Multi-modal teaching evaluation method and system based on artificial intelligence

Abstract

According to the multi-modal teaching evaluation method and system based on artificial intelligence disclosed by the invention, on the basis of teaching metadata, a school-level intelligent teaching cockpit and teaching portraits of students for teaching reaction of teachers are constructed, so that the first average evaluation score of teaching of the teachers is obtained, and the teaching state evaluation score of the teachers is combined to obtain the multi-modal teaching evaluation score of the teachers. Obtaining a second average evaluation score of teaching of the teacher, and evaluating the teaching of the teacher by taking a preset evaluation score range as an evaluation index; by comparing and analyzing the teaching mode of the teacher, the teaching characteristics and the learning state of the student, the method is not limited to an evaluation mode with score as the purpose, and quantification and evaluation are performed from multiple dimensions, so that the teaching method is further optimized, the professional growth of the teacher and the student and the improvement of the teaching method are better promoted, and the improvement of the education quality is promoted.

Description

Technical Field

[0001] The present invention relates to the field of education, and more specifically, to a multimodal teaching evaluation method and system based on artificial intelligence. Background Art

[0002] Artificial intelligence technology is constantly developing, for example, technologies such as speech synthesis, text translation, and image generation are constantly improving, providing a technical basis for multimodal teaching evaluation methods; traditional teaching evaluation methods have gradually exposed many problems. For example, the existing classroom teaching evaluation system often has problems such as a single evaluation dimension, a cumbersome evaluation process, and insufficiently objective evaluation results. Summary of the invention

[0003] In order to solve at least one of the above technical problems, the present invention provides a multimodal teaching evaluation method and system based on artificial intelligence, which quantifies and evaluates from multiple dimensions, so as to further optimize the teaching methods, better promote the professional growth of teachers and students and the improvement of teaching methods, thereby promoting the improvement of education quality.

[0004] The first aspect of the present invention provides a multimodal teaching evaluation method based on artificial intelligence, comprising: Get teaching metadata; Classify the teaching metadata to obtain multiple teaching sub-data, and build corresponding teaching sub-data models based on the teaching sub-data; Based on the teaching metadata and teaching sub-data model, a school-level smart teaching cockpit is constructed, and teaching driving data is obtained; Extract all teaching data of each student and teacher in the teaching driving data to form a data portrait of a single student's response to the teacher's teaching; Based on the data portrait of individual students' responses to the teacher's teaching, the corresponding student's evaluation score on the teacher's teaching is obtained; Calculate the average of all students' evaluation scores on the teacher's teaching to obtain the first average evaluation score of the teacher's teaching; Get the teacher's teaching status evaluation score; The first average evaluation score of the teacher's teaching and the teacher's teaching status evaluation score are averaged to obtain the second average evaluation score of the teacher's teaching; The teacher's teaching is evaluated based on the preset evaluation score range within which the second average evaluation score of the teacher's teaching falls.

[0005] In this solution, the step of forming a data profile of a single student's response to the teacher's teaching specifically includes: Obtain the knowledge points contained in the textbook in the teaching metadata; All teaching data of individual students and teachers are sliced ​​and sorted according to knowledge points to obtain the learning status under each knowledge point; Construct the sub-puzzle of the corresponding knowledge point according to the learning status of each knowledge point; Combine sub-puzzles of knowledge points of the same type to obtain puzzles of the corresponding knowledge point type; Combine all puzzle pieces of knowledge point types to obtain a data portrait of a single student's response to the teacher's teaching; The types of knowledge points include mastered knowledge points, unmastered knowledge points and untested knowledge points; and the learning status includes active learning status, passive learning status and general learning status.

[0006] In this solution, the step of obtaining the learning status at each knowledge point includes: Extract the time nodes when knowledge points appear in the teacher's teaching process and set them as the time nodes when knowledge points appear; Extract the status characteristics of students at the time nodes where knowledge points appear; Classify the state characteristics of students at the time nodes where knowledge points appear, and record the number of times the state characteristics appear under each state type; Accumulate the number of occurrences of state features under all state types to obtain the total number of occurrences; Divide the number of times the state feature appears under the state type by the total number of times to obtain the probability value of the corresponding state type appearing at the knowledge point time node; If the probability value of the corresponding state type at the knowledge point appearance time node is greater than or equal to the preset first probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the difference between the probability value of the corresponding state type at the knowledge point appearance time node and the probability value of any other state type at the knowledge point appearance time node is greater than or equal to the preset second probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the probability value of the corresponding state type at the time node when the knowledge point appears is less than the preset first probability threshold, and the difference between the probability value of the corresponding state type at the time node when the knowledge point appears and the probability value of any other state type at the time node when the knowledge point appears is less than the preset second probability threshold, then when the knowledge point is a mastered knowledge point, the active learning state is set to the current learning state; when the knowledge point is an unmastered knowledge point, the passive learning state is set to the current learning state; when the knowledge point is an untested knowledge point, the general learning state is set to the current learning state.

