Non-cognitive skills improvement system

Abstract

Easily obtain new activity proposals to improve non-cognitive skills. [Solution] When an episode record is input, the photos to be used are confirmed and stored (S11-S17). The language generation AI processing system 42 is instructed to estimate the interest from the episode record and the tags of the selected images (S18), and determines whether the instructed creation policy is "deep dive," "diversity," or "reconsideration" (S20). If the policy is "deep dive," the prompt "deep dive" is read out (S21), if it is "diversity," the prompt "diversity" is read out (S23), and if it is reconsideration, the prompt "reconsideration" is read out (S25). This is given to the language generation AI processing system 42 as an activity proposal creation prompt that incorporates the estimated interest. When a new proposal is given by the system, the result is stored and displayed on the screen. This makes it easy to obtain new proposals based on the creation policies of "deep dive," "diversity," and "reconsideration."

Description

Technical Field

[0001] This invention relates to a non-cognitive ability improvement system, and particularly to the proposal of childcare activities that can arouse the interest of infants and young children.

Background Art

[0002] "Non-cognitive skills" refer to abilities that have an important impact on daily life and social activities, such as ways of thinking, attitudes towards dealing with things, and behaviors. It is said that the foundation is cultivated from the age of 0 and develops particularly during the "infant and toddler period" from the age of 1 to 5 or 6 years old. Therefore, in developmental psychology, the importance of nurturing non-cognitive abilities through childcare is recognized. The inventor thought that if the points that infants and young children are interested in could be objectively found, it might improve non-cognitive abilities.

[0003] In Patent Document 1, in order to improve non-cognitive abilities, a system is disclosed that provides support for giving children internal stimuli and a sense of achievement, and improving self-control by allowing children to input plans by themselves and displaying the achievement information.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Since the system of Patent Document 1 aims to improve non-cognitive abilities by planning by oneself and achieving it, it cannot be applied to children who are not old enough to plan by themselves.

[0006] In addition, the ability to objectively analyze that such children are interested is not limited to infants and young children, and it also becomes a problem even when the age is older.

[0007] Incidentally, while it's now known that it's possible to ask questions to language-generating AI like ChatGPT (trademark) and receive answers, there has been no research into what kinds of questions should be asked to objectively identify the points of interest for infants and toddlers, as described above.

[0008] This invention aims to solve the above problems and provide a non-cognitive ability enhancement system that uses a language generation AI to improve the non-cognitive abilities of each infant. [Means for solving the problem]

[0009] (1) The non-cognitive ability improvement system according to the present invention comprises: an imaging means for imaging the childcare worker's childcare activities for infants and toddlers; an image data storage means for storing the imaging data captured by the imaging means as image data; an annotation means for adding annotations to the image data that include at least one subject or an object related to the subject; an image identification means for identifying an image that matches the childcare activity when given an observation record of the childcare activity, based on annotations added to the image data stored in the image data storage means; a suggestion request means for generating a suggestion request prompt to cause a language generation AI processing system to generate a new activity suggestion regarding the childcare activity, and providing the generated suggestion request prompt to the language generation AI processing system; a language generation AI processing system having a large-scale language model, which outputs a response in response to the suggestion request prompt; and a suggestion storage means for receiving and storing the response from the language generation AI processing system, wherein the generated prompt is a prompt that causes the language generation AI processing system to generate a new activity suggestion regarding the childcare activity using the observation record and the annotations of the extracted image. This will enable any childcare provider to easily propose new activities that improve non-cognitive skills.

[0010] (2) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and comprises: an image data storage means that stores as annotated image data an image of the caregiver's childcare activities for infants and toddlers and annotations relating to at least one subject or an object related to said subject included in the image data; an image identification means that, when an observation record of the childcare activity is given, identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means; and a suggestion request means that generates a suggestion request prompt that causes the language generation AI processing system to generate a new activity suggestion relating to the childcare activity, and provides the generated prompt to the language generation AI processing system, wherein the generated prompt is a prompt that causes the language generation AI processing system to generate a new activity suggestion relating to the childcare activity using the observation record and the annotations of the extracted image.Therefore, the caregiver can obtain a new activity suggestion relating to the childcare activity simply by creating the observation record.This makes it possible for any caregiver to easily propose new activities that improve non-cognitive abilities.

[0011] (3) In the non-cognitive ability improvement system according to the present invention, the image identification means presents candidate image data from the image data storage means, and when a selection instruction is given, it performs the identification. Therefore, the caregiver selects image data that matches the observation record.

[0012] (4) In the non-cognitive ability improvement system according to the present invention, the proposed new activity regarding the status of childcare activities is a proposed modification to the childcare activity in the observation record, which is considered to be an already performed childcare activity, and is a proposed modification that is thought to be of greater interest or attention to the infant or other infant. Therefore, the caregiver can obtain a new proposed modification that is thought to be of greater interest to the infant or other infant simply by creating the observation record.

[0013] (5) In the non-cognitive ability improvement system according to the present invention, the language generation AI processing system generates a proposal in response to a proposal request prompt given to the language generation AI processing system, based on one of the following creation policies: 1) Deepening the exploration of a theme that a child has shown interest in during the childcare activity by changing the approach and exploring it in more depth to deepen the fundamental understanding; 2) Diversifying the theme to a higher level, and expanding the scope of interest by selecting and exploring new activity targets within that broader concept; or 3) Reconsidering the theme by re-implementing it. Therefore, it is possible to obtain proposals based on the creation policies of Deepening, Diversifying, or Reconsidering.

[0014] (6) In the non-cognitive ability improvement system according to the present invention, the proposal request prompt requests new activity proposals from experts in early childhood education and developmental psychology, and from teachers who act as supporters in the field of early childhood education, and includes instructions to create proposals for activities and environments that enable each infant to act, learn, or grow independently, based on the interests and concerns of each infant. Therefore, new activity proposals can be obtained from the aforementioned perspectives.

[0015] (7) In the non-cognitive ability improvement system according to the present invention, the generation format of the proposal request prompt is classified into multiple items, and the multiple classifications include at least two items from themes, learning points, background, objectives, and preparation / environmental design. Therefore, a new proposal for an easy-to-understand format can be obtained.

[0016] (8) In the non-cognitive ability improvement system according to the present invention, the generation format of the suggestion request prompt is classified into multiple items, and the multiple classifications include at least one or more of the following: adult perspectives and interactions, adult questions, and children's questions. Therefore, it is possible to obtain new suggestions that include the above classifications.

[0017] (9) In the non-cognitive ability improvement system according to the present invention, the prompts generated by the suggestion request means are an interest acquisition prompt that estimates the interest or concern of the infant in the childcare activity situation using the observation record and the annotations of the extracted images, and a suggestion prompt that generates a new activity suggestion regarding the childcare activity situation using the interest or concern of the infant given by the language generation AI processing system through this interest acquisition prompt. Therefore, along with the new activity suggestion, it is possible to acquire the interest and concern that estimate the interest or concern of the infant.

[0018] (10) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and comprises: an image data storage means that stores as annotated image data an image of the caregiver's childcare activities for infants and toddlers, and annotations relating to at least one subject or an object related to said subject included in the image data; an image identification means that, when given an observation record of the childcare activities, identifies an image that matches the childcare activities based on annotations attached to the image data stored in the image data storage means; and an interest identification request means that gives the language generation AI processing system an interest acquisition prompt that causes the language generation AI processing system to estimate the infant's interests or concerns in the childcare activities using the observation record and the annotations of the extracted image.

[0019] (11) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, comprising: an observation record receiving means for receiving observation records of the caregiver's childcare activities for infants and toddlers; and a suggestion request means for using the observation records to generate a suggestion request prompt that causes the language generation AI processing system to generate a new activity suggestion regarding the childcare activities in the received observation records, and for providing the generated prompt to the language generation AI processing system, wherein the suggestion request prompt is a suggestion request prompt given to the language generation AI processing system from one of the following creation policies: 1) in-depth exploration of an activity that the child was interested in in the childcare activities, 2) diversification of the activity to a higher level and selection of another activity target within that general concept, or 3) reconsideration of the same theme to be implemented again.Therefore, it is possible to obtain suggestions according to the creation policies of in-depth exploration, diversification, and reconsideration.

