Class evaluation method and apparatus, device, and medium

Abstract

The present application relates to the technical field of teaching evaluation, and provides a class evaluation method and apparatus, a device, and a medium. The method comprises: acquiring class data and class objective information of a class to be evaluated; converting the class data into a text to be input; receiving an objective analysis instruction for said class; inputting said text and the class objective information into a preset model, so that the preset model processes said text and the class objective information according to the objective analysis instruction, to output an evaluation result of said class. By using the technical solution, the goal of executing a question-answer dialogue analysis task, a class interaction analysis task and a course standard implementation analysis task for classes can be achieved, without requiring in-person class attendance, thereby improving the class evaluation efficiency; additionally, according to the present solution, class evaluation is carried out on the basis of the class objective information, thereby ensuring the objectivity and accuracy of evaluation results.

Description

Classroom evaluation method, device, equipment and medium Technical Field

[0001] The present application belongs to the field of teaching evaluation technology, and specifically relates to a classroom evaluation method, device, equipment and medium. Background Art

[0002] During a teacher's lecture, classroom interactions (e.g., asking questions) can draw students' attention to key and difficult points in the textbook, deepening their understanding of the material. During classroom interactions, asking stimulating questions that are relevant to the teaching objectives is crucial. Therefore, classroom evaluation of teachers becomes crucial during classroom observations.

[0003] In related technologies, classroom evaluation of teachers often involves schools designing questionnaires or scales based on relevant theories before class, and organizing teachers or instructors who specialize in classroom observation to attend classes online or offline. Finally, the teachers or instructors who observe the classroom summarize, discuss and generalize the teacher's classroom interaction behavior based on the classroom data collected from the classroom context after class, and come up with suggestions for the current class.

[0004] However, the current classroom evaluation method is mainly through manual listening and evaluation by professional classroom observation teachers, which is time-consuming. At the same time, different observation teachers may have different views on the classroom interaction behavior of the same teacher. Therefore, when using relevant technologies for classroom evaluation, there are problems of low evaluation efficiency and highly subjective evaluation conclusions.

[0005] Summary of the Invention

[0006] The purpose of the embodiments of the present application is to provide a classroom evaluation method, device, equipment and medium, which can solve the problems of low evaluation efficiency and strong subjectivity of evaluation conclusions when using related technologies to evaluate classroom interactive behaviors. By converting the classroom data of the classroom to be evaluated into text to be input, and receiving the target analysis instructions of the classroom to be evaluated, the text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instructions, and outputs the evaluation results of the classroom to be evaluated. The purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the classroom can be achieved without manual listening, thereby improving the efficiency of classroom evaluation. At the same time, this scheme is a classroom evaluation based on classroom target information and the text to be input obtained by converting the classroom data, which ensures the objectivity and accuracy of the evaluation results.

[0007] In a first aspect, an embodiment of the present application provides a classroom evaluation method, the method comprising:

[0008] Obtaining classroom data and classroom target information of the class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information;

[0009] Converting the classroom data into text to be input;

[0010] Receive target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions;

[0011] The text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction.

[0012] This solution converts the classroom data of the class to be evaluated into text to be input, receives the target analysis instruction of the class to be evaluated, and inputs the text to be input and the class target information into a preset model, so that the preset model processes the text to be input and the class target information according to the target analysis instruction, and outputs the evaluation result of the class to be evaluated. It can achieve the purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the class without manual listening, thereby improving the efficiency of class evaluation. At the same time, this solution is a class evaluation based on class target information and the text to be input obtained by converting classroom data, which ensures the objectivity and accuracy of the evaluation results.

[0013] In one embodiment, the inputting of the text to be input and the classroom target information into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction and outputs an evaluation result of the classroom to be evaluated, includes:

[0014] When it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the question-answering dialogue analysis instruction, executes the question-answering dialogue analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the class to be evaluated under the question-answering dialogue analysis main task module;

[0015] When it is confirmed that the received target analysis instruction is a classroom interaction analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the classroom interaction analysis instruction, executes the classroom interaction analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the classroom to be evaluated under the classroom interaction analysis main task module;

[0016] When it is confirmed that the received target analysis instruction is a curriculum standard implementation analysis instruction, the text to be input and the curriculum standard information are input into the preset model, so that the preset model responds to the curriculum standard implementation analysis instruction, executes the curriculum standard implementation analysis main task module based on the text to be input and the curriculum standard information, and outputs the evaluation result of the class to be evaluated under the curriculum standard implementation analysis main task module.

[0017] The beneficial effect of this solution is that, according to the different types of target analysis instructions received, the text to be input and the classroom target information are input into the preset model, so that the preset model responds to different target analysis instructions, and executes the corresponding main task module based on the curriculum standard information or teaching target information in the classroom target information and the text to be input, and outputs the evaluation results of the classroom to be evaluated under the corresponding selected main task module, which can achieve the purpose of evaluating the classroom to be evaluated according to the evaluation tasks in different types of main task modules, improve the pertinence of the evaluation of classroom interactive behavior, and enhance the user experience of viewing the evaluation results.

[0018] In one embodiment, when it is confirmed that the received target analysis instruction is a question-answer dialogue analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the question-answer dialogue analysis instruction, executes the question-answer dialogue analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the class to be evaluated under the question-answer dialogue analysis main task module, including:

[0019] When it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, confirming that the preset model responds to the question-answering dialogue analysis instruction, so that the preset model is in a state of waiting to execute the target evaluation tasks under each type of evaluation dimension in the question-answering dialogue analysis main task module;

[0020] Determine target evaluation tasks;

[0021] The text to be input and the teaching objective information are input into a preset model, so that the question-answering dialogue analysis main task module performs the target evaluation task based on the text to be input and the teaching objective information, and outputs the evaluation result of the class to be evaluated.

[0022] The beneficial effect of this solution is that after confirming the receipt of the question-answer dialogue analysis instruction and confirming that the preset model responds to the instruction, the preset model is put into a state of waiting to execute the question-answer dialogue analysis main task module. At this time, the target evaluation tasks under each type of evaluation dimension in the question-answer dialogue analysis main task module are determined, so that the preset model executes the target evaluation task based on the text to be input and the teaching objective information, which can achieve the effect of analyzing the teacher's classroom questions based on different evaluation dimensions.

[0023] In one embodiment, the question-answering dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module;

[0024] The target evaluation task is determined, including:

[0025] Obtaining a first type of evaluation prompt instruction; the first type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the question effectiveness analysis; the evaluation dimension of the question effectiveness analysis includes evaluating the effectiveness of the teacher's questions in the class to be evaluated;

[0026] Determining, based on the first type of evaluation prompt instruction, that the first overall subtask in the overall evaluation subtask module and the first local subtask in the local evaluation subtask module together constitute a target evaluation task; the first overall subtask is used to perform an overall evaluation of the overall effectiveness of all questions asked by teachers in the class to be evaluated, and the first local subtask is used to perform a specific judgment on the effectiveness of each question asked by a teacher in the class to be evaluated;

[0027] and / or,

[0028] Obtaining a second type of evaluation prompt instruction; the second type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the knowledge objective; the evaluation dimension of the knowledge objective includes evaluating the relevance between the teacher's questions in the class to be evaluated and the teaching objectives;

[0029] According to the second type of evaluation prompt instruction, the second overall subtask in the overall evaluation subtask module and the second local subtask in the local evaluation subtask module are determined to jointly constitute a target evaluation task; the second overall subtask is used to perform an overall evaluation of the knowledge objectives preset in the class to be evaluated, and the second local subtask is used to perform a specific evaluation of the relevance between the teacher's questions in the class to be evaluated and the teaching objectives;

[0030] and / or,

[0031] Obtaining a third type of evaluation prompt instruction as the target evaluation task; the third type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to an evaluation dimension of question avoidance; the evaluation dimension of question avoidance includes finding teacher questions that can be avoided in the class to be evaluated;

[0032] According to the third type of evaluation prompt instruction, a first local subtask in the local evaluation subtask module and a second local subtask in the local evaluation subtask module are determined to constitute a target evaluation task together; the first local subtask is used to specifically judge the effectiveness of each teacher's question in the class to be evaluated, and the second local subtask is used to specifically evaluate the relevance between the teacher's question in the class to be evaluated and the teaching objectives;

[0033] and / or,

[0034] Obtaining a fourth type of evaluation prompt instruction as the target evaluation task; the fourth type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of question optimization; the evaluation dimension of question optimization includes finding teacher questions that can be optimized in the class to be evaluated;

[0035] According to the fourth type of evaluation prompt instructions, the third overall subtask in the overall evaluation subtask module and the third local subtask in the local evaluation subtask module together constitute the target evaluation task; the third overall subtask is used to make overall optimization suggestions for the teacher's questions in the classroom to be evaluated, and the third local subtask is used to optimize and rewrite the teacher's questions in the classroom to be evaluated.

[0036] The beneficial effect of this solution is that by simultaneously or selectively evaluating the content of classroom question-and-answer dialogues in the classroom to be evaluated based on different evaluation dimensions and different evaluation scopes, the teacher's questions in the classroom to be evaluated can be evaluated in different dimensions under the premise of using a unified model. This avoids the problem in related technologies that different evaluation dimensions need to be trained separately using classroom data, resulting in the isolation of training results of different evaluation dimensions. It can more effectively reduce a large amount of repetitive model training work, strengthen the correlation between training tasks of different evaluation dimensions, and further improve the objectivity of classroom evaluation.