[0007] In this solution, the step of obtaining the evaluation score of the teacher's teaching by the corresponding student based on the data portrait of the individual student's response to the teacher's teaching also includes: Set different scores for different sub-puzzles, and add up the scores of all sub-puzzles to get the total score of the student's response to the teacher's teaching; Get the ideal total score of students' response to teachers' teaching; Divide the total score of the student's response to the teacher's teaching by the ideal total score of the student's response to the teacher's teaching, and then multiply it by the preset benchmark evaluation score to obtain the student's evaluation score for the teacher's teaching.

[0008] This plan also includes: Acquire knowledge points not mentioned in teaching; If a knowledge point not mentioned in the teaching appears in the test and the corresponding student does not get a score, the quantity value of the knowledge point not mentioned is recorded and set as the first quantity value; Multiply the first quantity value by a preset first revision weight score to obtain a first revision score; Subtract the first revised score from the total score of the student's response to the teacher's teaching to obtain the revised total score of the student's response to the teacher's teaching; If the first quantity value is greater than or equal to the preset first quantity threshold, the teaching evaluation of the corresponding teacher will be set to unqualified.

[0009] In this solution, the steps of obtaining the teacher's teaching status evaluation score specifically include: Obtaining metadata for historical teaching; According to the history teaching metadata, the score ratio of knowledge points in different history teaching methods is obtained; Obtain the current teacher's teaching methods, teaching characteristics, and knowledge points under the current teacher's teaching methods and teaching characteristics; Based on the same knowledge points, determine the predicted score probability value of the current teacher’s teaching method; Extract teaching sounds and teaching actions from teaching features; Extract the student learning status corresponding to the teaching sound, and set the basic score of the corresponding teaching sound according to the corresponding student learning status; If there is no teaching action at the time node of the teaching sound, the basic score of the teaching sound is set to the corresponding teaching feature score; If there is a teaching action at the time node of the teaching sound, the teaching action is compared and analyzed with the preset action to obtain the action similarity value; If the action similarity value is greater than or equal to a preset first similarity threshold, the corresponding teaching action is set as a violation action and a violation score is generated; If the action similarity value is less than the preset first similarity threshold, the corresponding teaching action is set as an auxiliary action and an action additional score is generated; The basic score of the teaching sound is added to the additional score of the action or subtracted from the violation score to obtain the corresponding teaching feature score; The teaching feature score is set as the set score of the corresponding knowledge point; Based on the same knowledge point, multiply the set score of the corresponding knowledge point by the predicted score probability value of the current teacher's teaching method to obtain the predicted score value of the corresponding knowledge point; Get all the knowledge points contained in this teaching; Accumulate the predicted scores of the knowledge points in this teaching to get the predicted total score; Accumulate the set scores of all knowledge points in this teaching to get the ideal score value in this teaching; Divide the predicted total score by the ideal score in this teaching, and then multiply it by the preset benchmark evaluation score to get the current teacher's teaching status evaluation score.

[0010] This plan also includes: Subtract the teacher's teaching status evaluation score from any student's evaluation score of the teacher's teaching to obtain the first score difference; If the first score difference is greater than the preset first score difference threshold, the corresponding student's evaluation score for the teacher's teaching will be deleted, and the first average student's evaluation score for the teacher's teaching will be recalculated, and the number of the deleted student's evaluation scores for the teacher's teaching will be obtained and set as the second number value; If the second quantity value is greater than or equal to the preset second quantity threshold, the corresponding teacher's teaching will be evaluated as unqualified.

[0011] A second aspect of the present invention provides a multimodal teaching evaluation system based on artificial intelligence, comprising a memory and a processor, wherein the memory stores a multimodal teaching evaluation method program based on artificial intelligence, and when the multimodal teaching evaluation method program based on artificial intelligence is executed by the processor, the following steps are implemented: Get teaching metadata; Classify the teaching metadata to obtain multiple teaching sub-data, and build corresponding teaching sub-data models based on the teaching sub-data; Based on the teaching metadata and teaching sub-data model, a school-level smart teaching cockpit is constructed, and teaching driving data is obtained; Extract all teaching data of each student and teacher in the teaching driving data to form a data portrait of a single student's response to the teacher's teaching; Based on the data portrait of individual students' responses to the teacher's teaching, the corresponding student's evaluation score on the teacher's teaching is obtained; Calculate the average of all students' evaluation scores on the teacher's teaching to obtain the first average evaluation score of the teacher's teaching; Get the teacher's teaching status evaluation score; The first average evaluation score of the teacher's teaching and the teacher's teaching status evaluation score are averaged to obtain the second average evaluation score of the teacher's teaching; The teacher's teaching is evaluated based on the preset evaluation score range within which the second average evaluation score of the teacher's teaching falls.

