Information processing system, information processing method, and information processing program

Abstract

This system provides an information processing system that presents user-specific instructions on how to use a device when the user is using it. [Solution] In the information processing system 10, the server acquires user data, which is data about the user who uses the equipment, and inputs the user data into a pre-generated trained model. The server then acquires equipment usage method data output from the trained model and outputs the equipment usage method data.

Description

[Technical Field]

[0001] The disclosed technology relates to an information processing system, an information processing method, and an information processing program. [Background Art]

[0002] Patent Document 1 discloses a massage system that allows a business operator to have a user receive a specific massage, because control content for controlling the operation of a massage unit can be specified by a plurality of media or a plurality of identification codes.

[0003] Patent Document 2 discloses an electric potential therapy device that can apply low-frequency stimulation to a human body with a simple circuit configuration using an existing converter, and can be expected to provide a relaxing effect.

[0004] Patent Document 3 discloses that, with a compact device configuration, an electric field, a magnetic field, and an electric magnetic field or electromagnetic wave generated from an electrode can be adjusted to perform effective therapy or treatment. [Prior Art Documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2018-094249 [Patent Document 2] Japanese Patent No. 5235040 [Patent Document 3] Japanese Patent No. 7731618 [Summary of the Invention] [Problem to be Solved by the Invention]

[0006] By the way, an electric potential therapy device is known as an example of a device for adjusting a person's physical condition. An electric potential therapy device is a device that adjusts physical condition by applying a safe high-voltage electric field to the human body, and is recognized to be effective in relieving symptoms such as headache, stiff shoulders, insomnia, and chronic constipation.

[0007] However, the mechanisms of action of existing electrotherapy devices are often not fully understood. Therefore, there is a challenge in providing clear explanations regarding the appropriate use of electrotherapy devices according to the user's symptoms.

[0008] The disclosed technology was made in light of the above circumstances and provides an information processing system, method, and program that can present a user-specific method of using the device when the user is using the device. [Means for solving the problem]

[0009] To achieve the above objective, a first aspect of this disclosure is an information processing system comprising: a first acquisition unit that acquires user data, which is data relating to a user who uses a device; a second acquisition unit that inputs the user data into a pre-generated pre-trained model to acquire device usage method data output from the pre-trained model; and an output unit that outputs the device usage method data, wherein the pre-trained model is a model that has been pre-trained based on training data, in which training user data, which is data relating to a training user, and training usage method data, which is data relating to how a training user uses the device.

[0010] A second aspect of this disclosure is an information processing method in which a computer performs a process to acquire user data, which is data relating to a user who uses a device, input the user data into a pre-generated trained model, thereby acquiring device usage method data output from the trained model, and outputting the device usage method data, wherein the trained model is a model that has been pre-trained based on training data, in which training user data, which is data relating to a training user, and training usage method data, which is data relating to how a training user uses the device.

[0011] A third aspect of this disclosure is an information processing program for causing a computer to perform a process that acquires user data, which is data relating to a user who uses a device, inputs the user data into a pre-generated trained model, acquires device usage method data output from the trained model, and outputs the device usage method data, wherein the trained model is a model that has been pre-trained based on training data, in which training user data, which is data relating to a training user, and training usage method data, which is data relating to how the training user used the device. [Effects of the Invention]

[0012] According to the disclosed technology, the effect is that when a user uses the device, it can be presented with user-specific instructions on how to use the device. [Brief explanation of the drawing]

[0013] [Figure 1] This figure shows an example of equipment 1 used by user U. [Figure 2] This figure shows an example of a schematic configuration of an information processing system according to an embodiment. [Figure 3] This is a diagram illustrating this embodiment. [Figure 4] This is a diagram illustrating user data and classification methods. [Figure 5] This diagram illustrates the classification of standard usage forms based on user data. [Figure 6] This figure shows an example of standard usage form data. [Figure 7] This figure shows an example of standard usage form data. [Figure 8] This figure shows an example of how effects can be applied to standard usage form data. [Figure 9] This is a diagram to explain stratified classification. [Figure 10] This diagram illustrates the system configuration for presenting health plans to users. [Figure 11] It is a diagram for explaining multimodal processing. [Figure 12] It is a diagram showing an example of a computer constituting a user terminal or a server. [Figure 13] It is a diagram showing an example of information processing executed by the server of the embodiment. [Figure 14] It is a diagram showing an example of information processing executed by the server of the embodiment.

Mode for Carrying Out the Invention

[0014] Hereinafter, embodiments of the disclosed technology will be described in detail with reference to the drawings.

