System

Abstract

To provide a system that can be utilized for the learning of artificial intelligence while appropriately managing the data owned by an individual or a corporation.SOLUTION: A system includes: means for acquiring data that has obtained use permission from a user; means for supplying the data that has obtained the use permission from the user to a generative AI model to generate a personalized advertisement customized for the user; and means for outputting the generated personalized advertisement to the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] The data used by AI is mainly copyright-free data that is available to anyone on the Internet. However, when using AI in a more specialized way, it may be necessary to use data owned by individuals or corporations. In such cases, it is necessary to obtain permission to use the data and properly manage the input and output data. [Means for solving the problem]

[0005] The present invention includes a means for acquiring data authorized for use by a user, a means for supplying the authorized data to a generative AI model to generate a personalized advertisement customized for the user, and a means for outputting the generated personalized advertisement to the user. The authorized data includes at least one of the user's browsing history, purchase history, and social media activity. The personalized advertisement is output in combination with an emotion engine that estimates the user's emotions. This enables individuals and corporations to obtain output data from artificial intelligence while appropriately managing data they own. [Brief explanation of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment Example 2. [Figure 15] FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, the processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0011] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0012] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0018] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0022] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0023] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0024] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] One embodiment of the present invention provides an artificial intelligence system for learning data licensed for use by an individual or corporation. This system uses a generative AI (generative pretrained transformer) to learn the data. Specifically, the licensed data is supplied as training data to the generative AI, and the system outputs the learning results.

[0029] "Example 2"

[0030] Furthermore, a means for managing data input and output is provided. This means is for properly managing the input of data for which permission has been granted and the output of learning results. Specifically, data input is provided by data providers for which permission has been granted, and output is data obtained as a result of learning. This data is properly managed and fed back to the providers as necessary.

[0031] "Example 3"

[0032] An environment for providing data is also provided. This environment is designed to safely provide data for which permission has been granted and to appropriately output learning results. Specifically, data providers provide data through this environment, and learning results are output through this environment. This allows for smooth data exchange between data providers and the learning system.

[0033] The processing flow of each embodiment will be described below.

[0034] "Example 1"

[0035] Step 1: Collect data with permission from individuals or entities. This data will serve as input for the AI ​​system to learn from.

[0036] Step 2: The collected data is fed to a generative AI (generative pretrained transformer), which then learns from this data.

[0037] Step 3: The generative AI outputs the learning results, which are used for subsequent processing and analysis.

[0038] "Example 2"

[0039] Step 1: Accept the input of authorized data. This data will be used as input for the AI ​​system to learn from.

[0040] Step 2: Properly manage input data, including storing, organizing, and controlling access to the data.

[0041] Step 3: Manage the output of learning results. The output data is stored appropriately and fed back to the provider as needed.

[0042] "Example 3"

[0043] Step 1: Prepare an environment for providing licensed data. This environment facilitates data exchange between data providers and the learning system.

[0044] Step 2: The data provider provides data through the environment. The provided data becomes input data for the AI ​​system to learn.

[0045] Step 3: The learning results are output through the environment. The output data is stored appropriately and fed back to the provider as needed.

[0046] Example 1

[0047] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0048] Conventional AI systems have had difficulty efficiently learning data licensed for use by individuals or corporations and outputting the results appropriately. Furthermore, the data preprocessing, supply, and learning processes require a lot of manual operations, which takes time and effort. Furthermore, there has been a lack of concrete methods for improving the quality of learning results using generative AI models.

[0049] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0050] In this invention, the server includes an artificial intelligence system for learning data licensed from an individual or a corporation, means for managing the input and output of the data, means for preprocessing the data, means for supplying the data to a generative AI model, means for training the generative AI model, and means for outputting the learning results, thereby enabling efficient learning of data and output of high-quality learning results.

[0051] "Individual or legal entity" refers to the entity that provides permission to use data, and includes individuals, companies, organizations, etc.

[0052] "Licensed Data" means data that has been officially licensed for use by an individual or legal entity, and may include text, images, audio, or other formats.

[0053] An "artificial intelligence system" is a computer system that learns from data and outputs the results, including generative AI models.

[0054] "Means for managing inputs and outputs" refers to the functions and processes for receiving data and providing results.

[0055] "Preprocessing means" refers to functions or processes that process data to convert it into a format suitable for learning.

[0056] A "generative AI model" refers to a generative artificial intelligence model such as the Generative Pretrained Transformer (GPT), which has the ability to learn data and generate new information.

[0057] "Means of supply" refers to the functions and processes for passing preprocessed data to a generative AI model.

[0058] "Means of learning" refers to the functions and processes used to train a generative AI model on data.

[0059] "Means for outputting learning results" refers to the functions and processes for providing the results learned by the generative AI model to the user.

[0060] MODE FOR CARRYING OUT THE INVENTION

[0061] The present invention relates to an artificial intelligence system for learning data licensed for use by an individual or a corporation. Specific embodiments of this system will be described below.

[0062] System Overview

[0063] The server will build an artificial intelligence system to learn from data provided by individuals or corporations. This system will use a generative AI model (Generative Pretrained Transformer, hereafter referred to as GPT) to learn from the data and output the results.

[0064] Hardware and software used

[0065] The server uses the following hardware and software:

[0066] Hardware: Server with high-performance CPU and GPU

[0067] Software: Python scripts for data preprocessing, GPT as a generative AI model, database system for data management

[0068] Data collection and preprocessing

[0069] A user provides the server with data for which they have received permission to use from an individual or a corporation. The data may be in various formats, such as text, images, or audio. The server preprocesses the provided data. Specifically, for text data, the server performs the following processing:

[0070] Remove unnecessary spaces and special characters

[0071] Tokenization (the process of dividing words and sentences)

[0072] Normalization (unifying processes such as converting uppercase to lowercase)

[0073] Supplying data and training the model

[0074] The server supplies the preprocessed data to GPT. The data is batch-processed to ensure efficient learning. The server trains GPT using the supplied data. The learning process involves capturing the features of the data and adjusting the weights to recognize patterns.

[0075] Output of training results

[0076] The server stores the trained GPT model and provides it to users as needed. Users can use this trained model to operate chatbots for customer support.

[0077] Specific examples

[0078] A user wants to create a chatbot for customer support using their company's customer support data. The user provides the server with past chat logs of customer support. The server preprocesses this data and feeds it to GPT for training. The trained model is then used for the user's customer support operations.

[0079] Prompt Sentence Examples

[0080] "Please use the following customer support chat logs as learning data to create a customer support chatbot."

[0081] User: I haven't received my order, what should I do?

[0082] Support: Can I have your order number?

[0083] User: 123456.

[0084] Support: Please wait a moment while we check this.

[0085] Using this prompt, the server trains the GPT and generates a chatbot for customer support.

[0086] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0087] Step 1:

[0088] Users provide the server with data that they have obtained permission to use from individuals or corporations. The input data can be in various formats, such as text, images, and audio. The server receives this data and stores it in a database. Specifically, users upload data, and the server stores it in the appropriate folder.

[0089] Step 2:

[0090] The server preprocesses the provided data. The input data is the raw data saved in step 1. For text data, the server removes unnecessary whitespace and special characters, tokenizes, and normalizes the data. The output is the preprocessed, clean data. Specifically, the server runs a Python script to tokenize and normalize the text data.

[0091] Step 3:

[0092] The server feeds the preprocessed data to the generative AI model (GPT). The input data is the data preprocessed in step 2. The server batches the data and efficiently feeds it to the model. The output is the batch of data fed to the model. Specifically, the server divides the data into chunks of a certain size and passes them to the model.

[0093] Step 4:

[0094] The server trains the generative AI model using the supplied data. The input data is the batch of data supplied in step 3. The server captures the features of the data and adjusts the weights to recognize patterns. The output is a trained generative AI model. Specifically, the server runs the model training process and updates the model parameters every epoch.

[0095] Step 5:

[0096] The server saves the generative AI model after training and provides it to the user as needed. The input data is the model that completed training in step 4. The server saves the model and makes it accessible to the user. The output is a trained model that can be used by the user. Specifically, the server saves the model as a file and provides the user with a download link.

[0097] Step 6:

[0098] A user uses the trained model to operate a customer support chatbot. The input data is an inquiry from the user. The trained model generates an appropriate response to the user's inquiry. The output is a response message to the user. In concrete terms, the user inputs a question to the chatbot and receives the response generated by the model.

[0099] (Application example 1)

[0100] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0101] Conventional advertising systems have had difficulty effectively generating personalized ads based on users' interests. Furthermore, there was a lack of technology to appropriately utilize user data and generate ads using generative AI. This resulted in low advertising effectiveness and an inability to capture users' attention.

[0102] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0103] In this invention, the server includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, means for managing the input and output of the data, means including an environment for providing the data, means for generating personalized advertisements based on the data licensed by the user, and means for displaying the generated advertisements to the user, thereby enabling the generation of effective personalized advertisements based on the user's interests and concerns.

[0104] "Personal or legal entity licensed data" means information or data that is owned and licensed by a person or legal entity.

[0105] An "artificial intelligence system" is a collection of computer programs and algorithms that learn, analyze, and generate data.

[0106] "Means for managing data input and output" are hardware and software components for controlling the receipt, storage, processing of data, and output of results.

[0107] A "data serving environment" is an infrastructure or platform for collecting, storing, accessing, and processing data.

[0108] "Data with user permission" refers to data such as personal information and behavioral history that the user has explicitly given permission to use.

[0109] "Personalized advertising" refers to advertising that is customized based on a particular user's interests.

[0110] "Generative AI (Generative Pretrained Transformer)" is an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data and content.

[0111] "Means for generating advertisements" refers to algorithms or programs for creating advertising content based on collected data.

[0112] "Means for displaying advertisements to users" are the hardware and software components for displaying the generated advertisements on the user's device.

[0113] To implement the present invention, the following system configuration and processing procedures are required.

[0114] First, the server collects data with permission from individuals or corporations and prepares an AI system to learn from it. This AI system uses a generative pretrained transformer (generative AI) to learn from the data and perform analysis and generation.

[0115] The server then has a means for managing the input and output of data. This means is a hardware and software component for controlling the reception, storage, processing, and output of data. Specifically, it uses a database management system or an API.

[0116] Furthermore, the server comprises a means for providing the data, including the environment, which is the infrastructure or platform for collecting, storing, accessing, and processing the data, such as a cloud service or an on-premise server.

[0117] Based on data with the user's permission (e.g., browsing history, purchase history, social media activity, etc.), the server has the means to generate personalized ads. This means is an algorithm or program that creates advertising content based on the collected data. Generative AI models are used to generate ads optimized for the user.

[0118] The generated advertisements are sent from the server to the user's device and displayed through a means for displaying advertisements to the user. The means are hardware and software components for displaying the generated advertisements on the user's device, specifically, a smartphone app or a web browser.

[0119] For example, if a user recently purchased a smartphone or laptop and liked an article about AI on social media, the ad generated might look like this:

[0120] Introducing a smartphone equipped with the latest AI technology! Now available at a special price. Click here for details!

[0121] An example of a prompt sentence to input to the generative AI model is as follows:

[0122] "Generate personalized ads based on your data, including:

[0123] Products and services that may interest you

[0124] Advertising catchphrase

[0125] Ad visual ideas

[0126] User Data:

[0127] Browsing history: tech news, smartphones, AI advancements

[0128] Purchase history: smartphone, laptop

[0129] Social media activity: liked AI article, shared tech news

[0130] In this way, it becomes possible to generate effective personalized advertisements based on the user's interests.

[0131] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0132] Step 1:

[0133] The server collects data that is licensed for use by individuals or corporations.

[0134] Input: Data provided by individuals or entities

[0135] Output: Collected data

[0136] Specific operation: The server obtains data provided by individuals or corporations through APIs or database connections and stores it in a database.

[0137] Step 2:

[0138] The server generates prompt sentences to supply the collected data to the generative AI model.

[0139] Input: Collected data

[0140] Output: Prompt sentence to be fed to the generative AI model

[0141] How it works: The server analyzes the collected data and generates prompts based on the user's interests. For example, it creates prompts containing appropriate advertising content based on the user's browsing history and purchase history.

[0142] Step 3:

[0143] The server generates personalized advertisements using a generative AI model.

[0144] Input: Generated prompt text

[0145] Output: Personalized ads

[0146] How it works: The server provides prompts to a generative AI model (e.g., OpenAI's GPT-3®), which then generates advertising content. The generated ads are based on the user's interests.

[0147] Step 4:

[0148] The server transmits the generated advertisement to the user's terminal.

[0149] Input: Generated personalized ad

[0150] Output: Ad sent to user's device

[0151] How it works: The server sends the generated ad to the user's smartphone app or web browser using push notifications or APIs.

[0152] Step 5:

[0153] The device displays the received advertisement to the user.

[0154] Input: Advertisement sent by the server

[0155] Output: The ad shown to the user

[0156] What happens: The user's device (smartphone or PC) displays the received ad. The ad appears as a banner or pop-up within the app.

[0157] In this way, effective personalized advertisements based on the user's interests can be generated and displayed.

[0158] Example 2

[0159] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0160] Conventional AI systems have insufficient management of data input and output, making it difficult to properly manage data for which permission has been granted and to provide feedback on learning results. Feedback to data providers is often delayed, hindering efficient data use. This makes it difficult to build a relationship of trust between data providers and users, resulting in delays in data provision.

[0161] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0162] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, an environment for providing the data, a means for uploading the data to the server, a means for saving the data in a data management system, a means for inputting the data into a generative AI model for learning, a means for saving the learning results in the data management system, and a means for feeding back the learning results to the provider, thereby enabling appropriate management of data and rapid feedback.

[0163] An "artificial intelligence system" is a computer program that learns from data licensed by an individual or legal entity to perform a specific task.

[0164] "Means for managing data input and output" refers to a system that has the functionality to appropriately manage data input from data providers and the output of learning results.

[0165] The "data providing environment" is a system that includes an interface and a protocol for data providers to upload data to a server.

[0166] "Means for uploading to the server" refers to a system that has a function for users to send data obtained from providers to the server.

[0167] "Data Management System" means a system for storing uploaded data and verifying its integrity and permission to use.

[0168] A "generative AI model" is a model that learns from input data and performs specific tasks.

[0169] A "means for learning" is a system that inputs data into a generative AI model and runs a process to optimize the model's parameters.

[0170] "Means for storing learning results in a data management system" refers to a system that has the functionality to store the learning results obtained from a generative AI model in a data management system.

[0171] "Means for feeding back learning results to providers" refers to a system that has the function of reporting learning results to providers and providing feedback.

[0172] This invention is a system that includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, and an environment for providing the data. Specifically, the server, terminal, and user each play their respective roles, achieving appropriate management of data and rapid feedback.

[0173] Hardware and software used

[0174] Hardware: Servers, terminals

[0175] Software: Data management system, generative AI models

[0176] Data entry management

[0177] The user prepares the data for which they have received permission to use from the provider. Next, the user uses their device to upload the data they received from the provider to the server. The server receives the uploaded data and stores it in a data management system. Furthermore, the server checks the integrity and permission to use of the data to ensure that it is managed appropriately.

[0178] As a specific example of operation, a user uploads image data obtained from a provider to a server via a browser on their device. The server stores the uploaded image data in a database and records meta information about the data (such as the provider and usage permission details).

[0179] Data training

[0180] The server retrieves the data stored in the data management system. Then, the server inputs the retrieved data into the generative AI model to start the learning process. The generative AI model learns based on the input data and optimizes the model parameters. The server saves the generative AI model after learning is complete.

[0181] As a specific example of how it works, the server retrieves image data from a database and inputs it into a generative AI model. The generative AI model learns to classify cats and dogs based on the image data and adjusts the model parameters. The server then saves the model once it has completed learning, preparing it for the next step.

[0182] Output management of learning results

[0183] The server stores the learning results obtained from the generative AI model in a data management system. The device then feeds back the learning results to the provider as needed. The user then uses the device to check the learning results and report them to the provider.

[0184] As a specific example of how it works, the server stores the learning results (e.g., a cat / dog classification model) in a database. The user can then use their device to check the learning results and provide feedback to the provider via email or a dashboard.

[0185] Prompt Sentence Examples

[0186] Examples of prompts to be input to a generative AI model include:

[0187] "Use this image data to generate a model that classifies cats and dogs."

[0188] "Create a model that performs sentiment analysis based on the provided text data."

[0189] The above is an embodiment of the present invention.

[0190] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0191] Step 1:

[0192] Preparing and uploading data

[0193] The user prepares the data for which permission to use has been obtained from the provider. Next, the user uses the terminal to upload the data obtained from the provider to the server. Specifically, the user uploads the image data obtained from the provider to the server through the terminal's browser. The input is the image data obtained from the provider, and the output is the image data stored on the server.

[0194] Step 2:

[0195] Data storage and management

[0196] The server receives the uploaded data and stores it in the data management system. Furthermore, the server checks the data integrity and usage permissions to ensure that the data is managed appropriately. Specifically, the server stores the uploaded image data in a database and records the data's meta information (provider, usage permission details, etc.). The input is the uploaded image data, and the output is the data and its meta information stored in the database.

[0197] Step 3:

[0198] Obtaining data and starting training

[0199] The server retrieves data stored in the data management system. The server then inputs the retrieved data into the generative AI model and starts the learning process. Specifically, the server retrieves image data from the database and inputs it into the generative AI model. The input is the image data retrieved from the database, and the output is the data input into the generative AI model.

[0200] Step 4:

[0201] Training generative AI models

[0202] The generative AI model learns based on the input data and optimizes the model parameters. Specifically, the generative AI model learns to classify cats and dogs based on image data and adjusts the model parameters. The input is the image data entered into the generative AI model, and the output is the model after learning is complete.

[0203] Step 5:

[0204] Saving training results

[0205] The server stores the generative AI model after training. Specifically, the server stores the training results (for example, a cat / dog classification model) in a database. The input is the trained model, and the output is the training results stored in the database.

[0206] Step 6:

[0207] Feedback on learning results

[0208] The device feeds back the learning results to the provider as needed. The user uses the device to check the learning results and report them to the provider. Specifically, the user uses the device to check the learning results and provide feedback to the provider via email or dashboard. The input is the learning results stored in the database, and the output is the information fed back to the provider.

[0209] (Application example 2)

[0210] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0211] Conventional content recommendation systems have difficulty recommending optimal content to users because they are unable to effectively utilize user viewing history and rating data. Furthermore, there are problems with insufficient management of data input and output, and inappropriate feedback to data providers. This has hindered improvements to the user experience.

[0212] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0213] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, an environment for providing the data, a means for a generative AI model to recommend optimal content to users using data provided by the user, and a means for providing feedback on the recommended content to the user. This makes it possible to effectively utilize users' viewing history and rating data to recommend optimal content to users. Furthermore, the appropriate management of data input and output and appropriate feedback to data providers improves the user experience.

[0214] "Person or Legal Entity" means a natural person or a legally recognized entity.

[0215] "Permission to use" refers to obtaining formal permission from the provider to use the data.

[0216] "Data" means a collection of information that is collected, stored and processed for a specific purpose.

[0217] An "artificial intelligence system" refers to a system that uses technologies such as machine learning and deep learning to analyze data and automatically perform specific tasks.

[0218] "Means for managing data input and output" refers to methods and devices for appropriately collecting, storing, processing, and outputting data.

[0219] "Data providing environment" refers to the physical or virtual location or system where data is collected or provided.

[0220] A "generative AI model" refers to a model that uses a generative AI (generative pretrained transformer) to analyze data and generate new information and content.

[0221] "Means for recommending optimal content to users" refers to methods or devices that select and present the most appropriate content for a user based on user data.

[0222] "Feedback means" refers to a method or device for returning results or information generated by the system to the user or data provider.

[0223] The following system configuration will be described as an embodiment of the present invention.

[0224] First, the server is equipped with an artificial intelligence system for learning data licensed by individuals or corporations. This artificial intelligence system analyzes the data using a generative pretrained transformer (generative AI) and generates new information and content. The server also has a means for managing data input and output, and appropriately collects, stores, processes, and outputs data.

[0225] Data provided by users (viewing history, ratings, comments, etc.) is sent to the server. Based on this data, the server uses a generative AI model to recommend the most suitable content to the user. Specifically, the server converts the user's viewing history and rating data into TF-IDF vectors and calculates the cosine similarity to select the most suitable content for the user.

[0226] The server also has a means for providing feedback on the recommended content to the user, allowing the user to easily find content that matches their preferences.

[0227] The hardware used includes servers, users' smartphones, PCs, etc. The software used is Python and its libraries (scikit-learn, json).

[0228] As a concrete example, consider a system that uses data on the movies a user has watched to recommend the next movie to watch. Here is an example prompt:

[0229] User ID: user123

[0230] Viewing history: ["Movie A", "Movie B", "Movie C"]

[0231] Rating: {"Movie A": 5, "Movie B": 4, "Movie C": 3}

[0232] Comments: {"Movie A": "Great movie", "Movie B": "Good", "Movie C": "Okay"}

[0233] By inputting this prompt into a generative AI model, the system can recommend the most suitable movie to the user. This system effectively utilizes the user's viewing history and rating data to recommend the most suitable content to the user. In addition, the system properly manages data input and output, and provides appropriate feedback to data providers, improving the user experience.

[0234] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0235] Step 1:

[0236] The user inputs data such as viewing history, ratings, and comments from the terminal.

[0237] Input: User's viewing history, ratings, comments

[0238] Output: User data (viewing history, ratings, comments)

[0239] Specific operation: Users use their smartphones or PCs to enter their viewing history, ratings, and comments. This data is then sent from the device to the server.

[0240] Step 2:

[0241] The server receives the user data and stores it in JSON format.

[0242] Input: User data (viewing history, ratings, comments)

[0243] Output: User data in JSON format

[0244] Specific operation: The server receives user data sent from the device and stores it in JSON format, which ensures data consistency and readability.