[0037] In one embodiment, the question-answer dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module; during the supervised fine-tuning training of the question-answer dialogue analysis main task module by the preset model, the classroom text containing teaching objectives and teacher questions in the educational scene corpus is processed according to the overall evaluation subtask module and the local evaluation subtask module respectively to output an evaluation result for the classroom text.

[0038] The beneficial effect of this solution is that by setting the question-answer dialogue analysis main task module to include an overall evaluation subtask module and a local evaluation subtask module, and conducting supervised training on the overall evaluation subtask module and the local evaluation subtask module based on the educational scenario corpus, the purpose of conducting overall evaluation and specific evaluation of the teacher's question and answer questions can be achieved, thereby improving the accuracy of the evaluation results output by the preset model.

[0039] In one embodiment, the overall evaluation subtask module includes a first overall subtask, a second overall subtask, and a third overall subtask; the first overall subtask is used to conduct an overall evaluation of the overall effectiveness of all teacher questions in the classroom text during the supervised fine-tuning training process; the second overall subtask is used to conduct an overall evaluation of the preset knowledge objectives in the classroom text during the supervised fine-tuning training process; the third overall subtask is used to make overall optimization suggestions for teacher questions in the classroom text during the supervised fine-tuning training process; the local evaluation subtask module includes a first local subtask, a second local subtask, and a third local subtask; the first local subtask is used to conduct a specific judgment on the effectiveness of each teacher question in the classroom text during the supervised fine-tuning training process; the second local subtask is used to conduct a specific evaluation of the correlation between teacher questions in the classroom text and teaching objectives during the supervised fine-tuning training process; the third local subtask is used to optimize and rewrite teacher questions in the classroom text during the supervised fine-tuning training process.

[0040] The beneficial effect of this scheme is that by setting an overall evaluation subtask module including three overall subtasks, and setting a local evaluation subtask module including three local subtasks, the three overall subtasks and the three local subtasks, a total of six subtasks, are jointly supervised trained. During the training process, the first overall subtask and the first local subtask work together to output the evaluation results under the dimension of question effectiveness analysis, the second overall subtask and the second local subtask work together to output the evaluation results under the dimension of knowledge goals, the third overall subtask and the third local subtask work together to output the evaluation results under the dimension of question optimization, and at the same time, the first local subtask and the second local subtask work together to output the evaluation results under the dimension of question avoidance, that is, all subtasks jointly executed in the overall evaluation subtask module and the local evaluation subtask module can output the evaluation results of the teacher's questions in the evaluated classroom according to the four evaluation dimensions of question effectiveness analysis, knowledge goals, question avoidance and question optimization, thereby ensuring the objectivity of classroom evaluation and further improving the comprehensiveness of the analysis of teachers' classroom questions.

[0041] In one embodiment, obtaining classroom data and classroom objective information of the class to be evaluated includes:

[0042] Obtain classroom data of the class to be evaluated;

[0043] The class objective information that matches the class data is searched in a preset database, or the class data is input into a preset chapter matching model so that the chapter matching model outputs the class objective information that matches the class data.

[0044] The beneficial effect of this solution is that by querying the classroom objective information that matches the classroom data in a preset database, or inputting the classroom data into a preset chapter matching model so that the chapter matching model outputs the classroom objective information that matches the classroom data, the efficiency of determining the classroom objective information can be improved.

[0045] In one embodiment, converting the classroom data into text to be input includes:

[0046] Converting the classroom data into text to obtain classroom text;

[0047] extracting teacher question text from the classroom text;

[0048] The classroom text and the teacher's question text are used as texts to be input.

[0049] The beneficial effect of this solution is that the classroom text is obtained by converting the classroom data into text, the teacher's question text is extracted from the classroom text, and the classroom text and the teacher's question text are used as the text to be input, which is conducive to the subsequent specific analysis of the teacher's questions and the overall analysis combined with the context, thereby improving the comprehensiveness of the text to be input.

[0050] In one embodiment, extracting the teacher's question text from the classroom text includes:

[0051] Identify the dialogue characters with the most dialogue in the classroom text;

[0052] Determine the dialogue role with the most dialogues as the teacher role;

[0053] Extracting all the dialogue contents of the teacher role;

[0054] The teacher's question text is filtered out from the entire conversation content.

[0055] The beneficial effect of this solution is that by determining the dialogue role with the most dialogues in the classroom text as the teacher role and screening the teacher's question text in the dialogue content, the steps of extracting the teacher's question text can be simplified, and the efficiency and accuracy of extracting the teacher's question text can be improved.

[0056] In one embodiment, the step of filtering out the teacher's question text from the entire conversation content includes:

[0057] Identifying basic questions in the entire conversation content;

[0058] A denoising operation is performed on the basic question to obtain the teacher's question text.

[0059] The beneficial effect of this solution is that by identifying the basic questions in the entire conversation content of the teacher and performing a denoising operation on the basic questions, the teacher's question text is obtained, which can achieve the purpose of further refining the question text and avoid the influence of subsequent basic questions on the evaluation results of the class to be evaluated.

[0060] In a second aspect, an embodiment of the present application provides a classroom evaluation device, the device comprising:

[0061] A data acquisition module is used to acquire classroom data and classroom target information of the class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information;

[0062] A data conversion module, used for converting the classroom data into text to be input;

[0063] An instruction receiving module is used to receive target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions;

[0064] An evaluation module is used to input the text to be input and the classroom target information into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction.

[0065] This solution converts the classroom data of the class to be evaluated into text to be input, receives the target analysis instruction of the class to be evaluated, and inputs the text to be input and the class target information into a preset model, so that the preset model processes the text to be input and the class target information according to the target analysis instruction, and outputs the evaluation result of the class to be evaluated. It can achieve the purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the class without manual listening, thereby improving the efficiency of class evaluation. At the same time, this solution is a class evaluation based on class target information and the text to be input obtained by converting classroom data, which ensures the objectivity and accuracy of the evaluation results.

[0066] In a third aspect, an embodiment of the present application provides an electronic device comprising a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps of the method described in the first aspect.

[0067] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.

[0068] In a fifth aspect, an embodiment of the present application provides a chip, which includes a processor and a communication interface, the communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the method described in the first aspect.

[0069] In an embodiment of the present application, classroom data and classroom target information of a class to be evaluated are obtained; wherein, the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information; the classroom data is converted into a text to be input; a target analysis instruction for the class to be evaluated is received; the target analysis instruction includes a question-answer dialogue analysis instruction, a classroom interaction analysis instruction, and a curriculum standard implementation analysis instruction; the text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs an evaluation result of the class to be evaluated; wherein, the preset model uses educational scene corpus as a training sample, and uses the question-answer dialogue analysis main task module, the classroom interaction analysis main task module, and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answer dialogue analysis main task module is used to respond to the question-answer dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction. The above-mentioned classroom evaluation method can solve the problems of low evaluation efficiency and strong subjectivity of evaluation conclusions when relying on manual labor when evaluating classroom interactive behaviors using relevant technologies. By converting the classroom data of the classroom to be evaluated into text to be input, and receiving the target analysis instructions of the classroom to be evaluated, the text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instructions, and outputs the evaluation results of the classroom to be evaluated. This can achieve the purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the classroom without manual listening, thereby improving the efficiency of classroom evaluation. At the same time, this scheme is a classroom evaluation based on classroom target information and the text to be input obtained by converting classroom data, which ensures the objectivity and accuracy of the evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] FIG1 is a flow chart of a classroom evaluation method provided in an embodiment of the present application;

[0071] FIG2 is a flow chart of a classroom evaluation method provided in an embodiment of the present application;

[0072] FIG3 is a user interface provided in an embodiment of the present application;

[0073] FIG4 is a flowchart of a classroom question evaluation process provided by an embodiment of the present application;

[0074] FIG5 is a schematic diagram of the structure of a classroom evaluation device provided in an embodiment of the present application;

[0075] FIG6 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0076] In order to make the purpose, technical solutions and advantages of the present application clearer, the specific embodiments of the present application are described in detail below in conjunction with the accompanying drawings. It is understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. It should also be noted that, for ease of description, only parts related to the present application, not all of the contents, are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the accompanying drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0077] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0078] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0079] The following, in conjunction with the accompanying drawings, describes in detail the classroom evaluation method, device, equipment and medium provided in the embodiments of the present application through embodiments and their application scenarios.

[0080] In related technologies, classroom evaluations for teachers often involve schools designing questionnaires or scales based on relevant theories before class. They then organize teachers or instructors who specialize in classroom observation to observe the class online or offline. Finally, after class, the observing teachers or instructors, based on the classroom data collected from the classroom context, summarize and discuss the teacher's classroom interactions and draw recommendations for the current class. However, current classroom evaluation methods primarily rely on manual observations by professional classroom observers, which is time-consuming. Furthermore, different observers may have different perspectives on the same teacher's classroom interactions. Therefore, using related technologies for classroom evaluations can lead to low evaluation efficiency and highly subjective conclusions.