[0015] <Information processing system of embodiment> FIG. 1 shows an example of a device 1 used by a user U. The device 1 shown in FIG. 1 is an electric potential therapy device, which is provided with electrodes for generating an electric field at each position. The user U can set the electric field intensity and the duration for generating the electric field by operating an operation unit provided on the electric potential therapy device.

[0016] In the present embodiment, a usage instruction representing a usage method of the device 1 is generated according to user data of the user U. The user data includes, for example, the user's date of birth, age + number of months (number of days from the birth date), height, weight, gender, purpose of use, detailed purpose, and symptom degree, etc.

[0017] Then, in the present embodiment, the usage instruction is presented to the user U. The user uses the device 1 in accordance with the usage instruction. In addition, the user U evaluates the degree of remission of his / her own symptoms (e.g., headache, etc.) after using the device 1. The information processing system of the present embodiment causes a known machine learning model to learn what kind of usage method of the device 1 is effective for the user based on this evaluation result. This makes it possible to present a device usage method that suits the user.

[0018] A specific description will be given below.

[0019] Figure 2 is a schematic diagram of the information processing system 10 according to the embodiment. As shown in Figure 2, the information processing system 10 of the embodiment comprises a plurality of user terminals 12A, 12B, 12C, ... and a server 14 which is an example of an information processing system. Each device is connected to communicate via a network 20 such as the Internet. Hereinafter, any one of the plurality of user terminals 12A, 12B, 12C, ... will simply be referred to as user terminal 12.

[0020] (User terminal 12) The user terminal 12 is operated by the user using device 1. The user terminal 12 is, for example, a mobile device (e.g., a smartphone) held by the user. Alternatively, the user terminal 12 is a terminal that is integrated into device 1 and can be operated via the display unit (not shown) of device 1. As will be described later, the user terminal 12 is implemented by a computer.

[0021] (Server 14) Server 14 presents a usage slip for device 1 to the user operating user terminal 12. As will be described later, server 14 is implemented by a computer.

[0022] Figure 3 is a diagram illustrating this embodiment. As shown in Figure 3, in this embodiment, four types of models are used to generate usage slips for device 1 tailored to the user. Figure 3 shows a classification model, a first trained model, a second trained model, and a third trained model.

[0023] (Classification model) The classification model outputs usage prescription data, which is an example of device usage data tailored to a given user, by classifying user data that represents the user. Figure 4 is a diagram illustrating user data and the classification method. As shown in Figure 4, user data includes the user's date of birth, age + months (number of days since birthday), height, weight, gender, purpose of use, detailed purpose, and symptom severity. Note that data that can be calculated from the user data (e.g., BMI value) is also included in the user data.

[0024] First, the user operates their user terminal 12 to input their date of birth, age + months (number of days since birthday), height, weight, gender, purpose of use of device 1, details of the purpose, and severity of symptoms. Based on this, a suitable prescription is identified from 28 pre-classified standard prescriptions.

[0025] Figure 5 illustrates the classification of standard prescriptions based on user data. As shown in Figure 5, a suitable standard prescription is identified for the user based on user data, the purpose of use, detailed purpose, BMI / subjective symptom score, and symptom severity entered by the user, and that standard prescription data is output. BMI is calculated based on height and weight from the user data. The subjective symptom scores—headache HIT-6 score, stiff shoulder NDI score, insomnia ISI score, and chronic constipation Wexner score—are calculated based on the results of a medical interview or questionnaire given to the user. The 28 types of standard prescriptions shown in Figure 5 are prescriptions pre-created by physicians with clinical expertise. The classification model identifies the standard prescription data suitable for the user from among the 28 types of standard prescriptions pre-prepared according to the physician's clinical expertise. Therefore, it can be said to be a classification model based on clinical expertise.

[0026] Figures 6 and 7 show examples of standard prescription data. Figure 6 shows standard prescription 1. This standard prescription 1 is for users whose purpose is "maintaining health" and whose BMI is less than 18.5 [underweight]. Figure 7 shows standard prescription 17. This standard prescription 17 is for users whose purpose is "relieving stiff shoulders" and whose "degree of discomfort" is moderate.

[0027] As shown in Figures 6 and 7, the usage slip is marked with an asterisk, instructing the user to use Device 1 multiple times (from the first use onwards) in the manner corresponding to the asterisk. The column for the number of times Device 1 is used also shows "Date: Month / Day / Year," indicating that Device 1 will be used on different days. The usage slip also shows the electric field strength of Device 1 and the duration of its use (in minutes).