[0020] (12) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and includes: an image data storage means that stores as annotated image data an image of the caregiver's childcare activities for infants and toddlers and annotations relating to at least one subject or an object related to said subject included in the image data; an image identification means that, when an observation record of the childcare activity is given, identifies an image that matches the childcare activity based on the annotations attached to the image data stored in the image data storage means; and a report generation request means that generates a report generation request prompt that causes the language generation AI processing system to generate a childcare activity report for a third party other than the caregiver regarding the childcare activity, and provides the generated prompt to the language generation AI processing system.Therefore, the caregiver can obtain a childcare activity report for a third party based on the observation record and image data.

[0021] (13) In the non-cognitive ability improvement system according to the present invention, the proposal storage means stores the corresponding data of the observation record and the extracted image used when generating a prompt to be generated by the language generation AI processing system, and further includes a report generation request means that generates a report generation request prompt to generate a childcare activity report to the third party, and gives the language generation AI processing system a prompt to generate a childcare activity report to a third party other than the childcare worker regarding the status of the childcare activities, and when the observation record is given, the report generation request means generates the report generation request prompt using the corresponding data and gives the generated prompt to the language generation AI processing system.Therefore, the childcare activity report based on the image data at the time of proposal creation can be generated.

[0022] (14) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, comprising: a reference knowledge data storage means that stores image data of the caregiver's childcare activities for infants and toddlers and corresponding data of the childcare activities in the image data; an image data storage means that stores image data of the caregiver's childcare activities for infants and toddlers; and a suggestion request means that uses the image data stored in the image data storage means as input data to generate a suggestion request prompt for the language generation AI processing system to generate an activity report regarding the childcare activities in the input data, and provides the generated prompt to the language generation AI processing system, wherein the generated prompt is a suggestion request prompt given to the language generation AI processing system based on the corresponding data of the childcare activities stored in the reference knowledge data storage means, from one of the following creation policies: 1) in-depth exploration of activities that the child showed interest in during the childcare activities; 2) diversification of conceptualizing the activity and selecting another activity target within that general concept; or 3) reconsideration of repeating the same theme. Therefore, the caregiver can obtain proposals based on the aforementioned guidelines for in-depth analysis, diversity, and reconsideration.

[0023] (15) In the non-cognitive ability improvement system according to the present invention, the image data includes corresponding audio data, and further comprises text data generation means for generating text data from the audio data, and the proposal request means generates the proposal request prompt using the text data of the corresponding image data in addition to the observation record and the annotations of the extracted image. Therefore, the proposal can be generated using the text data generated from the audio data.

[0024] In the non-cognitive ability improvement system according to the present invention, the image data includes corresponding audio data, and further includes text data generation means for generating text data from the audio data. When the report generation request means refers to the corresponding data of the childcare activity status stored in the reference knowledge data storage means, it uses the text data of the image data given as the input data to generate the report generation request prompt. The report can be further generated by using the text data generated from the audio data.

[0025] (17) In the non-cognitive ability improvement system according to the present invention, the generated report generated by the language generation AI processing system is an individual report including the interests or concerns of each infant. There is individual transition output means for providing an individual transition output prompt to the language generation AI processing system, which, when an ID of an infant is given, extracts the interests or concerns of the infant from the individual report indicated by the ID and outputs the time-series change of the interests and concerns of the infant specified by the ID as individual transition data. Therefore, each of the infants can obtain the change of their own individual report.

[0026] (18) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and comprises: an image data storage means that stores as annotated image data an image of the educational activity situation of the student in the educational setting and annotations relating to at least one student or an object related to the student included in the image data; an image identification means that, when an educational record by the educator in the educational activity situation is given, identifies an image that matches the educational activity situation based on annotations attached to the image data stored in the image data storage means; and an interest identification request means that gives the language generation AI processing system an interest acquisition prompt that causes the language generation AI processing system to estimate the student's interests or concerns in the educational activity situation using the educational record and the annotations of the extracted image.Therefore, the educator can obtain the student's interests or concerns.This makes it possible for any educator to easily provide education that improves non-cognitive abilities based on the student's interests or concerns.

[0027] (19) The non-cognitive ability improvement method according to the present invention is a non-cognitive ability improvement method for connecting a computer with a language generation AI processing system having a large language model that outputs an answer according to the given proposal request prompt, and improving non-cognitive ability by the computer. The method includes an image data storage step of imaging the childcare activity status of a caregiver for an infant and storing, as annotated image data, the imaging data and annotations related to at least one subject or an object related to the subject included in the imaging image data; an image identification step of identifying an image that matches the childcare activity status based on the annotations given to the stored image data when an observation record in the childcare activity status is given; and a proposal request step of generating a proposal request prompt for generating a new activity proposal related to the childcare activity status in the language generation AI processing system and giving the generated prompt to the language generation AI processing system. The generated prompt is a prompt for causing the language generation AI processing system to generate a new activity proposal related to the childcare activity status using the observation record and the annotations of the extracted image. Therefore, the caregiver can obtain a new activity proposal related to the childcare activity status only by creating the observation record. As a result, any caregiver can easily obtain a new activity proposal for improving non-cognitive ability.

[0028] (20) The non-cognitive ability improvement program according to the present invention comprises a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and a non-cognitive ability improvement program that causes a connected computer to function as a non-cognitive ability improvement system, and performs the following steps: an image data storage step of storing as image data an image of the caregiver's childcare activities for infants and toddlers, and annotations relating to at least one subject or an object related to said subject included in the image data annotated image data; an image identification step of identifying an image that matches the childcare activity when an observation record of the childcare activity is given, based on the annotations attached to the stored image data; and a suggestion request step of generating a suggestion request prompt that causes the language generation AI processing system to generate a new activity suggestion relating to the childcare activity, and providing the generated prompt to the language generation AI processing system, wherein the generated prompt is a prompt that causes the language generation AI processing system to generate a new activity suggestion relating to the childcare activity using the observation record and the annotations of the extracted image.Therefore, the caregiver can obtain a new activity suggestion relating to the childcare activity simply by creating the observation record. This will enable any childcare provider to easily propose new activities that improve non-cognitive skills.

[0029] (21) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and comprises: an image data storage means that stores as annotated image data an image of the caregiver's childcare activities for infants and toddlers and annotations relating to at least one subject or an object related to said subject included in the image data; an image identification means that, when an observation record of the childcare activity is given, identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means; and an instruction means that gives instructions to the language generation AI processing system to generate a new activity suggestion relating to the childcare activity using the observation record and the annotations of the extracted image.Therefore, the caregiver can obtain a new activity suggestion relating to the childcare activity simply by creating the observation record.This makes it possible for any caregiver to easily propose a new activity that improves non-cognitive abilities.

[0030] (22) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and includes an image data storage means that stores imaging data of the caregiver's childcare activities for infants and toddlers, and annotations relating to at least one subject or an object related to said subject included in the imaging data as annotated image data, an image identification means that, when given observation records of the childcare activities, identifies an image that matches the childcare activities based on annotations attached to the image data stored in the image data storage means, and the observation records and The system includes: an interest identification request means that gives the language generation AI processing system an interest acquisition prompt to estimate the interests or concerns of the infants in the childcare activity using the annotations of the extracted images; an estimated interest report storage means that stores the estimated interests or concerns of the infants as an estimated interest report for each infant; and an individual change data generation request means that, when given an infant's ID, reads the history of the estimated interest report of the infant identified by that ID and gives the language generation AI processing system an individual change data acquisition prompt to output the temporal changes in the interests or concerns of the infant identified by that ID as individual change data. Therefore, each infant can acquire the changes in their own individual report.