[0081] This solution converts the classroom data of the class to be evaluated into text to be input, receives the target analysis instruction of the class to be evaluated, and inputs the text to be input and the class target information into a preset model, so that the preset model processes the text to be input and the class target information according to the target analysis instruction, and outputs the evaluation result of the class to be evaluated. It can achieve the purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the class without manual listening, thereby improving the efficiency of class evaluation. At the same time, this solution is a class evaluation based on class target information and the text to be input obtained by converting classroom data, which ensures the objectivity and accuracy of the evaluation results.

[0082] FIG1 is a flow chart of a classroom evaluation method provided in an embodiment of the present application. As shown in FIG1 , the method specifically includes the following steps:

[0083] S101, obtaining classroom data and classroom target information of a class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information;

[0084] First, the usage scenario of this solution can be a scenario for analyzing and evaluating teachers' question-and-answer dialogues, classroom interactions, and the implementation of curriculum standards during the teaching process.

[0085] Based on the above usage scenarios, it can be understood that the executor of this application can be a server used to process classroom data and evaluate teachers' classroom interactive behaviors. The server may include servers, terminal devices (such as interactive tablets, computers, smart terminals, mobile terminals, etc.), etc.

[0086] Among them, classroom interactive behaviors can be the behavior of teachers asking questions, talking to students, and teaching courses during the teaching process.

[0087] A classroom to be evaluated refers to a classroom awaiting evaluation. It can be a classroom whose teacher's classroom interactions are being evaluated, or a classroom whose teaching content is being evaluated. The classroom data for a classroom to be evaluated can be recorded or live data of the classroom to be evaluated. The classroom data can be data from the entire class to be evaluated, or it can be data from selected classroom segments of the classroom to be evaluated.

[0088] The classroom data includes: audio data and / or video data. The audio data may include both teacher audio data and student audio data. The video data may include both teacher video data and student video data.

[0089] The classroom goal information includes teaching goal information and curriculum standard information. The teaching goal information can be the expected result of the teaching activities, the concretization of educational purposes, teaching goals and curriculum goals, and the requirements and standards that teachers must meet to complete their teaching tasks. In the classroom teaching of a certain subject, teachers need to determine detailed teaching goals based on the curriculum goals and teaching content in order to select teaching content and determine the teaching effect. Curriculum standard information can be a teaching guidance document that stipulates the nature of the course, course goals, content goals, and implementation suggestions of a certain subject. Compared with the teaching syllabus, the curriculum standard is more detailed and clear in terms of the basic concepts of the course, course goals, curriculum implementation suggestions, etc., and puts forward basic learning requirements for all students. For example: the new curriculum standards for nine-year compulsory education.

[0090] In one embodiment, classroom data of a class to be evaluated can be received through a data transmission interface, and classroom objective information of the current class can be obtained by reading classroom data pre-stored or uploaded by the teacher, wherein the classroom data includes audio data and / or video data.

[0091] In a feasible embodiment, obtaining classroom data and classroom objective information of the class to be evaluated includes:

[0092] Obtain classroom data of the class to be evaluated;

[0093] The class objective information that matches the class data is searched in a preset database, or the class data is input into a preset chapter matching model so that the chapter matching model outputs the class objective information that matches the class data.

[0094] Among them, the preset database can be a classroom database constructed based on pre-collected educational corpus, or it can be a database constructed based on classroom materials uploaded by users. The preset database can store classroom objective information that matches the classroom data of each class or classroom objective information that matches the classroom type of each class in the form of a list. The classroom type can be a type corresponding to the classroom content. For example: lecture class, demonstration class, practice class, experimental class and review class. The classroom objective information that matches the classroom data can be classroom objective information that corresponds to at least one of the classroom name, classroom type and subject direction in the classroom data. The chapter matching model can be a model for matching classroom objective information for classroom data.

[0095] In one embodiment, classroom data of a class to be evaluated can be obtained through a preset interface. Class objective information matching the class data can be searched in a preset database based on at least one of the class name, class type, and subject area in the class data. Alternatively, the class data can be input into a preset chapter matching model, which then determines and outputs class objective information matching the class data through methods such as keyword extraction and similarity matching.

[0096] In an embodiment of the present application, the efficiency of determining classroom objective information can be improved by querying classroom objective information that matches the classroom data in a preset database, or inputting the classroom data into a preset chapter matching model so that the chapter matching model outputs classroom objective information that matches the classroom data.

[0097] S102, converting the classroom data into text to be input;

[0098] Among them, the text to be input can be the text obtained after the classroom data is converted into text, or it can be the teacher's question text extracted from the classroom text, or it can be a text that includes both classroom text and teacher's question text. The classroom text can be the text obtained after the classroom data is converted into text. The text to be input can also include: teacher's statement sentences, teacher's question sentences, student's statement sentences and student's question sentences during teaching, etc. The beginning of each sentence text in the text to be input can include the speaker's role prefix. For example: Speaker A: "Classmates, what kind of thoughts and feelings does this passage express?"; Speaker B: "It expresses the author's homesickness." It can be understood that different speakers who express each sentence text can be distinguished according to the role prefix.

[0099] In one embodiment, speech recognition technology such as ASR (Automatic Speech Recognition) can be used to convert the audio data in the classroom data into text to be input.

[0100] In a feasible embodiment, converting the classroom data into text to be input includes:

[0101] Converting the classroom data into text to obtain classroom text;

[0102] extracting teacher question text from the classroom text;

[0103] The classroom text and the teacher's question text are used as texts to be input.

[0104] The classroom text may be a text obtained by converting the classroom data into text, and the teacher's question text may be a teacher's question sentence in the classroom text.

[0105] In one embodiment, the classroom data can be converted into a classroom text. The beginning of each sentence text in the classroom text may also include a speaker's role prefix, and the end of each sentence text may include punctuation marks. The title noun in each sentence text in the classroom text can be identified, and whether the role prefix of each sentence text is a teacher role can be determined based on the title noun, and then the sentence text with the role prefix as a teacher role and the ending punctuation mark of the sentence text as a question mark is determined as the teacher question text in the classroom text. For example: if the title noun in the sentence text is identified as "teacher", then the role prefix corresponding to the sentence text is determined to be student, and if the title noun in the sentence text is identified as "XX classmate", then the role prefix corresponding to the sentence text is determined to be teacher. The classroom text and the teacher question text are used as the text to be input.

[0106] In an embodiment of the present application, the classroom data is converted into a classroom text, the teacher's question text is extracted from the classroom text, and the classroom text and the teacher's question text are used as the text to be input, which is conducive to the subsequent specific analysis of the teacher's questions and the overall analysis combined with the context, thereby improving the comprehensiveness of the text to be input.

[0107] In a feasible embodiment, extracting the teacher's question text from the classroom text includes:

[0108] Identify the dialogue characters with the most dialogue in the classroom text;

[0109] Determine the dialogue role with the most dialogues as the teacher role;

[0110] Extracting all the dialogue contents of the teacher role;

[0111] The teacher's question text is filtered out from the entire conversation content.

[0112] In one embodiment, the dialogue character with the most conversations in the classroom text can be identified and counted based on the number of occurrences of the role prefix of each speaker in the classroom text. Since the teaching process primarily consists of teacher presentations and teacher questions, the dialogue character with the most conversations can be determined to be the teacher. The entire dialogue content of the teacher character is extracted based on the role prefix corresponding to the teacher character, and the teacher's questions in the dialogue content are filtered based on the punctuation marks at the end of each sentence in the dialogue content.

[0113] In the embodiment of the present application, by determining the dialogue role with the most conversations in the classroom text as the teacher role and filtering the teacher question text in the dialogue content, the steps of extracting the teacher question text can be simplified, and the efficiency and accuracy of extracting the teacher question text can be improved.

[0114] In a feasible embodiment, the step of filtering out the teacher's question text from the entire conversation content includes:

[0115] Identifying basic questions in the entire conversation content;

[0116] A denoising operation is performed on the basic question to obtain the teacher's question text.

[0117] The basic questions may be overly simple and / or repetitive questions, such as questions of the form "Is this right?" and "Is this right?" The denoising operation may be an operation of deleting the basic questions.

[0118] In one embodiment, basic questions in the entire conversation content can be identified based on the punctuation marks at the end of each sentence in the entire conversation content and keyword search. For example, words such as "right or not" and "is it" are used as keywords, and the entire conversation content is searched for sentence texts containing the search keywords. It is identified whether the ending punctuation of the searched sentence texts is a question mark, and the sentence texts ending with a question mark are determined as basic questions. The basic questions in the entire conversation content are deleted, and the sentences ending with a period are deleted to obtain the teacher's question text.

[0119] In the embodiment of the present application, by identifying the basic questions in the entire conversation content of the teacher and performing a denoising operation on the basic questions, the teacher's question text is obtained, which can achieve the purpose of further refining the question text and avoid the influence of subsequent basic questions on the evaluation results of the class to be evaluated.

[0120] S103, receiving target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions;

[0121] The target analysis instruction for the class to be evaluated may be an instruction for performing question-and-answer dialogue analysis, classroom interaction analysis, and curriculum standard implementation analysis on the classroom content of the class to be evaluated. The target analysis instruction includes at least one of a question-and-answer dialogue analysis instruction, a classroom interaction analysis instruction, and a curriculum standard implementation analysis instruction. The question-and-answer dialogue analysis instruction is used to analyze whether the question-and-answer dialogue between the teacher and the students during the course meets the teaching objectives. The classroom interaction analysis is used to analyze whether the classroom interaction between the teacher and the students during the course meets the teaching objectives or whether it promotes the achievement of the teaching objectives. The curriculum standard implementation analysis instruction is used to analyze whether the teaching content and curriculum plan in the class to be evaluated meet the curriculum standards prescribed by the Ministry of Education.