[0028] The user uses device 1 according to the usage instructions marked with stars in Figures 6 and 7, and records the effects on the usage form. Specifically, if the user feels an improvement after using device 1, they add a circle to the corresponding star. In Figure 8, it can be seen that circles were added on the 1st, 2nd, 3rd, 6th, and 7th uses.

[0029] In this embodiment, the first trained model described above is generated by using the data on the effects of using the device 1, obtained as described above, as training data. Specifically, the first trained model is generated by conducting the clinical trial described above and using machine learning on the user data of the subject users and the usage patterns in which the effects were confirmed.

[0030] (First pre-trained model) As described above, the first pre-trained model is a machine learning model obtained by machine learning with user data from the subject users and data representing patterns of usage instructions in which effectiveness has been confirmed. When user data is input to the first pre-trained model, it outputs usage instruction data, which is data on how to use device 1. For example, a doctor presents multiple usage instructions selected by the classification model described above to the subject user, the user uses device 1 according to the provided usage instructions, and the user records the effect. The first pre-trained model is then generated using the data obtained in this way as training data. For this reason, the first pre-trained model can be said to be a model based on the results of clinical trials.

[0031] (Second pre-trained model) The second pre-trained model is a pre-trained model that, when user stratification data obtained from user data is input, outputs usage data for device 1 according to the user stratification. The first pre-trained model is a model that was trained using data indicating whether or not there was an effect from using device 1 when the subject user answered a questionnaire about the effect of device 1. In contrast, the second pre-trained model outputs usage data according to the user stratification when user stratification is input.

[0032] Figure 9 is a diagram illustrating stratified classification. As shown in Figure 9, in this embodiment, the classification model described above selects a prescription suitable for the user from 28 types of standard prescriptions pre-created by a physician. However, it may be possible to create a prescription that is more suitable for the user than the 28 types of standard prescriptions pre-created. As shown in Figure 9, if user data is stratified in detail, there are 3564 classifications, so it may be preferable for users to use prescriptions that correspond to their stratification. Therefore, the second pre-trained model learns the patterns of prescriptions that have proven effective for combinations of user gender, age, BMI, purpose of use, detailed purpose, and symptom severity, and generates prescription data that is expected to be most effective for each stratification. The second pre-trained model is pre-generated using known machine learning techniques or clustering algorithms, etc.

[0033] (Third pre-trained model) The third pre-trained model outputs usage instructions data, which are data on how to use Device 1, when user data is input. The third pre-trained model adds learning items such as the number of days between uses of Device 1, the total usage time of Device 1, the total usage voltage of Device 1, and the effectiveness rate to the aforementioned gender, age, BMI, purpose of use, detailed purpose, and symptom severity layers, and performs reinforcement learning tailored to each individual user subject, enabling it to generate usage instructions data that is expected to be most effective.

[0034] As described above, the first and second trained models are models trained using not only the training data of a specific individual user but also the training data of other users. Therefore, the prescription data presented to a user is often the prescription data that was effective for users with similar characteristics to that user. However, it is preferable to present prescription data that takes into account the individual circumstances of a user. Therefore, in this embodiment, a third trained model for individual users is generated by adding the number of days between uses of device 1, the total usage time of device 1, the total usage voltage of device 1, and the effectiveness rate to the learning items and performing known reinforcement learning. Adding the number of days between uses of device 1, the total usage time of device 1, the total usage voltage of device 1, and the effectiveness rate to the learning items means, for example, setting the effectiveness rate as the reward function in reinforcement learning and training the third trained model so that the value output from that reward function becomes large. In this case, the number of days between uses of device 1, the total usage time of device 1, and the total usage voltage of device 1 are set as actions in reinforcement learning.

[0035] It is also possible to generate prescription data suitable for the user by using the above classification model, first pre-trained model, second pre-trained model, and third pre-trained model as they are. However, in this embodiment, not only prescription data but also a health plan including improvements to the user's lifestyle is presented to the user.

[0036] Figure 10 is a diagram illustrating the system configuration for presenting a health plan to the user. The solid arrows in Figure 10 correspond to the content described above. In the "Overall Machine Learning" shown in Figure 10, the first and second pre-trained models described above are generated. In addition, in the "Individual Learning" shown in Figure 10, the third pre-trained model described above is generated.

[0037] On the other hand, the dashed-dotted arrow in Figure 10 represents the content presented to the user, which includes not only usage data but also a health plan that includes improvements to the user's lifestyle.