[0012] In this solution, the step of forming a data profile of a single student's response to the teacher's teaching specifically includes: Obtain the knowledge points contained in the textbook in the teaching metadata; All teaching data of individual students and teachers are sliced ​​and sorted according to knowledge points to obtain the learning status under each knowledge point; Construct the sub-puzzle of the corresponding knowledge point according to the learning status of each knowledge point; Combine sub-puzzles of knowledge points of the same type to obtain puzzles of the corresponding knowledge point type; Combine all puzzle pieces of knowledge point types to obtain a data portrait of a single student's response to the teacher's teaching; The types of knowledge points include mastered knowledge points, unmastered knowledge points and untested knowledge points; and the learning status includes active learning status, passive learning status and general learning status.

[0013] In this solution, the step of obtaining the learning status at each knowledge point includes: Extract the time nodes when knowledge points appear in the teacher's teaching process and set them as the time nodes when knowledge points appear; Extract the status characteristics of students at the time nodes where knowledge points appear; Classify the state characteristics of students at the time nodes where knowledge points appear, and record the number of times the state characteristics appear under each state type; Accumulate the number of occurrences of state features under all state types to obtain the total number of occurrences; Divide the number of times the state feature appears under the state type by the total number of times to obtain the probability value of the corresponding state type appearing at the knowledge point time node; If the probability value of the corresponding state type at the knowledge point appearance time node is greater than or equal to the preset first probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the difference between the probability value of the corresponding state type at the knowledge point appearance time node and the probability value of any other state type at the knowledge point appearance time node is greater than or equal to the preset second probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the probability value of the corresponding state type at the time node when the knowledge point appears is less than the preset first probability threshold, and the difference between the probability value of the corresponding state type at the time node when the knowledge point appears and the probability value of any other state type at the time node when the knowledge point appears is less than the preset second probability threshold, then when the knowledge point is a mastered knowledge point, the active learning state is set to the current learning state; when the knowledge point is an unmastered knowledge point, the passive learning state is set to the current learning state; when the knowledge point is an untested knowledge point, the general learning state is set to the current learning state.

[0014] The present invention discloses an artificial intelligence-based multimodal teaching evaluation method and system, which compares and analyzes teachers' teaching methods, teaching characteristics, and students' learning status. It is not limited to evaluation methods aimed at scoring, but quantifies and evaluates from multiple dimensions, thereby further optimizing teaching methods, better promoting the professional growth of teachers and students and the improvement of teaching methods, and thus promoting the improvement of education quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart of a multimodal teaching evaluation method based on artificial intelligence of the present invention is shown; Figure 2 A flow chart showing the steps of forming a data portrait of a single student's response to a teacher's teaching according to the present invention; Figure 3 A block diagram of a multimodal teaching evaluation system based on artificial intelligence of the present invention is shown. DETAILED DESCRIPTION

[0016] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0017] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0018] Figure 1 A flow chart of a multimodal teaching evaluation method based on artificial intelligence of the present invention is shown.

[0019] like Figure 1 As shown, the present invention discloses a multimodal teaching evaluation method based on artificial intelligence, comprising: S101, obtaining teaching metadata; S102, classifying the teaching metadata to obtain a plurality of teaching sub-data, and constructing a corresponding teaching sub-data model based on the teaching sub-data; S103, based on the teaching metadata and the teaching sub-data model, construct a school-level smart teaching cockpit and obtain teaching driving data; S104, extracting all teaching data of each student and teacher in the teaching driving data to form a data portrait of a single student's response to the teacher's teaching; S105, obtaining the evaluation score of the teacher's teaching by the corresponding student based on the data portrait of the individual student's reaction to the teacher's teaching; S106, calculating the average of all students' evaluation scores on the teacher's teaching, and obtaining a first average evaluation score of the teacher's teaching; S107, obtaining the teacher's teaching status evaluation score; S108, calculating the average of the first average evaluation score of the teacher's teaching and the teacher's teaching status evaluation score to obtain a second average evaluation score of the teacher's teaching; S109, evaluating the teacher's teaching according to the preset evaluation score range within which the second average evaluation score of the teacher's teaching falls.

[0020] According to an embodiment of the present invention, the teaching data generated by the teaching platform and the smart teacher are structuredly managed through intelligent metadata collection, support docking with third-party platforms, and can obtain teaching metadata, wherein the teaching metadata includes teaching content data, teaching methods, teaching characteristics, student behavior data and other teaching sub-data in the teaching process. For example, a student behavior model is constructed based on the student behavior data, and the student behavior data is classified and identified through the student behavior model. For example, student behavior is divided into active learning behavior, passive learning behavior and general learning behavior. The active learning behavior includes reading and writing, looking at the blackboard, standing up, etc., and the passive learning behavior includes lying on the table, playing with mobile phones, eating, etc.; the general behavior is for unrecognizable learning behavior; a school-level smart teaching cockpit is established through the teaching metadata and teaching sub-data model, and the online smart teaching platform data, offline smart teacher data and other data in the teaching process are analyzed and extracted to form teaching driving data, and presented in the form of icons to display all teaching data of students and teachers, and form data portraits of individual students and teachers; if the preset benchmark evaluation score is 100, the preset evaluation score range can be set to , ,in The corresponding teaching evaluation is unsatisfactory. The corresponding teaching evaluation is qualified. The corresponding teaching evaluation is good; The corresponding teaching evaluation is excellent. The corresponding teaching evaluation is excellent; the higher the second average evaluation score is, the better the corresponding teacher's teaching level is, and the higher the corresponding teacher's teaching evaluation is.