[0122] In one embodiment, a target analysis instruction for the class to be evaluated can be received through a human-computer interaction interface. For example, a target analysis instruction selected or directly input by a user for the class to be evaluated can be received through the human-computer interaction interface. A default target analysis instruction for a course of the class to be evaluated can also be read based on the course type of the class to be evaluated, which is not further limited here. The analysis task for the class to be evaluated can be determined based on the target analysis instruction.

[0123] S104, inputting the text to be input and the classroom target information into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks for joint training, and performs supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction.

[0124] Among them, the preset model can be a large language model that can evaluate the text to be input. The input limit of the preset model is greater than the average length of the text to be input, and the text to be input can be processed accordingly based on the classroom target information according to different target analysis instructions. The evaluation result can be the result obtained after the preset model processes the text to be input accordingly according to the preset evaluation dimension. The preset evaluation dimension includes the evaluation dimension of question effectiveness analysis, the evaluation dimension of knowledge objectives, the evaluation dimension of question avoidance, and the evaluation dimension of question optimization.

[0125] The preset model uses educational scene corpus as training samples, and uses question-answer dialogue analysis main task module, classroom interaction analysis main task module and curriculum standard implementation analysis main task module as the main tasks of joint training, and is obtained through supervised fine-tuning training. The educational scene corpus refers to corpus related to educational resources and teaching content (corpus refers to a collection of text resources with a certain quantity and scale), including classroom texts generated by actual offline classroom data, classroom texts generated by online recorded classes, texts containing teaching content and question-answer dialogues, teacher lesson plans, course design plan texts, curriculum standard texts, teaching objective texts, etc. Supervised fine-tuning training can be the use of data with real labels to train and adjust a pre-trained language model (LLM) to make it more suitable for training a specific task. When supervised fine-tuning is performed, the model weights will be adjusted according to the difference from the real labels. Through this fine-tuning process, the model can capture patterns and characteristics specific to a certain task in the labeled data, making the model more accurate and better adapted to a specific task.

[0126] The question-and-answer dialogue analysis main task module is used to respond to the question-and-answer dialogue analysis instruction and analyze whether the teacher's question-and-answer dialogue in the text to be input meets the teaching objective information. The classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction and analyze whether the teacher's interactive behavior in the text to be input meets the teaching objective information. The curriculum standards implementation analysis main task module is used to respond to the curriculum standards implementation analysis instruction and analyze whether the course lectures in the text to be input meet the curriculum standards information stipulated by the Ministry of Education.

[0127] In one embodiment, the preset model can determine the analysis task for the text to be input according to the target analysis instruction. The text to be input and the classroom target information are input into the preset model, so that the preset model processes the text to be input according to the analysis task based on the classroom target information, and outputs the evaluation result of the classroom to be evaluated according to the preset evaluation dimension.

[0128] In a feasible embodiment, the question-answering dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module; during the supervised fine-tuning training of the question-answering dialogue analysis main task module by the preset model, the classroom text containing teaching objectives and teacher questions in the educational scene corpus is processed according to the overall evaluation subtask module and the local evaluation subtask module respectively to output the evaluation results for the classroom text.

[0129] Among them, the overall evaluation subtask module can be a module that focuses on the experience of the entire class, takes into account the interaction between teachers and students and the context, and conducts an overall evaluation of the teacher's questions in the text to be input. In addition to the teacher's questions, the input of the overall evaluation subtask module can also include the content of the classroom teaching to be evaluated and the students' responses. Its input can also include the classroom text of the entire class. The local evaluation subtask module can be a module for evaluating the effectiveness of the current teacher's questions in the classroom. The local evaluation subtask module does not rely heavily on the context, and can only use the teacher's questions in the classroom as input data, and can be a module that specifically evaluates the teacher's questions according to different preset evaluation dimensions.

[0130] In one embodiment, the question-answering dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module. During the supervised fine-tuning training of the preset model for the question-answering dialogue analysis main task module, the classroom text containing teaching objectives and teacher questions in the educational scene corpus can be processed according to the overall evaluation subtask module and the local evaluation subtask module respectively. The overall evaluation subtask module can evaluate whether the teacher's questions meet the teaching objectives based on the context of the teacher's questions in the educational scene corpus. The local evaluation can evaluate whether the teacher's questions meet the teaching objectives for a single question asked by the current teacher, so as to output an evaluation result for the classroom text.

[0131] In an embodiment of the present application, a question-and-answer dialogue analysis main task module is set to include an overall evaluation subtask module and a local evaluation subtask module, and supervised training is performed on the overall evaluation subtask module and the local evaluation subtask module based on educational scenario corpus, thereby achieving the purpose of conducting overall evaluation and specific evaluation of the teacher's question and answer questions respectively, thereby improving the accuracy of the evaluation results output by the preset model.

[0132] In a feasible embodiment, the overall evaluation subtask module includes a first overall subtask, a second overall subtask, and a third overall subtask; the first overall subtask is used to conduct an overall evaluation of the overall effectiveness of all teacher questions in the classroom text during the supervised fine-tuning training process; the second overall subtask is used to conduct an overall evaluation of the preset knowledge objectives in the classroom text during the supervised fine-tuning training process; the third overall subtask is used to make overall optimization suggestions for teacher questions in the classroom text during the supervised fine-tuning training process; the local evaluation subtask module includes a first local subtask, a second local subtask, and a third local subtask; the first local subtask is used to conduct a specific judgment on the effectiveness of each teacher question in the classroom text during the supervised fine-tuning training process; the second local subtask is used to conduct a specific evaluation of the correlation between teacher questions in the classroom text and teaching objectives during the supervised fine-tuning training process; the third local subtask is used to optimize and rewrite teacher questions in the classroom text during the supervised fine-tuning training process.

[0133] Among them, the overall evaluation subtask module includes a first overall subtask, a second overall subtask, and a third overall subtask. The first overall subtask is used to conduct an overall evaluation of the overall effectiveness of all teacher questions in the classroom text during the supervised fine-tuning training process. The overall evaluation of the overall effectiveness focuses on whether the teacher's questions can guide students to think. The second overall subtask is used to conduct an overall evaluation of the preset knowledge goals in the classroom text during the supervised fine-tuning training process. The overall evaluation of the preset knowledge goals focuses on the relevance between teacher questions and knowledge. The third overall subtask is used to make overall optimization suggestions for the teacher's questions in the classroom text during the supervised fine-tuning training process. The overall optimization suggestions can be a combination of effectiveness and knowledge goals to give corresponding modification suggestions that can be optimized.

[0134] The local evaluation subtask module includes a first local subtask, a second local subtask, and a third local subtask. During the supervised fine-tuning training process, the first local subtask is used to make a specific judgment on the validity of each teacher's question in the classroom text, which is a classification task. The validity can be to judge whether the current teacher's question is an open-ended question and whether it will cause students to think. During the supervised fine-tuning training process, the second local subtask is used to make a specific evaluation of the relevance between the teacher's question in the classroom text and the teaching objectives, which is a classification and natural language generation task. The relevance requires the local evaluation subtask module to judge whether the current question is relevant to the teaching objectives, and if so, the reason needs to be explained. During the supervised fine-tuning training process, the third local subtask is used to optimize and rewrite the teacher's questions in the classroom text, which is a text rewriting task. The optimized rewriting of the teacher's questions can be to rewrite some invalid questions into valid questions, or to rewrite questions from closed questions into open questions. The above tasks all require the output of a text summary of preset fields (for example, 100-150 words), which can be output through the following framework:

[0135] The output results of the six subtasks obtained by executing the three subtasks corresponding to the above-mentioned overall evaluation subtask module and the local evaluation subtask module can be integrated and summarized according to the four evaluation dimensions of question effectiveness analysis, knowledge objectives, question avoidance, and question optimization, and then output according to the four evaluation dimensions. Among them, question effectiveness analysis can be used to evaluate the effectiveness of the questions asked by the teacher in the current course, that is, whether they can stimulate students' thinking; knowledge objectives can be used to evaluate the relevance of the teacher's questions in the current course to the teaching objectives; question avoidance can be used to find teacher questions that can be avoided in the current course; and question optimization can be used to find teacher question directions that can be optimized in the current course.

[0136] The evaluation dimension of question validity analysis can be combined with the overall analysis of question validity and the effectiveness judgment of specific evaluations. In an example, the framework can be as follows:

[0137] The evaluation dimension of knowledge objectives can be combined with the overall analysis of knowledge objectives and the relevance of specific evaluations to teaching objectives. In an example, the framework can be as follows:

[0138] For the evaluation dimension of question avoidance, inappropriate questions in the classroom can be listed. In this application, questions that need to be avoided are defined as questions that are "irrelevant to the teaching content" and "cannot arouse students' thinking". The judgment logic is that the final result is questions with low effectiveness and poor relevance to question-asking teaching in the specific analysis.