[0038] Specifically, as shown in Figure 10, when the first and second trained models generated in "Overall Machine Learning," along with user data, prescription data, effectiveness data, and other data, are input to the "Prescription / User Data Processing System," multimodal processing is performed in the "Prescription / User Data Processing System," and the various data are converted into data that can be input to the language model. The language model then outputs health plan data (e.g., health advice) corresponding to the input data.

[0039] Furthermore, as shown in Figure 10, when the third pre-trained model generated in "Individual Machine Learning," along with user data, prescription data, effectiveness data, and other data, are input to the "Prescription / User / Individual Effectiveness Data Processing System," multimodal processing is performed in the "Prescription / User / Individual Effectiveness Data Processing System," and the various data are converted into data that can be input to the language model. The language model then outputs health plan data (e.g., health advice) corresponding to the input data.

[0040] Figure 11 is a diagram illustrating multimodal processing in the "Prescription / User Data Processing System" and the "Prescription / User / Individual Effect Data Processing System" (hereinafter simply referred to as the "Data Integration / Standardization System"). As shown in Figure 11, when user data, prescription data, effect data, personal health records (an example of other data), and machine learning models (e.g., the first, second, and third trained models mentioned above) are input to the Data Integration / Standardization System, the system converts the input data into a format that can be processed by the language model. Specifically, the Data Integration / Standardization System converts this data into a format that can be processed by the language model by prompting it. For example, it uses pre-prepared prompt templates to convert the above data, which is simply a list of numerical data, into a format that is easy for the language model to understand.

[0041] As shown in Figure 11, the numerical data (e.g., time series data, graph data, machine learning model parameters, etc.) and image data included in the personal health record and machine learning model are converted into data that includes language by the language model. Even if the numerical data and image data themselves are input directly into the language model, accurate results may not be obtained. Therefore, the numerical data (e.g., time series data, graph data, machine learning model parameters, etc.) and image data included in the personal health record and machine learning model are first input into the language model to convert them into a format that the subsequent language model can process. The parameters of the machine learning model (the first pre-trained model, the second pre-trained model, and the third pre-trained model mentioned above) are tailored to the user data and other data. When these parameters are input into the language model, the language model interprets the structure and parameters of the machine learning model and can determine what kind of health plan is appropriate for the user. For this reason, as mentioned above, the machine learning model is also input into the data integration and standardization system.

[0042] As described above, in this embodiment, usage data suitable for the user is created by utilizing at least one of the first trained model, the second trained model, and the third trained model. Further details will be explained below.

[0043] As shown in Figure 2, the server 14 functionally comprises a first acquisition unit 140, a data storage unit 142, a model storage unit 144, a second acquisition unit 146, and an output unit 148.

[0044] The first acquisition unit 140 acquires user data, which is data about the user using the device 1. The first acquisition unit 140 acquires user data entered via the user terminal 12.

[0045] The data storage unit 142 stores user data and various related data acquired by the first acquisition unit 140. Specifically, it stores, for example, 28 types of standard usage form data, past user data, effectiveness data, user usage history data of device 1 (for example, total usage time of device 1, total usage voltage of device 1, and effectiveness rate, etc.), and personal health records.

[0046] The model memory unit 144 stores the classification model, the first pre-trained model, the second pre-trained model, the third pre-trained model, and the language model described above. Each of these models is a model that has been pre-generated using known machine learning and artificial intelligence techniques.

[0047] As described above, the classification model is designed to output usage slip data, which is an example of usage method data for device 1, when user data is input. Specifically, the classification model outputs standard usage slip data from 28 pre-prepared types of standard usage slip data, according to the classification of the user data.

[0048] Furthermore, the first pre-trained model is a model that outputs usage data for device 1 when user data is input. As described above, the first pre-trained model is a model that has been pre-trained based on training data, in which training user data, which is data about training users corresponding to subjects, and training usage data, which is usage data when the training users used device 1, are associated. The training usage data is effect data regarding usage methods, indicating whether or not there was an effect from using device 1 when the training users corresponding to subjects answered a questionnaire about the effect of device 1. For example, as described above, the training usage data is data with circles attached to the usage data.

[0049] Furthermore, the second pre-trained model is a model that, when user stratification data is input, outputs usage data for device 1 corresponding to the user's stratification. As described above, the second pre-trained model is a model that has been pre-trained based on training data in which the stratification data of the training users who are subjects and the training usage data when the training users corresponding to that stratification data used device 1 are associated.