[0021] Figure 2 A flow chart showing the steps of forming a data portrait of a single student's response to a teacher's teaching according to the present invention.

[0022] like Figure 2 As shown, the step of forming a data portrait of a single student's response to the teacher's teaching specifically includes: S201, obtaining knowledge points contained in the textbook in the teaching metadata; S202, slicing and arranging all teaching data of a single student and teacher according to knowledge points to obtain the learning status under each knowledge point; S203, constructing a sub-puzzle corresponding to each knowledge point according to the learning status of each knowledge point; S204, combining sub-puzzles of knowledge points of the same type to obtain puzzles of corresponding knowledge point types; S205, combine all puzzles of knowledge point types to obtain a data portrait of a single student's response to the teacher's teaching.

[0023] It should be noted that the knowledge point types include mastered knowledge points, unmastered knowledge points and untested knowledge points; the learning status includes active learning status, passive learning status and general learning status; the learning status is determined based on the student behavior, for example, if the student behavior is active learning behavior, the corresponding learning status is an active learning state; if the student behavior is passive learning behavior, the corresponding learning state is a passive learning state.

[0024] According to an embodiment of the present invention, the step of obtaining the learning status at each knowledge point includes: Extract the time nodes when knowledge points appear in the teacher's teaching process and set them as the time nodes when knowledge points appear; Extract the status characteristics of students at the time nodes where knowledge points appear; Classify the state characteristics of students at the time nodes where knowledge points appear, and record the number of times the state characteristics appear under each state type; Accumulate the number of occurrences of state features under all state types to obtain the total number of occurrences; Divide the number of times the state feature appears under the state type by the total number of times to obtain the probability value of the corresponding state type appearing at the knowledge point time node; If the probability value of the corresponding state type at the knowledge point appearance time node is greater than or equal to the preset first probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the difference between the probability value of the corresponding state type at the knowledge point appearance time node and the probability value of any other state type at the knowledge point appearance time node is greater than or equal to the preset second probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the probability value of the corresponding state type at the time node when the knowledge point appears is less than the preset first probability threshold, and the difference between the probability value of the corresponding state type at the time node when the knowledge point appears and the probability value of any other state type at the time node when the knowledge point appears is less than the preset second probability threshold, then when the knowledge point is a mastered knowledge point, the active learning state is set to the current learning state; when the knowledge point is an unmastered knowledge point, the passive learning state is set to the current learning state; when the knowledge point is an untested knowledge point, the general learning state is set to the current learning state.

[0025] It should be noted that the state characteristics of the students at the time nodes where the knowledge points appear are the students' learning behaviors at the corresponding time nodes. Each learning behavior corresponds to a state characteristic, and the state types are active learning behavior type, passive learning behavior type and general learning behavior type.

[0026] According to an embodiment of the present invention, the step of obtaining the evaluation score of the teacher's teaching by the corresponding student based on the data portrait of the individual student's reaction to the teacher's teaching also includes: Set different scores for different sub-puzzles, and add up the scores of all sub-puzzles to get the total score of the student's response to the teacher's teaching; Get the ideal total score of students' response to teachers' teaching; Divide the total score of the student's response to the teacher's teaching by the ideal total score of the student's response to the teacher's teaching, and then multiply it by the preset benchmark evaluation score to obtain the student's evaluation score for the teacher's teaching.

[0027] It should be noted that, in this embodiment, the sub-puzzles include an active learning state under mastering knowledge points, a passive learning state under mastering knowledge points, a general learning state under mastering knowledge points, an active learning state under not mastering knowledge points, a passive learning state under not mastering knowledge points, a general learning state under not mastering knowledge points, an active learning state under untested knowledge points, a passive learning state under untested knowledge points, and a general learning state under untested knowledge points; wherein the set score of the active learning state under mastering knowledge points is greater than the set score of the general learning state under mastering knowledge points, and the set score of the passive learning state under mastering knowledge points is greater than the set score of the passive learning state under mastering knowledge points, and the set score of the active ... The score set for the active learning state under the unmastered knowledge point is greater than zero, for example, the scores are set to 3, 2 and 1 respectively; the set score for the active learning state under the unmastered knowledge point is greater than the set score for the general learning state under the unmastered knowledge point, and is greater than the set score for the passive learning state under the unmastered knowledge point, and both are less than zero, for example, the scores are set to -1, -2 and -3 respectively; the set integral for the active learning state under the untested knowledge point is the average score of the mastered knowledge points and the unmastered knowledge points of the corresponding student in the active learning state. For example, if the student has 9 mastered knowledge points and 1 unmastered knowledge point in the active learning state, the set integral for the active learning state under the untested knowledge point is The set score for the general learning state under the untested knowledge points is the average score of the corresponding students' mastering and not mastering the knowledge points under the general learning state, and the set score for the passive learning state under the untested knowledge points is the average score of the corresponding students' mastering and not mastering the knowledge points under the passive learning state.