[0139] For the question optimization evaluation dimension, the optimization of the current question needs to be output, such as the question rewording part in the specific evaluation and the question optimization suggestions part in the overall evaluation. The final result is output in the modification-suggestion structure. In an example, the framework can be as follows:

[0140] Finally, the above four evaluation dimensions are analyzed and organized into an analysis report for final output.

[0141] In an embodiment of the present application, an overall evaluation subtask module is set to include three overall subtasks, and a local evaluation subtask module is set to include three local subtasks. The three overall subtasks and the three local subtasks, a total of six subtasks, are jointly supervised trained. During the training process, the first overall subtask and the first local subtask cooperate to output the evaluation results under the dimension of question effectiveness analysis, the second overall subtask and the second local subtask cooperate to output the evaluation results under the dimension of knowledge goals, the third overall subtask and the third local subtask cooperate to output the evaluation results under the dimension of question optimization, and at the same time, the first local subtask and the second local subtask cooperate to output the evaluation results under the dimension of question avoidance, that is, all subtasks jointly executed in the overall evaluation subtask module and the local evaluation subtask module can output the evaluation results of the teacher's questions in the evaluated classroom according to the four evaluation dimensions of question effectiveness analysis, knowledge goals, question avoidance and question optimization, thereby ensuring the objectivity of classroom evaluation and further improving the comprehensiveness of the analysis of teachers' classroom questions.

[0142] The technical solution provided in the embodiment of the present application obtains classroom data and classroom target information of a classroom to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information; the classroom data is converted into a text to be input; a target analysis instruction of the classroom to be evaluated is received; the target analysis instruction includes a question-answer dialogue analysis instruction, a classroom interaction analysis instruction, and a curriculum standard implementation analysis instruction; the text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs an evaluation result of the classroom to be evaluated; wherein the preset model uses educational scenario corpus as a training sample, and uses the question-answer dialogue analysis main task module, the classroom interaction analysis main task module, and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answer dialogue analysis main task module is used to respond to the question-answer dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction. The above-mentioned classroom evaluation method can solve the problems of low evaluation efficiency and strong subjectivity of evaluation conclusions when relying on manual labor when evaluating classroom interactive behaviors using relevant technologies. By converting the classroom data of the classroom to be evaluated into text to be input, and receiving the target analysis instructions of the classroom to be evaluated, the text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instructions, and outputs the evaluation results of the classroom to be evaluated. This can achieve the purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the classroom without manual listening, thereby improving the efficiency of classroom evaluation. At the same time, this scheme is a classroom evaluation based on classroom target information and the text to be input obtained by converting classroom data, which ensures the objectivity and accuracy of the evaluation results.

[0143] FIG2 is a flow chart of the classroom evaluation method provided in an embodiment of the present application. As shown in FIG2 , the method specifically includes the following steps:

[0144] S201, obtaining classroom data and classroom target information of a class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information;

[0145] S202, converting the classroom data into text to be input;

[0146] S203, receiving target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions;

[0147] S2041: When it is confirmed that the received target analysis instruction is a question-answer dialogue analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the question-answer dialogue analysis instruction, executes the question-answer dialogue analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the class to be evaluated under the question-answer dialogue analysis main task module;

[0148] In one embodiment, when it is confirmed that the received target analysis instruction is a question-and-answer dialogue analysis instruction, the text to be input and the teaching objective information can be input into a preset model, so that the preset model responds to the question-and-answer dialogue analysis instruction and executes the question-and-answer dialogue analysis main task module. The question-and-answer dialogue analysis main task module can process the text to be input based on the teaching objective information, analyze whether the teacher's question-and-answer dialogue in the text to be input meets the teaching objective information, and then output the evaluation result of the class to be evaluated under the question-and-answer dialogue analysis main task module.

[0149] In a feasible embodiment, when it is confirmed that the received target analysis instruction is a question-answer dialogue analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the question-answer dialogue analysis instruction, executes the question-answer dialogue analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the class to be evaluated under the question-answer dialogue analysis main task module, including:

[0150] When it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, confirming that the preset model responds to the question-answering dialogue analysis instruction, so that the preset model is in a state of waiting to execute the target evaluation tasks under each type of evaluation dimension in the question-answering dialogue analysis main task module;

[0151] Determine target evaluation tasks;

[0152] The text to be input and the teaching objective information are input into a preset model, so that the question-answering dialogue analysis main task module performs the target evaluation task based on the text to be input and the teaching objective information, and outputs the evaluation result of the class to be evaluated.

[0153] The evaluation dimensions may be question validity analysis, knowledge objectives, question avoidance, and question optimization. The target evaluation task may be a task corresponding to at least one of the four evaluation dimensions.

[0154] In one embodiment, when it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, it can be confirmed whether the preset model responds to the question-answering dialogue analysis instruction based on whether a response signal fed back by the preset model is received. When it is confirmed that the preset model responds to the question-answering dialogue analysis instruction, the preset model is placed in a state of waiting to execute the target evaluation tasks under various types of evaluation dimensions in the question-answering dialogue analysis main task module. The evaluation instruction input or selection by the user can be received through the human-computer interaction interface, and the target evaluation task is determined according to the evaluation instruction. The text to be input and the teaching objective information are input into the preset model, so that the question-answering dialogue analysis main task module executes the target evaluation task based on the text to be input and the teaching objective information, and outputs the evaluation results of the class to be evaluated.

[0155] Figure 3 is a user interface provided by an embodiment of the present application. As shown in Figure 3, the left side of the interface includes three major tasks: question-and-answer dialogue, classroom interaction, and implementation of the new curriculum standards. When the user selects "Question-and-answer dialogue" on the left side of the interface, the right side of the interface can display the evaluation results of the default evaluation dimension "Question effectiveness analysis". The options below the evaluation results displayed on the right include: which questions meet the knowledge objectives, which inappropriate questions, optimize my questions, and view all my questions. Users can select the corresponding options according to their viewing needs. When the user selects the "Optimize My Questions" option, it can be determined that the target evaluation task of the preset model is question optimization, and then the text to be input and the teaching objective information are input into the preset model, so that the question-and-answer dialogue analysis main task module performs the question optimization task based on the text to be input and the teaching objective information, and outputs the optimization suggestions for the classroom to be evaluated viewed by the user.

[0156] This solution, after confirming the receipt of the question-answer dialogue analysis instruction and confirming that the preset model responds to the instruction, puts the preset model in a state of waiting to execute the question-answer dialogue analysis main task module. At this time, the target evaluation tasks under various types of evaluation dimensions in the question-answer dialogue analysis main task module are determined, and the preset model executes the target evaluation tasks based on the text to be input and the teaching objective information, thereby achieving the effect of analyzing teachers' classroom questions based on different evaluation dimensions.

[0157] FIG4 is a flowchart of the classroom question evaluation provided by an embodiment of the present application. As shown in FIG4 , the data processing module performs ASR conversion on the classroom transcript data to obtain the classroom text. The classroom text and the teaching objectives can be input into a preset model, and the preset model can be used to perform an overall evaluation of the teacher's questions in the classroom text in combination with the context based on the three dimensions of overall evaluation of question effectiveness, overall evaluation of knowledge objectives, and question optimization suggestions. Speaker recognition can also be performed on the classroom text before or after the classroom text and the teaching objectives are input into the preset model to extract the teacher's dialogue, and sentence extraction and / or question filtering can be performed on the teacher's dialogue to obtain the effective questions asked by the teacher. The effective questions and the teaching objectives are input into the preset model, and the preset model can be used to perform a specific evaluation of the teacher's questions in the classroom text based on the three dimensions of effectiveness judgment, question teaching relevance, and question rewriting.

[0158] As shown in FIG4 , the question-answering dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module;

[0159] The target evaluation task is determined, including:

[0160] Obtaining a first type of evaluation prompt instruction; the first type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the question effectiveness analysis; the evaluation dimension of the question effectiveness analysis includes evaluating the effectiveness of the teacher's questions in the class to be evaluated;

[0161] Determining, based on the first type of evaluation prompt instruction, that the first overall subtask in the overall evaluation subtask module and the first local subtask in the local evaluation subtask module together constitute a target evaluation task; the first overall subtask is used to perform an overall evaluation of the overall effectiveness of all questions asked by teachers in the class to be evaluated, and the first local subtask is used to perform a specific judgment on the effectiveness of each question asked by a teacher in the class to be evaluated;

[0162] and / or,

[0163] Obtaining a second type of evaluation prompt instruction; the second type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the knowledge objective; the evaluation dimension of the knowledge objective includes evaluating the relevance between the teacher's questions in the class to be evaluated and the teaching objectives;

[0164] According to the second type of evaluation prompt instruction, the second overall subtask in the overall evaluation subtask module and the second local subtask in the local evaluation subtask module are determined to jointly constitute a target evaluation task; the second overall subtask is used to perform an overall evaluation of the knowledge objectives preset in the class to be evaluated, and the second local subtask is used to perform a specific evaluation of the relevance between the teacher's questions in the class to be evaluated and the teaching objectives;

[0165] and / or,

[0166] Obtaining a third type of evaluation prompt instruction as the target evaluation task; the third type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to an evaluation dimension of question avoidance; the evaluation dimension of question avoidance includes finding teacher questions that can be avoided in the class to be evaluated;

[0167] According to the third type of evaluation prompt instruction, a first local subtask in the local evaluation subtask module and a second local subtask in the local evaluation subtask module are determined to constitute a target evaluation task together; the first local subtask is used to specifically judge the effectiveness of each teacher's question in the class to be evaluated, and the second local subtask is used to specifically evaluate the relevance between the teacher's question in the class to be evaluated and the teaching objectives;

[0168] and / or,

[0169] Obtaining a fourth type of evaluation prompt instruction as the target evaluation task; the fourth type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of question optimization; the evaluation dimension of question optimization includes finding teacher questions that can be optimized in the class to be evaluated;

[0170] According to the fourth type of evaluation prompt instructions, the third overall subtask in the overall evaluation subtask module and the third local subtask in the local evaluation subtask module together constitute the target evaluation task; the third overall subtask is used to make overall optimization suggestions for the teacher's questions in the classroom to be evaluated, and the third local subtask is used to optimize and rewrite the teacher's questions in the classroom to be evaluated.