[0050] Furthermore, the third pre-trained model is a model that has been individually reinforced based on training data that associates past user data obtained from the user with past usage data from when the user used device 1. In this case, the user who uses device 1 and the training user who is the subject are the same user.

[0051] Furthermore, the language model is a model capable of language processing, such as a known large-scale language model. As mentioned above, the language model is used when outputting the user's health plan data. Note that each of the first pre-trained model, second pre-trained model, third pre-trained model, and language model described above may be a so-called AI agent.

[0052] In this embodiment, the system outputs prescription data and health plan data suitable for the user by using each model stored in the model storage unit 144 individually or in combination.

[0053] The second acquisition unit 146 inputs the user data acquired by the first acquisition unit 140 into a pre-generated, trained model stored in the model storage unit 144, thereby acquiring the usage data for device 1 output from the trained model. The second acquisition unit 146 generates usage data suitable for the user using one of the multiple models stored in the model storage unit 144.

[0054] The output unit 148 outputs the usage data for device 1 acquired by the second acquisition unit 146. The usage data output from the output unit 148 is displayed on the display unit of the user terminal 12. The user checks the usage data displayed on the display unit of the user terminal 12 and uses device 1 according to the usage method indicated by that usage data.

[0055] The output unit 148 displays a questionnaire regarding the effectiveness of device 1 on the display unit of the user terminal 12 when the user uses device 1. The first acquisition unit 140 acquires the results of the user's response to the questionnaire. This questionnaire result is effectiveness data regarding usage, indicating whether or not the user experienced any effect from using device 1. The first acquisition unit 140 stores the effectiveness data in the data storage unit 142. The effectiveness data is used to train each of the models described above.

[0056] The above describes the process by which server 14 simply generates prescription data from user data. In contrast, the following describes the case where server 14 outputs health plan data.

[0057] When outputting health plan data, server 14 uses a language model capable of natural language processing as a pre-trained model.

[0058] The first acquisition unit 140 acquires first user data and second user data as user data. The first user data includes the user's date of birth, age + number of months (number of days since birthday), height, weight, gender, purpose of use, detailed purpose, and symptom severity, as described above. Data that can be calculated from the user data (e.g., BMI value) is also included in the first user data. On the other hand, the second user data includes usage data for device 1, effect data representing the effect of using device 1, the user's personal health history data, and data that includes at least one of the predictive models that generate predictive data about the user (e.g., a predictive model that includes at least one of the first trained model, the second trained model, and the third trained model described above).

[0059] The second acquisition unit 146 generates prompt data corresponding to the first user data and the second user data, and inputs the prompt data into a language model, which is a trained model, thereby acquiring data related to the user's health plan output from the language model. As described above, if the first user data or the second user data contains data different from the language data, the second acquisition unit 146 uses an arbitrary language model to convert the data different from the language data into the language data, and inputs the converted language data into the language model. Instead of inputting data different from the language data directly into the language model, the quality of the data output from the language model is improved by first translating such data into language and then inputting it back into the language model.

[0060] The output unit 148 outputs data related to the health plan, including usage data for device 1. The health plan data includes daily health advice, etc. For example, the health plan data is displayed on the display unit of the user terminal 12.

[0061] The user terminal 12 and server 14 can be implemented, for example, by the computer 50 shown in Figure 12. The computer 50 includes a CPU 51, a memory 52 as a temporary storage area, and a non-volatile storage unit 53. The computer 50 also includes an input / output interface (I / F) 54 to which external devices and output devices (for example, the display unit of the user terminal 12) are connected, and a read / write (R / W) unit 55 that controls the reading and writing of data to the recording medium. The computer 50 also includes a network I / F 56 that connects to a network such as the Internet. The CPU 51, memory 52, storage unit 53, input / output I / F 54, R / W unit 55, and network I / F 56 are connected to each other via a bus 57.

[0062] The storage unit 53 can be implemented using a Hard Disk Drive (HDD), Solid State Drive (SSD), flash memory, etc. The storage unit 53, as a storage medium, stores a program that allows the computer 50 to function. The CPU 51 reads the program from the storage unit 53, loads it into memory 52, and sequentially executes the processes contained in the program.

[0063] [Operation of Server 14 in the Embodiment] Next, the specific operation of the server 14 in this embodiment will be described. The server 14 performs the information processing shown in Figure 13.

[0064] First, in step S100, the first acquisition unit 140 acquires user data entered via the user terminal 12.