[0245] Step 3:

[0246] The server analyzes the data using a generative AI model based on user data.

[0247] Input: User data in JSON format

[0248] Output: Analysis results (user profile)

[0249] Specific operation: The server uses Python and the scikit-learn library to convert user data into TF-IDF vectors and generate user profiles.

[0250] Step 4:

[0251] The server reads the content database and compares it to the user profile.

[0252] Input: User profile, content database

[0253] Output: Similarity score

[0254] Specific operation: The server reads the content database and compares the user profile with the profile of each content using cosine similarity.

[0255] Step 5:

[0256] The server recommends the most suitable content to the user based on the similarity score.

[0257] Input: Similarity score

[0258] Output: Recommended content list

[0259] Specific operation: The server selects the most suitable content for the user based on the similarity score and generates a recommended content list.

[0260] Step 6:

[0261] The server feeds back the recommended content list to the user.

[0262] Input: Recommended content list

[0263] Output: Feedback to the user (recommended content)

[0264] Specific operation: The server sends the generated recommended content list to the user's device and provides feedback to the user, who can then check the recommended content on the device.

[0265] Example 3

[0266] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0267] Conventional AI systems lack a consistent mechanism for data input from data providers, safe storage of data, learning from the data using generative AI models, saving of learning results, and providing results to users. This leads to problems such as data exchange not being smooth, and the reliability and security of learning results not being ensured.

[0268] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0269] In this invention, the server includes a means for the terminal to input data from a data provider, a means for the server to receive and store the data, a means for the server to input the data to the generative AI model, a means for the generative AI model to learn from the data and generate results, a means for the server to store the learning results, and a means for the user to request and obtain the learning results. This allows smooth data exchange between the data provider and the learning system, enabling the user to quickly obtain the learning results.

[0270] "Individual or Legal Entity" refers to the natural person or legal entity recognized as the data provider.

[0271] "Licensed Data" refers to data that has been officially licensed for use by the data provider.

[0272] "Artificial intelligence system" refers to a computer system for learning from data and generating analytical results.

[0273] "Means for managing data input and output" refers to the functionality for receiving, storing, analyzing data, and outputting the results.

[0274] "Data provision environment" refers to the system environment in which data providers can safely provide data and appropriately output learning results.

[0275] "Terminal" means the computer or device used by the Data Contributor to enter data.

[0276] "Server" refers to a central processing unit for receiving, storing, analyzing data, and outputting results.

[0277] A "generative AI model" refers to an artificial intelligence model that learns from data and generates analytical results.

[0278] "Means for receiving and storing data" refers to the function of receiving data sent from a terminal and storing it in a database, etc.

[0279] "Means for inputting data" refers to the interface or function that allows data providers to input data.

[0280] "Means of learning from data and generating results" refers to the function by which a generative AI model analyzes data and generates learning results.

[0281] "Means for saving learning results" refers to the function of saving the learning results generated by the generative AI model in a database, etc.

[0282] "Means for requesting and obtaining learning results" refers to the interface or functionality that allows users to request and obtain learning results.

[0283] "User" refers to an individual or legal entity that accesses the system to utilize the learning results.

[0284] This invention is an AI system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, and an environment for providing the data. Specifically, the system processes and calculates the data using terminals, a server, and a generative AI model.

[0285] Hardware and software used

[0286] Hardware: Servers, terminals

[0287] Software: Generative AI models (e.g., GPT-4®), data management systems (e.g., SQL databases)

[0288] System configuration

[0289] 1. Terminal: A device used by data providers to input data. For example, used by medical staff to input patient symptom data.

[0290] 2. Server: A central processing unit that receives, stores, analyzes, and outputs data. The server uses a data management system to store data and generative AI models to analyze the data.

[0291] 3. Generative AI model: An artificial intelligence model that learns from data and generates analytical results. For example, GPT-4 analyzes patient symptom data and generates a diagnosis.

[0292] System Operation

[0293] Data provider enters data: Data providers enter data using dedicated terminals. For example, medical institution staff enter information such as the patient's name, age, and symptoms.

[0294] The server receives and stores the data: The server receives the data sent from the device and stores it in the data management system. When storing the data, it checks the integrity and security of the data.

[0295] The server inputs data into the generative AI model: The server inputs the stored data into the generative AI model, for example, inputting patient symptom data into the GPT-4 model.

[0296] Generative AI models learn from data and generate results: Generative AI models learn from input data and generate results. For example, GPT-4 analyzes a patient's symptom data and generates a diagnosis.

[0297] The server stores the learning results: The server receives the learning results output from the generative AI model and stores them in a data management system.

[0298] User requests and obtains learning results: A user requests learning results using a dedicated device. For example, a doctor obtains diagnosis results.

[0299] Examples of concrete examples and prompts

[0300] Example: A medical institution provides patient data and uses a generative AI model (GPT-4) to generate a diagnosis. A doctor obtains the diagnosis.

[0301] Example prompt sentence:

[0302] Input prompt: "Enter patient symptom data and generate a diagnosis."

[0303] Output prompt: "The diagnostic results are as follows:..."

[0304] This system allows smooth data exchange between the data provider and the learning system, enabling users to quickly obtain learning results. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0305] Step 1:

[0306] The terminal inputs data from the data provider.

[0307] Data providers input data using a dedicated terminal. For example, a medical institution staff member inputs information such as the patient's name, age, and symptoms. The input data is structured based on the input form on the terminal. Specifically, the data provider inputs the required information into the input form on the terminal and clicks the "Submit" button.

[0308] Step 2:

[0309] The server receives and stores the data.

[0310] Data sent from the device is sent to the server using a secure communication protocol (e.g., HTTPS). The server stores the received data in a data management system (e.g., an SQL database). When storing the data, the server checks the integrity and security of the data. Specifically, the server receives the data and executes an SQL query to store it in the database.

[0311] Step 3:

[0312] The server inputs data into the generative AI model.

[0313] The server inputs the stored data into the generative AI model. For example, it inputs patient symptom data into the GPT-4 model. The input data is sent to the API of the generative AI model. Specifically, the server retrieves the necessary data from the database and creates a request to send the data to the API of the generative AI model.

[0314] Step 4:

[0315] Generative AI models learn from data and generate results.

[0316] A generative AI model learns from input data and generates results. For example, GPT-4 analyzes a patient's symptom data and generates a diagnosis. Specifically, the generative AI model analyzes the data and outputs the diagnosis and recommended treatment in text format.

[0317] Step 5:

[0318] The server stores the learning results.

[0319] The server receives the learning results output from the generative AI model and stores them in a data management system. When saving, it checks the integrity and security of the results. Specifically, the server receives the learning results and executes SQL queries to save them in a database.

[0320] Step 6:

[0321] The user requests and obtains the learning results.

[0322] A user requests learning results using a dedicated terminal. For example, a doctor obtains diagnosis results. The user enters a patient's ID into the search form on the terminal and clicks the "Search" button. The server receives the request and provides the learning results to the user. Specifically, the server retrieves the learning results from the database and generates a response to display the results on the user's terminal.

[0323] (Application example 3)

[0324] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0325] Conventional artificial intelligence systems have faced the problem of lacking an environment in which data providers can safely provide data and appropriately output learning results. Furthermore, data security is often not ensured during data provision and learning result output, creating the risk of data leaks and unauthorized use. This has led to a demand for an environment in which data providers can provide data with peace of mind.

[0326] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0327] In this invention, the server includes an artificial intelligence system for learning data licensed from an individual or a corporation, a means for managing the input and output of data, a means including an environment for providing data, a means for encrypting and providing data, and a means for decrypting the encrypted data and outputting the learning results. This enables data providers to provide data safely and output the learning results appropriately.

[0328] "Person or legal entity" means a natural person or a legal entity.

[0329] "Permission to use" refers to the permission given by the data provider to use the data.

[0330] "Data" refers to a collection of information, the material that artificial intelligence systems use to learn.

[0331] "Artificial intelligence system" refers to a computer program that learns from data and performs specific tasks.

[0332] "Means for managing input and output" refers to the functions for controlling the receipt of data and the delivery of results.

[0333] "Data provision environment" refers to the systems and infrastructure that enable data providers to provide data securely.

[0334] "Encryption" refers to the conversion of data using a specific algorithm to make it in a form that cannot be easily understood by third parties.

[0335] "Decryption" refers to restoring encrypted data to its original form.

[0336] "Learning results" refers to the output that an AI system generates after learning from data.

[0337] "Generative AI (Generative Pretrained Transformer)" refers to an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data.

[0338] As an embodiment of the present invention, the following system can be constructed.

[0339] System Overview

[0340] This system includes an artificial intelligence system for learning data that has been granted permission to be used by an individual or a corporation, a means for managing the input and output of data, an environment for providing the data, a means for encrypting and providing the data, and a means for decrypting the encrypted data and outputting the learning results.

[0341] Hardware and software used

[0342] Hardware: Smartphones, servers

[0343] Software: Python, requests library, cryptography library, Generative Pretrained Transformer

[0344] Program processing explanation

[0345] Data encryption and delivery

[0346] When a user provides data using their smartphone, the data is first encrypted using the Fernet class from the cryptography library. The encrypted data is then sent to the server using an HTTP POST request using the requests library.

[0347] Obtaining and decoding the training results

[0348] The server decrypts the encrypted data it receives and uses a generative pretrained transformer (Generative AI) to learn from the data. The learning results are then re-encrypted and decrypted when the user retrieves them using their smartphone.

[0349] Specific examples

[0350] For example, if a user is using a smartphone to provide "user behavior data," the following prompt sentence can be used:

[0351] Prompt Sentence Examples

[0352] "You provide user behavior data, and the AI ​​model learns from that data and outputs results that predict user behavior patterns."

[0353] In this way, a system can be realized that provides an environment in which data providers can safely provide data and appropriately output learning results.

[0354] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0355] Step 1:

[0356] A user enters data using a smartphone.

[0357] Input: User-provided data (e.g., user behavior data)

[0358] Specific operation: The user launches the smartphone application and enters the data to be provided.

[0359] Step 2:

[0360] The device encrypts the data.

[0361] Input: User-supplied data

[0362] Data processing: Encrypt the data using the Fernet class from the cryptography library.

[0363] Output: Encrypted data

[0364] Specific operation: The terminal encrypts the data entered by the user using the Fernet class to generate encrypted data.

[0365] Step 3:

[0366] The device sends the encrypted data to the server.

[0367] Input: Encrypted data

[0368] Data operations: Send data using an HTTP POST request.

[0369] Output: Data sent to the server

[0370] Specific operation: The device uses the requests library to send the encrypted data to the server via an HTTP POST request.

[0371] Step 4:

[0372] The server receives the encrypted data and decrypts it.

[0373] Input: Encrypted data

[0374] Data processing: Decrypt the data using the Fernet class from the cryptography library.

[0375] Output: Decrypted data

[0376] Specific operation: The server decrypts the received encrypted data using the Fernet class to obtain the original data.

[0377] Step 5:

[0378] The server inputs the decrypted data into a generative AI model for learning.

[0379] Input: Decrypted data

[0380] Data computation: Learn data using a generative AI (generative pretrained transformer).

[0381] Output: Learning results

[0382] Specific operation: The server inputs the decoded data into the generative AI model, which learns the data and generates the learning results.

[0383] Step 6:

[0384] The server encrypts the learning results.

[0385] Input: Learning results

[0386] Data processing: Encrypt the training results using the Fernet class from the cryptography library.

[0387] Output: Encrypted training results

[0388] Specific operation: The server encrypts the generated learning result using the Fernet class and generates the encrypted learning result.

[0389] Step 7:

[0390] The server sends the encrypted learning results to the terminal.

[0391] Input: Encrypted learning results

[0392] Data operation: Sends data using an HTTP GET request.

[0393] Output: Data sent to the terminal

[0394] Specific operation: The server uses the requests library to send the encrypted learning results to the terminal via an HTTP GET request.

[0395] Step 8:

[0396] The terminal receives the encrypted learning result and decrypts it.

[0397] Input: Encrypted learning results

[0398] Data processing: Decrypt the training results using the Fernet class from the cryptography library.

[0399] Output: Decoded training results

[0400] Specific operation: The terminal decrypts the received encrypted learning result using the Fernet class and obtains the original learning result.

[0401] Step 9:

[0402] The user checks the decoded learning results.

[0403] Input: Decoded training results

[0404] Output: Learning results for user confirmation

[0405] Specific operation: The user checks the decoded learning results through a smartphone application.

[0406] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0407] "Example 1"

[0408] As one embodiment of the present invention, an artificial intelligence system including an emotion engine is provided. This system learns from data licensed for use by individuals or corporations and manages the input and output of that data. The emotion engine also recognizes a user's emotion and adjusts learning based on that emotion. Specifically, when a user inputs text or voice, the emotion engine recognizes the user's emotion from that input. The AI ​​system then adjusts learning based on the recognized emotion. For example, if a user inputs text or voice that indicates the emotion of joy, the emotion engine recognizes the emotion of joy, and the AI ​​system learns based on that emotion.

[0409] "Example 2"

[0410] Another embodiment of the present invention provides a specific technique for an emotion engine to recognize a user's emotion. Specifically, the emotion engine uses natural language processing (NLP) technology to recognize emotions from a user's text input. Also, it uses speech recognition technology to recognize emotions from a user's voice input. For example, if a user inputs the text "I'm very happy," the emotion engine uses NLP technology to recognize the emotion of joy from the text. Also, if a user inputs a voice containing the emotion of joy, the emotion engine uses speech recognition technology to recognize the emotion of joy from the voice.

[0411] "Example 3"

[0412] In yet another embodiment of the present invention, a specific technique is provided for adjusting the learning of an AI system based on emotions recognized by an emotion engine. Specifically, the emotion engine feeds back the recognized emotions to the AI ​​system, and the AI ​​system adjusts its learning based on the feedback. For example, if the emotion engine recognizes the emotion of joy, it feeds back that information to the AI ​​system. Then, the AI ​​system strengthens its learning of data related to the emotion of joy based on the feedback.

[0413] The processing flow of each embodiment will be described below.

[0414] "Example 1"

[0415] Step 1: The user enters text or voice input.

[0416] Step 2: The emotion engine recognizes the user's emotion from the input.

[0417] Step 3: Adjust the AI ​​system's learning based on the recognized emotions.

[0418] "Example 2"

[0419] Step 1: The user enters text.

[0420] Step 2: The emotion engine uses natural language processing (NLP) techniques to recognize emotions from the text.

[0421] Step 3: The user provides voice input.

[0422] Step 4: The emotion engine uses speech recognition technology to recognize emotions from the speech.

[0423] "Example 3"

[0424] Step 1: The emotion engine recognizes the user's emotion.

[0425] Step 2: The recognized emotions are fed back to the AI ​​system.

[0426] Step 3: The AI ​​system adjusts its learning based on that feedback.

[0427] Example 1

[0428] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0429] Conventional AI systems have difficulty effectively learning from data that has been authorized for use by individuals or corporations, and have not adjusted their learning to take user emotions into account. As a result, they have been unable to obtain appropriate learning results based on user emotions, limiting the flexibility and accuracy of the system.

[0430] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0431] In this invention, the server includes an AI system for learning data licensed from an individual or a corporation, means for managing the input and output of the data, means for preprocessing the data, means for learning the data using a generative AI model, and means for recognizing a user's emotions and adjusting learning based on the emotions, thereby enabling flexible and highly accurate learning results that take the user's emotions into consideration.

[0432] "Individual or Legal Entity" refers to a specific individual or legal entity that provides permission to use data.

[0433] "Licensed data" refers to data that has been explicitly licensed by an individual or legal entity and is used to train an artificial intelligence system.

[0434] "Artificial intelligence system" refers to a computer program and associated hardware for learning from data and performing specific tasks.

[0435] "Means for managing input and output" refers to software and hardware for controlling the input and output of data and for appropriate processing.

[0436] "Pre-processing means" refers to software and hardware for pre-processing data, such as cleaning, formatting, and tokenization.

[0437] "Generative artificial intelligence model" refers to a generative AI model such as the Generative Pretrained Transformer, which has the ability to generate new information based on data.

[0438] "Means for recognizing emotions and adjusting learning" refers to software and hardware for analyzing a user's emotions and dynamically adjusting the learning process of an artificial intelligence system based on those emotions.

[0439] "Data serving environment" refers to the infrastructure for collecting, storing, processing, and serving data.

[0440] The present invention relates to an AI system for learning from data licensed by individuals or corporations. The system utilizes a generative AI model and has the ability to adjust learning based on user emotions.

[0441] System configuration

[0442] Data collection and permission acquisition

[0443] A user provides the system with data for which they have received permission to use from an individual or corporation. The terminal then sends the data provided by the user to the server. The server then checks the permission to use the received data and stores it in a database. Specifically, a database management system (e.g., MySQL (registered trademark) or PostgreSQL) is used.

[0444] Data Preprocessing

[0445] The server preprocesses the collected data. Specifically, it cleans the data (filling in missing values ​​and removing outliers) and converts the format (tokenizing text data, etc.). The server cleans the data using Python's pandas library and converts the numeric data using NumPy.

[0446] Learning generative AI models

[0447] The server supplies the preprocessed data to a generative AI model (Generative Pretrained Transformer). The server uses an NVIDIA® GPU (e.g., Tesla® V100) to train the generative AI model using TENSORFLOW® or PyTorch. The server monitors the model's performance during training and adjusts hyperparameters as necessary.

[0448] Use of emotion engine

[0449] When a user inputs text or voice, the device sends the input to a server. The server uses an emotion engine to recognize the user's emotions. Specifically, it uses the NLP library Hugging Face Transformers to extract emotions from text data. The server then adjusts the learning of the generative AI based on the recognized emotions. For example, if a user inputs text that expresses joy, it prioritizes positive data in the learning process.

[0450] Output of training results

[0451] The server outputs the learning results of the generative AI and provides them to the user via a terminal. The server generates the learning results in text and graph format and displays them in a way that is easy for the user to understand. Specifically, graphs are created using Matplotlib and Seaborn.

[0452] Specific examples

[0453] For example, if the user enters the text "I'm so happy today!", the specific behavior is as follows:

[0454] 1. The user types, "I'm so happy today!"

[0455] 2. The device sends this text to the server.

[0456] 3. The server uses an emotion engine to recognize the emotion "joy" from this text.

[0457] 4. Based on this emotion of "joy," the server adjusts the training dataset for the generative AI model, giving priority to positive data.

[0458] 5. The server generates the learning results and provides the user with "text generation results that reflect positive emotions" via the terminal.

[0459] Prompt Sentence Examples

[0460] Below are some example prompts to input to a generative AI model:

[0461] If a user types "I'm so happy today!", how would the emotion engine respond and adjust the generative AI's learning?

[0462] In this way, the system effectively learns from the data using generative AI while taking into account the user's emotions.

[0463] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0464] Step 1:

[0465] Data collection and permission acquisition

[0466] A user provides the system with data for which they have received permission to use from an individual or corporation. The terminal sends the data provided by the user to the server. The server checks the permission to use the received data and stores it in a database. Specifically, a database management system (e.g., MySQL (registered trademark) or PostgreSQL) is used. The input is the data provided by the user, and the output is the data stored in the database.

[0467] Step 2:

[0468] Data Preprocessing

[0469] The server preprocesses the collected data. Specifically, it cleans the data (filling in missing values ​​and removing outliers) and converts the format (tokenizing text data, etc.). The server cleans the data using Python's pandas library and converts numerical data using NumPy. The input is raw data obtained from the database, and the output is preprocessed data.

[0470] Step 3:

[0471] Learning generative AI models

[0472] The server supplies the preprocessed data to a generative AI model (Generative Pretrained Transformer). The server uses an NVIDIA GPU (e.g., Tesla V100) to train the generative AI model using TensorFlow or PyTorch. The server monitors the model's performance during training and adjusts hyperparameters as necessary. The input is the preprocessed data, and the output is a trained generative AI model.

[0473] Step 4:

[0474] Use of emotion engine

[0475] When a user inputs text or voice, the device sends the input to the server. The server uses an emotion engine to recognize the user's emotions. Specifically, it uses the NLP library Hugging Face Transformers to extract emotions from text data. The server then adjusts the learning of the generative AI based on the recognized emotions. For example, if a user inputs text that expresses the emotion of joy, it prioritizes learning positive data. The input is text or voice data from the user, and the output is the emotion recognition results and adjusted learning data.

[0476] Step 5:

[0477] Output of training results

[0478] The server outputs the learning results of the generative AI and provides them to the user via a terminal. The server generates the learning results in text or graph format and displays them in a way that is easy for the user to understand. Specifically, graphs are created using Matplotlib or Seaborn. The input is the trained generative AI model, and the output is the text generation results and graphs provided to the user.

[0479] (Application example 1)

[0480] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0481] Conventional AI systems lack the ability to recognize user emotions and adjust learning based on those emotions, making it difficult to recommend products that meet individual user needs and emotions. Furthermore, they lack the ability to integrate real-time emotion analysis and product recommendations to improve the shopping experience in brick-and-mortar stores.

[0482] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0483] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, means for managing the input and output of the data, means including an environment for providing the data, means including an emotion engine that recognizes a user's emotions and adjusts learning based on the emotions, means for recommending products based on the user's emotions and past purchase history, and means for analyzing the user's emotions when viewing or touching products in real time and providing appropriate product information and coupons. This enables personalized product recommendations based on the user's emotions and an improved real-time shopping experience in physical stores.

[0484] "Data licensed for use by an individual or corporation" means data owned by a specific individual or corporation, and for which formal permission has been given to use the data.

[0485] An "artificial intelligence system" is a collection of computer programs or hardware designed to learn from data and perform specific tasks automatically.

[0486] "Means for managing data input and output" refers to functions or devices for controlling the receipt, processing, and output of results of data.

[0487] A "data serving environment" is the physical or virtual infrastructure in which data is properly collected, stored, and accessed.