[0028] According to an embodiment of the present invention, it also includes: Acquire knowledge points not mentioned in teaching; If a knowledge point not mentioned in the teaching appears in the test and the corresponding student does not get a score, the quantity value of the knowledge point not mentioned is recorded and set as the first quantity value; Multiply the first quantity value by a preset first revision weight score to obtain a first revision score; Subtract the first revised score from the total score of the student's response to the teacher's teaching to obtain the revised total score of the student's response to the teacher's teaching; If the first quantity value is greater than or equal to the preset first quantity threshold, the teaching evaluation of the corresponding teacher will be set to unqualified.

[0029] It should be noted that the learning status of knowledge points not mentioned in teaching is set to general learning status.

[0030] According to an embodiment of the present invention, the step of obtaining the teacher's teaching status evaluation score specifically includes: Obtaining metadata for historical teaching; According to the history teaching metadata, the score ratio of knowledge points in different history teaching methods is obtained; Obtain the current teacher's teaching methods, teaching characteristics, and knowledge points under the current teacher's teaching methods and teaching characteristics; Based on the same knowledge points, determine the predicted score probability value of the current teacher’s teaching method; Extract teaching sounds and teaching actions from teaching features; Extract the student learning status corresponding to the teaching sound, and set the basic score of the corresponding teaching sound according to the corresponding student learning status; If there is no teaching action at the time node of the teaching sound, the basic score of the teaching sound is set to the corresponding teaching feature score; If there is a teaching action at the time node of the teaching sound, the teaching action is compared and analyzed with the preset action to obtain the action similarity value; If the action similarity value is greater than or equal to a preset first similarity threshold, the corresponding teaching action is set as a violation action and a violation score is generated; If the action similarity value is less than the preset first similarity threshold, the corresponding teaching action is set as an auxiliary action and an action additional score is generated; The basic score of the teaching sound is added to the additional score of the action or subtracted from the violation score to obtain the corresponding teaching feature score; The teaching feature score is set as the set score of the corresponding knowledge point; Based on the same knowledge point, multiply the set score of the corresponding knowledge point by the predicted score probability value of the current teacher's teaching method to obtain the predicted score value of the corresponding knowledge point; Get all the knowledge points contained in this teaching; Accumulate the predicted scores of the knowledge points in this teaching to get the predicted total score; Accumulate the set scores of all knowledge points in this teaching to get the ideal score value in this teaching; Divide the predicted total score by the ideal score in this teaching, and then multiply it by the preset benchmark evaluation score to get the current teacher's teaching status evaluation score.

[0031] It should be noted that the predicted score probability value of the current teacher's teaching method is the score ratio value of the corresponding teaching method in history teaching. For example, in the history teaching of knowledge point 1, there are teaching methods a and teaching methods b. If the score of knowledge point 1 after history teaching is 10, of which 6 times are teaching method a and 4 times are teaching method b, then the score ratio value of teaching method a is 0.6, and the score ratio value of teaching method b is 0.4. If the current teacher teaches knowledge point 1 through teaching method a, the predicted score probability value of the corresponding teacher's teaching method is 0.6; students are actively learning The basic score of the teaching sound set in the positive state is greater than the basic score of the teaching sound set in the general learning state, which is greater than the basic score of the teaching sound set in the negative learning state, for example, they are set to 1, 0 and (-1) respectively; the preset actions include illegal actions such as knocking on the table and knocking on the students. When the action similarity value is greater than or equal to the preset first similarity threshold, the corresponding teacher's teaching action is set as an illegal action. For example, if the teaching feature score of knowledge point 2 is 1, the set score of the corresponding knowledge point 2 is set to 1; if the teacher's teaching evaluation is based on a percentage system, the corresponding preset benchmark evaluation score is 100.

[0032] According to an embodiment of the present invention, it also includes: Subtract the teacher's teaching status evaluation score from any student's evaluation score of the teacher's teaching to obtain the first score difference; If the first score difference is greater than the preset first score difference threshold, the corresponding student's evaluation score for the teacher's teaching will be deleted, and the first average student's evaluation score for the teacher's teaching will be recalculated, and the number of the deleted student's evaluation scores for the teacher's teaching will be obtained and set as the second number value; If the second quantity value is greater than or equal to the preset second quantity threshold, the corresponding teacher's teaching will be evaluated as unqualified.