[0065] Next, in step S102, the second acquisition unit 146 inputs the user data acquired in step S100 into a pre-generated trained model stored in the model storage unit 144, thereby acquiring the usage data for device 1 output from the trained model. The second acquisition unit 146 generates usage data suitable for the user using one of the multiple models stored in the model storage unit 144.

[0066] In step S104, the output unit 148 outputs the usage data for device 1 acquired by the second acquisition unit 146. The usage data output from the output unit 148 is displayed on the display unit of the user terminal 12. The user checks the usage data displayed on the display unit of the user terminal 12 and uses device 1 according to the usage method indicated by the usage data.

[0067] Furthermore, when server 14 outputs health plan data, it performs the information processing shown in Figure 14.

[0068] In step S200, the first acquisition unit 140 acquires the first user data input via the user terminal 12. Also in step S200, the second user data stored in the data storage unit 142 is acquired.

[0069] In step S202, the second acquisition unit 146 generates prompt data corresponding to the first user data and the second user data acquired in step S200. If the data acquired in step S200 includes numerical data or other data different from language data, it converts that data into appropriate data.

[0070] In step S204, the second acquisition unit 146 inputs the processed data obtained in step S202 into the language model stored in the model storage unit 144, thereby acquiring the health plan data output from the language model. The health plan data includes health advice expressed in language. This health plan data is tailored to the individual user.

[0071] In step S206, the output unit 148 outputs data related to the health plan, including the usage data of the device 1.

[0072] As described above, the server in this embodiment acquires user data, which is data about the user using the device. The server inputs the user data into a pre-generated trained model, and acquires device usage method data output from the trained model. The server outputs the device usage method data. The trained model is a model that has been pre-trained based on training data, in which training user data, which is data about the training user, and training usage method data, which is data about how the training user used the device, are associated. This makes it possible to present user-specific device usage methods when a user uses the device.

[0073] Furthermore, this embodiment generates usage methods and health guidance that are expected to be effective according to the user's purpose of use and the severity of their symptoms for medical devices and health equipment / devices. As a result, it is expected to be used in the medical and health-related industries as a whole. Moreover, it generates usage methods and health guidance that are expected to be most effective for each user based on the user's biometric data, purpose of use, severity of symptoms, effect information, and usage information, and the more the user uses it, the more effective usage methods and health guidance can be generated for that user.

[0074] Furthermore, this embodiment can be described as an AI agent system that can autonomously support the achievement of the user's desired goals by combining different data processing methods, such as a natural language model, the user's biometric data, the method of using medical devices and health equipment / devices, and information on the effects of using medical devices and health equipment / devices.

[0075] Furthermore, this embodiment can be described as an AI agent system that can autonomously help users achieve their desired goals and improve their health and quality of life by combining different data processing methods, such as natural language models, user biometric data, methods of using medical devices and health equipment, and information on the effects of using medical devices and health equipment.

[0076] Furthermore, this embodiment goes beyond simply using the user's biometric data; by also utilizing the user's purpose of use, the severity of symptoms, and effectiveness data, it is possible to generate usage methods and health guidance that are expected to be more effective for the user.

[0077] Furthermore, this embodiment makes it possible to generate more personalized and effective usage methods and health guidance for medical devices and health equipment / devices by reinforcing learning based on the effectiveness data of individual users.

[0078] Furthermore, according to this embodiment, it becomes possible to realize generative AI and AI agents through language processing that combine not only combinations of still images, videos, 3D shape data, audio data and language processing, but also different data processing such as language models and user biometric data, methods of using medical devices and health devices / equipment, information on the effects of using medical devices and health devices / equipment, learning models and personal health data (medical, health, and lifestyle information).

[0079] Furthermore, the technology disclosed herein is not limited to the embodiments described above, and various modifications and applications are possible without departing from the spirit of this invention.

[0080] For example, although the present specification describes an embodiment in which the program is pre-installed, it is also possible to provide the program by storing it on a computer-readable recording medium.

[0081] Furthermore, although the above embodiment was described using the example where device 1 is an electrotherapy device, it is not limited to this. Any device can be used as the target device, as long as the user can evaluate its effectiveness.

[0082] In addition, the processing that the CPU reads and executes in the above embodiment may be executed by various processors other than the CPU. Examples of such processors include PLDs (Programmable Logic Devices) such as FPGAs (Field-Programmable Gate Arrays) whose circuit configuration can be changed after manufacturing, and dedicated electrical circuits that are processors with circuit configurations specifically designed to execute specific processing, such as ASICs (Application Specific Integrated Circuits). Alternatively, a GPGPU (General-purpose graphics processing unit) may be used as the processor. Furthermore, each processing may be executed by one of these various processors, or by a combination of two or more processors of the same or different types (for example, multiple FPGAs, and a combination of a CPU and an FPGA). More specifically, the hardware structure of these various processors is an electrical circuit that combines circuit elements such as semiconductor elements.

