JP2020187713A - Ai-teacher - Google Patents

Abstract

To solve a problem that, in a constitution of a lesson in a school education, a genus element depending on a teacher is strong, thus causing the generation of a large number of problems by the teachers resulting from a concentration of a right, in particular, a lesson evaluation of "an interest, an intension and an attitude" and "a thinking, a determination and an expression" in the lesson is most liable to be varied depending on the teachers, and unfairness generated at schoolchildren and guardians is also strong because the evaluation is surely executed by human beings, on the other hand, in a conventional education, it is mandatory for the teacher to grasp and evaluate learning activities of the schoolchildren one by one in parallel with the progress of the lesson, however, it is impossible for one teacher to simultaneously perform all the processing in terms of human resources.SOLUTION: Since the work of a teacher is manualized in fact, an evaluation of school children which should be performed by a school teacher is substituted by an AI-teacher using artificial intelligence. A sound collector is introduced into a school lecture room. Voiceprints of the school children and the teachers are registered in advance, and minutes of utterances are created by voice recognition. Feelings which can be estimated from the number of utterances and voices at a voice analysis are recorded, and a distribution to a lecture participation and an argument is digitalized by natural language processing. Score evaluations by the digitalized lessen data of the school children and feelings of the teachers are accumulated as big data, and the big data are used as training data which are necessary for a determination process of the artificial intelligence. The accuracy of the analogical reasoning of the artificial intelligence is enhanced, and made to further approximate an expert determination which is performed by the teacher, and a genus evaluation method depending on a teacher individual is changed to a standardized and equal evaluation method. By automating the evaluation of the school children, human resources of the school teachers are saved.SELECTED DRAWING: None

Description

Electricity / electronics

School teachers perform by using voice analysis (minute minutes creation, voiceprint authentication, emotion grasping), natural language processing (language understanding), and professional judgment (analogous processing from training data) made possible by the development of artificial intelligence. The role is replaced by an artificial intelligence evaluation processing device.

In school education, the maximum number of students per teacher is 40. Because one teacher handles a large number of students, the lessons are not wide and learning activities are fixed. It is necessary to secure the number of teachers in order to realize desirable education for students, but human resources are limited due to the budget available by the national and local governments. Therefore, it is difficult to realize educational reform that suits the changing times.
In conventional education, teachers are obliged to grasp and evaluate the learning activities of each student in parallel with the progress of the lesson. However, based on my own experience as a teacher, I conclude that it is impossible for each teacher to grasp the learning activities of each student and evaluate the learning activities individually while proceeding with the lesson, considering human resources. It was.
Due to the number of class hours, there are very few opportunities for students to express their intentions in the current class structure, and teachers cannot grasp the learning activities of each student. Therefore, "interest, motivation, attitude" and "thinking, judgment, expression", which are the viewpoints of evaluation, are evaluated by analogy with the results of the paper test, which appears as clear numbers in the measures against complaints from students, children and parents. .. However, the paper test is not originally a material for determining the evaluation of "interest, motivation, attitude" and "thinking, judgment, expression" in the lesson subject.
From the viewpoint of evaluation conducted throughout all subjects, there is a strong personal element that the evaluation fluctuates greatly depending on the teacher except for "knowledge / understanding" by the paper test, and it is not possible because the evaluation is performed by human beings from the viewpoint of students, children and parents. In many cases, a sense of fairness is likely to occur and it is not convincing. However, since the authority to evaluate grades is left to the faculty members, it is difficult to disagree, and there are many cases in which the faculty members themselves become a problem due to the concentration of authority. Furthermore, it is extremely difficult for any teacher to evaluate "morality" and "patriotism" in the moral subject that was the subject of grade evaluation as a special subject.
Evaluation materials are needed to determine the evaluation, and with the current class structure, there are few opportunities for students to speak, and it is difficult to obtain manifestations of the students' intentions. In addition, the fact that there are only the number of complaints made by parents as materials for evaluating not only the grades of students and children but also the abilities of teachers is a big problem that distorts school education.
Therefore, in order to realize educational reform, we think that tools for solving problems that have arisen in conventional education are needed. This proposal is an invention for reducing the burden on teachers and disseminating high-quality school education at a low price compared to labor costs to society.

The means to solve the problem is to change from a personal evaluation method that depends on the individual teacher who is the evaluator to an objective evaluation method that is standardized by artificial intelligence.
Introduce a sound collector to the school classroom. Voiceprint authentication of students and teachers is performed in advance, and minutes of remarks during class are created by voice recognition.
During voice analysis, the number of remarks and emotions that can be inferred from the voice are recorded, and the contribution to class participation and discussion is quantified by natural language processing.
The digitized lesson data of students and the grade evaluation by the teacher's sense are converted into big data and used as training data of artificial intelligence.
The accuracy of artificial intelligence analogy processing will be brought closer to the professional judgment made by teachers, and the evaluation method will be changed from personal to standardized.
The reason why parents make complaints to teachers is that they can only know the unclear grade evaluation criteria and the actual living conditions of children entrusted to the school with only a few lines at the time of grade notification three times a year. By automatically detecting abnormal words and behaviors when creating the minutes of remarks by voice recognition, problematic remarks by teachers and the actual living conditions of students and children are automatically reported. With the introduction of AI teachers, we will reform the school situation that is closed from society and guarantee the human rights of students and children.
From the current class structure, it is difficult to obtain data for grade judgment because the number of times students speak is limited. Therefore, students can answer easily during class, use a button-type input function to vote during breaks that teachers do not control, and use the same voice recognition technology as the sound collector to individually speak the speaker's voice. Distribute a small remote control device with a function to input.

A sound collector will be installed in the school classroom, and voiceprint authentication data will be prepared so that individuals can be identified. Record student and teacher remarks by voice recognition.
Understand what you said during class by natural language processing. Participation in class and contribution to discussion are measured from the number of remarks, emotions, and remarks. The faculty member evaluates the grades using the measured data as a reference value. By using the big data accumulated by each school and local government that introduced the AI teacher as training data, the accuracy of professional judgment will be closer to that of a real teacher. In the future, AI teacher will automatically judge and execute the evaluation.

Claims (1)

A device that standardizes the evaluation method of students, children and teachers in school education and automates grade evaluation using voice recognition, natural language understanding, and analogy processing by artificial intelligence.