[0033] It should be noted that the first score difference is determined by presetting the first score difference threshold, thereby eliminating errors in individual students' evaluation of teachers.

[0034] Figure 3 A block diagram of a multimodal teaching evaluation system based on artificial intelligence of the present invention is shown.

[0035] like Figure 3 As shown, the second aspect of the present invention provides a multimodal teaching evaluation system 3 based on artificial intelligence, including a memory 31 and a processor 32, wherein the memory stores a multimodal teaching evaluation method program based on artificial intelligence, and when the multimodal teaching evaluation method program based on artificial intelligence is executed by the processor, the following steps are implemented: Get teaching metadata; Classify the teaching metadata to obtain multiple teaching sub-data, and build corresponding teaching sub-data models based on the teaching sub-data; Based on the teaching metadata and teaching sub-data model, a school-level smart teaching cockpit is constructed, and teaching driving data is obtained; Extract all teaching data of each student and teacher in the teaching driving data to form a data portrait of a single student's response to the teacher's teaching; Based on the data portrait of individual students' responses to the teacher's teaching, the corresponding student's evaluation score on the teacher's teaching is obtained; Calculate the average of all students' evaluation scores on the teacher's teaching to obtain the first average evaluation score of the teacher's teaching; Get the teacher's teaching status evaluation score; The first average evaluation score of the teacher's teaching and the teacher's teaching status evaluation score are averaged to obtain the second average evaluation score of the teacher's teaching; The teacher's teaching is evaluated based on the preset evaluation score range within which the second average evaluation score of the teacher's teaching falls.

[0036] In this solution, the step of forming a data profile of a single student's response to the teacher's teaching specifically includes: Obtain the knowledge points contained in the textbook in the teaching metadata; All teaching data of individual students and teachers are sliced ​​and sorted according to knowledge points to obtain the learning status under each knowledge point; Construct the sub-puzzle of the corresponding knowledge point according to the learning status of each knowledge point; Combine the sub-puzzles of the knowledge point to obtain the puzzle of the corresponding knowledge point; Combine the puzzles of different knowledge points to obtain a data portrait of a single student's response to the teacher's teaching; The types of knowledge points include mastered knowledge points, unmastered knowledge points and untested knowledge points; and the learning status includes active learning status, passive learning status and general learning status.

[0037] In this solution, the step of obtaining the learning status at each knowledge point includes: Extract the time nodes when knowledge points appear in the teacher's teaching process and set them as the time nodes when knowledge points appear; Extract the status characteristics of students at the time nodes where knowledge points appear; Classify the state characteristics of students at the time nodes where knowledge points appear, and record the number of times the state characteristics appear under each state type; Accumulate the number of occurrences of state features under all state types to obtain the total number of occurrences; Divide the number of times the state feature appears under the state type by the total number of times to obtain the probability value of the corresponding state type appearing at the knowledge point time node; If the probability value of the corresponding state type at the knowledge point appearance time node is greater than or equal to the preset first probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the difference between the probability value of the corresponding state type at the knowledge point appearance time node and the probability value of any other state type at the knowledge point appearance time node is greater than or equal to the preset second probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the probability value of the corresponding state type at the time node when the knowledge point appears is less than the preset first probability threshold, and the difference between the probability value of the corresponding state type at the time node when the knowledge point appears and the probability value of any other state type at the time node when the knowledge point appears is less than the preset second probability threshold, then when the knowledge point is a mastered knowledge point, the active learning state is set to the current learning state; when the knowledge point is an unmastered knowledge point, the passive learning state is set to the current learning state; when the knowledge point is an untested knowledge point, the general learning state is set to the current learning state.

[0038] The present invention discloses an artificial intelligence-based multimodal teaching evaluation method and system, which compares and analyzes teachers' teaching methods, teaching characteristics, and students' learning status. It is not limited to evaluation methods aimed at scoring, but quantifies and evaluates from multiple dimensions, thereby further optimizing teaching methods, better promoting the professional growth of teachers and students and the improvement of teaching methods, and thus promoting the improvement of education quality.

[0039] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0040] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0041] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0042] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0043] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

Claims

1. A multimodal teaching evaluation method based on artificial intelligence, characterized in that: include: Get teaching metadata; Classify the teaching metadata to obtain multiple teaching sub-data, and build corresponding teaching sub-data models based on the teaching sub-data; Based on the teaching metadata and teaching sub-data model, a school-level smart teaching cockpit is constructed, and teaching driving data is obtained; Extract all teaching data of each student and teacher in the teaching driving data to form a data portrait of a single student's response to the teacher's teaching; Based on the data portrait of individual students' responses to the teacher's teaching, the corresponding student's evaluation score on the teacher's teaching is obtained; Calculate the average of all students' evaluation scores on the teacher's teaching to obtain the first average evaluation score of the teacher's teaching; Get the teacher's teaching status evaluation score; The first average evaluation score of the teacher's teaching and the teacher's teaching status evaluation score are averaged to obtain the second average evaluation score of the teacher's teaching; The teacher's teaching is evaluated based on the preset evaluation score range within which the second average evaluation score of the teacher's teaching falls.