[0488] An "emotion engine" is a software or hardware component that recognizes a user's emotions and adjusts the system's behavior based on those emotions.

[0489] "Means for recommending products based on a user's emotions and past purchase history" refers to an algorithm or system that analyzes a user's emotional state and past purchase data and suggests appropriate products based on that.

[0490] "Means for analyzing in real time and providing appropriate product information and coupons" refers to a function or device that instantly evaluates a user's current behavior and emotions and instantly provides appropriate product information and discount coupons.

[0491] "Generative AI (Generative Pretrained Transformer)" is an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data.

[0492] As an embodiment of the present invention, a "smart shopping assistant" system that improves the shopping experience in brick-and-mortar stores will be described as an example.

[0493] System Configuration

[0494] The system consists of the following main components:

[0495] 1. Server:

[0496] This includes artificial intelligence systems for learning from data licensed for use by individuals or entities.

[0497] It uses generative AI (Generative Pretrained Transformer) to learn from data and recommend products based on the user's emotions and past purchase history.

[0498] It includes an emotion engine that recognizes the user's emotions and adjusts learning based on those emotions.

[0499] 2. Device (smartphone):

[0500] It uses a camera and microphone to capture the user's facial expressions and voice and recognize emotions.

[0501] It analyzes users' emotions in real time when they look at or touch a product, and provides appropriate product information and coupons.

[0502] Program processing explanation

[0503] The server processes data using the following hardware and software:

[0504] Hardware: Smartphone camera and microphone

[0505] Software: OpenCV (image processing), Transformers (generative AI model), EmotionRecognizer (emotion recognition)

[0506] Data processing and calculation

[0507] 1. Emotion recognition:

[0508] The device's camera and microphone are used to capture the user's facial expressions and voice, and emotions are recognized using EmotionRecognizer.

[0509] 2. Product recommendation:

[0510] Based on the recognized emotions and the user's past purchase history data, a generative AI model (GPT-3) is used to recommend products.

[0511] 3. Real-time feedback:

[0512] It analyzes users' emotions in real time when they look at or touch a product, and provides appropriate product information and coupons.

[0513] Specific examples

[0514] For example, when a user uses a smartphone in a physical store, the camera captures the user's facial expressions and performs emotion recognition. If the user is smiling, the system recognizes the emotion of "joy" and recommends "new fashion items" based on past product purchase data.

[0515] Prompt Sentence Examples

[0516] The user's emotion is joy. Their past purchase history is "fashion items, accessories." Please recommend appropriate products.

[0517] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0518] Step 1:

[0519] It uses the device's camera and microphone to capture the user's facial expressions and voice.

[0520] Input: User's facial expression images and voice data

[0521] Output: Captured facial expression images and audio data

[0522] How it works: The device's camera captures a picture of the user's face and the microphone records the user's voice. These data are collected in real time.

[0523] Step 2:

[0524] The device inputs the captured facial expression images and voice data into EmotionRecognizer to recognize the user's emotions.

[0525] Input: Captured facial expression images and audio data

[0526] Output: Recognized user emotion (e.g., happy, sad, surprised, etc.)

[0527] How it works: EmotionRecognizer analyzes facial images and voice data to identify the user's emotions. For example, it recognizes the emotion "joy" from an image of a smiling face.

[0528] Step 3:

[0529] The server uses a generative AI model (GPT-3) to recommend products based on the recognized emotions and the user's past purchase history data.

[0530] Input: Recognized user sentiment, user's past purchase history data

[0531] Output: Recommended product information

[0532] Specific operation: The server inputs a prompt into the generative AI model and recommends appropriate products. For example, the prompt might be, "The user's emotion is joy. Their past purchase history is 'fashion items, accessories.' Please recommend appropriate products."

[0533] Step 4:

[0534] The server transmits the recommended product information to the terminal, which then displays the information to the user.

[0535] Input: Recommended product information

[0536] Output: Product information displayed to the user

[0537] Specific operation: The server sends the recommended product information to the terminal, which then displays the information on the user's screen. The user can then check the displayed product information.

[0538] Step 5:

[0539] The device recaptures and analyzes the user's emotions in real time when they look at or touch the product.

[0540] Input: User's facial expression images and voice data (recaptured)

[0541] Output: Re-recognized user emotion

[0542] Specific operation: The device's camera and microphone capture the user's facial expressions and voice again, and EmotionRecognizer recognizes emotions again.

[0543] Step 6:

[0544] The server generates appropriate product information and coupons based on the re-recognized emotion and transmits them to the terminal.

[0545] Input: Re-recognized user emotion

[0546] Output: Generated product information and coupons

[0547] Specific operation: Based on the re-recognized emotions, the server uses a generative AI model to generate new product information and coupons, which are then sent to the device, which then displays them to the user.

[0548] Example 2

[0549] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0550] Conventional AI systems have had problems with insufficient management of data input and output, resulting in low accuracy, especially in emotion recognition. Furthermore, there is a lack of means to accurately recognize user emotions, making it difficult to recognize emotions from text or voice.

[0551] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes an artificial intelligence system for learning data whose use is permitted by an individual or a corporation, means for managing the input of the data, means for managing the output of the data, means including an emotion engine for recognizing a user's emotion, means for recognizing emotion from text input using natural language processing technology, and means for recognizing emotion from voice input using voice recognition technology. This allows appropriate management of data input and output, making it possible to accurately recognize a user's emotion.

[0552] "Person or legal entity" means a natural person or a legal entity.

[0553] "Licensed Data" means data used with explicit permission from the data provider.

[0554] An "artificial intelligence system" refers to a system that analyzes data and learns using technologies such as machine learning and deep learning.

[0555] "Means for managing data input" refers to the means for properly storing and processing data received from data providers.

[0556] "Means for managing data output" refers to the means for properly storing learning results and analysis results and providing feedback to providers as necessary.

[0557] "Emotion engine" refers to a software or hardware component for recognizing a user's emotions.

[0558] "Natural language processing technology" refers to technology for analyzing text data and understanding meaning and emotions.

[0559] "Speech recognition technology" refers to technology for analyzing voice data and converting it into text.

[0560] A "generative artificial intelligence model" refers to an artificial intelligence technology that generates new data using a model that has been trained in advance using large amounts of data.

[0561] "Copyright-free data that can be viewed by anyone on the Internet" refers to data that is publicly available on the Internet and can be used freely because it does not have copyright.

[0562] The present invention is a system that includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, a means for managing the input and output of the data, and an emotion engine for recognizing a user's emotions.

[0563] System configuration

[0564] 1. Managing data input and output

[0565] Enter the data:

[0566] The server receives data for which permission to use has been granted from the data provider, such as a user's behavior log or survey results.

[0567] Specific operation: The server receives data in JSON format from the data provider via an HTTP request, and stores the received data in a MySQL database.

[0568] Data output:

[0569] The server uses the received data to train a machine learning model and generate learning results.

[0570] Specific operation: The server trains a machine learning model using Python's scikit-learn library and generates results, which are then sent to the data provider via email.

[0571] 2. Emotion Recognition by Emotion Engine

[0572] Emotion recognition from text input:

[0573] The user inputs text containing an emotion, for example, "I'm so happy."

[0574] What happens: A user enters text into a web form and clicks the submit button. The server uses Python's NLTK library to analyze the text and recognize the emotion "joy."

[0575] Emotion recognition from speech input:

[0576] The user inputs a voice containing an emotion, for example, a voice containing an emotion of joy.

[0577] Specific operation: The user records voice using a microphone and uploads the audio file. The server converts the voice into text using the Google® Cloud Speech-to-Text API, and then analyzes the text using NLP technology to recognize the emotion of "joy."

[0578] Examples and prompts

[0579] Specific examples

[0580] Managing data input and output:

[0581] Input: The server receives the survey results that the user agreed to via an HTTP request and stores them in a MySQL database in JSON format.

[0582] Output: The server trains a machine learning model using Python's scikit-learn library and sends the results to the data contributor via email.

[0583] Emotion Recognition with Emotion Engine:

[0584] Text input: A user types "I had a lot of fun today" into a web form, and the server uses the NLTK library to recognize the emotion "joy."

[0585] Voice input: The user uses a microphone to record "Today was amazing!", and the server converts the speech to text using the Google Cloud Speech-to-Text API and recognizes the emotion "joy."

[0586] Prompt Sentence Examples

[0587] Managing data input and output:

[0588] Please explain how the server receives data from data providers who have given permission to use it, properly manages the learning results, and provides feedback to the providers.

[0589] Emotion Recognition with Emotion Engine:

[0590] "Explain how a server can use NLP and speech recognition techniques to recognize emotions when a user inputs text or voice."

[0591] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0592] Step 1:

[0593] Entering data

[0594] The server receives data for which permission to use has been granted from the data provider.

[0595] Input: JSON formatted data from data provider.

[0596] Specific operation: The server receives data in JSON format from the data provider via an HTTP request.

[0597] Output: The received data is saved on the server.

[0598] Step 2:

[0599] Data storage

[0600] The server stores the received data in a database.

[0601] Input: Received data in JSON format.

[0602] Specific operation: The server parses the received JSON data and stores it in a MySQL database.

[0603] Output: Data stored in a database.

[0604] Step 3:

[0605] Generating learning results

[0606] The server uses the received data to train a machine learning model and generate learning results.

[0607] Input: Data stored in a database.

[0608] What it does: The server trains a machine learning model using Python's scikit-learn library and generates results.

[0609] Output: The generated training results.

[0610] Step 4:

[0611] Feedback on learning results

[0612] The server feeds back the generated learning results to the data provider.

[0613] Input: The generated training results.

[0614] Specific operation: The server sends the generated learning results to the data provider by email.

[0615] Output: The training results sent to the data provider.

[0616] Step 5:

[0617] Receiving text input

[0618] The user inputs text containing the emotion.

[0619] Input: The text entered by the user.

[0620] What happens: A user enters text into a web form and clicks the submit button.

[0621] Output: The text data sent to the server.

[0622] Step 6:

[0623] Text Analysis

[0624] The server analyzes the received text using natural language processing (NLP) technology to recognize emotions.

[0625] Input: Text data sent to the server.

[0626] Specific operation: The server uses Python's NLTK library to analyze the text and recognize the emotion "joy."

[0627] Output: Recognized emotion data.

[0628] Step 7:

[0629] Receiving audio input

[0630] The user inputs a voice containing emotion.

[0631] Input: The voice data entered by the user.

[0632] What happens: A user records audio using a microphone and uploads the audio file.

[0633] Output: The audio data sent to the server.

[0634] Step 8:

[0635] Audio analysis

[0636] The server analyzes the received voice using voice recognition technology and recognizes emotions.

[0637] Input: The audio data sent to the server.

[0638] Specific operation: The server uses the Google Cloud Speech-to-Text API to convert speech into text, and then analyzes the text using NLP technology to recognize the emotion of "joy."

[0639] Output: Recognized emotion data.

[0640] (Application example 2)

[0641] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0642] In traditional brick-and-mortar stores, it was difficult to grasp customers' emotions in real time, and it was often difficult to respond appropriately. In addition, there was a lack of means to provide feedback based on customers' emotions to store staff, making it difficult to improve customer satisfaction.

[0643] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0644] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, means for managing the input and output of the data, an environment for providing the data, an emotion engine for recognizing a user's emotion, means for recognizing an emotion from a user's voice input using speech recognition technology, means for recognizing an emotion from a user's text input using natural language processing technology, and means for providing feedback on the recognized emotion, thereby making it possible to recognize a customer's emotion in real time and provide appropriate feedback to a store clerk.

[0645] "Person or Legal Entity" means a natural person or a legally recognized entity.

[0646] "Permission to Use" refers to formal authorization from a provider to use specific data.

[0647] "Data" means a collection of information that is collected, stored and processed for a specific purpose.

[0648] An "artificial intelligence system" refers to a computer system that mimics human intelligence through machine learning and data analysis.

[0649] "Means for managing input and output" refers to methods and devices for appropriately controlling the receipt of data and the provision of results.

[0650] "Environment" refers to the physical or virtual conditions or settings required for a system to operate.

[0651] "Emotion engine" refers to a software or hardware component for recognizing and analyzing a user's emotions.

[0652] "Voice recognition technology" refers to the technology that analyzes voice data and converts it into text data.

[0653] "Natural language processing technology" refers to technology that allows computers to understand and process human language.

[0654] "Feedback" refers to the system returning information or results it has obtained to the user or to another system.

[0655] As an embodiment of the present invention, a smart customer service assistant system in a brick-and-mortar store will be described as an example.

[0656] System configuration

[0657] The system consists of the following major components:

[0658] 1. Server: Includes an artificial intelligence system for learning data licensed from individuals or corporations. It analyzes the data using a generative pretrained transformer (Generative AI) and outputs the learning results.

[0659] 2. Smart glasses: A device worn by store clerks that uses voice recognition and natural language processing technologies to recognize users' emotions in real time.

[0660] 3. Emotion engine: Software installed on the server that recognizes emotions from the user's voice and text input.

[0661] 4. Feedback system: A means of providing real-time feedback to store associates about perceived emotions.

[0662] Program processing

[0663] The server runs an artificial intelligence system built using Python. Specifically, it uses the SpeechRecognition library to convert voice data into text and the Transformers library to recognize emotions. The smart glasses, similar to Google Glass®, allow store staff to check customer emotions in real time.

[0664] Data processing and calculation

[0665] 1. Speech recognition: The server converts the voice data acquired from the microphone of the smart glasses into text using the SpeechRecognition library.

[0666] 2. Emotion Recognition: The converted text is fed into an emotion recognition model from the Transformers library to recognize emotions.

[0667] 3. Feedback: The recognized emotion is displayed on the smart glasses display and provided as feedback to the store clerk.

[0668] Specific examples

[0669] For example, if a customer says, "I want to know more about this product," the voice recognition technology converts it into text, and the emotion recognition model recognizes "interest." The salesperson then looks at the feedback displayed on the smart glasses' display and provides the customer with a detailed explanation.

[0670] Prompt Sentence Examples

[0671] Create a program that converts the customer's voice input into text and recognizes emotions from that text. Use the Hugging Face Transformers library as the emotion recognition model. Voice input will be in Japanese.

[0672] In this way, it becomes possible to recognize customer sentiment in real time in physical stores and provide appropriate feedback.

[0673] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0674] Step 1:

[0675] The server receives voice data from the microphone of the smart glasses, where the input is the customer's voice and the output is voice data, which is used in subsequent processing steps.

[0676] Step 2:

[0677] The server converts the acquired voice data into text data using the SpeechRecognition library. The input is voice data and the output is text data. This conversion allows the voice information to be treated as text information.

[0678] Step 3:

[0679] The server inputs the converted text data into an emotion recognition model in the Transformers library to recognize emotions. The input is text data, and the output is emotion labels (e.g., "interested," "happy," "frustrated," etc.). This process identifies the user's emotion from the text.

[0680] Step 4:

[0681] The server then feeds the recognized emotion labels back to the smart glasses display. The input is the emotion labels, and the output is the feedback information displayed on the display. This allows store clerks to check the customer's emotions in real time.

[0682] Step 5:

[0683] The terminal (smart glasses) provides emotional feedback displayed on the display to the store clerk. The input is feedback information, and the output is emotional information that the store clerk visually confirms. This allows the store clerk to take appropriate action.

[0684] Step 6:

[0685] The user (store clerk) responds appropriately to the customer based on the feedback information from the smart glasses. The input is feedback information, and the output is the customer response action. This is expected to improve customer satisfaction.

[0686] Example 3

[0687] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0688] Conventional AI systems do not adequately ensure the safe provision of data from data providers and the appropriate output of learning results. Furthermore, they lack the ability to recognize emotions and adjust learning accordingly, which means they are unable to perform advanced learning based on the user's emotions.

[0689] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0690] In this invention, the server includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, means for managing the input and output of the data, means including an environment for providing the data, an emotion engine for recognizing emotions, means for feeding back the emotions recognized by the emotion engine to the artificial intelligence system, and means for adjusting the learning of the artificial intelligence system based on the feedback. This ensures the safe provision of data from data providers and the appropriate output of learning results, and also enables advanced learning based on emotions.

[0691] "Person or Legal Entity" means the natural person or legal entity recognized as the data provider.

[0692] "Licensed data" means data whose use has been officially authorized by the data provider.

[0693] An "artificial intelligence system" is a computer system that learns data and outputs analytical results.

[0694] "Means for managing data input and output" refers to a system that has the functionality to receive, store, analyze, and output data.

[0695] The "data provision environment" refers to the infrastructure that enables data providers to provide data safely and output learning results appropriately.

[0696] An "emotion engine" is software or hardware that recognizes emotions from data and outputs the results.

[0697] "Feedback means" is a function for transmitting the emotions recognized by the emotion engine to the artificial intelligence system.

[0698] The "means for adjusting learning" is a function for dynamically changing the learning process of an artificial intelligence system based on fed-back emotional information.

[0699] A "generative AI model" is a generative artificial intelligence model that has been trained in advance using large amounts of data.

[0700] "Copyright-free data that can be viewed by anyone on the Internet" refers to data that is freely accessible on the Internet and is free of copyright.

[0701] This invention relates to an AI system that includes an AI system for learning data licensed for use by an individual or a corporation, a means for managing the input and output of the data, and an environment for providing the data. The system further includes an emotion engine for recognizing emotions, a means for feeding back the emotions recognized by the emotion engine to the AI ​​system, and a means for adjusting the learning of the AI ​​system based on the feedback.

[0702] System configuration

[0703] Data provision environment

[0704] Users upload data using a dedicated interface, such as a CSV file containing customer reviews, via a web browser (general browser software). The uploaded data is received by the server and securely stored in a cloud storage service (general cloud storage).

[0705] Emotion Engine

[0706] The server then sends the stored data to the emotion engine, which uses high-speed processing units such as NVIDIA GPUs to analyze the data and recognize emotions. For example, it uses text analysis algorithms to analyze each sentence in a review and recognize emotions such as joy, anger, and sadness.

[0707] Feedback and learning adjustments

[0708] The emotion engine feeds the recognized emotions back to the AI ​​system in real time via an API. The AI ​​system then adjusts its learning based on the fed-back emotion information. For example, it strengthens learning of data related to the emotion of joy.

[0709] Output of training results

[0710] The server outputs the learning results of the AI ​​system to the data provider. The results are provided in a format that the data provider can access (e.g., CSV file). The data provider can download the learning results from the dashboard.

[0711] Specific examples

[0712] For example, if a user uploads a CSV file containing customer reviews, the server receives the CSV file and stores it in cloud storage. The server then sends the stored CSV file to an emotion engine, which analyzes the review text to recognize emotions. The recognized emotions are fed back to the AI ​​system in real time, and the AI ​​system adjusts its learning based on the feedback. Finally, the server outputs the learning results to the data provider, who can download the results from a dashboard.

[0713] Prompt Sentence Examples

[0714] "Design a system that analyzes customer reviews, recognizes emotions, and provides feedback. For example, identify the emotion of happiness in the review text, and use that information to adjust the AI ​​system's learning."

[0715] In this way, a system can be implemented in which the data provider, server, emotion engine, and artificial intelligence system work in cooperation with each other. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0716] Step 1:

[0717] The user uploads data using a dedicated interface. Specifically, they open a web browser, access the upload form, drag and drop a CSV file containing customer reviews, and click the "Upload" button. The input is the CSV file, and the output is data transmission to the server.

[0718] Step 2:

[0719] The server receives the CSV file sent by the user. The received data is checked for file integrity and then securely stored in a cloud storage service. The input is the CSV file sent by the user, and the output is the data stored in the cloud storage.

[0720] Step 3:

[0721] The server sends the saved data to the emotion engine. Specifically, it obtains the path of the saved CSV file and sends the data to the emotion engine's API endpoint via an HTTP POST request. The input is the path of the data saved in cloud storage, and the output is the data sent to the emotion engine.

[0722] Step 4:

[0723] The emotion engine analyzes the received data and recognizes emotions. Specifically, it uses NVIDIA GPUs to perform high-speed text analysis, analyzes each sentence of the review, and calculates an emotion score. The input is the data sent from the server, and the output is the recognized emotion score.

[0724] Step 5:

[0725] The emotion engine feeds back the recognized emotion to the AI ​​system. Specifically, the emotion score is sent in JSON format to the AI ​​system's API endpoint via an HTTP POST request. The input is the recognized emotion score, and the output is feedback to the AI ​​system.

[0726] Step 6:

[0727] The AI ​​system adjusts learning based on feedback from the emotion engine. Specifically, it analyzes the received emotion scores and dynamically modifies the learning algorithm. For example, it sets it to prioritize learning data with a high happiness emotion score. The input is feedback from the emotion engine, and the output is the adjusted learning model.

[0728] Step 7:

[0729] The server outputs the learning results of the AI ​​system to the data provider. Specifically, it obtains the learning results and displays them on a dashboard that the data provider can access. The data provider can download the learning results from the dashboard as a CSV file. The input is the result of the adjusted learning model, and the output is the provision of the learning results to the data provider.

[0730] (Application example 3)

[0731] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0732] In conventional artificial intelligence systems, data exchange between data providers and learning systems is sometimes not smooth, and it is difficult to propose services and products based on customer emotions. The present invention aims to solve these problems, facilitate data exchange between data providers and learning systems, and enable services and product proposals based on customer emotions.

[0733] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0734] In this invention, the server includes an artificial intelligence system for learning data licensed for use by individuals or corporations, means for managing the input and output of the data, means including an environment for providing the data, an emotion engine for recognizing emotions, means for adjusting learning based on the emotions recognized by the emotion engine, and means for proposing services and products based on the customer's emotions. This allows smooth data exchange between the data provider and the learning system, making it possible to propose services and products based on the customer's emotions.

[0735] "Individual or Legal Entity" means a particular natural person or legal entity.

[0736] "Licensed data" refers to data that has been officially licensed for use by the data provider.