[0031] (23) The non-cognitive ability improvement system according to the present invention is a non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, and includes: an image data storage means that stores annotations relating to at least one subject or an object related to the subject, which are included in imaging data of the caregiver's childcare activities for infants and toddlers as annotated image data; an observation record storage means that stores a plurality of observation records of the childcare activities, which include the ID of an infant or toddler appearing in the observation record; and when an ID of an infant or toddler is given, the system extracts from the plurality of observation records the observation record in which the given ID of the infant or toddler appears. The system includes an interest identification request means that outputs the observation record, extracts annotations from the annotated image data of the infants appearing in the observation record, and gives the language generation AI processing system an interest acquisition prompt to estimate the infant's interests or concerns in the childcare activity using the observation record and the extracted annotations; and an individual change data generation request means that stores the infant's interests or concerns as an estimated interest report, and gives the language generation AI processing system an individual change data acquisition prompt to output the temporal changes in the infant's interests or concerns as individual change data using multiple estimated interest reports for the infant.Therefore, each infant can acquire the changes in their own individual report.

[0032] This section explains the terminology used in the patent claims.

[0033] The term "infant and toddler" is a concept that includes both infants and toddlers. In this embodiment, it refers to preschool children.

[0034] "Imaging the caregiver's activities for infants and toddlers" refers to imaging the caregiver's activities for infants and toddlers, and includes, for example, the infant's appearance, the objects used in the caregiving activities, the location, etc.

[0035] A "prompt" refers to a command given to an AI in a language generation AI processing system.

[0036] A "large-scale language model" is a language model constructed using a large amount of data and deep learning technology, which generates text based on a given "prompt". In this embodiment, OpenAI's GPT-4 is used, but the model is not limited to this.

[0037] "Childcare provider" refers to an entity responsible for the care or education of infants and young children. In this embodiment, this includes childcare workers, but also includes other individuals such as nursery school teachers, kindergarten teachers and their assistants, and even parents.

[0038] In this embodiment, "observation records" refer to the child's episode records. Episode records include the content of the activity, the caregiver's wishes or impressions, or the child's remarks, expressions, attitudes, or circumstances. Episode records are often written in a way that conveys the scene to someone who was not present.

[0039] "New activity proposals regarding childcare activities" are proposals generated by the language generation AI processing system 42 from "observation records" regarding childcare activities, and in the embodiment, these proposals are those shown in Figures 15, 17, and 19. [Brief explanation of the drawing]

[0040] [Figure 1] This diagram shows the overall structure (proposed new activity) of Non-Cognitive Ability Improvement System 1. [Figure 2] This diagram shows the hardware configuration of the Non-Cognitive Ability Improvement System 1. [Figure 3] This describes the data structure of various types of data. [Figure 4] This is a flowchart of the tagging process. [Figure 5] This is a flowchart for creating a proposal. [Figure 6] This is an example of a proposal creation screen. [Figure 7] Figure 7 shows an example of an episode record displayed on the screen. [Figure 8] This is an example of a prompt for estimating interests. [Figure 9] This is an example of developmental and activity data by age in months that is provided to the language generation AI processing system 42. [Figure 10] This is an example of developmental and activity data by age in months that is provided to the language generation AI processing system 42. [Figure 11] This is an example of a response using an interest generation prompt. [Figure 12] This is an example of a prompt to "deepen" the discussion based on the interests and concerns that were identified. [Figure 13] This is an example of a common format for creating proposals. [Figure 14] This is an example of a proposal creation screen. [Figure 15] Figure 14 shows an example of a proposal (in-depth analysis) displayed on the screen. [Figure 16] This is an example of a prompt to make diverse suggestions based on the interests and concerns that were gathered. [Figure 17] Figure 14 shows examples of the various proposals displayed on the screen. [Figure 18] This is an example of a prompt to "reconsider" a proposal based on the interests and concerns that were raised. [Figure 19] Figure 14 shows an example of a proposal (reconsideration) displayed on the screen. [Figure 20] This diagram shows the overall configuration (partially omitted) of the Non-Cognitive Ability Improvement System 1. [Figure 21] This is a flowchart for creating a report. [Figure 22] This is an example of a report creation screen. [Figure 23] This is an example of a prompt (class log) and a generated report. [Figure 24] This is an example of a prompt (childcare documentation) and a generated report. [Figure 25] This is an example of a prompt (individual contact log) and a generated report. [Figure 26] This is an example of a prompt for generating a record of upbringing. [Figure 27]This is an example of a prompt that "deepens" on a proposal without any interest estimation processing. [Modes for carrying out the invention]

[0041] Embodiments of the present invention will be described below with reference to the drawings.

[0042] (1. Overall structure) Figure 1 shows the configuration of the non-cognitive ability improvement system 1 according to the present invention. The non-cognitive ability improvement system 1 is used in connection with a language generation AI processing system 13 which has a large-scale language model and outputs a corresponding response when given the suggestion request prompt. The non-cognitive ability improvement system 1 comprises an imaging means 3, an image data storage means 5, an annotation request means 7, an annotation means 10, an image identification means 9, a suggestion request means 12, an observation record input means 4, an observation record storage means 8, and a suggestion storage means 14.

[0043] The imaging means 3 captures images of the childcare activities provided by the caregiver to the infant. The image data storage means 5 stores the image data captured by the imaging means 3 as image data. The annotation request means 7 provides an annotation request to the annotation means 10. The annotation means 10 adds annotations to the image data relating to at least one subject or an object related to that subject.

[0044] Observation record input means 4 receives observation records of the childcare activity. The entered observation records are stored in observation record storage means 8.

[0045] The image identification means 9 includes an extracted image display selection means 9b and an image extraction means 9a. When the image extraction means 9a is given observation records of the childcare activity, it extracts images that match the childcare activity based on annotations attached to the image data stored in the image data storage means 5. The extracted images are displayed by the extracted image display selection means 9b, and when a selection command is given, it identifies the image that matches the childcare activity. The suggestion request means 12 generates a suggestion request prompt to cause the language generation AI processing system to generate new activity suggestions related to the childcare activity, and provides the generated suggestion request prompt to the language generation AI processing system 13. When the language generation AI processing system 13 receives a new suggestion generated based on the suggestion request prompt, the suggestion storage means 14 stores the new suggestion.

[0046] Thus, the generated prompt is a prompt that causes the language generation AI processing system 13 to generate new activity suggestions regarding the childcare activity situation, using the observation record and the annotations of the extracted images.

[0047] Furthermore, the generated new activity proposals are proposed modifications to the childcare activities already performed, as recorded in the observation records, and are considered to be of greater interest to the infant or other infants. Such proposed modifications to the childcare activities are generated by the language generation AI processing system 13 in response to a proposal request prompt, based on one of the following creation policies: 1) further exploration of activities that the children showed interest in during the childcare activities; 2) conceptualization of the activity and selection of a different activity target within that general concept; or 3) reconsideration of performing the same theme again.

[0048] Furthermore, the aforementioned prompt for proposals requests new activity proposals from the perspective of a supporter of teachers in the field of early childhood education, and may include instructions to create new activity ideas that start from the interests and concerns of each infant and child, enabling them to learn independently. In addition, it may also request new activity proposals from the perspective of experts in early childhood education and developmental psychology.

[0049] Furthermore, the format for generating the solicitation prompt is categorized into multiple items, and these categories may include at least two items from the following: background, objectives, learning points, preparation, and environmental design. In addition, the solicitation prompt may include at least one or more items from the following: adult perspectives and involvement during inquiry activities, adult questions, and children's questions.

[0050] (2. Hardware configuration of Non-Cognitive Ability Improvement System 1) Using Figure 2, we will explain the hardware configuration of the non-cognitive ability improvement system 1, which is configured using a CPU.

[0051] The non-cognitive ability improvement system 1 comprises a CPU 23, memory 27, hard disk 26, input device 28, optical drive 25, monitor 30, communication unit 32, imaging unit 44, and bus line 29. The CPU 23 controls each unit via the bus line 29 according to the programs stored in the hard disk 26.