[0171] In a first embodiment, a first type of evaluation prompt instruction can be obtained through a user interface. This first type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result for the class to be evaluated based on the evaluation dimension of question effectiveness analysis. This evaluation dimension of question effectiveness analysis includes evaluating the effectiveness of questions asked by teachers in the class to be evaluated. Based on the first type of evaluation prompt instruction, the first overall subtask in the overall evaluation subtask module and the first local subtask in the local evaluation subtask module are determined to jointly constitute the target evaluation task. As indicated by the arrows "101" and "201" in Figure 4, "Overall Question Effectiveness Analysis" in the figure represents the first overall subtask. The first overall subtask is used to conduct an overall evaluation of the overall effectiveness of all questions asked by teachers in the class to be evaluated, while the first local subtask is used to conduct a specific assessment of the effectiveness of each teacher's question in the class to be evaluated. "Effectiveness Assessment" in the figure represents the first local subtask. The first local subtask is used to determine whether each teacher's question can stimulate student reflection. Under this target evaluation task, the output overall evaluation result and local evaluation results of the class to be evaluated can be combined to form the first type of evaluation result. The first type of evaluation results is the evaluation results of the class to be evaluated output by the preset model according to the evaluation dimension of question effectiveness analysis, and can be output according to the output framework of the above-mentioned question effectiveness analysis dimension, which will not be repeated here.

[0172] In a second embodiment, a second type of evaluation prompt instruction can be obtained through the user interface. The second type of evaluation prompt instruction is used to prompt the preset model to output the evaluation results of the class to be evaluated according to the evaluation dimension of the knowledge objectives. The evaluation dimension of the knowledge objectives includes evaluating the correlation between the teacher's questions in the class to be evaluated and the teaching objectives. According to the second type of evaluation prompt instruction, the second overall subtask in the overall evaluation subtask module and the second local subtask in the local evaluation subtask module are determined to jointly constitute the target evaluation task. As shown by the arrows "102" and "202" in Figure 4, the "overall analysis of knowledge objectives" in the figure represents the second overall subtask. The second overall subtask is used to conduct an overall evaluation of the preset knowledge objectives in the class to be evaluated. The "relevance of teaching objectives" in the figure represents the second local subtask. The second local subtask is used to specifically evaluate the correlation between the teacher's questions in the class to be evaluated and the teaching objectives. Under the above-mentioned target evaluation task, the output overall evaluation results and local evaluation results of the class to be evaluated can be combined to form the second type of evaluation results. The second type of evaluation results is the evaluation results of the class to be evaluated output by the preset model according to the evaluation dimension of the knowledge goal, and can be output according to the above-mentioned output framework according to the analysis dimension of the knowledge goal, which will not be repeated here.

[0173] In a third embodiment, a third type of evaluation prompt instruction can be obtained through the user interface as the target evaluation task. The third type of evaluation prompt instruction is used to prompt the preset model to output the evaluation results of the class to be evaluated according to the evaluation dimension of question avoidance. The evaluation dimension of question avoidance includes finding out the teacher's questions that can be avoided in the class to be evaluated. In this embodiment, avoidable questions are defined as questions that are "irrelevant to the teaching content" and "cannot arouse students' thinking". In one embodiment, the judgment logic is that the questions whose effectiveness is judged to be low and poorly relevant to question-asking teaching in the analysis results of the local subtask are used as the final question avoidance results. According to the third type of evaluation prompt instruction, the first local subtask in the local evaluation subtask module and the second local subtask in the local evaluation subtask module are determined to constitute the target evaluation task together; the first local subtask is used to make a specific judgment on the effectiveness of each teacher's questions in the class to be evaluated, and the second local subtask is used to make a specific evaluation of the relevance between the teacher's questions in the class to be evaluated and the teaching objectives.

[0174] In a fourth embodiment, a fourth type of evaluation prompt instruction can be obtained through the user interface as the target evaluation task. The fourth type of evaluation prompt instruction is used to prompt the preset model to output the evaluation results of the class to be evaluated according to the evaluation dimension of question optimization. The evaluation dimension of question optimization includes finding the teacher's questions in the class to be evaluated that can be optimized and outputting the optimization results of the current teacher's questions. According to the fourth type of evaluation prompt instruction, the third overall subtask in the overall evaluation subtask module and the third local subtask in the local evaluation subtask module are determined to jointly constitute the target evaluation task. As indicated by the arrows "103" and "203" in Figure 4, "Question Optimization Suggestions" in the figure represents the third overall subtask. The third overall subtask is used to provide overall optimization suggestions for the teacher's questions in the class to be evaluated. "Question Rewriting" in the figure represents the third local subtask. The third local subtask is used to optimize and rewrite the teacher's questions in the class to be evaluated. Under the above-mentioned target evaluation task, the output overall evaluation results and local evaluation results of the class to be evaluated can be combined to form the third type of evaluation results. The third type of evaluation results is the optimization result of the classroom to be evaluated output by the preset model according to the evaluation dimension optimized according to the questions, and can be output according to the output framework of the above-mentioned question optimization dimension, which will not be repeated here.

[0175] It is understood that this application can evaluate the classroom to be evaluated according to at least one dimension through the above embodiments, or it can evaluate the classroom to be evaluated according to the above four dimensions at the same time. That is, the application has fifteen schemes for evaluating the classroom to be evaluated based on different dimensions.

[0176] In the embodiments of the present application, by simultaneously or selectively evaluating the content of classroom question-and-answer dialogues in the classroom to be evaluated based on different evaluation dimensions and different evaluation scopes, the teacher's questions in the classroom to be evaluated can be evaluated in different dimensions under the premise of using a unified model. This avoids the problem in related technologies that different evaluation dimensions need to be trained separately using classroom data, resulting in isolation between training results of different evaluation dimensions. It can more effectively reduce a large amount of repetitive model training work, strengthen the correlation between training tasks of different evaluation dimensions, and further improve the objectivity of classroom evaluation.

[0177] S2042: When it is confirmed that the received target analysis instruction is a classroom interaction analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the classroom interaction analysis instruction, executes the classroom interaction analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the classroom to be evaluated under the classroom interaction analysis main task module;

[0178] In one embodiment, when it is confirmed that the received target analysis instruction is a classroom interaction analysis instruction, the text to be input and the teaching objective information can be input into a preset model so that the preset model responds to the classroom interaction analysis instruction. The preset model can execute the classroom interaction analysis main task module based on the text to be input and the teaching objective information. The classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, analyze whether the teacher's interactive behavior in the text to be input meets the teaching objective information, and then output the evaluation result of the classroom to be evaluated under the classroom interaction analysis main task module.

[0179] S2043, when it is confirmed that the received target analysis instruction is a curriculum standard implementation analysis instruction, the text to be input and the curriculum standard information are input into the preset model, so that the preset model responds to the curriculum standard implementation analysis instruction, executes the curriculum standard implementation analysis main task module based on the text to be input and the curriculum standard information, and outputs the evaluation result of the class to be evaluated under the curriculum standard implementation analysis main task module.

[0180] In one embodiment, when it is confirmed that the received target analysis instruction is a curriculum standards implementation analysis instruction, the text to be input and the curriculum standards information are input into a preset model, so that the preset model responds to the curriculum standards implementation analysis instruction. The curriculum standards implementation analysis main task module is used to respond to the curriculum standards implementation analysis instruction, analyze whether the course teaching in the text to be input implements the curriculum standards information prescribed by the Ministry of Education, and then output the evaluation result of the class to be evaluated under the curriculum standards implementation analysis main task module.

[0181] The technical solution provided in the embodiment of the present application inputs the text to be input and the classroom target information into a preset model according to different types of target analysis instructions received, so that the preset model responds to different target analysis instructions, executes the corresponding analysis main task module based on the curriculum standard information or teaching target information in the classroom target information and the text to be input, and outputs the evaluation result of the classroom to be evaluated under the corresponding selected main task module, which can achieve the purpose of evaluating the classroom to be evaluated according to the evaluation tasks in different types of main task modules, improves the pertinence of the evaluation of classroom interactive behavior, and enhances the user experience of viewing the evaluation results.

[0182] FIG5 is a schematic diagram of the structure of the classroom evaluation device provided in an embodiment of the present application. As shown in FIG5 , it specifically includes the following:

[0183] The data acquisition module 501 is used to acquire classroom data and classroom target information of the class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information;

[0184] A data conversion module 502 is used to convert the classroom data into text to be input;

[0185] The instruction receiving module 503 is configured to receive the target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions;

[0186] The evaluation module 504 is used to input the text to be input and the classroom target information into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction.