[0737] An "artificial intelligence system" refers to a computer system that learns from data and performs analysis and predictions.

[0738] "Means for managing data input and output" refers to functions for controlling the reception of data and the output of results.

[0739] "Data provision environment" refers to the system environment in which data providers can safely provide data and receive learning results.

[0740] An "emotion engine" refers to software or hardware that recognizes and analyzes a user's emotions.

[0741] "Means to adjust learning" refers to the ability to modify or optimize the learning process of the artificial intelligence system based on feedback from the emotion engine.

[0742] "Means for proposing services and products" refers to the function for proposing appropriate services and products based on customer sentiment.

[0743] "Generative AI (Generative Pretrained Transformer)" refers to a generative model that has been pre-trained on a large dataset.

[0744] "Copyright-free data that is publicly viewable on the Internet" refers to data that is freely accessible on the Internet and is free of copyright.

[0745] As an embodiment of the present invention, a smart glasses application for improving customer experience in a brick-and-mortar store will be described as an example.

[0746] System Program

[0747] The system uses a camera mounted on smart glasses to capture customers' faces in real time and includes a program that analyzes their emotions using an emotion recognition model. Based on the emotions recognized by the emotion engine, the AI ​​system adjusts its learning and proposes services and products according to the customer's emotions.

[0748] Hardware and software used

[0749] Hardware: Smart glasses (e.g., Google Glass)

[0750] Software: OpenCV, Keras, emotion recognition model

[0751] Data processing and calculation

[0752] 1. Camera initialization: Initialize the camera on the smart glasses and capture video in real time.

[0753] 2. Face detection: Use OpenCV to detect faces in the captured frames.

[0754] 3. Emotion Recognition: The detected face regions are input into the emotion recognition model to predict the emotion. Run the emotion recognition model using Keras.

[0755] 4. Proposal Generation: Generate service or product proposals for customers based on the recognized emotions.

[0756] Specific examples

[0757] For example, if a customer smiles, the system generates a suggestion such as, "Let's introduce new products to the customer!" Conversely, if the customer looks sad, the system generates a suggestion such as, "Let's suggest products that will help the customer relax."

[0758] Prompt Sentence Examples

[0759] The following prompts might be fed to a generative AI model:

[0760] "If the customer smiles, generate a suggestion to introduce them to a new product."

[0761] By inputting this prompt into a generative AI model, specific suggestions can be generated that correspond to the customer's emotions.

[0762] As described above, this invention provides a system that uses smart glasses to analyze customer emotions in real time and proposes appropriate services and products based on the results, which is expected to improve the customer experience.

[0763] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0764] Step 1:

[0765] The server initializes the camera on the smart glasses and captures video in real time. The input is video data from the camera, and the output is the captured frames. Specifically, the server starts the camera device and continuously acquires frames.

[0766] Step 2:

[0767] The server detects faces from captured frames. The input is the captured frame, and the output is the coordinate information of the detected face area. Specifically, it converts the frame to grayscale using OpenCV and applies a face detection algorithm.

[0768] Step 3:

[0769] The server inputs the detected face region into an emotion recognition model to predict the emotion. The input is the image data of the face region, and the output is the predicted emotion label. Specifically, the server resizes the face region to an appropriate size and inputs it into the emotion recognition model using Keras.

[0770] Step 4:

[0771] The server generates service and product proposals for customers based on the recognized emotions. The input is the predicted emotion label, and the output is the proposal content. Specifically, the server retrieves the proposal content corresponding to the emotion label from the database and inputs a prompt sentence into the generative AI model to generate a specific proposal.

[0772] Step 5:

[0773] The server displays the generated suggestion on the display of the smart glasses. The input is the suggestion, and the output is text to be displayed on the display of the smart glasses. Specifically, the server converts the suggestion into text format and sends it to the display of the smart glasses.

[0774] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0775] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0776] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are listed.

[0777] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0778] [Second embodiment]

[0779] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0780] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0781] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0782] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0783] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0784] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0785] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0786] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0787] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0788] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0789] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0790] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0791] "Example 1"

[0792] One embodiment of the present invention provides an artificial intelligence system for learning data licensed for use by an individual or corporation. This system uses a generative AI (generative pretrained transformer) to learn the data. Specifically, the licensed data is supplied as training data to the generative AI, and the system outputs the learning results.

[0793] "Example 2"

[0794] Furthermore, a means for managing data input and output is provided. This means is for properly managing the input of data for which permission has been granted and the output of learning results. Specifically, data input is provided by data providers for which permission has been granted, and output is data obtained as a result of learning. This data is properly managed and fed back to the providers as necessary.

[0795] "Example 3"

[0796] An environment for providing data is also provided. This environment is designed to safely provide data for which permission has been granted and to appropriately output learning results. Specifically, data providers provide data through this environment, and learning results are output through this environment. This allows for smooth data exchange between data providers and the learning system.

[0797] The processing flow of each embodiment will be described below.

[0798] "Example 1"

[0799] Step 1: Collect data with permission from individuals or entities. This data will serve as input for the AI ​​system to learn from.

[0800] Step 2: The collected data is fed to a generative AI (generative pretrained transformer), which then learns from this data.

[0801] Step 3: The generative AI outputs the learning results, which are used for subsequent processing and analysis.

[0802] "Example 2"

[0803] Step 1: Accept the input of authorized data. This data will be used as input for the AI ​​system to learn from.

[0804] Step 2: Properly manage input data, including storing, organizing, and controlling access to the data.

[0805] Step 3: Manage the output of learning results. The output data is stored appropriately and fed back to the provider as needed.

[0806] "Example 3"

[0807] Step 1: Prepare an environment for providing licensed data. This environment facilitates data exchange between data providers and the learning system.

[0808] Step 2: The data provider provides data through the environment. The provided data becomes input data for the AI ​​system to learn.

[0809] Step 3: The learning results are output through the environment. The output data is stored appropriately and fed back to the provider as needed.

[0810] Example 1

[0811] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0812] Conventional AI systems have had difficulty efficiently learning data licensed for use by individuals or corporations and outputting the results appropriately. Furthermore, the data preprocessing, supply, and learning processes require a lot of manual operations, which takes time and effort. Furthermore, there has been a lack of concrete methods for improving the quality of learning results using generative AI models.

[0813] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0814] In this invention, the server includes an artificial intelligence system for learning data licensed from an individual or a corporation, means for managing the input and output of the data, means for preprocessing the data, means for supplying the data to a generative AI model, means for training the generative AI model, and means for outputting the learning results, thereby enabling efficient learning of data and output of high-quality learning results.

[0815] "Individual or legal entity" refers to the entity that provides permission to use data, and includes individuals, companies, organizations, etc.

[0816] "Licensed Data" means data that has been officially licensed for use by an individual or legal entity, and may include text, images, audio, or other formats.

[0817] An "artificial intelligence system" is a computer system that learns from data and outputs the results, including generative AI models.

[0818] "Means for managing inputs and outputs" refers to the functions and processes for receiving data and providing results.

[0819] "Preprocessing means" refers to functions or processes that process data to convert it into a format suitable for learning.

[0820] A "generative AI model" refers to a generative artificial intelligence model such as the Generative Pretrained Transformer (GPT), which has the ability to learn data and generate new information.

[0821] "Means of supply" refers to the functions and processes for passing preprocessed data to a generative AI model.

[0822] "Means of learning" refers to the functions and processes used to train a generative AI model on data.

[0823] "Means for outputting learning results" refers to the functions and processes for providing the results learned by the generative AI model to the user.

[0824] MODE FOR CARRYING OUT THE INVENTION

[0825] The present invention relates to an artificial intelligence system for learning data licensed for use by an individual or a corporation. Specific embodiments of this system will be described below.

[0826] System Overview

[0827] The server will build an artificial intelligence system to learn from data provided by individuals or corporations. This system will use a generative AI model (Generative Pretrained Transformer, hereafter referred to as GPT) to learn from the data and output the results.

[0828] Hardware and software used

[0829] The server uses the following hardware and software:

[0830] Hardware: Server with high-performance CPU and GPU

[0831] Software: Python scripts for data preprocessing, GPT as a generative AI model, database system for data management

[0832] Data collection and preprocessing

[0833] A user provides the server with data for which they have received permission to use from an individual or a corporation. The data may be in various formats, such as text, images, or audio. The server preprocesses the provided data. Specifically, for text data, the server performs the following processing:

[0834] Remove unnecessary spaces and special characters

[0835] Tokenization (the process of dividing words and sentences)

[0836] Normalization (unifying processes such as converting uppercase to lowercase)

[0837] Supplying data and training the model

[0838] The server supplies the preprocessed data to GPT. The data is batch-processed to ensure efficient learning. The server trains GPT using the supplied data. The learning process involves capturing the features of the data and adjusting the weights to recognize patterns.

[0839] Output of training results

[0840] The server stores the trained GPT model and provides it to users as needed. Users can use this trained model to operate chatbots for customer support.

[0841] Specific examples

[0842] A user wants to create a chatbot for customer support using their company's customer support data. The user provides the server with past chat logs of customer support. The server preprocesses this data and feeds it to GPT for training. The trained model is then used for the user's customer support operations.

[0843] Prompt Sentence Examples

[0844] "Please use the following customer support chat logs as learning data to create a customer support chatbot."

[0845] User: I haven't received my order, what should I do?

[0846] Support: Can I have your order number?

[0847] User: 123456.

[0848] Support: Please wait a moment while we check this.

[0849] Using this prompt, the server trains the GPT and generates a chatbot for customer support.

[0850] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0851] Step 1:

[0852] Users provide the server with data that they have obtained permission to use from individuals or corporations. The input data can be in various formats, such as text, images, and audio. The server receives this data and stores it in a database. Specifically, users upload data, and the server stores it in the appropriate folder.

[0853] Step 2:

[0854] The server preprocesses the provided data. The input data is the raw data saved in step 1. For text data, the server removes unnecessary whitespace and special characters, tokenizes, and normalizes the data. The output is the preprocessed, clean data. Specifically, the server runs a Python script to tokenize and normalize the text data.

[0855] Step 3:

[0856] The server feeds the preprocessed data to the generative AI model (GPT). The input data is the data preprocessed in step 2. The server batches the data and efficiently feeds it to the model. The output is the batch of data fed to the model. Specifically, the server divides the data into chunks of a certain size and passes them to the model.

[0857] Step 4:

[0858] The server trains the generative AI model using the supplied data. The input data is the batch of data supplied in step 3. The server captures the features of the data and adjusts the weights to recognize patterns. The output is a trained generative AI model. Specifically, the server runs the model training process and updates the model parameters every epoch.

[0859] Step 5:

[0860] The server saves the generative AI model after training and provides it to the user as needed. The input data is the model that completed training in step 4. The server saves the model and makes it accessible to the user. The output is a trained model that can be used by the user. Specifically, the server saves the model as a file and provides the user with a download link.

[0861] Step 6:

[0862] A user uses the trained model to operate a customer support chatbot. The input data is an inquiry from the user. The trained model generates an appropriate response to the user's inquiry. The output is a response message to the user. In concrete terms, the user inputs a question to the chatbot and receives the response generated by the model.

[0863] (Application example 1)

[0864] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0865] Conventional advertising systems have had difficulty effectively generating personalized ads based on users' interests. Furthermore, there was a lack of technology to appropriately utilize user data and generate ads using generative AI. This resulted in low advertising effectiveness and an inability to capture users' attention.

[0866] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0867] In this invention, the server includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, means for managing the input and output of the data, means including an environment for providing the data, means for generating personalized advertisements based on the data licensed by the user, and means for displaying the generated advertisements to the user, thereby enabling the generation of effective personalized advertisements based on the user's interests and concerns.

[0868] "Personal or legal entity licensed data" means information or data that is owned and licensed by a person or legal entity.

[0869] An "artificial intelligence system" is a collection of computer programs and algorithms that learn, analyze, and generate data.

[0870] "Means for managing data input and output" are hardware and software components for controlling the receipt, storage, processing of data, and output of results.

[0871] A "data serving environment" is an infrastructure or platform for collecting, storing, accessing, and processing data.

[0872] "Data with user permission" refers to data such as personal information and behavioral history that the user has explicitly given permission to use.

[0873] "Personalized advertising" refers to advertising that is customized based on a particular user's interests.

[0874] "Generative AI (Generative Pretrained Transformer)" is an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data and content.

[0875] "Means for generating advertisements" refers to algorithms or programs for creating advertising content based on collected data.

[0876] "Means for displaying advertisements to users" are the hardware and software components for displaying the generated advertisements on the user's device.

[0877] To implement the present invention, the following system configuration and processing procedures are required.

[0878] First, the server collects data with permission from individuals or corporations and prepares an AI system to learn from it. This AI system uses a generative pretrained transformer (generative AI) to learn from the data and perform analysis and generation.

[0879] The server then has a means for managing the input and output of data. This means is a hardware and software component for controlling the reception, storage, processing, and output of data. Specifically, it uses a database management system or an API.

[0880] Furthermore, the server comprises a means for providing the data, including the environment, which is the infrastructure or platform for collecting, storing, accessing, and processing the data, such as a cloud service or an on-premise server.

[0881] Based on data with the user's permission (e.g., browsing history, purchase history, social media activity, etc.), the server has the means to generate personalized ads. This means is an algorithm or program that creates advertising content based on the collected data. Generative AI models are used to generate ads optimized for the user.

[0882] The generated advertisements are sent from the server to the user's device and displayed through a means for displaying advertisements to the user. The means are hardware and software components for displaying the generated advertisements on the user's device, specifically, a smartphone app or a web browser.

[0883] For example, if a user recently purchased a smartphone or laptop and liked an article about AI on social media, the ad generated might look like this:

[0884] Introducing a smartphone equipped with the latest AI technology! Now available at a special price. Click here for details!

[0885] An example of a prompt sentence to input to the generative AI model is as follows:

[0886] "Generate personalized ads based on your data, including:

[0887] Products and services that may interest you

[0888] Advertising catchphrase

[0889] Ad visual ideas

[0890] User Data:

[0891] Browsing history: tech news, smartphones, AI advancements

[0892] Purchase history: smartphone, laptop

[0893] Social media activity: liked AI article, shared tech news

[0894] In this way, it becomes possible to generate effective personalized advertisements based on the user's interests.

[0895] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0896] Step 1:

[0897] The server collects data that is licensed for use by individuals or corporations.

[0898] Input: Data provided by individuals or entities

[0899] Output: Collected data

[0900] Specific operation: The server obtains data provided by individuals or corporations through APIs or database connections and stores it in a database.

[0901] Step 2:

[0902] The server generates prompt sentences to supply the collected data to the generative AI model.

[0903] Input: Collected data

[0904] Output: Prompt sentence to be fed to the generative AI model

[0905] How it works: The server analyzes the collected data and generates prompts based on the user's interests. For example, it creates prompts containing appropriate advertising content based on the user's browsing history and purchase history.

[0906] Step 3:

[0907] The server generates personalized advertisements using a generative AI model.

[0908] Input: Generated prompt text

[0909] Output: Personalized ads

[0910] How it works: The server provides prompts to a generative AI model (e.g., OpenAI's GPT-3), which then generates ad content based on the user's interests.

[0911] Step 4:

[0912] The server transmits the generated advertisement to the user's terminal.

[0913] Input: Generated personalized ad

[0914] Output: Ad sent to user's device

[0915] How it works: The server sends the generated ad to the user's smartphone app or web browser using push notifications or APIs.

[0916] Step 5:

[0917] The device displays the received advertisement to the user.

[0918] Input: Advertisement sent by the server

[0919] Output: The ad shown to the user

[0920] What happens: The user's device (smartphone or PC) displays the received ad. The ad appears as a banner or pop-up within the app.

[0921] In this way, effective personalized advertisements based on the user's interests can be generated and displayed.

[0922] Example 2

[0923] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0924] Conventional AI systems have insufficient management of data input and output, making it difficult to properly manage data for which permission has been granted and to provide feedback on learning results. Feedback to data providers is often delayed, hindering efficient data use. This makes it difficult to build a relationship of trust between data providers and users, resulting in delays in data provision.

[0925] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0926] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, an environment for providing the data, a means for uploading the data to the server, a means for saving the data in a data management system, a means for inputting the data into a generative AI model for learning, a means for saving the learning results in the data management system, and a means for feeding back the learning results to the provider, thereby enabling appropriate management of data and rapid feedback.

[0927] An "artificial intelligence system" is a computer program that learns from data licensed by an individual or legal entity to perform a specific task.

[0928] "Means for managing data input and output" refers to a system that has the functionality to appropriately manage data input from data providers and the output of learning results.

[0929] The "data providing environment" is a system that includes an interface and a protocol for data providers to upload data to a server.

[0930] "Means for uploading to the server" refers to a system that has a function for users to send data obtained from providers to the server.

[0931] "Data Management System" means a system for storing uploaded data and verifying its integrity and permission to use.

[0932] A "generative AI model" is a model that learns from input data and performs specific tasks.

[0933] A "means for learning" is a system that inputs data into a generative AI model and runs a process to optimize the model's parameters.

[0934] "Means for storing learning results in a data management system" refers to a system that has the functionality to store the learning results obtained from a generative AI model in a data management system.

[0935] "Means for feeding back learning results to providers" refers to a system that has the function of reporting learning results to providers and providing feedback.

[0936] This invention is a system that includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, and an environment for providing the data. Specifically, the server, terminal, and user each play their respective roles, achieving appropriate management of data and rapid feedback.

[0937] Hardware and software used

[0938] Hardware: Servers, terminals

[0939] Software: Data management system, generative AI models

[0940] Data entry management

[0941] The user prepares the data for which they have received permission to use from the provider. Next, the user uses their device to upload the data they received from the provider to the server. The server receives the uploaded data and stores it in a data management system. Furthermore, the server checks the integrity and permission to use of the data to ensure that it is managed appropriately.

[0942] As a specific example of operation, a user uploads image data obtained from a provider to a server via a browser on their device. The server stores the uploaded image data in a database and records meta information about the data (such as the provider and usage permission details).

[0943] Data training

[0944] The server retrieves the data stored in the data management system. Then, the server inputs the retrieved data into the generative AI model to start the learning process. The generative AI model learns based on the input data and optimizes the model parameters. The server saves the generative AI model after learning is complete.

[0945] As a specific example of how it works, the server retrieves image data from a database and inputs it into a generative AI model. The generative AI model learns to classify cats and dogs based on the image data and adjusts the model parameters. The server then saves the model once it has completed learning, preparing it for the next step.

[0946] Output management of learning results

[0947] The server stores the learning results obtained from the generative AI model in a data management system. The device then feeds back the learning results to the provider as needed. The user then uses the device to check the learning results and report them to the provider.

[0948] As a specific example of how it works, the server stores the learning results (e.g., a cat / dog classification model) in a database. The user can then use their device to check the learning results and provide feedback to the provider via email or a dashboard.

[0949] Prompt Sentence Examples

[0950] Examples of prompts to be input to a generative AI model include:

[0951] "Use this image data to generate a model that classifies cats and dogs."

[0952] "Create a model that performs sentiment analysis based on the provided text data."

[0953] The above is an embodiment of the present invention.

[0954] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0955] Step 1:

[0956] Preparing and uploading data

[0957] The user prepares the data for which permission to use has been obtained from the provider. Next, the user uses the terminal to upload the data obtained from the provider to the server. Specifically, the user uploads the image data obtained from the provider to the server through the terminal's browser. The input is the image data obtained from the provider, and the output is the image data stored on the server.

[0958] Step 2:

[0959] Data storage and management

[0960] The server receives the uploaded data and stores it in the data management system. Furthermore, the server checks the data integrity and usage permissions to ensure that the data is managed appropriately. Specifically, the server stores the uploaded image data in a database and records the data's meta information (provider, usage permission details, etc.). The input is the uploaded image data, and the output is the data and its meta information stored in the database.

[0961] Step 3:

[0962] Obtaining data and starting training

[0963] The server retrieves data stored in the data management system. The server then inputs the retrieved data into the generative AI model and starts the learning process. Specifically, the server retrieves image data from the database and inputs it into the generative AI model. The input is the image data retrieved from the database, and the output is the data input into the generative AI model.

[0964] Step 4:

[0965] Training generative AI models

[0966] The generative AI model learns based on the input data and optimizes the model parameters. Specifically, the generative AI model learns to classify cats and dogs based on image data and adjusts the model parameters. The input is the image data entered into the generative AI model, and the output is the model after learning is complete.

[0967] Step 5:

[0968] Saving training results

[0969] The server stores the generative AI model after training. Specifically, the server stores the training results (for example, a cat / dog classification model) in a database. The input is the trained model, and the output is the training results stored in the database.

[0970] Step 6:

[0971] Feedback on learning results

[0972] The device feeds back the learning results to the provider as needed. The user uses the device to check the learning results and report them to the provider. Specifically, the user uses the device to check the learning results and provide feedback to the provider via email or dashboard. The input is the learning results stored in the database, and the output is the information fed back to the provider.

[0973] (Application example 2)

[0974] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0975] Conventional content recommendation systems have difficulty recommending optimal content to users because they are unable to effectively utilize user viewing history and rating data. Furthermore, there are problems with insufficient management of data input and output, and inappropriate feedback to data providers. This has hindered improvements to the user experience.

[0976] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0977] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, an environment for providing the data, a means for a generative AI model to recommend optimal content to users using data provided by the user, and a means for providing feedback on the recommended content to the user. This makes it possible to effectively utilize users' viewing history and rating data to recommend optimal content to users. Furthermore, the appropriate management of data input and output and appropriate feedback to data providers improves the user experience.

[0978] "Person or Legal Entity" means a natural person or a legally recognized entity.

[0979] "Permission to use" refers to obtaining formal permission from the provider to use the data.

[0980] "Data" means a collection of information that is collected, stored and processed for a specific purpose.

[0981] An "artificial intelligence system" refers to a system that uses technologies such as machine learning and deep learning to analyze data and automatically perform specific tasks.

