0091] It should be noted that the AI-based homework grading system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the AI-based homework grading system provided in the above embodiment and the AI-based homework grading method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0092] Figure 5 FIG. 1 shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present disclosure. It should be noted that: Figure 5 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0093] like Figure 5 As shown, the electronic device includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage part 1008 into a random access memory (RAM) 1003. In the RAM 1003, various programs and data required for system operation are also stored. The CPU 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0094] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage section 1008 as needed.

[0095] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A homework grading method based on AI, characterized in that: include: Acquire multiple jobs under each cluster, and acquire the job content of each job according to the job ID of each job in the multiple jobs; The AI ​​homework grading parameters are encapsulated based on each cluster, and the AI ​​big model is used to batch grade the homework content of each homework based on the AI ​​homework grading parameters.

2. The method according to claim 1, characterized in that Based on the AI ​​homework correction parameters, the AI ​​big model is used to batch correct the homework content of each homework, including: Based on the AI ​​job correction parameters and the job content of each job, batch call the architecture AI job correction interface; The dialogue interface of the AI ​​big model is called through the architecture AI homework correction interface, and the AI ​​big model is used to correct the homework content of each homework based on the AI ​​homework correction parameters.

3. The method according to claim 2, characterized in that Using the AI ​​big model to correct the homework content of each homework based on the AI ​​homework correction parameters includes: Calling the dialogue interface of the AI ​​big model, and inputting the homework content of each homework and the AI ​​homework correction parameters into the AI ​​big model through the dialogue interface; The AI ​​big model parses and grades each assignment based on the AI ​​assignment grading parameters, and returns the assignment grading results.

4. The method according to claim 3, characterized in that After returning the homework correction result, the method further includes: Associating the homework correction result with its corresponding homework ID to generate correction feedback data with the homework ID; The correction feedback data is stored in a preset job correction database, and the correction status of the corresponding job is updated based on the job ID and the architecture message ID, wherein the correction status includes initiated, failed to initiate, or correction completed.

5. The method according to claim 4, characterized in that After returning the homework correction result, the method further includes: Generate a student feedback report based on the correction results in the correction feedback data, wherein the student feedback report includes error point markings, correct answer suggestions and homework scores; The student feedback report is pushed to the corresponding student account.

6. The method according to claim 5, characterized in that After generating the student feedback report, the method further includes: Summarizing and analyzing all assignments based on the correction feedback data and the student feedback report; Based on the summary analysis, the grading accuracy is calculated and the AI ​​homework grading parameters are optimized.

7. An AI-based homework grading system, characterized in that: include: The CRM system is configured to obtain multiple jobs under each cluster, and obtain the job content of each job in the multiple jobs according to the job ID of each job; The AI ​​big model is configured to encapsulate AI homework correction parameters based on each cluster, and use the AI ​​big model to batch correct the homework content of each homework based on the AI ​​homework correction parameters.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

9. A computer device, characterized in that: include: Memory and processor, The memory stores a computer program; The processor is used to execute the computer program stored in the memory, and when the computer program is run, the processor is enabled to execute the method according to any one of claims 1 to 6.

10. A computer program product, comprising 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 6 are implemented.

Images