[0187] In one embodiment, the evaluation module 504 includes:

[0188] a question-and-answer dialogue analysis unit configured to, upon confirming that the received target analysis instruction is a question-and-answer dialogue analysis instruction, input the to-be-input text and the teaching objective information into a preset model, so that the preset model responds to the question-and-answer dialogue analysis instruction, executes the question-and-answer dialogue analysis main task module based on the to-be-input text and the teaching objective information, and outputs an evaluation result of the to-be-evaluated class under the question-and-answer dialogue analysis main task module;

[0189] a classroom interaction analysis unit configured to, upon confirming that the received target analysis instruction is a classroom interaction analysis instruction, input the text to be input and the teaching objective information into a preset model, so that the preset model responds to the classroom interaction analysis instruction, executes the classroom interaction analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the classroom to be evaluated under the classroom interaction analysis main task module;

[0190] The curriculum standards implementation analysis unit is used to input the text to be input and the curriculum standards information into a preset model when it is confirmed that the received target analysis instruction is a curriculum standards implementation analysis instruction, so that the preset model responds to the curriculum standards implementation analysis instruction, executes the curriculum standards implementation analysis main task module based on the text to be input and the curriculum standards information, and outputs the evaluation result of the class to be evaluated under the curriculum standards implementation analysis main task module.

[0191] In one embodiment, the question-answer dialogue analysis unit is specifically configured to:

[0192] When it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, confirming that the preset model responds to the question-answering dialogue analysis instruction, so that the preset model is in a state of waiting to execute the target evaluation tasks under each type of evaluation dimension in the question-answering dialogue analysis main task module;

[0193] Determine target evaluation tasks;

[0194] The text to be input and the teaching objective information are input into a preset model, so that the question-answering dialogue analysis main task module performs the target evaluation task based on the text to be input and the teaching objective information, and outputs the evaluation result of the class to be evaluated.

[0195] In one embodiment, the question-answering dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module;

[0196] The question-answer dialogue analysis unit is specifically used to:

[0197] Obtaining a first type of evaluation prompt instruction; the first type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the question effectiveness analysis; the evaluation dimension of the question effectiveness analysis includes evaluating the effectiveness of the teacher's questions in the class to be evaluated;

[0198] Determining, based on the first type of evaluation prompt instruction, that the first overall subtask in the overall evaluation subtask module and the first local subtask in the local evaluation subtask module together constitute a target evaluation task; the first overall subtask is used to perform an overall evaluation of the overall effectiveness of all questions asked by teachers in the class to be evaluated, and the first local subtask is used to perform a specific judgment on the effectiveness of each question asked by a teacher in the class to be evaluated;

[0199] and / or,

[0200] Obtaining a second type of evaluation prompt instruction; the second type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the knowledge objective; the evaluation dimension of the knowledge objective includes evaluating the relevance between the teacher's questions in the class to be evaluated and the teaching objectives;

[0201] According to the second type of evaluation prompt instruction, the second overall subtask in the overall evaluation subtask module and the second local subtask in the local evaluation subtask module are determined to jointly constitute a target evaluation task; the second overall subtask is used to perform an overall evaluation of the knowledge objectives preset in the class to be evaluated, and the second local subtask is used to perform a specific evaluation of the relevance between the teacher's questions in the class to be evaluated and the teaching objectives;

[0202] and / or,

[0203] Obtaining a third type of evaluation prompt instruction as the target evaluation task; the third type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to an evaluation dimension of question avoidance; the evaluation dimension of question avoidance includes finding teacher questions that can be avoided in the class to be evaluated;

[0204] According to the third type of evaluation prompt instruction, a first local subtask in the local evaluation subtask module and a second local subtask in the local evaluation subtask module are determined to constitute a target evaluation task together; the first local subtask is used to specifically judge the effectiveness of each teacher's question in the class to be evaluated, and the second local subtask is used to specifically evaluate the relevance between the teacher's question in the class to be evaluated and the teaching objectives;

[0205] and / or,

[0206] Obtaining a fourth type of evaluation prompt instruction as the target evaluation task; the fourth type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of question optimization; the evaluation dimension of question optimization includes finding teacher questions that can be optimized in the class to be evaluated;

[0207] According to the fourth type of evaluation prompt instructions, the third overall subtask in the overall evaluation subtask module and the third local subtask in the local evaluation subtask module together constitute the target evaluation task; the third overall subtask is used to make overall optimization suggestions for the teacher's questions in the classroom to be evaluated, and the third local subtask is used to optimize and rewrite the teacher's questions in the classroom to be evaluated.

[0208] In one embodiment, the question-answer dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module; during the supervised fine-tuning training of the question-answer dialogue analysis main task module by the preset model, the classroom text containing teaching objectives and teacher questions in the educational scene corpus is processed according to the overall evaluation subtask module and the local evaluation subtask module respectively to output an evaluation result for the classroom text.

[0209] In one embodiment, the overall evaluation subtask module includes a first overall subtask, a second overall subtask, and a third overall subtask; the first overall subtask is used to conduct an overall evaluation of the overall effectiveness of all teacher questions in the classroom text during the supervised fine-tuning training process; the second overall subtask is used to conduct an overall evaluation of the preset knowledge objectives in the classroom text during the supervised fine-tuning training process; the third overall subtask is used to make overall optimization suggestions for teacher questions in the classroom text during the supervised fine-tuning training process; the local evaluation subtask module includes a first local subtask, a second local subtask, and a third local subtask; the first local subtask is used to conduct a specific judgment on the effectiveness of each teacher question in the classroom text during the supervised fine-tuning training process; the second local subtask is used to conduct a specific evaluation of the correlation between teacher questions in the classroom text and teaching objectives during the supervised fine-tuning training process; the third local subtask is used to optimize and rewrite teacher questions in the classroom text during the supervised fine-tuning training process.

[0210] In one embodiment, the data acquisition module 501 is specifically configured to:

[0211] Obtain classroom data of the class to be evaluated;

[0212] The class objective information that matches the class data is searched in a preset database, or the class data is input into a preset chapter matching model so that the chapter matching model outputs the class objective information that matches the class data.

[0213] In one embodiment, the data conversion module 502 is specifically configured to:

[0214] Converting the classroom data into text to obtain classroom text;

[0215] extracting teacher question text from the classroom text;

[0216] The classroom text and the teacher's question text are used as texts to be input.

[0217] In one embodiment, the data conversion module 502 is specifically configured to:

[0218] Identify the dialogue characters with the most dialogue in the classroom text;

[0219] Determine the dialogue role with the most dialogues as the teacher role;

[0220] Extracting all the dialogue contents of the teacher role;

[0221] The teacher's question text is filtered out from the entire conversation content.

[0222] In one embodiment, the data conversion module 502 is specifically configured to:

[0223] Identifying basic questions in the entire conversation content;

[0224] A denoising operation is performed on the basic question to obtain the teacher's question text.

[0225] The technical solution provided in the embodiment of the present application is a data acquisition module for acquiring classroom data and classroom target information of a class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information; a data conversion module for converting the classroom data into a text to be input; an instruction receiving module for receiving target analysis instructions of the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions and curriculum standard implementation analysis instructions; an evaluation module for inputting the text to be input and the classroom target information into a preset model so that the preset The model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks of joint training, and is obtained through supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction. The above-mentioned classroom evaluation device can solve the problems of low evaluation efficiency and strong subjectivity of evaluation conclusions when relying on manual labor when evaluating classroom interactive behaviors using relevant technologies. By converting the classroom data of the classroom to be evaluated into text to be input, and receiving the target analysis instructions of the classroom to be evaluated, the text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instructions, and outputs the evaluation results of the classroom to be evaluated. This can achieve the purpose of performing question-and-answer dialogue analysis tasks, classroom interaction analysis tasks and curriculum standard implementation analysis tasks on the classroom without manual listening, thereby improving the efficiency of classroom evaluation. At the same time, this scheme is a classroom evaluation based on classroom target information and text to be input, which ensures the objectivity and accuracy of the evaluation results.

[0226] The classroom evaluation device in the embodiment of the present application can be a device, or a component, integrated circuit, or chip in a terminal. The device can be a mobile electronic device or a non-mobile electronic device. For example, the mobile electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle electronic device, a wearable device, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), etc. The non-mobile electronic device can be a server, a network attached storage (NAS), a personal computer (PC), a television (TV), an ATM, or an kiosks, etc., which are not specifically limited in the embodiment of the present application.

[0227] The classroom evaluation device in the embodiment of the present application may be a device having an operating system. The operating system may be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0228] The classroom evaluation device provided in the embodiment of the present application can implement each process implemented in the above method embodiment. To avoid repetition, it will not be described here.

[0229] FIG6 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. As shown in FIG6 , an embodiment of the present application further provides an electronic device 600, comprising a processor 601, a memory 602, and a program or instruction stored in the memory 602 and executable on the processor 601. When executed by the processor 601, the program or instruction implements each process of the above-mentioned classroom evaluation method embodiment and can achieve the same technical effect. To avoid repetition, the details will not be repeated here.