[0982] "Means for managing data input and output" refers to methods and devices for appropriately collecting, storing, processing, and outputting data.

[0983] "Data providing environment" refers to the physical or virtual location or system where data is collected or provided.

[0984] A "generative AI model" refers to a model that uses a generative AI (generative pretrained transformer) to analyze data and generate new information and content.

[0985] "Means for recommending optimal content to users" refers to methods or devices that select and present the most appropriate content for a user based on user data.

[0986] "Feedback means" refers to a method or device for returning results or information generated by the system to the user or data provider.

[0987] The following system configuration will be described as an embodiment of the present invention.

[0988] First, the server is equipped with an artificial intelligence system for learning data licensed by individuals or corporations. This artificial intelligence system analyzes the data using a generative pretrained transformer (generative AI) and generates new information and content. The server also has a means for managing data input and output, and appropriately collects, stores, processes, and outputs data.

[0989] Data provided by users (viewing history, ratings, comments, etc.) is sent to the server. Based on this data, the server uses a generative AI model to recommend the most suitable content to the user. Specifically, the server converts the user's viewing history and rating data into TF-IDF vectors and calculates the cosine similarity to select the most suitable content for the user.

[0990] The server also has a means for providing feedback on the recommended content to the user, allowing the user to easily find content that matches their preferences.

[0991] The hardware used includes servers, users' smartphones, PCs, etc. The software used is Python and its libraries (scikit-learn, json).

[0992] As a concrete example, consider a system that uses data on the movies a user has watched to recommend the next movie to watch. Here is an example prompt:

[0993] User ID: user123

[0994] Viewing history: ["Movie A", "Movie B", "Movie C"]

[0995] Rating: {"Movie A": 5, "Movie B": 4, "Movie C": 3}

[0996] Comments: {"Movie A": "Great movie", "Movie B": "Good", "Movie C": "Okay"}

[0997] By inputting this prompt into a generative AI model, the system can recommend the most suitable movie to the user. This system effectively utilizes the user's viewing history and rating data to recommend the most suitable content to the user. In addition, the system properly manages data input and output, and provides appropriate feedback to data providers, improving the user experience.

[0998] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0999] Step 1:

[1000] The user inputs data such as viewing history, ratings, and comments from the terminal.

[1001] Input: User's viewing history, ratings, comments

[1002] Output: User data (viewing history, ratings, comments)

[1003] Specific operation: Users use their smartphones or PCs to enter their viewing history, ratings, and comments. This data is then sent from the device to the server.

[1004] Step 2:

[1005] The server receives the user data and stores it in JSON format.

[1006] Input: User data (viewing history, ratings, comments)

[1007] Output: User data in JSON format

[1008] Specific operation: The server receives user data sent from the device and stores it in JSON format, which ensures data consistency and readability.

[1009] Step 3:

[1010] The server analyzes the data using a generative AI model based on user data.

[1011] Input: User data in JSON format

[1012] Output: Analysis results (user profile)

[1013] Specific operation: The server uses Python and the scikit-learn library to convert user data into TF-IDF vectors and generate user profiles.

[1014] Step 4:

[1015] The server reads the content database and compares it to the user profile.

[1016] Input: User profile, content database

[1017] Output: Similarity score

[1018] Specific operation: The server reads the content database and compares the user profile with the profile of each content using cosine similarity.

[1019] Step 5:

[1020] The server recommends the most suitable content to the user based on the similarity score.

[1021] Input: Similarity score

[1022] Output: Recommended content list

[1023] Specific operation: The server selects the most suitable content for the user based on the similarity score and generates a recommended content list.

[1024] Step 6:

[1025] The server feeds back the recommended content list to the user.

[1026] Input: Recommended content list

[1027] Output: Feedback to the user (recommended content)

[1028] Specific operation: The server sends the generated recommended content list to the user's device and provides feedback to the user, who can then check the recommended content on the device.

[1029] Example 3

[1030] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1031] Conventional AI systems lack a consistent mechanism for data input from data providers, safe storage of data, learning from the data using generative AI models, saving of learning results, and providing results to users. This leads to problems such as data exchange not being smooth, and the reliability and security of learning results not being ensured.

[1032] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1033] In this invention, the server includes a means for the terminal to input data from a data provider, a means for the server to receive and store the data, a means for the server to input the data to the generative AI model, a means for the generative AI model to learn from the data and generate results, a means for the server to store the learning results, and a means for the user to request and obtain the learning results. This allows smooth data exchange between the data provider and the learning system, enabling the user to quickly obtain the learning results.

[1034] "Individual or Legal Entity" refers to the natural person or legal entity recognized as the data provider.

[1035] "Licensed Data" refers to data that has been officially licensed for use by the data provider.

[1036] "Artificial intelligence system" refers to a computer system for learning from data and generating analytical results.

[1037] "Means for managing data input and output" refers to the functionality for receiving, storing, analyzing data, and outputting the results.

[1038] "Data provision environment" refers to the system environment in which data providers can safely provide data and appropriately output learning results.

[1039] "Terminal" means the computer or device used by the Data Contributor to enter data.

[1040] "Server" refers to a central processing unit for receiving, storing, analyzing data, and outputting results.

[1041] A "generative AI model" refers to an artificial intelligence model that learns from data and generates analytical results.

[1042] "Means for receiving and storing data" refers to the function of receiving data sent from a terminal and storing it in a database, etc.

[1043] "Means for inputting data" refers to the interface or function that allows data providers to input data.

[1044] "Means of learning from data and generating results" refers to the function by which a generative AI model analyzes data and generates learning results.

[1045] "Means for saving learning results" refers to the function of saving the learning results generated by the generative AI model in a database, etc.

[1046] "Means for requesting and obtaining learning results" refers to the interface or functionality that allows users to request and obtain learning results.

[1047] "User" refers to an individual or legal entity that accesses the system to utilize the learning results.

[1048] This invention is an AI system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, and an environment for providing the data. Specifically, the system processes and calculates the data using terminals, a server, and a generative AI model.

[1049] Hardware and software used

[1050] Hardware: Servers, terminals

[1051] Software: Generative AI models (e.g., GPT-4), data management systems (e.g., SQL databases)

[1052] System configuration

[1053] 1. Terminal: A device used by data providers to input data. For example, used by medical staff to input patient symptom data.

[1054] 2. Server: A central processing unit that receives, stores, analyzes, and outputs data. The server uses a data management system to store data and generative AI models to analyze the data.

[1055] 3. Generative AI model: An artificial intelligence model that learns from data and generates analytical results. For example, GPT-4 analyzes patient symptom data and generates a diagnosis.

[1056] System Operation

[1057] Data provider enters data: Data providers enter data using dedicated terminals. For example, medical institution staff enter information such as the patient's name, age, and symptoms.

[1058] The server receives and stores the data: The server receives the data sent from the device and stores it in the data management system. When storing the data, it checks the integrity and security of the data.

[1059] The server inputs data into the generative AI model: The server inputs the stored data into the generative AI model, for example, inputting patient symptom data into the GPT-4 model.

[1060] Generative AI models learn from data and generate results: Generative AI models learn from input data and generate results. For example, GPT-4 analyzes a patient's symptom data and generates a diagnosis.

[1061] The server stores the learning results: The server receives the learning results output from the generative AI model and stores them in a data management system.

[1062] User requests and obtains learning results: A user requests learning results using a dedicated device. For example, a doctor obtains diagnosis results.

[1063] Examples of concrete examples and prompts

[1064] Example: A medical institution provides patient data and uses a generative AI model (GPT-4) to generate a diagnosis. A doctor obtains the diagnosis.

[1065] Example prompt sentence:

[1066] Input prompt: "Enter patient symptom data and generate a diagnosis."

[1067] Output prompt: "The diagnostic results are as follows:..."

[1068] This system allows smooth data exchange between the data provider and the learning system, enabling users to quickly obtain learning results. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1069] Step 1:

[1070] The terminal inputs data from the data provider.

[1071] Data providers input data using a dedicated terminal. For example, a medical institution staff member inputs information such as the patient's name, age, and symptoms. The input data is structured based on the input form on the terminal. Specifically, the data provider inputs the required information into the input form on the terminal and clicks the "Submit" button.

[1072] Step 2:

[1073] The server receives and stores the data.

[1074] Data sent from the device is sent to the server using a secure communication protocol (e.g., HTTPS). The server stores the received data in a data management system (e.g., an SQL database). When storing the data, the server checks the integrity and security of the data. Specifically, the server receives the data and executes an SQL query to store it in the database.

[1075] Step 3:

[1076] The server inputs data into the generative AI model.

[1077] The server inputs the stored data into the generative AI model. For example, it inputs patient symptom data into the GPT-4 model. The input data is sent to the API of the generative AI model. Specifically, the server retrieves the necessary data from the database and creates a request to send the data to the API of the generative AI model.

[1078] Step 4:

[1079] Generative AI models learn from data and generate results.

[1080] A generative AI model learns from input data and generates results. For example, GPT-4 analyzes a patient's symptom data and generates a diagnosis. Specifically, the generative AI model analyzes the data and outputs the diagnosis and recommended treatment in text format.

[1081] Step 5:

[1082] The server stores the learning results.

[1083] The server receives the learning results output from the generative AI model and stores them in a data management system. When saving, it checks the integrity and security of the results. Specifically, the server receives the learning results and executes SQL queries to save them in a database.

[1084] Step 6:

[1085] The user requests and obtains the learning results.

[1086] A user requests learning results using a dedicated terminal. For example, a doctor obtains diagnosis results. The user enters a patient's ID into the search form on the terminal and clicks the "Search" button. The server receives the request and provides the learning results to the user. Specifically, the server retrieves the learning results from the database and generates a response to display the results on the user's terminal.

[1087] (Application example 3)

[1088] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1089] Conventional artificial intelligence systems have faced the problem of lacking an environment in which data providers can safely provide data and appropriately output learning results. Furthermore, data security is often not ensured during data provision and learning result output, creating the risk of data leaks and unauthorized use. This has led to a demand for an environment in which data providers can provide data with peace of mind.

[1090] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1091] In this invention, the server includes an artificial intelligence system for learning data licensed from an individual or a corporation, a means for managing the input and output of data, a means including an environment for providing data, a means for encrypting and providing data, and a means for decrypting the encrypted data and outputting the learning results. This enables data providers to provide data safely and output the learning results appropriately.

[1092] "Person or legal entity" means a natural person or a legal entity.

[1093] "Permission to use" refers to the permission given by the data provider to use the data.

[1094] "Data" refers to a collection of information, the material that artificial intelligence systems use to learn.

[1095] "Artificial intelligence system" refers to a computer program that learns from data and performs specific tasks.

[1096] "Means for managing input and output" refers to the functions for controlling the receipt of data and the delivery of results.

[1097] "Data provision environment" refers to the systems and infrastructure that enable data providers to provide data securely.

[1098] "Encryption" refers to the conversion of data using a specific algorithm to make it in a form that cannot be easily understood by third parties.

[1099] "Decryption" refers to restoring encrypted data to its original form.

[1100] "Learning results" refers to the output that an AI system generates after learning from data.

[1101] "Generative AI (Generative Pretrained Transformer)" refers to an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data.

[1102] As an embodiment of the present invention, the following system can be constructed.

[1103] System Overview

[1104] This system includes an artificial intelligence system for learning data that has been granted permission to be used by an individual or a corporation, a means for managing the input and output of data, an environment for providing the data, a means for encrypting and providing the data, and a means for decrypting the encrypted data and outputting the learning results.

[1105] Hardware and software used

[1106] Hardware: Smartphones, servers

[1107] Software: Python, requests library, cryptography library, Generative Pretrained Transformer

[1108] Program processing explanation

[1109] Data encryption and delivery

[1110] When a user provides data using their smartphone, the data is first encrypted using the Fernet class from the cryptography library. The encrypted data is then sent to the server using an HTTP POST request using the requests library.

[1111] Obtaining and decoding the training results

[1112] The server decrypts the encrypted data it receives and uses a generative pretrained transformer (Generative AI) to learn from the data. The learning results are then re-encrypted and decrypted when the user retrieves them using their smartphone.

[1113] Specific examples

[1114] For example, if a user is using a smartphone to provide "user behavior data," the following prompt sentence can be used:

[1115] Prompt Sentence Examples

[1116] "You provide user behavior data, and the AI ​​model learns from that data and outputs results that predict user behavior patterns."

[1117] In this way, a system can be realized that provides an environment in which data providers can safely provide data and appropriately output learning results.

[1118] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1119] Step 1:

[1120] A user enters data using a smartphone.

[1121] Input: User-provided data (e.g., user behavior data)

[1122] Specific operation: The user launches the smartphone application and enters the data to be provided.

[1123] Step 2:

[1124] The device encrypts the data.

[1125] Input: User-supplied data

[1126] Data processing: Encrypt the data using the Fernet class from the cryptography library.

[1127] Output: Encrypted data

[1128] Specific operation: The terminal encrypts the data entered by the user using the Fernet class to generate encrypted data.

[1129] Step 3:

[1130] The device sends the encrypted data to the server.

[1131] Input: Encrypted data

[1132] Data operations: Send data using an HTTP POST request.

[1133] Output: Data sent to the server

[1134] Specific operation: The device uses the requests library to send the encrypted data to the server via an HTTP POST request.

[1135] Step 4:

[1136] The server receives the encrypted data and decrypts it.

[1137] Input: Encrypted data

[1138] Data processing: Decrypt the data using the Fernet class from the cryptography library.

[1139] Output: Decrypted data

[1140] Specific operation: The server decrypts the received encrypted data using the Fernet class to obtain the original data.

[1141] Step 5:

[1142] The server inputs the decrypted data into a generative AI model for learning.

[1143] Input: Decrypted data

[1144] Data computation: Learn data using a generative AI (generative pretrained transformer).

[1145] Output: Learning results

[1146] Specific operation: The server inputs the decoded data into the generative AI model, which learns the data and generates the learning results.

[1147] Step 6:

[1148] The server encrypts the learning results.

[1149] Input: Learning results

[1150] Data processing: Encrypt the training results using the Fernet class from the cryptography library.

[1151] Output: Encrypted training results

[1152] Specific operation: The server encrypts the generated learning result using the Fernet class and generates the encrypted learning result.

[1153] Step 7:

[1154] The server sends the encrypted learning results to the terminal.

[1155] Input: Encrypted learning results

[1156] Data operation: Sends data using an HTTP GET request.

[1157] Output: Data sent to the terminal

[1158] Specific operation: The server uses the requests library to send the encrypted learning results to the terminal via an HTTP GET request.

[1159] Step 8:

[1160] The terminal receives the encrypted learning result and decrypts it.

[1161] Input: Encrypted learning results

[1162] Data processing: Decrypt the training results using the Fernet class from the cryptography library.

[1163] Output: Decoded training results

[1164] Specific operation: The terminal decrypts the received encrypted learning result using the Fernet class and obtains the original learning result.

[1165] Step 9:

[1166] The user checks the decoded learning results.

[1167] Input: Decoded training results

[1168] Output: Learning results for user confirmation

[1169] Specific operation: The user checks the decoded learning results through a smartphone application.

[1170] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1171] "Example 1"

[1172] As one embodiment of the present invention, an artificial intelligence system including an emotion engine is provided. This system learns from data licensed for use by individuals or corporations and manages the input and output of that data. The emotion engine also recognizes a user's emotion and adjusts learning based on that emotion. Specifically, when a user inputs text or voice, the emotion engine recognizes the user's emotion from that input. The AI ​​system then adjusts learning based on the recognized emotion. For example, if a user inputs text or voice that indicates the emotion of joy, the emotion engine recognizes the emotion of joy, and the AI ​​system learns based on that emotion.

[1173] "Example 2"

[1174] Another embodiment of the present invention provides a specific technique for an emotion engine to recognize a user's emotion. Specifically, the emotion engine uses natural language processing (NLP) technology to recognize emotions from a user's text input. Also, it uses speech recognition technology to recognize emotions from a user's voice input. For example, if a user inputs the text "I'm very happy," the emotion engine uses NLP technology to recognize the emotion of joy from the text. Also, if a user inputs a voice containing the emotion of joy, the emotion engine uses speech recognition technology to recognize the emotion of joy from the voice.

[1175] "Example 3"

[1176] In yet another embodiment of the present invention, a specific technique is provided for adjusting the learning of an AI system based on emotions recognized by an emotion engine. Specifically, the emotion engine feeds back the recognized emotions to the AI ​​system, and the AI ​​system adjusts its learning based on the feedback. For example, if the emotion engine recognizes the emotion of joy, it feeds back that information to the AI ​​system. Then, the AI ​​system strengthens its learning of data related to the emotion of joy based on the feedback.

[1177] The processing flow of each embodiment will be described below.

[1178] "Example 1"

[1179] Step 1: The user enters text or voice input.

[1180] Step 2: The emotion engine recognizes the user's emotion from the input.

[1181] Step 3: Adjust the AI ​​system's learning based on the recognized emotions.

[1182] "Example 2"

[1183] Step 1: The user enters text.

[1184] Step 2: The emotion engine uses natural language processing (NLP) techniques to recognize emotions from the text.

[1185] Step 3: The user provides voice input.

[1186] Step 4: The emotion engine uses speech recognition technology to recognize emotions from the speech.

[1187] "Example 3"

[1188] Step 1: The emotion engine recognizes the user's emotion.

[1189] Step 2: The recognized emotions are fed back to the AI ​​system.

[1190] Step 3: The AI ​​system adjusts its learning based on that feedback.

[1191] Example 1

[1192] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1193] Conventional AI systems have difficulty effectively learning from data that has been authorized for use by individuals or corporations, and have not adjusted their learning to take user emotions into account. As a result, they have been unable to obtain appropriate learning results based on user emotions, limiting the flexibility and accuracy of the system.

[1194] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1195] In this invention, the server includes an AI system for learning data licensed from an individual or a corporation, means for managing the input and output of the data, means for preprocessing the data, means for learning the data using a generative AI model, and means for recognizing a user's emotions and adjusting learning based on the emotions, thereby enabling flexible and highly accurate learning results that take the user's emotions into consideration.

[1196] "Individual or Legal Entity" refers to a specific individual or legal entity that provides permission to use data.

[1197] "Licensed data" refers to data that has been explicitly licensed by an individual or legal entity and is used to train an artificial intelligence system.

[1198] "Artificial intelligence system" refers to a computer program and associated hardware for learning from data and performing specific tasks.

[1199] "Means for managing input and output" refers to software and hardware for controlling the input and output of data and for appropriate processing.

[1200] "Pre-processing means" refers to software and hardware for pre-processing data, such as cleaning, formatting, and tokenization.

[1201] "Generative artificial intelligence model" refers to a generative AI model such as the Generative Pretrained Transformer, which has the ability to generate new information based on data.

[1202] "Means for recognizing emotions and adjusting learning" refers to software and hardware for analyzing a user's emotions and dynamically adjusting the learning process of an artificial intelligence system based on those emotions.

[1203] "Data serving environment" refers to the infrastructure for collecting, storing, processing, and serving data.

[1204] The present invention relates to an AI system for learning from data licensed by individuals or corporations. The system utilizes a generative AI model and has the ability to adjust learning based on user emotions.

[1205] System configuration

[1206] Data collection and permission acquisition

[1207] The user provides the system with data for which they have received permission to use from an individual or corporation. The terminal then sends the data provided by the user to the server. The server then checks the permission to use the received data and stores it in a database. Specifically, a database management system (e.g., MySQL or PostgreSQL) is used.

[1208] Data Preprocessing

[1209] The server preprocesses the collected data. Specifically, it cleans the data (filling in missing values ​​and removing outliers) and converts the format (tokenizing text data, etc.). The server cleans the data using Python's pandas library and converts the numeric data using NumPy.

[1210] Learning generative AI models

[1211] The server supplies the preprocessed data to a generative AI model (Generative Pretrained Transformer). The server uses an NVIDIA GPU (e.g., Tesla V100) to train the generative AI model using TensorFlow or PyTorch. The server monitors the model's performance during training and adjusts hyperparameters as needed.

[1212] Use of emotion engine

[1213] When a user inputs text or voice, the device sends the input to a server. The server uses an emotion engine to recognize the user's emotions. Specifically, it uses the NLP library Hugging Face Transformers to extract emotions from text data. The server then adjusts the learning of the generative AI based on the recognized emotions. For example, if a user inputs text that expresses joy, it prioritizes positive data in the learning process.

[1214] Output of training results

[1215] The server outputs the learning results of the generative AI and provides them to the user via a terminal. The server generates the learning results in text and graph format and displays them in a way that is easy for the user to understand. Specifically, graphs are created using Matplotlib and Seaborn.

[1216] Specific examples

[1217] For example, if the user enters the text "I'm so happy today!", the specific behavior is as follows:

[1218] 1. The user types, "I'm so happy today!"

[1219] 2. The device sends this text to the server.

[1220] 3. The server uses an emotion engine to recognize the emotion "joy" from this text.

[1221] 4. Based on this emotion of "joy," the server adjusts the training dataset for the generative AI model, giving priority to positive data.

[1222] 5. The server generates the learning results and provides the user with "text generation results that reflect positive emotions" via the terminal.

[1223] Prompt Sentence Examples

[1224] Below are some example prompts to input to a generative AI model:

[1225] If a user types "I'm so happy today!", how would the emotion engine respond and adjust the generative AI's learning?

[1226] In this way, the system effectively learns from the data using generative AI while taking into account the user's emotions.

[1227] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1228] Step 1:

[1229] Data collection and permission acquisition

[1230] A user provides the system with data for which they have received permission to use from an individual or corporation. The terminal sends the data provided by the user to the server. The server checks the permission to use the received data and stores it in a database. Specifically, a database management system (e.g., MySQL or PostgreSQL) is used. The input is the data provided by the user, and the output is the data stored in the database.