[0083] Furthermore, while the above embodiments describe a configuration in which the program is pre-stored (installed) on storage, the invention is not limited to this. The program may be provided in a form stored on a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), DVD-ROM (Digital Versatile Disk Read Only Memory), or USB (Universal Serial Bus) memory. Alternatively, the program may be downloaded from an external device via a network.

[0084] Furthermore, each process in this embodiment may be configured by a computer or server equipped with a general-purpose processing unit and storage device, and each process may be executed by a program. This program is stored in the storage device and can be recorded on a recording medium such as a magnetic disk, optical disk, or semiconductor memory, or provided over a network. Of course, none of the other components have to be implemented by a single computer or server; they may be implemented in a distributed manner across multiple computers connected by a network.

[0085] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0086] (Note) The following is an addendum regarding the nature of this disclosure.

[0087] (Note 1) A first acquisition unit acquires user data, which is data about the user using the device, A second acquisition unit inputs the user data into a pre-generated trained model to acquire the device usage method data output from the trained model, An output unit that outputs usage method data for the aforementioned device, An information processing system comprising, The aforementioned pre-trained model is a model that has been pre-trained based on training data, which is a combination of training user data, which is data about the training user, and training usage data, which is data about how the training user used the device. Information processing system. (Note 2) The aforementioned learning method data is effect data regarding the method of use, indicating whether or not there was an effect from using the device, based on the learning user's response to a questionnaire about the device's effectiveness. The information processing system described in Appendix 1. (Note 3) The aforementioned device is an electrotherapy device, The aforementioned usage method data is data relating to the usage prescription for the electrotherapy device. The information processing system described in Appendix 1 or Appendix 2. (Note 4) The aforementioned information processing system includes a user terminal, The output unit causes the user terminal's display unit to display the questionnaire when the user uses the device. The first acquisition unit acquires the answers to the questionnaire entered by the user by operating the user terminal. An information processing system described in any one of the items in Appendix 1 to Appendix 3. (Note 5) The aforementioned trained model is a first trained model that outputs device usage method data when user data is input. An information processing system described in any one of the items in Appendix 1 to Appendix 4. (Note 6) The aforementioned trained model is This is a second trained model that, upon inputting stratified user data, outputs data on how to use the device according to the user's stratification. The second pre-trained model is This is a model that has been pre-trained based on training data, in which stratified data of the training users and training usage data of the training users when they used the equipment, corresponding to the stratified data, are associated. An information processing system described in any one of the items in Appendix 1 to Appendix 5. (Note 7) The aforementioned user and the aforementioned learning user are the same user. The aforementioned trained model is This is a third pre-trained model that outputs usage method data for the device when the aforementioned user data is input. The third pre-trained model is This model is individually trained based on training data that associates past user data obtained from the user with past usage data from when the user used the device. An information processing system described in any one of the items in Appendix 1 to Appendix 6. (Note 8) The aforementioned trained model comprises a language model capable of language processing, The user data includes first user data and second user data representing at least one of the following: data on how to use the device, effect data representing the effects of using the device, the user's personal health history data, and predictive data about the user. The first acquisition unit acquires the first user data and the second user data, The second acquisition unit generates prompt data corresponding to the first user data and the second user data, and inputs the prompt data to the trained model, thereby acquiring data related to the user's health plan output from the trained model. The output unit outputs data relating to the health plan, including data on how to use the device. An information processing system described in any one of the items in Appendix 1 to Appendix 7. (Note 9) The aforementioned second acquisition unit is, If the first user data or the second user data contains data different from the language data, the language model is used to convert the data different from the language data into the language data, and the converted language data is input into the trained model. The information processing system described in Appendix 8. (Note 10) We obtain user data, which is data about the users who use the device. By inputting the user data into a pre-generated trained model, the device usage method data output from the trained model is obtained. Outputs usage method data for the aforementioned device. An information processing method in which a computer performs the processing, The aforementioned pre-trained model is a model that has been pre-trained based on training data, which is a combination of training user data, which is data about the training user, and training usage data, which is data about how the training user used the device. Information processing methods. (Note 11) We obtain user data, which is data about the users who use the device. By inputting the user data into a pre-generated trained model, the device usage method data output from the trained model is obtained. Outputs usage method data for the aforementioned device. An information processing program that causes a computer to perform a process, The aforementioned pre-trained model is a model that has been pre-trained based on training data, which is a combination of training user data, which is data about the training user, and training usage data, which is data about how the training user used the device. Information processing program. [Explanation of Symbols]