[0052] The hard disk 26 includes an operating system program 26o (hereinafter abbreviated as OS), a main program 26p, a child data storage unit 26k, an image data storage unit 26u, an episode recording data storage unit 26e, a prompt storage unit 26t, and an output data storage unit 26d.

[0053] The processing of the main program 26p will be described later.

[0054] Figure 3A shows the data structure of the child data stored in the child data storage unit 26a. In this embodiment, the child data consists of an ID, name, and facial photograph. Figure 3B shows the data structure of the image data stored in the image data storage unit 26u. In this embodiment, the image data consists of a file name (file ID), shooting date, and tags. Figure 3D shows the data structure of the episode recording data stored in the episode recording data storage unit 26e. In this embodiment, the episode recording data consists of a file name (file ID), creation date, creator, and content. Figure 3E shows the data structure of the output data stored in the output data storage unit 26k. In this embodiment, the output data consists of a file name (file ID), creation date, creation policy, and content.

[0055] In this embodiment, the output data includes new activity proposals related to existing childcare activities and childcare activity reports related to existing childcare activities. The process for generating such new activity proposals and childcare activity reports will be described later. Furthermore, the prompt memory unit 26t stores various prompts. This will be explained later.

[0056] Multiple cameras 45a to 45n located within the park are connected to the imaging processing unit 44, and the image data captured by the cameras 45a to 45n is stored in the image data storage unit 26u.

[0057] The communication unit 32 is connected to an externally installed image / language AI processing system 41 and a language AI processing system 42. When image data is provided to the image / language AI processing system 41, it tags (annotates) objects, people, places, etc., present in the image. Specifically, if "three kindergarten children are playing on a swing in the playground," it assigns tags such as "swing," "playground," and "children." In this embodiment, the image / language AI processing system 41 uses the AI ​​cloud service "Amazon Rekognition" (trademark), but is not limited to this.

[0058] Furthermore, the language AI processing system 42 generates text according to a given prompt. In this embodiment, OpenAI's GPT-4™ is used, but the system is not limited to this.

[0059] Furthermore, in this embodiment, Windows (registered trademark or trademark) is used as the operating system program (OS) 26o, but it is not limited to this.

[0060] The above programs were read from the DVD-ROM 25a containing the programs via the optical drive 25 and installed on the hard disk 26. Alternatively, programs may be installed on the hard disk from a computer-readable storage medium such as a USB memory stick. Furthermore, they may be downloaded using a communication line.

[0061] In this embodiment, the program stored on the DVD-ROM is indirectly executed by the computer by installing the program from the DVD-ROM to the hard disk 26. However, the invention is not limited to this, and the program stored on the DVD-ROM may also be executed directly from the optical drive 25. Note that executable programs by the computer include not only those that can be executed directly after installation, but also those that require conversion to another form (for example, decompressing data that has been compressed), and even those that can be executed in combination with other module parts.

[0062] (3. New Activity Proposal Creation Process) The non-cognitive ability improvement system 1 performs image data acquisition and tagging processing independently of the login process described below. This process will be explained using Figure 4. The CPU 23 determines whether or not it has received new image data (step S101). If it has received it, it stores it in the image data storage unit 26u (step S102). The CPU 23 requests the image / language AI processing system 41 to tag the image data (step S103). When the image / language AI processing system 41 receives this request, it analyzes the image data and assigns one or more tags to the image data.

[0063] The CPU 23 determines whether or not to receive the tag (step S104), and if it receives it, it appends it to the image data storage unit 26u (step S105). Here, as shown in Figure 3C, a tag is attached to the image data. In this case, the file name "2024090602.jpg" is assumed to have the tags "pomegranate", "garden", and "Tanaka Kaede" attached.

[0064] Regarding the children's names, the facial photograph data shown in Figure 3A can be referenced. Multiple such reference data sets may be stored. Furthermore, a photograph of each child's appearance may be taken upon arrival at the nursery and used as reference photograph data for that day.

[0065] Furthermore, the age of each child may be entered manually during episode recording. Alternatively, the child data storage unit 26k may store the birth date of each child, and when the name of the child in question is entered during episode recording, the birth date may be automatically read, the number of months elapsed since birth may be calculated, and the date may be automatically displayed.

[0066] The above tagging process will be executed as needed.

[0067] Next, we will explain the process of creating a new activity proposal using Figure 5.

[0068] Each childcare worker logs in using their own ID and password. The login process is the same as usual, so we will omit the explanation. The following explanation will describe the process when childcare worker "Yoshiko Fujimoto" logs in.

[0069] CPU23 displays an initial screen (not shown) (step S1). The initial screen displays the creator, creation date, etc. A button to select either the proposal creation process or the report creation process is displayed (not shown).

[0070] The creator, childcare worker "Yoshiko Fujimoto," will now select the proposal creation process. The CPU 23 has determined that the proposal creation process has been selected (step S3), and once this process is selected, it switches from the initial screen to the proposal creation screen shown in Figure 6 (step S5). On this proposal screen, below the episode record input field 61, candidate photo display field 78, and proposal creation button 63, three buttons are displayed as a menu: a "Deep Dive" button 64, a "Variation" button 65, and a "Reconsider" button 66.

[0071] Episode record input field 61 is where childcare worker "Yoshiko Fujimoto" enters episode records from today's childcare activities.

[0072] In the candidate display area 78, one or more options provided by the image / language AI processing system 41 are displayed. This will be explained later.

[0073] This section describes the "Deep Dive" button 64, the "Diverse" button 65, and the "Reconsider" button 66, which are located below the "Proposal Creation" button 63. These three buttons are used to specify the policy for creating childcare activities when instructing the language generation AI processing system 42 to create new childcare activity proposals. The "Deep Dive," "Diverse," and "Reconsider" buttons will be explained later.

[0074] Here, we will explain using the case where the "Deep Dive" button 64 is selected as an example. The CPU 23 determines whether there is a selection for the "Policy Decision" button (step S6), and if the "Deep Dive" button 64 is selected, it stores the selected policy (step S7).

[0075] Next, the creator, childcare worker "Yoshiko Fujimoto," enters an episode record about today's activities into the episode record field 61. Here, we assume that the episode record shown in Figure 7 has been entered.

[0076] When childcare worker Yoshiko Fujimoto finishes entering text into the episode record input field 61, she selects the confirmation button 71 shown in Figure 6. The CPU 23 determines whether or not the episode record input should be confirmed (Step S9 in Figure 5), and if confirmed, it requests the image / language AI processing system 41 to select a photograph related to the entered text data (Step S11). Specifically, it requests the system to extract related photographs using the image data storage unit 26u as the search range along with the entered text data.

[0077] When the image / language AI processing system 41 receives such a request, it determines the file ID of the photo candidate associated with the text data based on the text data and the tags attached to the image data in the image data storage unit 26u. A well-known method can be used to determine the tags associated with such text data.

[0078] When CPU23 receives the file ID of a candidate photo, it displays it as a candidate in the candidate photo display area 78 (see Figure 6) (Step S13, Figure 5). CPU23 determines whether the user has finished selecting a photo (Step S15). When childcare worker "Yoshiko Fujimoto" selects a photo that she deems appropriate from the displayed candidate photos and selects the confirmation button 79, it confirms the photo to be used and stores it in the image data storage unit 26u, associating it with the episode record (Step S17). This association is used when creating a report, which will be described later. Through this photo selection by the childcare worker, a more appropriate photo is selected than a photo identified solely by the assigned tags.

[0079] Once the photograph to be used is determined, the CPU 23 reads the tags of the photograph and provides these tags and episodes to the language generation AI processing system 42, along with an interest identification prompt (step S18). For example, if the image data 2024090601.jpg shown in Figure 3C is identified, tags such as "pomegranate", "garden", and "Tanaka Kaede" are provided.