[1231] Step 2:

[1232] Data Preprocessing

[1233] The server preprocesses the collected data. Specifically, it cleans the data (filling in missing values ​​and removing outliers) and converts the format (tokenizing text data, etc.). The server cleans the data using Python's pandas library and converts numerical data using NumPy. The input is raw data obtained from the database, and the output is preprocessed data.

[1234] Step 3:

[1235] Learning generative AI models

[1236] The server supplies the preprocessed data to a generative AI model (Generative Pretrained Transformer). The server uses an NVIDIA GPU (e.g., Tesla V100) to train the generative AI model using TensorFlow or PyTorch. The server monitors the model's performance during training and adjusts hyperparameters as necessary. The input is the preprocessed data, and the output is a trained generative AI model.

[1237] Step 4:

[1238] Use of emotion engine

[1239] When a user inputs text or voice, the device sends the input to the server. The server uses an emotion engine to recognize the user's emotions. Specifically, it uses the NLP library Hugging Face Transformers to extract emotions from text data. The server then adjusts the learning of the generative AI based on the recognized emotions. For example, if a user inputs text that expresses the emotion of joy, it prioritizes learning positive data. The input is text or voice data from the user, and the output is the emotion recognition results and adjusted learning data.

[1240] Step 5:

[1241] Output of training results

[1242] The server outputs the learning results of the generative AI and provides them to the user via a terminal. The server generates the learning results in text or graph format and displays them in a way that is easy for the user to understand. Specifically, graphs are created using Matplotlib or Seaborn. The input is the trained generative AI model, and the output is the text generation results and graphs provided to the user.

[1243] (Application example 1)

[1244] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1245] Conventional AI systems lack the ability to recognize user emotions and adjust learning based on those emotions, making it difficult to recommend products that meet individual user needs and emotions. Furthermore, they lack the ability to integrate real-time emotion analysis and product recommendations to improve the shopping experience in brick-and-mortar stores.

[1246] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1247] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, means for managing the input and output of the data, means including an environment for providing the data, means including an emotion engine that recognizes a user's emotions and adjusts learning based on the emotions, means for recommending products based on the user's emotions and past purchase history, and means for analyzing the user's emotions when viewing or touching products in real time and providing appropriate product information and coupons. This enables personalized product recommendations based on the user's emotions and an improved real-time shopping experience in physical stores.

[1248] "Data licensed for use by an individual or corporation" means data owned by a specific individual or corporation, and for which formal permission has been given to use the data.

[1249] An "artificial intelligence system" is a collection of computer programs or hardware designed to learn from data and perform specific tasks automatically.

[1250] "Means for managing data input and output" refers to functions or devices for controlling the receipt, processing, and output of results of data.

[1251] A "data serving environment" is the physical or virtual infrastructure in which data is properly collected, stored, and accessed.

[1252] An "emotion engine" is a software or hardware component that recognizes a user's emotions and adjusts the system's behavior based on those emotions.

[1253] "Means for recommending products based on a user's emotions and past purchase history" refers to an algorithm or system that analyzes a user's emotional state and past purchase data and suggests appropriate products based on that.

[1254] "Means for analyzing in real time and providing appropriate product information and coupons" refers to a function or device that instantly evaluates a user's current behavior and emotions and instantly provides appropriate product information and discount coupons.

[1255] "Generative AI (Generative Pretrained Transformer)" is an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data.

[1256] As an embodiment of the present invention, a "smart shopping assistant" system that improves the shopping experience in brick-and-mortar stores will be described as an example.

[1257] System Configuration

[1258] The system consists of the following main components:

[1259] 1. Server:

[1260] This includes artificial intelligence systems for learning from data licensed for use by individuals or entities.

[1261] It uses generative AI (Generative Pretrained Transformer) to learn from data and recommend products based on the user's emotions and past purchase history.

[1262] It includes an emotion engine that recognizes the user's emotions and adjusts learning based on those emotions.

[1263] 2. Device (smartphone):

[1264] It uses a camera and microphone to capture the user's facial expressions and voice and recognize emotions.

[1265] It analyzes users' emotions in real time when they look at or touch a product, and provides appropriate product information and coupons.

[1266] Program processing explanation

[1267] The server processes data using the following hardware and software:

[1268] Hardware: Smartphone camera and microphone

[1269] Software: OpenCV (image processing), Transformers (generative AI model), EmotionRecognizer (emotion recognition)

[1270] Data processing and calculation

[1271] 1. Emotion recognition:

[1272] The device's camera and microphone are used to capture the user's facial expressions and voice, and emotions are recognized using EmotionRecognizer.

[1273] 2. Product recommendation:

[1274] Based on the recognized emotions and the user's past purchase history data, a generative AI model (GPT-3) is used to recommend products.

[1275] 3. Real-time feedback:

[1276] It analyzes users' emotions in real time when they look at or touch a product, and provides appropriate product information and coupons.

[1277] Specific examples

[1278] For example, when a user uses a smartphone in a physical store, the camera captures the user's facial expressions and performs emotion recognition. If the user is smiling, the system recognizes the emotion of "joy" and recommends "new fashion items" based on past product purchase data.

[1279] Prompt Sentence Examples

[1280] The user's emotion is joy. Their past purchase history is "fashion items, accessories." Please recommend appropriate products.

[1281] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1282] Step 1:

[1283] It uses the device's camera and microphone to capture the user's facial expressions and voice.

[1284] Input: User's facial expression images and voice data

[1285] Output: Captured facial expression images and audio data

[1286] How it works: The device's camera captures a picture of the user's face and the microphone records the user's voice. These data are collected in real time.

[1287] Step 2:

[1288] The device inputs the captured facial expression images and voice data into EmotionRecognizer to recognize the user's emotions.

[1289] Input: Captured facial expression images and audio data

[1290] Output: Recognized user emotion (e.g., happy, sad, surprised, etc.)

[1291] How it works: EmotionRecognizer analyzes facial images and voice data to identify the user's emotions. For example, it recognizes the emotion "joy" from an image of a smiling face.

[1292] Step 3:

[1293] The server uses a generative AI model (GPT-3) to recommend products based on the recognized emotions and the user's past purchase history data.

[1294] Input: Recognized user sentiment, user's past purchase history data

[1295] Output: Recommended product information

[1296] Specific operation: The server inputs a prompt into the generative AI model and recommends appropriate products. For example, the prompt might be, "The user's emotion is joy. Their past purchase history is 'fashion items, accessories.' Please recommend appropriate products."

[1297] Step 4:

[1298] The server transmits the recommended product information to the terminal, which then displays the information to the user.

[1299] Input: Recommended product information

[1300] Output: Product information displayed to the user

[1301] Specific operation: The server sends the recommended product information to the terminal, which then displays the information on the user's screen. The user can then check the displayed product information.

[1302] Step 5:

[1303] The device recaptures and analyzes the user's emotions in real time when they look at or touch the product.

[1304] Input: User's facial expression images and voice data (recaptured)

[1305] Output: Re-recognized user emotion

[1306] Specific operation: The device's camera and microphone capture the user's facial expressions and voice again, and EmotionRecognizer recognizes emotions again.

[1307] Step 6:

[1308] The server generates appropriate product information and coupons based on the re-recognized emotion and transmits them to the terminal.

[1309] Input: Re-recognized user emotion

[1310] Output: Generated product information and coupons

[1311] Specific operation: Based on the re-recognized emotions, the server uses a generative AI model to generate new product information and coupons, which are then sent to the device, which then displays them to the user.

[1312] Example 2

[1313] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1314] Conventional AI systems have had problems with insufficient management of data input and output, resulting in low accuracy, especially in emotion recognition. Furthermore, there is a lack of means to accurately recognize user emotions, making it difficult to recognize emotions from text or voice.

[1315] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes an artificial intelligence system for learning data whose use is permitted by an individual or a corporation, means for managing the input of the data, means for managing the output of the data, means including an emotion engine for recognizing a user's emotion, means for recognizing emotion from text input using natural language processing technology, and means for recognizing emotion from voice input using voice recognition technology. This allows appropriate management of data input and output, making it possible to accurately recognize a user's emotion.

[1316] "Person or legal entity" means a natural person or a legal entity.

[1317] "Licensed Data" means data used with explicit permission from the data provider.

[1318] An "artificial intelligence system" refers to a system that analyzes data and learns using technologies such as machine learning and deep learning.

[1319] "Means for managing data input" refers to the means for properly storing and processing data received from data providers.

[1320] "Means for managing data output" refers to the means for properly storing learning results and analysis results and providing feedback to providers as necessary.

[1321] "Emotion engine" refers to a software or hardware component for recognizing a user's emotions.

[1322] "Natural language processing technology" refers to technology for analyzing text data and understanding meaning and emotions.

[1323] "Speech recognition technology" refers to technology for analyzing voice data and converting it into text.

[1324] A "generative artificial intelligence model" refers to an artificial intelligence technology that generates new data using a model that has been trained in advance using large amounts of data.

[1325] "Copyright-free data that can be viewed by anyone on the Internet" refers to data that is publicly available on the Internet and can be used freely because it does not have copyright.

[1326] The present invention is a system that includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, a means for managing the input and output of the data, and an emotion engine for recognizing a user's emotions.

[1327] System configuration

[1328] 1. Managing data input and output

[1329] Enter the data:

[1330] The server receives data for which permission to use has been granted from the data provider, such as a user's behavior log or survey results.

[1331] Specific operation: The server receives data in JSON format from the data provider via an HTTP request, and stores the received data in a MySQL database.

[1332] Data output:

[1333] The server uses the received data to train a machine learning model and generate learning results.

[1334] Specific operation: The server trains a machine learning model using Python's scikit-learn library and generates results, which are then sent to the data provider via email.

[1335] 2. Emotion Recognition by Emotion Engine

[1336] Emotion recognition from text input:

[1337] The user inputs text containing an emotion, for example, "I'm so happy."

[1338] What happens: A user enters text into a web form and clicks the submit button. The server uses Python's NLTK library to analyze the text and recognize the emotion "joy."

[1339] Emotion recognition from speech input:

[1340] The user inputs a voice containing an emotion, for example, a voice containing an emotion of joy.

[1341] How it works: Users record their voice using a microphone and upload the audio file. The server converts the voice into text using the Google Cloud Speech-to-Text API, and then analyzes the text using NLP technology to recognize the emotion of "joy."

[1342] Examples and prompts

[1343] Specific examples

[1344] Managing data input and output:

[1345] Input: The server receives the survey results that the user agreed to via an HTTP request and stores them in a MySQL database in JSON format.

[1346] Output: The server trains a machine learning model using Python's scikit-learn library and sends the results to the data contributor via email.

[1347] Emotion Recognition with Emotion Engine:

[1348] Text input: A user types "I had a lot of fun today" into a web form, and the server uses the NLTK library to recognize the emotion "joy."

[1349] Voice input: The user uses a microphone to record "Today was amazing!", and the server converts the speech to text using the Google Cloud Speech-to-Text API and recognizes the emotion "joy."

[1350] Prompt Sentence Examples

[1351] Managing data input and output:

[1352] Please explain how the server receives data from data providers who have given permission to use it, properly manages the learning results, and provides feedback to the providers.

[1353] Emotion Recognition with Emotion Engine:

[1354] "Explain how a server can use NLP and speech recognition techniques to recognize emotions when a user inputs text or voice."

[1355] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1356] Step 1:

[1357] Entering data

[1358] The server receives data for which permission to use has been granted from the data provider.

[1359] Input: JSON formatted data from data provider.

[1360] Specific operation: The server receives data in JSON format from the data provider via an HTTP request.

[1361] Output: The received data is saved on the server.

[1362] Step 2:

[1363] Data storage

[1364] The server stores the received data in a database.

[1365] Input: Received data in JSON format.

[1366] Specific operation: The server parses the received JSON data and stores it in a MySQL database.

[1367] Output: Data stored in a database.

[1368] Step 3:

[1369] Generating learning results

[1370] The server uses the received data to train a machine learning model and generate learning results.

[1371] Input: Data stored in a database.

[1372] What it does: The server trains a machine learning model using Python's scikit-learn library and generates results.

[1373] Output: The generated training results.

[1374] Step 4:

[1375] Feedback on learning results

[1376] The server feeds back the generated learning results to the data provider.

[1377] Input: The generated training results.

[1378] Specific operation: The server sends the generated learning results to the data provider by email.

[1379] Output: The training results sent to the data provider.

[1380] Step 5:

[1381] Receiving text input

[1382] The user inputs text containing the emotion.

[1383] Input: The text entered by the user.

[1384] What happens: A user enters text into a web form and clicks the submit button.

[1385] Output: The text data sent to the server.

[1386] Step 6:

[1387] Text Analysis

[1388] The server analyzes the received text using natural language processing (NLP) technology to recognize emotions.

[1389] Input: Text data sent to the server.

[1390] Specific operation: The server uses Python's NLTK library to analyze the text and recognize the emotion "joy."

[1391] Output: Recognized emotion data.

[1392] Step 7:

[1393] Receiving audio input

[1394] The user inputs a voice containing emotion.

[1395] Input: The voice data entered by the user.

[1396] What happens: A user records audio using a microphone and uploads the audio file.

[1397] Output: The audio data sent to the server.

[1398] Step 8:

[1399] Audio analysis

[1400] The server analyzes the received voice using voice recognition technology and recognizes emotions.

[1401] Input: The audio data sent to the server.

[1402] Specific operation: The server uses the Google Cloud Speech-to-Text API to convert speech into text, and then analyzes the text using NLP technology to recognize the emotion of "joy."

[1403] Output: Recognized emotion data.

[1404] (Application example 2)

[1405] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1406] In traditional brick-and-mortar stores, it was difficult to grasp customers' emotions in real time, and it was often difficult to respond appropriately. In addition, there was a lack of means to provide feedback based on customers' emotions to store staff, making it difficult to improve customer satisfaction.

[1407] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1408] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, means for managing the input and output of the data, an environment for providing the data, an emotion engine for recognizing a user's emotion, means for recognizing an emotion from a user's voice input using speech recognition technology, means for recognizing an emotion from a user's text input using natural language processing technology, and means for providing feedback on the recognized emotion, thereby making it possible to recognize a customer's emotion in real time and provide appropriate feedback to a store clerk.

[1409] "Person or Legal Entity" means a natural person or a legally recognized entity.

[1410] "Permission to Use" refers to formal authorization from a provider to use specific data.

[1411] "Data" means a collection of information that is collected, stored and processed for a specific purpose.

[1412] An "artificial intelligence system" refers to a computer system that mimics human intelligence through machine learning and data analysis.

[1413] "Means for managing input and output" refers to methods and devices for appropriately controlling the receipt of data and the provision of results.

[1414] "Environment" refers to the physical or virtual conditions or settings required for a system to operate.

[1415] "Emotion engine" refers to a software or hardware component for recognizing and analyzing a user's emotions.

[1416] "Voice recognition technology" refers to the technology that analyzes voice data and converts it into text data.

[1417] "Natural language processing technology" refers to technology that allows computers to understand and process human language.

[1418] "Feedback" refers to the system returning information or results it has obtained to the user or to another system.

[1419] As an embodiment of the present invention, a smart customer service assistant system in a brick-and-mortar store will be described as an example.

[1420] System configuration

[1421] The system consists of the following major components:

[1422] 1. Server: Includes an artificial intelligence system for learning data licensed from individuals or corporations. It analyzes the data using a generative pretrained transformer (Generative AI) and outputs the learning results.

[1423] 2. Smart glasses: A device worn by store clerks that uses voice recognition and natural language processing technologies to recognize users' emotions in real time.

[1424] 3. Emotion engine: Software installed on the server that recognizes emotions from the user's voice and text input.

[1425] 4. Feedback system: A means of providing real-time feedback to store associates about perceived emotions.

[1426] Program processing

[1427] The server runs an artificial intelligence system built using Python, specifically the SpeechRecognition library to convert voice data to text and the Transformers library to recognize emotions. The smart glasses, a Google Glass-like device, allow store staff to see customers' emotions in real time.

[1428] Data processing and calculation

[1429] 1. Speech recognition: The server converts the voice data acquired from the microphone of the smart glasses into text using the SpeechRecognition library.

[1430] 2. Emotion Recognition: The converted text is fed into an emotion recognition model from the Transformers library to recognize emotions.

[1431] 3. Feedback: The recognized emotion is displayed on the smart glasses display and provided as feedback to the store clerk.

[1432] Specific examples

[1433] For example, if a customer says, "I want to know more about this product," the voice recognition technology converts it into text, and the emotion recognition model recognizes "interest." The salesperson then looks at the feedback displayed on the smart glasses' display and provides the customer with a detailed explanation.

[1434] Prompt Sentence Examples

[1435] Create a program that converts the customer's voice input into text and recognizes emotions from that text. Use the Hugging Face Transformers library as the emotion recognition model. Voice input will be in Japanese.

[1436] In this way, it becomes possible to recognize customer sentiment in real time in physical stores and provide appropriate feedback.

[1437] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1438] Step 1:

[1439] The server receives voice data from the microphone of the smart glasses, where the input is the customer's voice and the output is voice data, which is used in subsequent processing steps.

[1440] Step 2:

[1441] The server converts the acquired voice data into text data using the SpeechRecognition library. The input is voice data and the output is text data. This conversion allows the voice information to be treated as text information.

[1442] Step 3:

[1443] The server inputs the converted text data into an emotion recognition model in the Transformers library to recognize emotions. The input is text data, and the output is emotion labels (e.g., "interested," "happy," "frustrated," etc.). This process identifies the user's emotion from the text.

[1444] Step 4:

[1445] The server then feeds the recognized emotion labels back to the smart glasses display. The input is the emotion labels, and the output is the feedback information displayed on the display. This allows store clerks to check the customer's emotions in real time.

[1446] Step 5:

[1447] The terminal (smart glasses) provides emotional feedback displayed on the display to the store clerk. The input is feedback information, and the output is emotional information that the store clerk visually confirms. This allows the store clerk to take appropriate action.

[1448] Step 6:

[1449] The user (store clerk) responds appropriately to the customer based on the feedback information from the smart glasses. The input is feedback information, and the output is the customer response action. This is expected to improve customer satisfaction.

[1450] Example 3

[1451] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1452] Conventional AI systems do not adequately ensure the safe provision of data from data providers and the appropriate output of learning results. Furthermore, they lack the ability to recognize emotions and adjust learning accordingly, which means they are unable to perform advanced learning based on the user's emotions.

[1453] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1454] In this invention, the server includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, means for managing the input and output of the data, means including an environment for providing the data, an emotion engine for recognizing emotions, means for feeding back the emotions recognized by the emotion engine to the artificial intelligence system, and means for adjusting the learning of the artificial intelligence system based on the feedback. This ensures the safe provision of data from data providers and the appropriate output of learning results, and also enables advanced learning based on emotions.

[1455] "Person or Legal Entity" means the natural person or legal entity recognized as the data provider.

[1456] "Licensed data" means data whose use has been officially authorized by the data provider.

[1457] An "artificial intelligence system" is a computer system that learns data and outputs analytical results.

[1458] "Means for managing data input and output" refers to a system that has the functionality to receive, store, analyze, and output data.

[1459] The "data provision environment" refers to the infrastructure that enables data providers to provide data safely and output learning results appropriately.

[1460] An "emotion engine" is software or hardware that recognizes emotions from data and outputs the results.

[1461] "Feedback means" is a function for transmitting the emotions recognized by the emotion engine to the artificial intelligence system.

[1462] The "means for adjusting learning" is a function for dynamically changing the learning process of an artificial intelligence system based on fed-back emotional information.

[1463] A "generative AI model" is a generative artificial intelligence model that has been trained in advance using large amounts of data.

[1464] "Copyright-free data that can be viewed by anyone on the Internet" refers to data that is freely accessible on the Internet and is free of copyright.

[1465] This invention relates to an AI system that includes an AI system for learning data licensed for use by an individual or a corporation, a means for managing the input and output of the data, and an environment for providing the data. The system further includes an emotion engine for recognizing emotions, a means for feeding back the emotions recognized by the emotion engine to the AI ​​system, and a means for adjusting the learning of the AI ​​system based on the feedback.

[1466] System configuration

[1467] Data provision environment

[1468] Users upload data using a dedicated interface, such as a CSV file containing customer reviews, via a web browser (general browser software). The uploaded data is received by the server and securely stored in a cloud storage service (general cloud storage).

[1469] Emotion Engine

[1470] The server then sends the stored data to the emotion engine, which uses high-speed processing units such as NVIDIA GPUs to analyze the data and recognize emotions. For example, it uses text analysis algorithms to analyze each sentence in a review and recognize emotions such as joy, anger, and sadness.

[1471] Feedback and learning adjustments

[1472] The emotion engine feeds the recognized emotions back to the AI ​​system in real time via an API. The AI ​​system then adjusts its learning based on the fed-back emotion information. For example, it strengthens learning of data related to the emotion of joy.

[1473] Output of training results

[1474] The server outputs the learning results of the AI ​​system to the data provider. The results are provided in a format that the data provider can access (e.g., CSV file). The data provider can download the learning results from the dashboard.

[1475] Specific examples

[1476] For example, if a user uploads a CSV file containing customer reviews, the server receives the CSV file and stores it in cloud storage. The server then sends the stored CSV file to an emotion engine, which analyzes the review text to recognize emotions. The recognized emotions are fed back to the AI ​​system in real time, and the AI ​​system adjusts its learning based on the feedback. Finally, the server outputs the learning results to the data provider, who can download the results from a dashboard.

[1477] Prompt Sentence Examples

[1478] "Design a system that analyzes customer reviews, recognizes emotions, and provides feedback. For example, identify the emotion of happiness in the review text, and use that information to adjust the AI ​​system's learning."

[1479] In this way, a system can be implemented in which the data provider, server, emotion engine, and artificial intelligence system work in cooperation with each other. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1480] Step 1:

[1481] The user uploads data using a dedicated interface. Specifically, they open a web browser, access the upload form, drag and drop a CSV file containing customer reviews, and click the "Upload" button. The input is the CSV file, and the output is data transmission to the server.