[0230] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0231] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned classroom evaluation method embodiment are implemented and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0232] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk.

[0233] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned classroom evaluation method embodiment and achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0234] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0235] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element. In addition, it should be noted that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the opposite order according to the functions involved. For example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0236] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the relevant technology, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0237] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

[0238] The above are only preferred embodiments of the present application and the technical principles employed. The present application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that are possible for those skilled in the art will not depart from the scope of protection of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments and may include more other equivalent embodiments without departing from the concept of the present application. The scope of the present application is determined by the scope of the claims.

Claims

1. A classroom evaluation method, wherein: The method comprises: Obtaining classroom data and classroom target information of the class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information; Converting the classroom data into text to be input; Receive target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions; The text to be input and the classroom target information are input into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction.

2. The method according to claim 1, wherein The step of inputting the text to be input and the classroom target information into a preset model, causing the preset model to process the text to be input and the classroom target information according to the target analysis instruction, and outputting an evaluation result of the classroom to be evaluated, includes: When it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the question-answering dialogue analysis instruction, executes the question-answering dialogue analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the class to be evaluated under the question-answering dialogue analysis main task module; When it is confirmed that the received target analysis instruction is a classroom interaction analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the classroom interaction analysis instruction, executes the classroom interaction analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the classroom to be evaluated under the classroom interaction analysis main task module; When it is confirmed that the received target analysis instruction is a curriculum standard implementation analysis instruction, the text to be input and the curriculum standard information are input into the preset model, so that the preset model responds to the curriculum standard implementation analysis instruction, executes the curriculum standard implementation analysis main task module based on the text to be input and the curriculum standard information, and outputs the evaluation result of the class to be evaluated under the curriculum standard implementation analysis main task module.

3. The method according to claim 2, wherein: When it is confirmed that the received target analysis instruction is a question-answer dialogue analysis instruction, the text to be input and the teaching objective information are input into a preset model, so that the preset model responds to the question-answer dialogue analysis instruction, executes the question-answer dialogue analysis main task module based on the text to be input and the teaching objective information, and outputs an evaluation result of the class to be evaluated under the question-answer dialogue analysis main task module, including: When it is confirmed that the received target analysis instruction is a question-answering dialogue analysis instruction, confirming that the preset model responds to the question-answering dialogue analysis instruction, so that the preset model is in a state of waiting to execute the target evaluation tasks under each type of evaluation dimension in the question-answering dialogue analysis main task module; Determine target evaluation tasks; The text to be input and the teaching objective information are input into a preset model, so that the question-answering dialogue analysis main task module performs the target evaluation task based on the text to be input and the teaching objective information, and outputs the evaluation result of the class to be evaluated.

4. The method according to claim 3, wherein: The question-answering dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module; The target evaluation task is determined, including: Obtaining a first type of evaluation prompt instruction; the first type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the question effectiveness analysis; the evaluation dimension of the question effectiveness analysis includes evaluating the effectiveness of the teacher's questions in the class to be evaluated; Determining, based on the first type of evaluation prompt instruction, that the first overall subtask in the overall evaluation subtask module and the first local subtask in the local evaluation subtask module together constitute a target evaluation task; the first overall subtask is used to perform an overall evaluation of the overall effectiveness of all questions asked by teachers in the class to be evaluated, and the first local subtask is used to perform a specific judgment on the effectiveness of each question asked by a teacher in the class to be evaluated; and / or, Obtaining a second type of evaluation prompt instruction; the second type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of the knowledge objective; the evaluation dimension of the knowledge objective includes evaluating the relevance between the teacher's questions in the class to be evaluated and the teaching objectives; According to the second type of evaluation prompt instruction, the second overall subtask in the overall evaluation subtask module and the second local subtask in the local evaluation subtask module are determined to jointly constitute a target evaluation task; the second overall subtask is used to perform an overall evaluation of the knowledge objectives preset in the class to be evaluated, and the second local subtask is used to perform a specific evaluation of the relevance between the teacher's questions in the class to be evaluated and the teaching objectives; and / or, Obtaining a third type of evaluation prompt instruction as the target evaluation task; the third type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to an evaluation dimension of question avoidance; the evaluation dimension of question avoidance includes finding teacher questions that can be avoided in the class to be evaluated; According to the third type of evaluation prompt instruction, a first local subtask in the local evaluation subtask module and a second local subtask in the local evaluation subtask module are determined to constitute a target evaluation task together; the first local subtask is used to specifically judge the effectiveness of each teacher's question in the class to be evaluated, and the second local subtask is used to specifically evaluate the relevance between the teacher's question in the class to be evaluated and the teaching objectives; and / or, Obtaining a fourth type of evaluation prompt instruction as the target evaluation task; the fourth type of evaluation prompt instruction is used to prompt the preset model to output an evaluation result of the class to be evaluated according to the evaluation dimension of question optimization; the evaluation dimension of question optimization includes finding teacher questions that can be optimized in the class to be evaluated; According to the fourth type of evaluation prompt instructions, the third overall subtask in the overall evaluation subtask module and the third local subtask in the local evaluation subtask module together constitute the target evaluation task; the third overall subtask is used to make overall optimization suggestions for the teacher's questions in the classroom to be evaluated, and the third local subtask is used to optimize and rewrite the teacher's questions in the classroom to be evaluated.

5. The method according to any one of claims 1 to 3, wherein The question-answer dialogue analysis main task module includes an overall evaluation subtask module and a local evaluation subtask module; during the supervised fine-tuning training of the question-answer dialogue analysis main task module by the preset model, the classroom text containing teaching objectives and teacher questions in the educational scene corpus is processed according to the overall evaluation subtask module and the local evaluation subtask module respectively to output the evaluation results for the classroom text.

6. The method according to claim 5, wherein: The overall evaluation subtask module includes a first overall subtask, a second overall subtask, and a third overall subtask; the first overall subtask is used to perform an overall evaluation of the overall effectiveness of all teacher questions in the classroom text during the supervised fine-tuning training process; the second overall subtask is used to perform an overall evaluation of the knowledge objectives preset in the classroom text during the supervised fine-tuning training process; the third overall subtask is used to provide overall optimization suggestions for the teacher questions in the classroom text during the supervised fine-tuning training process; The local evaluation subtask module includes a first local subtask, a second local subtask, and a third local subtask; during the supervised fine-tuning training process, the first local subtask is used to make a specific judgment on the effectiveness of each teacher's questions in the classroom text; during the supervised fine-tuning training process, the second local subtask is used to make a specific evaluation of the relevance between the teacher's questions in the classroom text and the teaching objectives; during the supervised fine-tuning training process, the third local subtask is used to optimize and rewrite the teacher's questions in the classroom text.

7. The method according to any one of claims 1 to 4, wherein The acquisition of classroom data and classroom target information of the class to be evaluated includes: Obtain classroom data of the class to be evaluated; The class objective information that matches the class data is searched in a preset database, or the class data is input into a preset chapter matching model so that the chapter matching model outputs the class objective information that matches the class data.

8. The method according to any one of claims 1 to 4, wherein The converting of the classroom data into text to be input includes: Converting the classroom data into text to obtain classroom text; extracting teacher question text from the classroom text; The classroom text and the teacher's question text are used as texts to be input.

9. The method according to claim 8, wherein The step of extracting the teacher's question text from the classroom text includes: Identify the dialogue characters with the most dialogue in the classroom text; Determine the dialogue role with the most dialogues as the teacher role; Extracting all the dialogue contents of the teacher role; The teacher's question text is filtered out from the entire conversation content.

10. The method according to claim 9, wherein: The step of filtering out the teacher's question text from the entire conversation content includes: Identifying basic questions in the entire conversation content; A denoising operation is performed on the basic question to obtain the teacher's question text.

11. A classroom evaluation device, wherein: The device comprises: A data acquisition module is used to acquire classroom data and classroom target information of the class to be evaluated; wherein the classroom data includes: audio data and / or video data; the classroom target information includes teaching target information and curriculum standard information; A data conversion module, used for converting the classroom data into text to be input; An instruction receiving module is used to receive target analysis instructions for the class to be evaluated; the target analysis instructions include question-answer dialogue analysis instructions, classroom interaction analysis instructions, and curriculum standard implementation analysis instructions; An evaluation module is used to input the text to be input and the classroom target information into a preset model, so that the preset model processes the text to be input and the classroom target information according to the target analysis instruction, and outputs the evaluation result of the classroom to be evaluated; wherein, the preset model uses educational scene corpus as training samples, and uses the question-answering dialogue analysis main task module, the classroom interaction analysis main task module and the curriculum standard implementation analysis main task module as the main tasks for joint training, and is obtained through supervised fine-tuning training; the question-answering dialogue analysis main task module is used to respond to the question-answering dialogue analysis instruction, the classroom interaction analysis main task module is used to respond to the classroom interaction analysis instruction, and the curriculum standard implementation analysis main task module is used to respond to the curriculum standard implementation analysis instruction.

12. An electronic device, wherein: It includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor. When the program or instruction is executed by the processor, the steps of the classroom evaluation method as described in any one of claims 1 to 10 are implemented.

13. A readable storage medium, wherein: The readable storage medium stores a program or instruction, and when the program or instruction is executed by a processor, the steps of the classroom evaluation method according to any one of claims 1 to 10 are implemented.

Images