[0080] An example of a given interest identification prompt is shown in Figure 8. Let's explain the interest identification prompt. The "role" is described as "a specialist in early childhood education and developmental psychology, and a supporter of teachers in the field of early childhood education," and the instructions are "starting with the child's episode record, estimate the interests and concerns the child is showing," and the points to note are "considering the behaviors and reactions included in the episode record and the developmental stage according to the child's age, flexibly select the most relevant "interest identification perspective," analyze the child's tendency of interest accordingly, and describe it according to the "format."

[0081] Furthermore, the "perspectives for identifying interests" are described as follows: "1. Science (objects and phenomena): If the episode includes focusing on, repeating, or observing specific objects, events, or phenomena with interest, estimate the child's tendency toward curiosity based on this." "2. Art (expression): If the episode includes self-expression or expression of emotions through facial expressions, voice, and gestures, estimate the development of emotions and tendencies toward self-expression based on this." "3. Humanity (sociality): If the episode includes interaction with others, imitative behavior, and communication such as gaze and smiles, estimate the tendency toward development of social skills based on this." "4. Mobility (physicality): If the episode includes fine motor skills or gross motor skills, estimate the direction of interests and concerns related to manual dexterity and physical development based on this." Thus, in this embodiment, interests are estimated from four perspectives: "science," "art," "humanity," and "mobility."

[0082] Figures 9 and 10 show reference data for developmental stages at each age in months. The language generation AI processing system 42 uses this data as reference when generating text based on the prompts.

[0083] The CPU 23 provides the above prompt, along with the episode recording and the tags of the selected images, to the language generation AI processing system 42 (Figure 5, step S18).

[0084] The language generation AI processing system 42 generates estimated results of the child's interests and concerns (hereinafter referred to as interest data) based on the prompt and transmits them to the non-cognitive ability improvement system 1. The CPU 23 receives this and stores it in the output data storage unit 26d (see Figure 2) of the hard disk 26. The results may also be displayed on the screen. This allows the childcare worker to know what aspects the child is interested in.

[0085] Figure 11 shows the estimated results of the received interests and concerns.

[0086] The CPU 23 determines whether the creation policy stored in step S7 is "deep dive," "diversity," or "reconsideration" (step S20). In this case, since the policy "deep dive" is stored (see step S7), the CPU 23 reads the proposal generation prompt "deep dive" (step S21). An example of the proposal generation prompt "deep dive" is shown in Figure 12. In this embodiment, the prompt is designed to propose activities and environments that allow children to act, learn, and grow proactively, starting from the child's interests and concerns stored in step S19, using the "perspective of activity proposals."

[0087] Furthermore, the concept of "in-depth exploration" in activity proposals was defined as "taking a theme that a child is interested in and exploring it more deeply by changing the approach, thereby deepening their fundamental understanding." The relationship between themes and activity proposals was also defined as "aiming to propose activities that take the child's interests and concerns, estimated from the episode records as themes, and explore those themes in depth from different angles."

[0088] Furthermore, the guidelines for activity proposals stipulate that they should be mindful of encouraging questions and dialogues to broaden or deepen the childcare worker's thinking, and that the appropriateness of the proposal should be evaluated in relation to the characteristics of the children's age and developmental stage. In addition, it is specified that proposals should be appropriate for each child's age and developmental stage, taking into account "development and activities by age," and looking ahead to the next stage of development.

[0089] Furthermore, as a "format," as shown in Figure 13, it was specified that attention should be paid to "learning points," "background," "objectives," "preparation and environmental design," "adult perspectives and involvement," "adult questions," and "children's questions." Detailed instructions were given to the language generation AI processing system 42 for each item. In this embodiment, in particular, from the perspective of development and learning, prompts were made to describe points that lead to growth as the reason for the proposal, the background of the episode record and interest estimation, and the objectives of what results are expected through this activity. In addition, prompts were made to describe how to interact with children during the activity and what aspects of the children to pay attention to as observation points. Furthermore, prompts were made to create proposals for questions to ask children during the activity, and to describe the wishes of the childcare worker through this activity.

[0090] Furthermore, this common format for creating proposals was adopted as a common format used in other proposal generation prompts as well.

[0091] The CPU 23 provides the read proposal generation prompt and the interest data stored in step S19 to the language generation AI processing system 42 (step S27).

[0092] In this way, a new prompt is generated by combining the previously stored suggestion generation prompt with the interest data, and this prompt is provided to the language generation AI processing system 42.

[0093] When such a prompt is given, the language generation AI processing system 42 generates a new proposal. When the CPU 23 receives such a new proposal, it stores the result and displays it on the screen (step S29). In this embodiment, the screen changes as shown in Figure 14, and the received new proposal is displayed in the proposal display area 88. An example of a proposal (in-depth analysis) displayed in the proposal display area 88 is shown in Figure 15.

[0094] In this way, proposals are created according to the policy for each item: "theme," "background," "objective," "learning points," "preparation and environmental design," "children's perspective," "adults' perspective and involvement," and "adults' questions." Specifically, in the episode record in Figure 7, the children were interested in the color and texture of the pomegranate seeds, so in the proposal in Figure 15, activities are suggested to deepen their understanding of differences in texture and color using different natural materials.

[0095] If the language generation AI processing system 42 lacks sufficient knowledge of childcare, knowledge of childcare activities may be stored separately as data, and the language generation AI processing system 42 may be instructed to generate suggestions by referring to the data using Search and Expand Generation (RAG).

[0096] Furthermore, if the "Diversity" button 65 was selected in step S6 of Figure 5, the process proceeds from step S20 to step S23, and the prompt for policy "Diversity" is read (step S23). The prompt for policy "Diversity" is shown in Figure 16. In this case, unlike policy "In-depth Exploration," it is specified that "Diversity means conceptualizing the themes that children are interested in, and expanding their interests by selecting new activity targets within that broad concept and exploring them. Aim to propose activities that use the children's interests estimated from the episode records as themes and explore those concepts through different objects and events." Everything else is the same as policy "In-depth Exploration."

[0097] This generates a prompt for the policy "diversity," and the language generation AI processing system 42, upon receiving this prompt, generates the new proposal shown in Figure 17. Specifically, in the episode record in Figure 7, the children were interested in the color and texture of the pomegranate seeds, so the proposal shown in Figure 17 suggests activities that use different natural materials to deepen their understanding of differences in texture and color.

[0098] Furthermore, if the reconsider button 66 was selected in step S6 of Figure 5, the process proceeds from step S20 to step S25, and the prompt for policy "reconsider" is read (step S25). The prompt for policy "reconsider" is shown in Figure 18. In this case, unlike policy "deep dive," it specifies, "Reconsidering means repeating a theme that the child has shown interest in. Based on the episode record, use the child's estimated interests as a theme and propose that activity again as the next activity." Everything else is the same as policy "deep dive."

[0099] This generates a prompt to "reconsider" the policy, and the language generation AI processing system 42, upon receiving this, generates a new proposal as shown in Figure 19. Specifically, in the episode record in Figure 7, the children were interested in the color and texture of the pomegranate seeds, so the proposal in Figure 19 generates a suggestion that will further stimulate the children's curiosity and cultivate their observation and expression skills by providing sensory experiences obtained by directly touching the pomegranate seeds and by repeatedly engaging in art activities using pomegranates as a theme.

[0100] In this embodiment, when a childcare worker specifies one of three policies—"deep dive," "diversity," or "reconsideration"—a prompt for estimating interests is read, and a prompt is generated along with the tags of the episode record and image data. Furthermore, using the estimated interests, one of the three types of suggestion creation prompts is read, and a new suggestion is made. Therefore, the language generation AI processing system 42 can make suggestions corresponding to any of these. This enables childcare activities that are not limited by past experience, not only for new childcare workers but also for veterans. The relationship between such suggestions and the improvement of non-cognitive abilities will now be explained.

[0101] In childcare settings, various activities are implemented to improve children's non-cognitive abilities through everyday play. However, new staff members may be unfamiliar with these methods of stimulating children's interests. Even experienced staff members may rely solely on outdated methods. Therefore, the Non-Cognitive Ability Improvement System 1 utilizes language generation AI to enable childcare activities that stimulate the individual interests of infants and toddlers through the various suggestions mentioned above.