[1482] Step 2:

[1483] The server receives the CSV file sent by the user. The received data is checked for file integrity and then securely stored in a cloud storage service. The input is the CSV file sent by the user, and the output is the data stored in the cloud storage.

[1484] Step 3:

[1485] The server sends the saved data to the emotion engine. Specifically, it obtains the path of the saved CSV file and sends the data to the emotion engine's API endpoint via an HTTP POST request. The input is the path of the data saved in cloud storage, and the output is the data sent to the emotion engine.

[1486] Step 4:

[1487] The emotion engine analyzes the received data and recognizes emotions. Specifically, it uses NVIDIA GPUs to perform high-speed text analysis, analyzes each sentence of the review, and calculates an emotion score. The input is the data sent from the server, and the output is the recognized emotion score.

[1488] Step 5:

[1489] The emotion engine feeds back the recognized emotion to the AI ​​system. Specifically, the emotion score is sent in JSON format to the AI ​​system's API endpoint via an HTTP POST request. The input is the recognized emotion score, and the output is feedback to the AI ​​system.

[1490] Step 6:

[1491] The AI ​​system adjusts learning based on feedback from the emotion engine. Specifically, it analyzes the received emotion scores and dynamically modifies the learning algorithm. For example, it sets it to prioritize learning data with a high happiness emotion score. The input is feedback from the emotion engine, and the output is the adjusted learning model.

[1492] Step 7:

[1493] The server outputs the learning results of the AI ​​system to the data provider. Specifically, it obtains the learning results and displays them on a dashboard that the data provider can access. The data provider can download the learning results from the dashboard as a CSV file. The input is the result of the adjusted learning model, and the output is the provision of the learning results to the data provider.

[1494] (Application example 3)

[1495] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1496] In conventional artificial intelligence systems, data exchange between data providers and learning systems is sometimes not smooth, and it is difficult to propose services and products based on customer emotions. The present invention aims to solve these problems, facilitate data exchange between data providers and learning systems, and enable services and product proposals based on customer emotions.

[1497] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1498] In this invention, the server includes an artificial intelligence system for learning data licensed for use by individuals or corporations, means for managing the input and output of the data, means including an environment for providing the data, an emotion engine for recognizing emotions, means for adjusting learning based on the emotions recognized by the emotion engine, and means for proposing services and products based on the customer's emotions. This allows smooth data exchange between the data provider and the learning system, making it possible to propose services and products based on the customer's emotions.

[1499] "Individual or Legal Entity" means a particular natural person or legal entity.

[1500] "Licensed data" refers to data that has been officially licensed for use by the data provider.

[1501] An "artificial intelligence system" refers to a computer system that learns from data and performs analysis and predictions.

[1502] "Means for managing data input and output" refers to functions for controlling the reception of data and the output of results.

[1503] "Data provision environment" refers to the system environment in which data providers can safely provide data and receive learning results.

[1504] An "emotion engine" refers to software or hardware that recognizes and analyzes a user's emotions.

[1505] "Means to adjust learning" refers to the ability to modify or optimize the learning process of the artificial intelligence system based on feedback from the emotion engine.

[1506] "Means for proposing services and products" refers to the function for proposing appropriate services and products based on customer sentiment.

[1507] "Generative AI (Generative Pretrained Transformer)" refers to a generative model that has been pre-trained on a large dataset.

[1508] "Copyright-free data that is publicly viewable on the Internet" refers to data that is freely accessible on the Internet and is free of copyright.

[1509] As an embodiment of the present invention, a smart glasses application for improving customer experience in a brick-and-mortar store will be described as an example.

[1510] System Program

[1511] The system uses a camera mounted on smart glasses to capture customers' faces in real time and includes a program that analyzes their emotions using an emotion recognition model. Based on the emotions recognized by the emotion engine, the AI ​​system adjusts its learning and proposes services and products according to the customer's emotions.

[1512] Hardware and software used

[1513] Hardware: Smart glasses (e.g., Google Glass)

[1514] Software: OpenCV, Keras, emotion recognition model

[1515] Data processing and calculation

[1516] 1. Camera initialization: Initialize the camera on the smart glasses and capture video in real time.

[1517] 2. Face detection: Use OpenCV to detect faces in the captured frames.

[1518] 3. Emotion Recognition: The detected face regions are input into the emotion recognition model to predict the emotion. Run the emotion recognition model using Keras.

[1519] 4. Proposal Generation: Generate service or product proposals for customers based on the recognized emotions.

[1520] Specific examples

[1521] For example, if a customer smiles, the system generates a suggestion such as, "Let's introduce new products to the customer!" Conversely, if the customer looks sad, the system generates a suggestion such as, "Let's suggest products that will help the customer relax."

[1522] Prompt Sentence Examples

[1523] The following prompts might be fed to a generative AI model:

[1524] "If the customer smiles, generate a suggestion to introduce them to a new product."

[1525] By inputting this prompt into a generative AI model, specific suggestions can be generated that correspond to the customer's emotions.

[1526] As described above, this invention provides a system that uses smart glasses to analyze customer emotions in real time and proposes appropriate services and products based on the results, which is expected to improve the customer experience.

[1527] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1528] Step 1:

[1529] The server initializes the camera on the smart glasses and captures video in real time. The input is video data from the camera, and the output is the captured frames. Specifically, the server starts the camera device and continuously acquires frames.

[1530] Step 2:

[1531] The server detects faces from captured frames. The input is the captured frame, and the output is the coordinate information of the detected face area. Specifically, it converts the frame to grayscale using OpenCV and applies a face detection algorithm.

[1532] Step 3:

[1533] The server inputs the detected face region into an emotion recognition model to predict the emotion. The input is the image data of the face region, and the output is the predicted emotion label. Specifically, the server resizes the face region to an appropriate size and inputs it into the emotion recognition model using Keras.

[1534] Step 4:

[1535] The server generates service and product proposals for customers based on the recognized emotions. The input is the predicted emotion label, and the output is the proposal content. Specifically, the server retrieves the proposal content corresponding to the emotion label from the database and inputs a prompt sentence into the generative AI model to generate a specific proposal.

[1536] Step 5:

[1537] The server displays the generated suggestion on the display of the smart glasses. The input is the suggestion, and the output is text to be displayed on the display of the smart glasses. Specifically, the server converts the suggestion into text format and sends it to the display of the smart glasses.

[1538] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1539] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1540] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are listed.

[1541] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1542] [Third embodiment]

[1543] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1544] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[1545] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1546] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1547] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1548] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1549] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1550] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1551] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1552] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1553] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1554] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1555] "Example 1"

[1556] One embodiment of the present invention provides an artificial intelligence system for learning data licensed for use by an individual or corporation. This system uses a generative AI (generative pretrained transformer) to learn the data. Specifically, the licensed data is supplied as training data to the generative AI, and the system outputs the learning results.

[1557] "Example 2"

[1558] Furthermore, a means for managing data input and output is provided. This means is for properly managing the input of data for which permission has been granted and the output of learning results. Specifically, data input is provided by data providers for which permission has been granted, and output is data obtained as a result of learning. This data is properly managed and fed back to the providers as necessary.

[1559] "Example 3"

[1560] An environment for providing data is also provided. This environment is designed to safely provide data for which permission has been granted and to appropriately output learning results. Specifically, data providers provide data through this environment, and learning results are output through this environment. This allows for smooth data exchange between data providers and the learning system.

[1561] The processing flow of each embodiment will be described below.

[1562] "Example 1"

[1563] Step 1: Collect data with permission from individuals or entities. This data will serve as input for the AI ​​system to learn from.

[1564] Step 2: The collected data is fed to a generative AI (generative pretrained transformer), which then learns from this data.

[1565] Step 3: The generative AI outputs the learning results, which are used for subsequent processing and analysis.

[1566] "Example 2"

[1567] Step 1: Accept the input of authorized data. This data will be used as input for the AI ​​system to learn from.

[1568] Step 2: Properly manage input data, including storing, organizing, and controlling access to the data.

[1569] Step 3: Manage the output of learning results. The output data is stored appropriately and fed back to the provider as needed.

[1570] "Example 3"

[1571] Step 1: Prepare an environment for providing licensed data. This environment facilitates data exchange between data providers and the learning system.

[1572] Step 2: The data provider provides data through the environment. The provided data becomes input data for the AI ​​system to learn.

[1573] Step 3: The learning results are output through the environment. The output data is stored appropriately and fed back to the provider as needed.

[1574] Example 1

[1575] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1576] Conventional AI systems have had difficulty efficiently learning data licensed for use by individuals or corporations and outputting the results appropriately. Furthermore, the data preprocessing, supply, and learning processes require a lot of manual operations, which takes time and effort. Furthermore, there has been a lack of concrete methods for improving the quality of learning results using generative AI models.

[1577] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1578] In this invention, the server includes an artificial intelligence system for learning data licensed from an individual or a corporation, means for managing the input and output of the data, means for preprocessing the data, means for supplying the data to a generative AI model, means for training the generative AI model, and means for outputting the learning results, thereby enabling efficient learning of data and output of high-quality learning results.

[1579] "Individual or legal entity" refers to the entity that provides permission to use data, and includes individuals, companies, organizations, etc.

[1580] "Licensed Data" means data that has been officially licensed for use by an individual or legal entity, and may include text, images, audio, or other formats.

[1581] An "artificial intelligence system" is a computer system that learns from data and outputs the results, including generative AI models.

[1582] "Means for managing inputs and outputs" refers to the functions and processes for receiving data and providing results.

[1583] "Preprocessing means" refers to functions or processes that process data to convert it into a format suitable for learning.

[1584] A "generative AI model" refers to a generative artificial intelligence model such as the Generative Pretrained Transformer (GPT), which has the ability to learn data and generate new information.

[1585] "Means of supply" refers to the functions and processes for passing preprocessed data to a generative AI model.

[1586] "Means of learning" refers to the functions and processes used to train a generative AI model on data.

[1587] "Means for outputting learning results" refers to the functions and processes for providing the results learned by the generative AI model to the user.

[1588] MODE FOR CARRYING OUT THE INVENTION

[1589] The present invention relates to an artificial intelligence system for learning data licensed for use by an individual or a corporation. Specific embodiments of this system will be described below.

[1590] System Overview

[1591] The server will build an artificial intelligence system to learn from data provided by individuals or corporations. This system will use a generative AI model (Generative Pretrained Transformer, hereafter referred to as GPT) to learn from the data and output the results.

[1592] Hardware and software used

[1593] The server uses the following hardware and software:

[1594] Hardware: Server with high-performance CPU and GPU

[1595] Software: Python scripts for data preprocessing, GPT as a generative AI model, database system for data management

[1596] Data collection and preprocessing

[1597] A user provides the server with data for which they have received permission to use from an individual or a corporation. The data may be in various formats, such as text, images, or audio. The server preprocesses the provided data. Specifically, for text data, the server performs the following processing:

[1598] Remove unnecessary spaces and special characters

[1599] Tokenization (the process of dividing words and sentences)

[1600] Normalization (unifying processes such as converting uppercase to lowercase)

[1601] Supplying data and training the model

[1602] The server supplies the preprocessed data to GPT. The data is batch-processed to ensure efficient learning. The server trains GPT using the supplied data. The learning process involves capturing the features of the data and adjusting the weights to recognize patterns.

[1603] Output of training results

[1604] The server stores the trained GPT model and provides it to users as needed. Users can use this trained model to operate chatbots for customer support.

[1605] Specific examples

[1606] A user wants to create a chatbot for customer support using their company's customer support data. The user provides the server with past chat logs of customer support. The server preprocesses this data and feeds it to GPT for training. The trained model is then used for the user's customer support operations.

[1607] Prompt Sentence Examples

[1608] "Please use the following customer support chat logs as learning data to create a customer support chatbot."

[1609] User: I haven't received my order, what should I do?

[1610] Support: Can I have your order number?

[1611] User: 123456.

[1612] Support: Please wait a moment while we check this.

[1613] Using this prompt, the server trains the GPT and generates a chatbot for customer support.

[1614] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1615] Step 1:

[1616] Users provide the server with data that they have obtained permission to use from individuals or corporations. The input data can be in various formats, such as text, images, and audio. The server receives this data and stores it in a database. Specifically, users upload data, and the server stores it in the appropriate folder.

[1617] Step 2:

[1618] The server preprocesses the provided data. The input data is the raw data saved in step 1. For text data, the server removes unnecessary whitespace and special characters, tokenizes, and normalizes the data. The output is the preprocessed, clean data. Specifically, the server runs a Python script to tokenize and normalize the text data.

[1619] Step 3:

[1620] The server feeds the preprocessed data to the generative AI model (GPT). The input data is the data preprocessed in step 2. The server batches the data and efficiently feeds it to the model. The output is the batch of data fed to the model. Specifically, the server divides the data into chunks of a certain size and passes them to the model.

[1621] Step 4:

[1622] The server trains the generative AI model using the supplied data. The input data is the batch of data supplied in step 3. The server captures the features of the data and adjusts the weights to recognize patterns. The output is a trained generative AI model. Specifically, the server runs the model training process and updates the model parameters every epoch.

[1623] Step 5:

[1624] The server saves the generative AI model after training and provides it to the user as needed. The input data is the model that completed training in step 4. The server saves the model and makes it accessible to the user. The output is a trained model that can be used by the user. Specifically, the server saves the model as a file and provides the user with a download link.

[1625] Step 6:

[1626] A user uses the trained model to operate a customer support chatbot. The input data is an inquiry from the user. The trained model generates an appropriate response to the user's inquiry. The output is a response message to the user. In concrete terms, the user inputs a question to the chatbot and receives the response generated by the model.

[1627] (Application example 1)

[1628] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1629] Conventional advertising systems have had difficulty effectively generating personalized ads based on users' interests. Furthermore, there was a lack of technology to appropriately utilize user data and generate ads using generative AI. This resulted in low advertising effectiveness and an inability to capture users' attention.

[1630] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1631] In this invention, the server includes an artificial intelligence system for learning data licensed for use by an individual or a corporation, means for managing the input and output of the data, means including an environment for providing the data, means for generating personalized advertisements based on the data licensed by the user, and means for displaying the generated advertisements to the user, thereby enabling the generation of effective personalized advertisements based on the user's interests and concerns.

[1632] "Personal or legal entity licensed data" means information or data that is owned and licensed by a person or legal entity.

[1633] An "artificial intelligence system" is a collection of computer programs and algorithms that learn, analyze, and generate data.

[1634] "Means for managing data input and output" are hardware and software components for controlling the receipt, storage, processing of data, and output of results.

[1635] A "data serving environment" is an infrastructure or platform for collecting, storing, accessing, and processing data.

[1636] "Data with user permission" refers to data such as personal information and behavioral history that the user has explicitly given permission to use.

[1637] "Personalized advertising" refers to advertising that is customized based on a particular user's interests.

[1638] "Generative AI (Generative Pretrained Transformer)" is an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data and content.

[1639] "Means for generating advertisements" refers to algorithms or programs for creating advertising content based on collected data.

[1640] "Means for displaying advertisements to users" are the hardware and software components for displaying the generated advertisements on the user's device.

[1641] To implement the present invention, the following system configuration and processing procedures are required.

[1642] First, the server collects data with permission from individuals or corporations and prepares an AI system to learn from it. This AI system uses a generative pretrained transformer (generative AI) to learn from the data and perform analysis and generation.

[1643] The server then has a means for managing the input and output of data. This means is a hardware and software component for controlling the reception, storage, processing, and output of data. Specifically, it uses a database management system or an API.

[1644] Furthermore, the server comprises a means for providing the data, including the environment, which is the infrastructure or platform for collecting, storing, accessing, and processing the data, such as a cloud service or an on-premise server.

[1645] Based on data with the user's permission (e.g., browsing history, purchase history, social media activity, etc.), the server has the means to generate personalized ads. This means is an algorithm or program that creates advertising content based on the collected data. Generative AI models are used to generate ads optimized for the user.

[1646] The generated advertisements are sent from the server to the user's device and displayed through a means for displaying advertisements to the user. The means are hardware and software components for displaying the generated advertisements on the user's device, specifically, a smartphone app or a web browser.

[1647] For example, if a user recently purchased a smartphone or laptop and liked an article about AI on social media, the ad generated might look like this:

[1648] Introducing a smartphone equipped with the latest AI technology! Now available at a special price. Click here for details!

[1649] An example of a prompt sentence to input to the generative AI model is as follows:

[1650] "Generate personalized ads based on your data, including:

[1651] Products and services that may interest you

[1652] Advertising catchphrase

[1653] Ad visual ideas

[1654] User Data:

[1655] Browsing history: tech news, smartphones, AI advancements

[1656] Purchase history: smartphone, laptop

[1657] Social media activity: liked AI article, shared tech news

[1658] In this way, it becomes possible to generate effective personalized advertisements based on the user's interests.

[1659] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1660] Step 1:

[1661] The server collects data that is licensed for use by individuals or corporations.

[1662] Input: Data provided by individuals or entities

[1663] Output: Collected data

[1664] Specific operation: The server obtains data provided by individuals or corporations through APIs or database connections and stores it in a database.

[1665] Step 2:

[1666] The server generates prompt sentences to supply the collected data to the generative AI model.

[1667] Input: Collected data

[1668] Output: Prompt sentence to be fed to the generative AI model

[1669] How it works: The server analyzes the collected data and generates prompts based on the user's interests. For example, it creates prompts containing appropriate advertising content based on the user's browsing history and purchase history.

[1670] Step 3:

[1671] The server generates personalized advertisements using a generative AI model.

[1672] Input: Generated prompt text

[1673] Output: Personalized ads

[1674] How it works: The server provides prompts to a generative AI model (e.g., OpenAI's GPT-3), which then generates ad content based on the user's interests.

[1675] Step 4:

[1676] The server transmits the generated advertisement to the user's terminal.

[1677] Input: Generated personalized ad

[1678] Output: Ad sent to user's device

[1679] How it works: The server sends the generated ad to the user's smartphone app or web browser using push notifications or APIs.

[1680] Step 5:

[1681] The device displays the received advertisement to the user.

[1682] Input: Advertisement sent by the server

[1683] Output: The ad shown to the user

[1684] What happens: The user's device (smartphone or PC) displays the received ad. The ad appears as a banner or pop-up within the app.

[1685] In this way, effective personalized advertisements based on the user's interests can be generated and displayed.

[1686] Example 2

[1687] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1688] Conventional AI systems have insufficient management of data input and output, making it difficult to properly manage data for which permission has been granted and to provide feedback on learning results. Feedback to data providers is often delayed, hindering efficient data use. This makes it difficult to build a relationship of trust between data providers and users, resulting in delays in data provision.

[1689] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1690] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, an environment for providing the data, a means for uploading the data to the server, a means for saving the data in a data management system, a means for inputting the data into a generative AI model for learning, a means for saving the learning results in the data management system, and a means for feeding back the learning results to the provider, thereby enabling appropriate management of data and rapid feedback.

[1691] An "artificial intelligence system" is a computer program that learns from data licensed by an individual or legal entity to perform a specific task.

[1692] "Means for managing data input and output" refers to a system that has the functionality to appropriately manage data input from data providers and the output of learning results.

[1693] The "data providing environment" is a system that includes an interface and a protocol for data providers to upload data to a server.

[1694] "Means for uploading to the server" refers to a system that has a function for users to send data obtained from providers to the server.

[1695] "Data Management System" means a system for storing uploaded data and verifying its integrity and permission to use.

[1696] A "generative AI model" is a model that learns from input data and performs specific tasks.

[1697] A "means for learning" is a system that inputs data into a generative AI model and runs a process to optimize the model's parameters.

[1698] "Means for storing learning results in a data management system" refers to a system that has the functionality to store the learning results obtained from a generative AI model in a data management system.

[1699] "Means for feeding back learning results to providers" refers to a system that has the function of reporting learning results to providers and providing feedback.

[1700] This invention is a system that includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, and an environment for providing the data. Specifically, the server, terminal, and user each play their respective roles, achieving appropriate management of data and rapid feedback.

[1701] Hardware and software used

[1702] Hardware: Servers, terminals

[1703] Software: Data management system, generative AI models

[1704] Data entry management

[1705] The user prepares the data for which they have received permission to use from the provider. Next, the user uses their device to upload the data they received from the provider to the server. The server receives the uploaded data and stores it in a data management system. Furthermore, the server checks the integrity and permission to use of the data to ensure that it is managed appropriately.

[1706] As a specific example of operation, a user uploads image data obtained from a provider to a server via a browser on their device. The server stores the uploaded image data in a database and records meta information about the data (such as the provider and usage permission details).

[1707] Data training

[1708] The server retrieves the data stored in the data management system. Then, the server inputs the retrieved data into the generative AI model to start the learning process. The generative AI model learns based on the input data and optimizes the model parameters. The server saves the generative AI model after learning is complete.

[1709] As a specific example of how it works, the server retrieves image data from a database and inputs it into a generative AI model. The generative AI model learns to classify cats and dogs based on the image data and adjusts the model parameters. The server then saves the model once it has completed learning, preparing it for the next step.

[1710] Output management of learning results

[1711] The server stores the learning results obtained from the generative AI model in a data management system. The device then feeds back the learning results to the provider as needed. The user then uses the device to check the learning results and report them to the provider.

[1712] As a specific example of how it works, the server stores the learning results (e.g., a cat / dog classification model) in a database. The user can then use their device to check the learning results and provide feedback to the provider via email or a dashboard.

[1713] Prompt Sentence Examples

[1714] Examples of prompts to be input to a generative AI model include:

[1715] "Use this image data to generate a model that classifies cats and dogs."

[1716] "Create a model that performs sentiment analysis based on the provided text data."

[1717] The above is an embodiment of the present invention.