2. According to the multimodal teaching evaluation method based on artificial intelligence in claim 1, it is characterized in that: The step of forming a data portrait of a single student's response to the teacher's teaching specifically includes: Obtain the knowledge points contained in the textbook in the teaching metadata; All teaching data of individual students and teachers are sliced ​​and sorted according to knowledge points to obtain the learning status under each knowledge point; Construct the sub-puzzle of the corresponding knowledge point according to the learning status of each knowledge point; Combine sub-puzzles of knowledge points of the same type to obtain puzzles of the corresponding knowledge point type; Combine all puzzle pieces of knowledge point types to obtain a data portrait of a single student's response to the teacher's teaching; The types of knowledge points include mastered knowledge points, unmastered knowledge points and untested knowledge points; and the learning status includes active learning status, passive learning status and general learning status.

3. The multimodal teaching evaluation method based on artificial intelligence according to claim 2 is characterized in that: The step of obtaining the learning status at each knowledge point includes: Extract the time nodes when knowledge points appear in the teacher's teaching process and set them as the time nodes when knowledge points appear; Extract the status characteristics of students at the time nodes where knowledge points appear; Classify the state characteristics of students at the time nodes where knowledge points appear, and record the number of times the state characteristics appear under each state type; Accumulate the number of occurrences of state features under all state types to obtain the total number of occurrences; Divide the number of times the state feature appears under the state type by the total number of times to obtain the probability value of the corresponding state type appearing at the knowledge point time node; If the probability value of the corresponding state type at the knowledge point appearance time node is greater than or equal to the preset first probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the difference between the probability value of the corresponding state type at the knowledge point appearance time node and the probability value of any other state type at the knowledge point appearance time node is greater than or equal to the preset second probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the probability value of the corresponding state type at the time node when the knowledge point appears is less than the preset first probability threshold, and the difference between the probability value of the corresponding state type at the time node when the knowledge point appears and the probability value of any other state type at the time node when the knowledge point appears is less than the preset second probability threshold, then when the knowledge point is a mastered knowledge point, the active learning state is set to the current learning state; when the knowledge point is an unmastered knowledge point, the passive learning state is set to the current learning state; when the knowledge point is an untested knowledge point, the general learning state is set to the current learning state.

4. The multimodal teaching evaluation method based on artificial intelligence according to claim 1 is characterized in that: The step of obtaining the evaluation score of the teacher's teaching by the corresponding student based on the data portrait of the individual student's response to the teacher's teaching also includes: Set different scores for different sub-puzzles, and add up the scores of all sub-puzzles to get the total score of the student's response to the teacher's teaching; Get the ideal total score of students' response to teachers' teaching; Divide the total score of the student's response to the teacher's teaching by the ideal total score of the student's response to the teacher's teaching, and then multiply it by the preset benchmark evaluation score to obtain the student's evaluation score for the teacher's teaching.

5. The multimodal teaching evaluation method based on artificial intelligence according to claim 4 is characterized in that: Also includes: Acquire knowledge points not mentioned in teaching; If a knowledge point not mentioned in the teaching appears in the test and the corresponding student does not get a score, the quantity value of the knowledge point not mentioned is recorded and set as the first quantity value; Multiply the first quantity value by a preset first revision weight score to obtain a first revision score; Subtract the first revised score from the total score of the student's response to the teacher's teaching to obtain the revised total score of the student's response to the teacher's teaching; If the first quantity value is greater than or equal to the preset first quantity threshold, the teaching evaluation of the corresponding teacher will be set to unqualified.

6. The multimodal teaching evaluation method based on artificial intelligence according to claim 1 is characterized in that: The step of obtaining the teacher's teaching status evaluation score specifically includes: Obtaining metadata for historical teaching; According to the history teaching metadata, the score ratio of knowledge points in different history teaching methods is obtained; Obtain the current teacher's teaching methods, teaching characteristics, and knowledge points under the current teacher's teaching methods and teaching characteristics; Based on the same knowledge points, determine the predicted score probability value of the current teacher’s teaching method; Extract teaching sounds and teaching actions from teaching features; Extract the student learning status corresponding to the teaching sound, and set the basic score of the corresponding teaching sound according to the corresponding student learning status; If there is no teaching action at the time node of the teaching sound, the basic score of the teaching sound is set to the corresponding teaching feature score; If there is a teaching action at the time node of the teaching sound, the teaching action is compared and analyzed with the preset action to obtain the action similarity value; If the action similarity value is greater than or equal to a preset first similarity threshold, the corresponding teaching action is set as a violation action and a violation score is generated; If the action similarity value is less than the preset first similarity threshold, the corresponding teaching action is set as an auxiliary action and an action additional score is generated; The basic score of the teaching sound is added to the additional score of the action or subtracted from the violation score to obtain the corresponding teaching feature score; The teaching feature score is set as the set score of the corresponding knowledge point; Based on the same knowledge point, multiply the set score of the corresponding knowledge point by the predicted score probability value of the current teacher's teaching method to obtain the predicted score value of the corresponding knowledge point; Get all the knowledge points contained in this teaching; Accumulate the predicted scores of the knowledge points in this teaching to get the predicted total score; Accumulate the set scores of all knowledge points in this teaching to get the ideal score value in this teaching; Divide the predicted total score by the ideal score in this teaching, and then multiply it by the preset benchmark evaluation score to get the current teacher's teaching status evaluation score.