[0102] In this embodiment, the childcare worker selects one of the three options above, but all three may be generated simultaneously.

[0103] As explained above, childcare workers can obtain suggestions for new activities (modified versions of episode records) that they believe the infant or other infants would be more interested in, as recorded in the childcare activity report as completed childcare activities.

[0104] In this embodiment, regarding the activities, events, and objects that children are interested in within the activity records, the "deep dive" option allows for deeper analysis and investigation; the "diversity" option conceptualizes them to a higher level, allowing for the selection of other activity targets within that general concept; and the "reconsideration" option proposes repeating the activities. This enables more multifaceted childcare activities related to the activities, events, and objects that children are interested in.

[0105] (4. Report preparation process) The non-cognitive ability improvement system 1 may be further modified by adding a report request means, a report storage means, and an individual change output means to enable the creation of reports and the output of individual changes. Figure 20 shows a functional block diagram in such a case. Note that in Figure 20, all functional blocks described in Figure 1 except for the image extraction means 9, the suggestion request means 11, and the suggestion storage means 13 are omitted.

[0106] The report request means 15, report storage means 17, and individual change output means 19 shown in Figure 20 will be described below.

[0107] The report request means 15 generates a report generation request prompt that causes the language generation AI processing system 13 to generate a report on the status of childcare activities for a third party other than the childcare provider, and provides the generated prompt to the language generation AI processing system 13.

[0108] The suggestion storage means 14 stores the corresponding data of the observation record and the extracted image used when generating prompts to be generated by the language generation AI processing system 13. When the observation record is provided, the report generation request means 15 uses the corresponding data to generate a report generation request prompt for generating a childcare activity report to the third party, and also gives the language generation AI processing system 13 a prompt to generate a childcare activity report to a third party other than the childcare provider regarding the status of the childcare activities.

[0109] Furthermore, the generation report generated by the language generation AI processing system 13 may be generated as an individual report for each infant, with a separate ID. In this way, when the individual change output means 16 is given an infant's ID, it may extract from the individual report for that ID and use the individual report stored in the report storage means 16 to output the temporal changes in the infant's area of ​​interest identified by that ID as individual change data.

[0110] The report creation process will be explained using Figure 21. The following explanation uses the example of creating a class log from the initial screen after logging in.

[0111] As already explained, the initial screen displays buttons to select either the proposal creation process or the report creation process (not shown). Here, the logged-in childcare worker selects report creation (not shown). The CPU 23 determines whether or not a report creation command is given (Figure 21, step S41), and if a report creation command is given, it displays the report creation screen shown in Figure 22 (step S43).

[0112] The report creation screen displays the following: episode record selection field 97, report display field 98, report creation button 93, class log creation button 94, childcare documentation creation button 95, and individual contact book creation button 96.

[0113] The episode record selection field 97 displays a list of episode records entered by the logged-in childcare worker, along with their creation date (not shown). The childcare worker selects the episode record to be used for report creation from this display, confirms it with the confirmation button 91, and also selects the class log creation button 94. The CPU 23 determines whether the episode record selection has been confirmed (Figure 21, step S44), and if the episode record selection has been confirmed, it displays the episode record in the episode record selection field. The CPU 23 also determines the type of report to be created (step S45). Here, the class log creation button 94 is selected, so the prompt for class log creation is read (step S47), and this, along with the episode record, is output to the language generation AI processing system 42 (step S51).

[0114] In this embodiment, the prompt for creating a class log is stored as shown in Figure 23A, read from it, and output to the language generation AI processing system 42 along with the episode record.

[0115] Figure 23B shows the class log generated by the language generation AI processing system 42. The episode record is generated without personal names so that it objectively covers the entire class.

[0116] The CPU 23 displays the results in the report field 98 (see Figure 22) and stores them in the output data storage unit 26d (see Figure 2).

[0117] Furthermore, if "childcare documentation" is specified as the report type in step S45 of Figure 21, the CPU 23 reads a prompt to create childcare documentation (step S48) and outputs it to the language generation AI processing system 42 along with the episode record (step S51). The prompt to create childcare documentation is shown in Figure 24A.

[0118] Figure 24B shows the childcare documentation generated by the language generation AI processing system 42. The episode record has been reconstructed as a parental communication log for each child appearing in the episode. In addition, a one-sentence activity description is included at the beginning so that parents can understand the overview of the activity, and the child's lines and behavior are described, along with the caregiver's impressions, imagining what the child might be thinking based on their behavior. The processing in step S55 is the same.

[0119] Furthermore, in step S45 of Figure 21, if "Individual Contact Log" is specified as the type of report, the CPU 23 determines whether or not the child to be reported is identified (step S46). If the childcare worker identifies the child to be reported, the CPU 23 reads a prompt to create an individual contact log (step S49) and outputs it to the language generation AI processing system 42 along with the episode record (step S51). The prompt to create an individual contact log is shown in Figure 25A. The following describes the case where the child "Kaede" is identified.

[0120] Figure 25B shows the individual communication notebooks generated by the language generation AI processing system 42. The episode records have been reconstructed into communication notebook texts for parents of each child appearing in the episode. In addition, a one-sentence activity description is included at the beginning so that parents can understand the overview of the activity, and the child's lines and behavior are described, along with the caregiver's impressions, imagining what the child might be thinking based on their behavior. The processing in step S55 is the same.

[0121] If other children named "Itsuki" or "Sakura" are specified, individual communication notebooks are generated as shown in Figures 25C and 25D. Comparing the individual communication notebooks in Figures 25B, C, and D, the initial activity description, "Today we worked on creative activities in the studio," is the same, but each notebook contains a report that reveals what aspects the child found interesting.

[0122] (5. Process for creating individual change reports) The generated individual contact logs are stored in the output data storage unit 26d, so by arranging this data in chronological order, it is also possible to generate an individual change report that shows the changes in each individual's interests.

[0123] Specifically, the system should store prompts to read the individual contact log of the person in question and generate an individual change report summarizing these entries, and then provide these prompts to the language generation AI processing system 42. Figure 26 shows an example of prompts for generating the individual change report.

[0124] In this embodiment, past individual contact logs are stored, and prompts are used to extract information about the development of childcare and education and the child's growth from these logs.

[0125] Additionally, as a point of caution, we made sure to highlight episodes that symbolize the growth and changes the individual has undergone. Furthermore, we aimed to provide an overview of the growth process, showing how it connects to their current state and how their interests and passions are linked to their development.

[0126] This allows us to obtain an archive summary of the changes in the individual's interests over time.

[0127] In this embodiment, past individual contact logs are stored and individual change reports are generated by referring to them. However, instead of using individual contact logs, estimated interests and concerns may be stored and a prompt is given to the language generation AI processing system to generate an individual change report using these.

[0128] Alternatively, the language generation AI processing system 42 may be given a prompt to generate an individual change report from past episode records, rather than based on the aforementioned interests. In this case, the language generation AI processing system 42 may be given a prompt to extract episode records that show the childcare activity status of a specific person, estimate the interests of that person from those episode records, and summarize the history in chronological order.

[0129] (6. Second Embodiment) In the first embodiment, tags estimated from episodes and photographs are used as input to identify interests using an interest identification prompt, and then a prompt is given to generate further suggestions using the identified interests. However, the system is not limited to this, and the identified interests may be output separately. This allows caregivers to learn about the children's interests.

[0130] Alternatively, suggestions can be generated directly using tags estimated from the episode and photographs. In this case, the suggestion generation prompt should include a prompt that considers the child's interests. An example of this is shown in Figure 27. Figure 27 shows the case where in-depth suggestions are generated. In this example, the prompt is designed to make suggestions based on estimated interests, stating: "Starting from the episode record, estimate the child's interests and concerns, and propose the next activity based on that. Based on the behaviors and reactions included in the episode record and the child's developmental stage according to their age, flexibly select the most relevant 'interest-specific perspective,' and then analyze the child's interest tendencies accordingly."