[0088] 10. Information Processing Systems 12 User terminals 14 Servers 140 First acquisition part 142 Data Storage Unit 144 Model Memory Unit 146 Second Acquisition Department 148 Output section

Claims

1. A first acquisition unit acquires user data, which is data about the user using the device, A second acquisition unit inputs the user data into a pre-generated trained model to acquire the device usage method data output from the trained model, An output unit that outputs usage method data for the aforementioned device, An information processing system comprising, The aforementioned trained model is a model that has been pre-trained based on training data, which is a model that associates training user data, which is data about the training user, with training usage data, which is data on how the training user used the device. The pre-trained model is a pre-trained model that, upon input of stratification data representing stratification when classifying the user's attributes, stratification when classifying the user's purpose of using the device for health improvement, and stratification when classifying the degree of symptoms the user wishes to improve, outputs usage method data for the device according to the user's stratification, and is a model that has been pre-trained based on training data in which the training user's stratification data and the training usage method data when the training user used the device corresponding to the stratification data are associated. The user data includes the stratified data of the user. Information processing system.

2. The aforementioned learning method data is effect data regarding the method of use, indicating whether or not there was an effect from using the device, based on the learning user's response to a questionnaire about the device's effectiveness. The information processing system according to claim 1.

3. The aforementioned device is an electrotherapy device, The aforementioned usage method data is data relating to the usage prescription for the electrotherapy device. The information processing system according to claim 1 or claim 2.

4. The aforementioned information processing system includes a user terminal, The output unit causes the user terminal's display unit to display the questionnaire when the user uses the device. The first acquisition unit acquires the answers to the questionnaire entered by the user by operating the user terminal. The information processing system according to claim 2.

5. The aforementioned trained model comprises a language model capable of language processing, The user data includes first user data representing the stratified data of the user, and second user data representing at least one of the predictive models that generate data on how to use the device, effect data representing the effects of using the device, the user's personal health history data, and predictive data about the user. The first acquisition unit acquires the first user data and the second user data, The second acquisition unit generates prompt data corresponding to the first user data and the second user data by incorporating the first user data and the second user data into a pre-prepared prompt template, and acquires data related to the user's health plan output from the trained model by inputting the prompt data into the trained model. The output unit outputs data relating to the health plan, including data on how to use the device. The information processing system according to claim 1 or claim 2.

6. The aforementioned second acquisition unit is, If the first user data or the second user data contains data different from the language data, the language model is used to convert the data different from the language data into the language data, and the converted language data is input into the trained model. The information processing system according to claim 5.

7. We obtain user data, which is data about the users who use the device. By inputting the user data into a pre-generated trained model, the device usage method data output from the trained model is obtained. Outputs usage method data for the aforementioned device. An information processing method in which a computer performs the processing, The aforementioned trained model is a model that has been pre-trained based on training data, which is a model that associates training user data, which is data about the training user, with training usage data, which is data on how the training user used the device. The pre-trained model is a pre-trained model that, upon input of stratification data representing stratification when classifying the user's attributes, stratification when classifying the user's purpose of using the device for health improvement, and stratification when classifying the degree of symptoms the user wishes to improve, outputs usage method data for the device according to the user's stratification, and is a model that has been pre-trained based on training data in which the training user's stratification data and the training usage method data when the training user used the device corresponding to the stratification data are associated. The user data includes the stratified data of the user. Information processing methods.

8. We obtain user data, which is data about the users who use the device. By inputting the user data into a pre-generated trained model, the device usage method data output from the trained model is obtained. Outputs usage method data for the aforementioned device. An information processing program that causes a computer to perform a process, The aforementioned trained model is a model that has been pre-trained based on training data, which is a model that associates training user data, which is data about the training user, with training usage data, which is data on how the training user used the device. The pre-trained model is a pre-trained model that, upon input of stratification data representing stratification when classifying the user's attributes, stratification when classifying the user's purpose of using the device for health improvement, and stratification when classifying the degree of symptoms the user wishes to improve, outputs usage method data for the device according to the user's stratification, and is a model that has been pre-trained based on training data in which the training user's stratification data and the training usage method data when the training user used the device corresponding to the stratification data are associated. The user data includes the stratified data of the user. Information processing program.

Images