7. The multimodal teaching evaluation method based on artificial intelligence according to claim 1 is characterized in that: Also includes: Subtract the teacher's teaching status evaluation score from any student's evaluation score of the teacher's teaching to obtain the first score difference; If the first score difference is greater than the preset first score difference threshold, the corresponding student's evaluation score for the teacher's teaching will be deleted, and the first average student's evaluation score for the teacher's teaching will be recalculated, and the number of the deleted student's evaluation scores for the teacher's teaching will be obtained and set as the second number value; If the second quantity value is greater than or equal to the preset second quantity threshold, the corresponding teacher's teaching will be evaluated as unqualified.

8. A multimodal teaching evaluation system based on artificial intelligence, characterized in that: The invention comprises a memory and a processor, wherein the memory stores a multimodal teaching evaluation method program based on artificial intelligence, and when the multimodal teaching evaluation method program based on artificial intelligence is executed by the processor, the following steps are implemented: Get teaching metadata; Classify the teaching metadata to obtain multiple teaching sub-data, and build corresponding teaching sub-data models based on the teaching sub-data; Based on the teaching metadata and teaching sub-data model, a school-level smart teaching cockpit is constructed, and teaching driving data is obtained; Extract all teaching data of each student and teacher in the teaching driving data to form a data portrait of a single student's response to the teacher's teaching; Based on the data portrait of individual students' responses to the teacher's teaching, the corresponding student's evaluation score on the teacher's teaching is obtained; Calculate the average of all students' evaluation scores on the teacher's teaching to obtain the first average evaluation score of the teacher's teaching; Get the teacher's teaching status evaluation score; The first average evaluation score of the teacher's teaching and the teacher's teaching status evaluation score are averaged to obtain the second average evaluation score of the teacher's teaching; The teacher's teaching is evaluated based on the preset evaluation score range within which the second average evaluation score of the teacher's teaching falls.

9. The multimodal teaching evaluation system based on artificial intelligence according to claim 8 is characterized in that: The step of forming a data portrait of a single student's response to the teacher's teaching specifically includes: Obtain the knowledge points contained in the textbook in the teaching metadata; All teaching data of individual students and teachers are sliced ​​and sorted according to knowledge points to obtain the learning status under each knowledge point; Construct the sub-puzzle of the corresponding knowledge point according to the learning status of each knowledge point; Combine sub-puzzles of knowledge points of the same type to obtain puzzles of the corresponding knowledge point type; Combine all puzzle pieces of knowledge point types to obtain a data portrait of a single student's response to the teacher's teaching; The types of knowledge points include mastered knowledge points, unmastered knowledge points and untested knowledge points; and the learning status includes active learning status, passive learning status and general learning status.

10. The multimodal teaching evaluation system based on artificial intelligence according to claim 9 is characterized in that: The step of obtaining the learning status at each knowledge point includes: Extract the time nodes when knowledge points appear in the teacher's teaching process and set them as the time nodes when knowledge points appear; Extract the status characteristics of students at the time nodes where knowledge points appear; Classify the state characteristics of students at the time nodes where knowledge points appear, and record the number of times the state characteristics appear under each state type; Accumulate the number of occurrences of state features under all state types to obtain the total number of occurrences; Divide the number of times the state feature appears under the state type by the total number of times to obtain the probability value of the corresponding state type appearing at the knowledge point time node; If the probability value of the corresponding state type at the knowledge point appearance time node is greater than or equal to the preset first probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the difference between the probability value of the corresponding state type at the knowledge point appearance time node and the probability value of any other state type at the knowledge point appearance time node is greater than or equal to the preset second probability threshold, the corresponding state type is set to the learning state under the corresponding knowledge point; If the probability value of the corresponding state type at the time node when the knowledge point appears is less than the preset first probability threshold, and the difference between the probability value of the corresponding state type at the time node when the knowledge point appears and the probability value of any other state type at the time node when the knowledge point appears is less than the preset second probability threshold, then when the knowledge point is a mastered knowledge point, the active learning state is set to the current learning state; when the knowledge point is an unmastered knowledge point, the passive learning state is set to the current learning state; when the knowledge point is an untested knowledge point, the general learning state is set to the current learning state.

Images