[0131] (7. Other Embodiments) Each of the above embodiments can also be understood as the following inventions.

[0132] A non-cognitive ability improvement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, comprising: an observation record receiving means for receiving observation records of the caregiver's childcare activities for infants and toddlers; a suggestion request means for generating a suggestion request prompt that causes the language generation AI processing system to generate a new activity suggestion regarding the childcare activities in the received observation records using the observation records, and providing the generated prompt to the language generation AI processing system, wherein the suggestion request prompt is a suggestion request prompt given to the language generation AI processing system from one of the creation policies of 1) in-depth analysis, 2) diversification, or 3) reconsideration.

[0133] In this embodiment, the image identification means 9 is composed of an extracted image display selection means 9b and an image extraction means 9a. However, the non-cognitive ability improvement system 1 may determine the extracted image without such a selection process.

[0134] In this embodiment, the individual communication log for a designated child is created by specifying the child to be reported to. However, it is also possible to extract the children who appear in the episode record and automatically generate logs for all of them.

[0135] In this embodiment, the case in which the non-cognitive ability improvement system 1 includes an image / language generation AI processing system 41 as an annotation means has been described. However, similar to the language generation AI processing system 42, these may also be configured as external image / language generation AI systems.

[0136] In this embodiment, the user logs in from the input device 28 of the non-cognitive ability improvement system 1, but the user may also log in from a computer or mobile terminal connected to the network with the non-cognitive ability improvement system 1. In this case, the connection is made via the communication unit 32.

[0137] In this embodiment, three types of reports (class log, childcare documentation, and individual contact book) are created from the childcare worker's episode records and corresponding images. However, the system is not limited to this, and a draft report may be generated from images alone. In this case, image data capturing the childcare worker's childcare activities with infants and toddlers, and corresponding data of the episode records in said image data, may be stored as reference knowledge data, and a prompt may be generated for the language generation AI processing system to generate a draft report by referring to the reference knowledge data.

[0138] In this embodiment, image data tags are used to generate new activity proposals. However, instead of tags, audio data may be converted into text data and used thereafter. Alternatively, the text data may be displayed as reference data when inputting episode records. The same applies to the report creation process.

[0139] In this embodiment, still images are used as image data, but individual frame data constituting video data, or even the video data itself, may also be used.

[0140] Furthermore, although image data tags are used in this embodiment, it is also possible to perform tasks such as identifying interests, making suggestions, and creating reports based solely on episode recordings without using such tags.

[0141] In this embodiment, the caregiver is described as a childcare worker at a children's center. However, if parents (for example, both parents) create video recordings and episode records at home, the same method can be applied to home education for infants and toddlers. In this case, by converting audio data and other information from the video data into text data, it becomes possible to more easily engage infants and toddlers in their activities.

[0142] Furthermore, by linking the childcare activities at the nursery school with the activities at home, as described above, and by communicating the higher level of application to the nursery school, it is possible to further improve non-cognitive skills.

[0143] In this embodiment, we explained the example of how to stimulate children's interest through childcare activities in a nursery school, but similar proposals can be made in after-school care programs at elementary schools and other settings from the same three perspectives of "in-depth exploration, diversity, and reconsideration."

[0144] Furthermore, for elementary, middle, high school, and university students, the system can use images from educational settings and anecdotal records from teachers as input to propose educational policies for each student from three perspectives: "in-depth analysis, diversity, and reconsideration." In this case, the teachers should generate anecdotal records about the educational activities of the students they are educating.

[0145] Furthermore, for elementary, junior high, high school, and university students, the interest estimation prompt preceding the proposal may be used to identify and output each student's interests. This allows students to articulate their interests, even those they may not be aware of themselves, and communicate them not only to teachers but also to other students. Consequently, this information can be used in subsequent activities such as further education or job hunting.

Claims

1. An imaging device for capturing images of the caregiver's activities with infants and toddlers. Image data storage means that stores the image data captured by the imaging means as image data, Annotation means for adding annotations to the image data relating to at least one subject or an object related to said subject, Given observation records of the childcare activity, an image identification means identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means. A proposal request means that generates a proposal request prompt for a language generation AI processing system to generate new activity proposals regarding the aforementioned childcare activity status, and provides the generated proposal request prompt to the language generation AI processing system. A language generation AI processing system having a large-scale language model, which outputs a corresponding response when given the aforementioned solicitation prompt. The system includes a suggestion storage means that receives and stores the response from the language generation AI processing system, The generated prompt is a prompt that causes the language generation AI processing system to generate new activity suggestions regarding the childcare activity situation, using the observation record and the annotations of the extracted images. A non-cognitive skills improvement system characterized by the following.

2. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with said subject, as annotated image data. Given observation records of the childcare activity, an image identification means identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means. A proposal request means that generates a proposal request prompt to cause the language generation AI processing system to generate a new activity proposal regarding the aforementioned childcare activity status, and provides the generated prompt to the language generation AI processing system. Equipped with, The generated prompt is a prompt that causes the language generation AI processing system to generate new activity suggestions regarding the childcare activity situation, using the observation record and the annotations of the extracted images. A non-cognitive skills improvement system characterized by the following.

3. In the non-cognitive ability improvement system of claim 2, The image identification means presents candidate image data from the image data stored in the image data storage means, and when a selection instruction is given, it performs the identification. A non-cognitive skills improvement system characterized by the following.

4. In the non-cognitive ability improvement system of claim 2 or 3, The proposed new activity regarding the aforementioned childcare activity status is a proposed change to the previously completed childcare activity, as recorded in the observation record, and is a proposed change that is likely to be of greater interest or attention to the infant or other infants. A non-cognitive skills improvement system characterized by the following.

5. In the non-cognitive ability improvement system of claim 4, the language generation AI processing system generates a proposal in response to a suggestion request prompt given to the language generation AI processing system, based on one of the following creation policies: 1) Deepening the exploration of a theme that a child has shown interest or concern in the childcare activity situation, by changing the approach and delving deeper to deepen the fundamental understanding; 2) Diversifying the theme to a higher level, and expanding the scope of interest by conceptualizing the theme and selecting and exploring new activity targets within that broader concept; or 3) Reconsidering the theme by re-implementing it. A non-cognitive skills improvement system characterized by the following.

6. In the non-cognitive ability improvement system of claim 2 or 3, The aforementioned request for proposals prompts for new activity proposals from experts in early childhood education and developmental psychology, as well as from those who support teachers in early childhood education settings, and includes instructions to create proposals for activities and environments that enable each infant to actively engage, learn, or grow, based on their individual interests and concerns. A non-cognitive skills improvement system characterized by the following.

7. In the non-cognitive ability improvement system of claim 2 or 3, The format for generating the aforementioned solicitation prompt is categorized into multiple items, and each of these categories includes at least two items from the following: theme, learning points, background, objectives, and preparation / environmental design. A non-cognitive skills improvement system characterized by the following.

8. In the non-cognitive ability improvement system of claim 2 or 3, The format for generating the aforementioned solicitation prompt is categorized into multiple items, and these multiple categories include at least one or more of the following: adult perspective, adult involvement, adult question, and child's question. A non-cognitive skills improvement system characterized by the following.

9. In the non-cognitive ability improvement system of claim 1 or 2, The prompts generated by the suggestion request means are an interest acquisition prompt that estimates the infant's interest or concern in the childcare activity using the observation record and the annotations of the extracted images, and a suggestion prompt that generates new activity suggestions regarding the childcare activity using the infant's interest or concern provided by the language generation AI processing system through this interest acquisition prompt. A non-cognitive skills improvement system characterized by the following.

10. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with said subject, as annotated image data. Given observation records of the childcare activity, an image identification means identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means. Interest identification request means that provides the language generation AI processing system with an interest identification prompt to estimate the interests or concerns of the infants in the childcare activity using the observation records and annotations of the extracted images, A system for improving non-cognitive abilities, characterized by having the following features.

11. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, A means of receiving observation records regarding the caregiver's activities with infants and toddlers. A proposal request means that generates a proposal request prompt to cause the language generation AI processing system to generate a new activity proposal regarding the childcare activity status in the received observation record using the observation record, and provides the generated prompt to the language generation AI processing system. Equipped with, The aforementioned suggestion request prompt is a suggestion request prompt given to the language generation AI processing system based on one of the following creation policies: 1) to delve deeper into the activities that the children showed interest in during the aforementioned childcare activities, 2) to conceptualize the activity and diversify by selecting a different activity target within that general concept, or 3) to reconsider and implement the same theme again. A non-cognitive skills improvement system characterized by the following.

12. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with said subject, as annotated image data. Given observation records of the childcare activity, an image identification means identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means. Regarding the aforementioned childcare activity status, a report generation request means generates a report generation request prompt that causes the language generation AI processing system to generate a childcare activity report for a third party other than the childcare provider, and provides the generated prompt to the language generation AI processing system. A system for improving non-cognitive abilities, characterized by having the following features.

13. A non-cognitive ability improvement system according to claim 2, The proposed storage means stores the corresponding data of the observation record and the extracted image used when generating prompts to be generated by the language generation AI processing system, and further, The system includes a report generation request means that generates a report generation request prompt for generating a childcare activity report to the third party, and also provides the language generation AI processing system with a prompt to generate a childcare activity report to a third party other than the childcare provider regarding the status of the childcare activities, The report generation request means, upon receiving the observation record, generates the report generation request prompt using the corresponding data and provides the generated prompt to the language generation AI processing system. A non-cognitive skills improvement system characterized by the following.

14. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data capturing the childcare activities of a caregiver with respect to infants and toddlers, and a reference knowledge data storage means that stores corresponding data for the childcare activities in said image data. Image data storage means for storing image data of the caregiver's activities with infants and toddlers. A suggestion request means generates a suggestion request prompt for the language generation AI processing system to generate an activity report regarding the status of childcare activities in the input data, using the image data stored in the image data storage means as input data, and provides the generated prompt to the language generation AI processing system. Equipped with, The generated prompt is, The request for suggestions given to the language generation AI processing system is based on the following creation policies, based on the corresponding data of childcare activity situations stored in the aforementioned reference knowledge data storage means: 1) to delve deeper into the activities that the children were interested in during the childcare activity situations; 2) to conceptualize the activity and diversify by selecting a different activity target within that general concept; or 3) to reconsider by repeating the same theme. A non-cognitive skills improvement system characterized by the following.

15. A non-cognitive ability improvement system according to claim 2 or claim 14, The aforementioned image data includes corresponding audio data, moreover, The system includes a text data generation means for generating text data from the aforementioned audio data, The solicitation means generates the solicitation prompt using the text data of the corresponding image data, in addition to the observation record and the annotations of the extracted image. A non-cognitive skills improvement system characterized by the following.

16. A non-cognitive ability improvement system according to claim 14, The aforementioned image data includes corresponding audio data, moreover, The system includes a text data generation means for generating text data from the aforementioned audio data, The report generation request means, when referring to the corresponding data of childcare activity status stored in the reference knowledge data storage means, generates the report generation request prompt using the text data of the image data provided as input data. A non-cognitive skills improvement system characterized by the following.

17. A non-cognitive ability improvement system according to claim 12, The generated report produced by the language generation AI processing system is an individual report that includes the interests or concerns of each infant, Individual report storage means for storing the aforementioned individual reports, Individual change output means that, upon being given an infant ID, extracts the infant's interests or concerns from the individual report indicated by that ID, and provides the language generation AI processing system with an individual change output prompt that outputs the chronological changes in the infant's interests and concerns identified by that ID as individual change data. A system for improving non-cognitive abilities, characterized by having the following features.

18. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing image data of the educational activities of a student in the educational setting, and annotations relating to at least one student or an object related to the student, included in the image data, as annotated image data. When educational records by educators regarding the aforementioned educational activity are provided, an image identification means identifies an image that matches the aforementioned childcare activity based on annotations attached to the image data stored in the image data storage means. Interest identification request means that provides the language generation AI processing system with an interest identification prompt to cause the language generation AI processing system to estimate the interests or concerns of the student in the educational activity situation, using the educational record and the annotations of the extracted images. A system for improving non-cognitive abilities, characterized by having the following features.

19. A method for improving non-cognitive abilities by connecting a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt with a computer, the computer improving non-cognitive abilities, Image data storage step: Stores imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with that subject, as annotated image data. Given observation records of the childcare activity, an image identification step is performed to identify an image that matches the childcare activity based on annotations attached to the stored image data. A proposal request step that generates a proposal request prompt causing the language generation AI processing system to generate a new activity proposal regarding the aforementioned childcare activity status, and provides the generated prompt to the language generation AI processing system. Equipped with, The generated prompt is a prompt that causes the language generation AI processing system to generate new activity suggestions regarding the childcare activity situation, using the observation record and the annotations of the extracted images. A method for improving non-cognitive skills characterized by the following.

20. A language generation AI processing system having a large-scale language model that outputs a response in response to a given solicitation prompt, and a non-cognitive skills improvement program that causes a connected computer to function as a non-cognitive skills improvement system, the program performing the following steps: Image data storage step: Stores imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with that subject, as annotated image data. Given observation records of the childcare activity, an image identification step is performed to identify an image that matches the childcare activity based on annotations attached to the stored image data. A proposal request step that generates a proposal request prompt causing the language generation AI processing system to generate a new activity proposal regarding the aforementioned childcare activity status, and provides the generated prompt to the language generation AI processing system. Equipped with, The generated prompt is a prompt that causes the language generation AI processing system to generate new activity suggestions regarding the childcare activity situation, using the observation record and the annotations of the extracted images. A non-cognitive skills improvement program characterized by the following:

21. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with said subject, as annotated image data. Given observation records of the childcare activity, an image identification means identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means. Instruction means for giving instructions to the language generation AI processing system to generate new activity proposals regarding the childcare activity status using the observation records and annotations of the extracted images, A system for improving non-cognitive abilities, characterized by having the following features.

22. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing imaging data of the caregiver's activities with infants and toddlers, and annotations related to at least one subject or object associated with said subject, as annotated image data. Given observation records of the childcare activity, an image identification means identifies an image that matches the childcare activity based on annotations attached to the image data stored in the image data storage means. Interest identification request means that provides the language generation AI processing system with an interest identification prompt to estimate the interests or concerns of the infants in the childcare activity using the observation records and annotations of the extracted images, Estimated interest report storage means for storing the estimated interests or concerns of each infant as an estimated interest report, A means for generating individual change data, which, upon being given an infant's ID, reads the history of estimated interest reports for the infant identified by that ID and provides the language generation AI processing system with an individual change data acquisition prompt to output the temporal changes in the infant's interests or concerns identified by that ID as individual change data. A system for improving non-cognitive abilities, characterized by having the following features.

23. A non-cognitive ability enhancement system connected to a language generation AI processing system having a large-scale language model that outputs a response in response to a given suggestion request prompt, Image data storage means for storing annotations regarding at least one subject or an object related to said subject, which are included in imaging data of the childcare activities of a caregiver for infants and toddlers, as annotated image data. Observation record storage means that stores multiple observation records, each including the ID of an infant appearing in the observation record, When an infant's ID is given, the interest identification request means extracts observation records from among the multiple observation records in which the given infant's ID appears, extracts annotations that have been attached to the infant appearing in those observation records from the annotated image data, and uses the observation records and the extracted annotations to provide the language generation AI processing system with an interest identification prompt that causes the language generation AI processing system to estimate the infant's interests or concerns in the childcare activity situation. A means for generating individual change data, which stores the interests or concerns of the infant as an estimated interest report, and uses multiple estimated interest reports for the infant to provide the language generation AI processing system with an individual change data acquisition prompt to output the temporal changes in the infant's interests or concerns as individual change data. A system for improving non-cognitive abilities, characterized by having the following features.

Images