[1718] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1719] Step 1:

[1720] Preparing and uploading data

[1721] The user prepares the data for which permission to use has been obtained from the provider. Next, the user uses the terminal to upload the data obtained from the provider to the server. Specifically, the user uploads the image data obtained from the provider to the server through the terminal's browser. The input is the image data obtained from the provider, and the output is the image data stored on the server.

[1722] Step 2:

[1723] Data storage and management

[1724] The server receives the uploaded data and stores it in the data management system. Furthermore, the server checks the data integrity and usage permissions to ensure that the data is managed appropriately. Specifically, the server stores the uploaded image data in a database and records the data's meta information (provider, usage permission details, etc.). The input is the uploaded image data, and the output is the data and its meta information stored in the database.

[1725] Step 3:

[1726] Obtaining data and starting training

[1727] The server retrieves data stored in the data management system. The server then inputs the retrieved data into the generative AI model and starts the learning process. Specifically, the server retrieves image data from the database and inputs it into the generative AI model. The input is the image data retrieved from the database, and the output is the data input into the generative AI model.

[1728] Step 4:

[1729] Training generative AI models

[1730] The generative AI model learns based on the input data and optimizes the model parameters. Specifically, the generative AI model learns to classify cats and dogs based on image data and adjusts the model parameters. The input is the image data entered into the generative AI model, and the output is the model after learning is complete.

[1731] Step 5:

[1732] Saving training results

[1733] The server stores the generative AI model after training. Specifically, the server stores the training results (for example, a cat / dog classification model) in a database. The input is the trained model, and the output is the training results stored in the database.

[1734] Step 6:

[1735] Feedback on learning results

[1736] The device feeds back the learning results to the provider as needed. The user uses the device to check the learning results and report them to the provider. Specifically, the user uses the device to check the learning results and provide feedback to the provider via email or dashboard. The input is the learning results stored in the database, and the output is the information fed back to the provider.

[1737] (Application example 2)

[1738] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1739] Conventional content recommendation systems have difficulty recommending optimal content to users because they are unable to effectively utilize user viewing history and rating data. Furthermore, there are problems with insufficient management of data input and output, and inappropriate feedback to data providers. This has hindered improvements to the user experience.

[1740] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1741] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, an environment for providing the data, a means for a generative AI model to recommend optimal content to users using data provided by the user, and a means for providing feedback on the recommended content to the user. This makes it possible to effectively utilize users' viewing history and rating data to recommend optimal content to users. Furthermore, the appropriate management of data input and output and appropriate feedback to data providers improves the user experience.

[1742] "Person or Legal Entity" means a natural person or a legally recognized entity.

[1743] "Permission to use" refers to obtaining formal permission from the provider to use the data.

[1744] "Data" means a collection of information that is collected, stored and processed for a specific purpose.

[1745] An "artificial intelligence system" refers to a system that uses technologies such as machine learning and deep learning to analyze data and automatically perform specific tasks.

[1746] "Means for managing data input and output" refers to methods and devices for appropriately collecting, storing, processing, and outputting data.

[1747] "Data providing environment" refers to the physical or virtual location or system where data is collected or provided.

[1748] A "generative AI model" refers to a model that uses a generative AI (generative pretrained transformer) to analyze data and generate new information and content.

[1749] "Means for recommending optimal content to users" refers to methods or devices that select and present the most appropriate content for a user based on user data.

[1750] "Feedback means" refers to a method or device for returning results or information generated by the system to the user or data provider.

[1751] The following system configuration will be described as an embodiment of the present invention.

[1752] First, the server is equipped with an artificial intelligence system for learning data licensed by individuals or corporations. This artificial intelligence system analyzes the data using a generative pretrained transformer (generative AI) and generates new information and content. The server also has a means for managing data input and output, and appropriately collects, stores, processes, and outputs data.

[1753] Data provided by users (viewing history, ratings, comments, etc.) is sent to the server. Based on this data, the server uses a generative AI model to recommend the most suitable content to the user. Specifically, the server converts the user's viewing history and rating data into TF-IDF vectors and calculates the cosine similarity to select the most suitable content for the user.

[1754] The server also has a means for providing feedback on the recommended content to the user, allowing the user to easily find content that matches their preferences.

[1755] The hardware used includes servers, users' smartphones, PCs, etc. The software used is Python and its libraries (scikit-learn, json).

[1756] As a concrete example, consider a system that uses data on the movies a user has watched to recommend the next movie to watch. Here is an example prompt:

[1757] User ID: user123

[1758] Viewing history: ["Movie A", "Movie B", "Movie C"]

[1759] Rating: {"Movie A": 5, "Movie B": 4, "Movie C": 3}

[1760] Comments: {"Movie A": "Great movie", "Movie B": "Good", "Movie C": "Okay"}

[1761] By inputting this prompt into a generative AI model, the system can recommend the most suitable movie to the user. This system effectively utilizes the user's viewing history and rating data to recommend the most suitable content to the user. In addition, the system properly manages data input and output, and provides appropriate feedback to data providers, improving the user experience.

[1762] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1763] Step 1:

[1764] The user inputs data such as viewing history, ratings, and comments from the terminal.

[1765] Input: User's viewing history, ratings, comments

[1766] Output: User data (viewing history, ratings, comments)

[1767] Specific operation: Users use their smartphones or PCs to enter their viewing history, ratings, and comments. This data is then sent from the device to the server.

[1768] Step 2:

[1769] The server receives the user data and stores it in JSON format.

[1770] Input: User data (viewing history, ratings, comments)

[1771] Output: User data in JSON format

[1772] Specific operation: The server receives user data sent from the device and stores it in JSON format, which ensures data consistency and readability.

[1773] Step 3:

[1774] The server analyzes the data using a generative AI model based on user data.

[1775] Input: User data in JSON format

[1776] Output: Analysis results (user profile)

[1777] Specific operation: The server uses Python and the scikit-learn library to convert user data into TF-IDF vectors and generate user profiles.

[1778] Step 4:

[1779] The server reads the content database and compares it to the user profile.

[1780] Input: User profile, content database

[1781] Output: Similarity score

[1782] Specific operation: The server reads the content database and compares the user profile with the profile of each content using cosine similarity.

[1783] Step 5:

[1784] The server recommends the most suitable content to the user based on the similarity score.

[1785] Input: Similarity score

[1786] Output: Recommended content list

[1787] Specific operation: The server selects the most suitable content for the user based on the similarity score and generates a recommended content list.

[1788] Step 6:

[1789] The server feeds back the recommended content list to the user.

[1790] Input: Recommended content list

[1791] Output: Feedback to the user (recommended content)

[1792] Specific operation: The server sends the generated recommended content list to the user's device and provides feedback to the user, who can then check the recommended content on the device.

[1793] Example 3

[1794] Next, a third embodiment of the third embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1795] Conventional AI systems lack a consistent mechanism for data input from data providers, safe storage of data, learning from the data using generative AI models, saving of learning results, and providing results to users. This leads to problems such as data exchange not being smooth, and the reliability and security of learning results not being ensured.

[1796] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1797] In this invention, the server includes a means for the terminal to input data from a data provider, a means for the server to receive and store the data, a means for the server to input the data to the generative AI model, a means for the generative AI model to learn from the data and generate results, a means for the server to store the learning results, and a means for the user to request and obtain the learning results. This allows smooth data exchange between the data provider and the learning system, enabling the user to quickly obtain the learning results.

[1798] "Individual or Legal Entity" refers to the natural person or legal entity recognized as the data provider.

[1799] "Licensed Data" refers to data that has been officially licensed for use by the data provider.

[1800] "Artificial intelligence system" refers to a computer system for learning from data and generating analytical results.

[1801] "Means for managing data input and output" refers to the functionality for receiving, storing, analyzing data, and outputting the results.

[1802] "Data provision environment" refers to the system environment in which data providers can safely provide data and appropriately output learning results.

[1803] "Terminal" means the computer or device used by the Data Contributor to enter data.

[1804] "Server" refers to a central processing unit for receiving, storing, analyzing data, and outputting results.

[1805] A "generative AI model" refers to an artificial intelligence model that learns from data and generates analytical results.

[1806] "Means for receiving and storing data" refers to the function of receiving data sent from a terminal and storing it in a database, etc.

[1807] "Means for inputting data" refers to the interface or function that allows data providers to input data.

[1808] "Means of learning from data and generating results" refers to the function by which a generative AI model analyzes data and generates learning results.

[1809] "Means for saving learning results" refers to the function of saving the learning results generated by the generative AI model in a database, etc.

[1810] "Means for requesting and obtaining learning results" refers to the interface or functionality that allows users to request and obtain learning results.

[1811] "User" refers to an individual or legal entity that accesses the system to utilize the learning results.

[1812] This invention is an AI system for learning data licensed from individuals or corporations, a means for managing the input and output of the data, and an environment for providing the data. Specifically, the system processes and calculates the data using terminals, a server, and a generative AI model.

[1813] Hardware and software used

[1814] Hardware: Servers, terminals

[1815] Software: Generative AI models (e.g., GPT-4), data management systems (e.g., SQL databases)

[1816] System configuration

[1817] 1. Terminal: A device used by data providers to input data. For example, used by medical staff to input patient symptom data.

[1818] 2. Server: A central processing unit that receives, stores, analyzes, and outputs data. The server uses a data management system to store data and generative AI models to analyze the data.

[1819] 3. Generative AI model: An artificial intelligence model that learns from data and generates analytical results. For example, GPT-4 analyzes patient symptom data and generates a diagnosis.

[1820] System Operation

[1821] Data provider enters data: Data providers enter data using dedicated terminals. For example, medical institution staff enter information such as the patient's name, age, and symptoms.

[1822] The server receives and stores the data: The server receives the data sent from the device and stores it in the data management system. When storing the data, it checks the integrity and security of the data.

[1823] The server inputs data into the generative AI model: The server inputs the stored data into the generative AI model, for example, inputting patient symptom data into the GPT-4 model.

[1824] Generative AI models learn from data and generate results: Generative AI models learn from input data and generate results. For example, GPT-4 analyzes a patient's symptom data and generates a diagnosis.

[1825] The server stores the learning results: The server receives the learning results output from the generative AI model and stores them in a data management system.

[1826] User requests and obtains learning results: A user requests learning results using a dedicated device. For example, a doctor obtains diagnosis results.

[1827] Examples of concrete examples and prompts

[1828] Example: A medical institution provides patient data and uses a generative AI model (GPT-4) to generate a diagnosis. A doctor obtains the diagnosis.

[1829] Example prompt sentence:

[1830] Input prompt: "Enter patient symptom data and generate a diagnosis."

[1831] Output prompt: "The diagnostic results are as follows:..."

[1832] This system allows smooth data exchange between the data provider and the learning system, enabling users to quickly obtain learning results. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1833] Step 1:

[1834] The terminal inputs data from the data provider.

[1835] Data providers input data using a dedicated terminal. For example, a medical institution staff member inputs information such as the patient's name, age, and symptoms. The input data is structured based on the input form on the terminal. Specifically, the data provider inputs the required information into the input form on the terminal and clicks the "Submit" button.

[1836] Step 2:

[1837] The server receives and stores the data.

[1838] Data sent from the device is sent to the server using a secure communication protocol (e.g., HTTPS). The server stores the received data in a data management system (e.g., an SQL database). When storing the data, the server checks the integrity and security of the data. Specifically, the server receives the data and executes an SQL query to store it in the database.

[1839] Step 3:

[1840] The server inputs data into the generative AI model.

[1841] The server inputs the stored data into the generative AI model. For example, it inputs patient symptom data into the GPT-4 model. The input data is sent to the API of the generative AI model. Specifically, the server retrieves the necessary data from the database and creates a request to send the data to the API of the generative AI model.

[1842] Step 4:

[1843] Generative AI models learn from data and generate results.

[1844] A generative AI model learns from input data and generates results. For example, GPT-4 analyzes a patient's symptom data and generates a diagnosis. Specifically, the generative AI model analyzes the data and outputs the diagnosis and recommended treatment in text format.

[1845] Step 5:

[1846] The server stores the learning results.

[1847] The server receives the learning results output from the generative AI model and stores them in a data management system. When saving, it checks the integrity and security of the results. Specifically, the server receives the learning results and executes SQL queries to save them in a database.

[1848] Step 6:

[1849] The user requests and obtains the learning results.

[1850] A user requests learning results using a dedicated terminal. For example, a doctor obtains diagnosis results. The user enters a patient's ID into the search form on the terminal and clicks the "Search" button. The server receives the request and provides the learning results to the user. Specifically, the server retrieves the learning results from the database and generates a response to display the results on the user's terminal.

[1851] (Application example 3)

[1852] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1853] Conventional artificial intelligence systems have faced the problem of lacking an environment in which data providers can safely provide data and appropriately output learning results. Furthermore, data security is often not ensured during data provision and learning result output, creating the risk of data leaks and unauthorized use. This has led to a demand for an environment in which data providers can provide data with peace of mind.

[1854] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1855] In this invention, the server includes an artificial intelligence system for learning data licensed from an individual or a corporation, a means for managing the input and output of data, a means including an environment for providing data, a means for encrypting and providing data, and a means for decrypting the encrypted data and outputting the learning results. This enables data providers to provide data safely and output the learning results appropriately.

[1856] "Person or legal entity" means a natural person or a legal entity.

[1857] "Permission to use" refers to the permission given by the data provider to use the data.

[1858] "Data" refers to a collection of information, the material that artificial intelligence systems use to learn.

[1859] "Artificial intelligence system" refers to a computer program that learns from data and performs specific tasks.

[1860] "Means for managing input and output" refers to the functions for controlling the receipt of data and the delivery of results.

[1861] "Data provision environment" refers to the systems and infrastructure that enable data providers to provide data securely.

[1862] "Encryption" refers to the conversion of data using a specific algorithm to make it in a form that cannot be easily understood by third parties.

[1863] "Decryption" refers to restoring encrypted data to its original form.

[1864] "Learning results" refers to the output that an AI system generates after learning from data.

[1865] "Generative AI (Generative Pretrained Transformer)" refers to an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data.

[1866] As an embodiment of the present invention, the following system can be constructed.

[1867] System Overview

[1868] This system includes an artificial intelligence system for learning data that has been granted permission to be used by an individual or a corporation, a means for managing the input and output of data, an environment for providing the data, a means for encrypting and providing the data, and a means for decrypting the encrypted data and outputting the learning results.

[1869] Hardware and software used

[1870] Hardware: Smartphones, servers

[1871] Software: Python, requests library, cryptography library, Generative Pretrained Transformer

[1872] Program processing explanation

[1873] Data encryption and delivery

[1874] When a user provides data using their smartphone, the data is first encrypted using the Fernet class from the cryptography library. The encrypted data is then sent to the server using an HTTP POST request using the requests library.

[1875] Obtaining and decoding the training results

[1876] The server decrypts the encrypted data it receives and uses a generative pretrained transformer (Generative AI) to learn from the data. The learning results are then re-encrypted and decrypted when the user retrieves them using their smartphone.

[1877] Specific examples

[1878] For example, if a user is using a smartphone to provide "user behavior data," the following prompt sentence can be used:

[1879] Prompt Sentence Examples

[1880] "You provide user behavior data, and the AI ​​model learns from that data and outputs results that predict user behavior patterns."

[1881] In this way, a system can be realized that provides an environment in which data providers can safely provide data and appropriately output learning results.

[1882] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1883] Step 1:

[1884] A user enters data using a smartphone.

[1885] Input: User-provided data (e.g., user behavior data)

[1886] Specific operation: The user launches the smartphone application and enters the data to be provided.

[1887] Step 2:

[1888] The device encrypts the data.

[1889] Input: User-supplied data

[1890] Data processing: Encrypt the data using the Fernet class from the cryptography library.

[1891] Output: Encrypted data

[1892] Specific operation: The terminal encrypts the data entered by the user using the Fernet class to generate encrypted data.

[1893] Step 3:

[1894] The device sends the encrypted data to the server.

[1895] Input: Encrypted data

[1896] Data operations: Send data using an HTTP POST request.

[1897] Output: Data sent to the server

[1898] Specific operation: The device uses the requests library to send the encrypted data to the server via an HTTP POST request.

[1899] Step 4:

[1900] The server receives the encrypted data and decrypts it.

[1901] Input: Encrypted data

[1902] Data processing: Decrypt the data using the Fernet class from the cryptography library.

[1903] Output: Decrypted data

[1904] Specific operation: The server decrypts the received encrypted data using the Fernet class to obtain the original data.

[1905] Step 5:

[1906] The server inputs the decrypted data into a generative AI model for learning.

[1907] Input: Decrypted data

[1908] Data computation: Learn data using a generative AI (generative pretrained transformer).

[1909] Output: Learning results

[1910] Specific operation: The server inputs the decoded data into the generative AI model, which learns the data and generates the learning results.

[1911] Step 6:

[1912] The server encrypts the learning results.

[1913] Input: Learning results

[1914] Data processing: Encrypt the training results using the Fernet class from the cryptography library.

[1915] Output: Encrypted training results

[1916] Specific operation: The server encrypts the generated learning result using the Fernet class and generates the encrypted learning result.

[1917] Step 7:

[1918] The server sends the encrypted learning results to the terminal.

[1919] Input: Encrypted learning results

[1920] Data operation: Sends data using an HTTP GET request.

[1921] Output: Data sent to the terminal

[1922] Specific operation: The server uses the requests library to send the encrypted learning results to the terminal via an HTTP GET request.

[1923] Step 8:

[1924] The terminal receives the encrypted learning result and decrypts it.

[1925] Input: Encrypted learning results

[1926] Data processing: Decrypt the training results using the Fernet class from the cryptography library.

[1927] Output: Decoded training results

[1928] Specific operation: The terminal decrypts the received encrypted learning result using the Fernet class and obtains the original learning result.

[1929] Step 9:

[1930] The user checks the decoded learning results.

[1931] Input: Decoded training results

[1932] Output: Learning results for user confirmation

[1933] Specific operation: The user checks the decoded learning results through a smartphone application.

[1934] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1935] "Example 1"

[1936] As one embodiment of the present invention, an artificial intelligence system including an emotion engine is provided. This system learns from data licensed for use by individuals or corporations and manages the input and output of that data. The emotion engine also recognizes a user's emotion and adjusts learning based on that emotion. Specifically, when a user inputs text or voice, the emotion engine recognizes the user's emotion from that input. The AI ​​system then adjusts learning based on the recognized emotion. For example, if a user inputs text or voice that indicates the emotion of joy, the emotion engine recognizes the emotion of joy, and the AI ​​system learns based on that emotion.

[1937] "Example 2"

[1938] Another embodiment of the present invention provides a specific technique for an emotion engine to recognize a user's emotion. Specifically, the emotion engine uses natural language processing (NLP) technology to recognize emotions from a user's text input. Also, it uses speech recognition technology to recognize emotions from a user's voice input. For example, if a user inputs the text "I'm very happy," the emotion engine uses NLP technology to recognize the emotion of joy from the text. Also, if a user inputs a voice containing the emotion of joy, the emotion engine uses speech recognition technology to recognize the emotion of joy from the voice.

[1939] "Example 3"

[1940] In yet another embodiment of the present invention, a specific technique is provided for adjusting the learning of an AI system based on emotions recognized by an emotion engine. Specifically, the emotion engine feeds back the recognized emotions to the AI ​​system, and the AI ​​system adjusts its learning based on the feedback. For example, if the emotion engine recognizes the emotion of joy, it feeds back that information to the AI ​​system. Then, the AI ​​system strengthens its learning of data related to the emotion of joy based on the feedback.

[1941] The processing flow of each embodiment will be described below.

[1942] "Example 1"

[1943] Step 1: The user enters text or voice input.

[1944] Step 2: The emotion engine recognizes the user's emotion from the input.

[1945] Step 3: Adjust the AI ​​system's learning based on the recognized emotions.

[1946] "Example 2"

[1947] Step 1: The user enters text.

[1948] Step 2: The emotion engine uses natural language processing (NLP) techniques to recognize emotions from the text.

[1949] Step 3: The user provides voice input.

[1950] Step 4: The emotion engine uses speech recognition technology to recognize emotions from the speech.

[1951] "Example 3"

[1952] Step 1: The emotion engine recognizes the user's emotion.

[1953] Step 2: The recognized emotions are fed back to the AI ​​system.

[1954] Step 3: The AI ​​system adjusts its learning based on that feedback.

[1955] Example 1

[1956] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1957] Conventional AI systems have difficulty effectively learning from data that has been authorized for use by individuals or corporations, and have not adjusted their learning to take user emotions into account. As a result, they have been unable to obtain appropriate learning results based on user emotions, limiting the flexibility and accuracy of the system.

[1958] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1959] In this invention, the server includes an AI system for learning data licensed from an individual or a corporation, means for managing the input and output of the data, means for preprocessing the data, means for learning the data using a generative AI model, and means for recognizing a user's emotions and adjusting learning based on the emotions, thereby enabling flexible and highly accurate learning results that take the user's emotions into consideration.

[1960] "Individual or Legal Entity" refers to a specific individual or legal entity that provides permission to use data.

[1961] "Licensed data" refers to data that has been explicitly licensed by an individual or legal entity and is used to train an artificial intelligence system.

[1962] "Artificial intelligence system" refers to a computer program and associated hardware for learning from data and performing specific tasks.

[1963] "Means for managing input and output" refers to software and hardware for controlling the input and output of data and for appropriate processing.

[1964] "Pre-processing means" refers to software and hardware for pre-processing data, such as cleaning, formatting, and tokenization.

[1965] "Generative artificial intelligence model" refers to a generative AI model such as the Generative Pretrained Transformer, which has the ability to generate new information based on data.

[1966] "Means for recognizing emotions and adjusting learning" refers to software and hardware for analyzing a user's emotions and dynamically adjusting the learning process of an artificial intelligence system based on those emotions.

[1967] "Data serving environment" refers to the infrastructure for collecting, storing, processing, and serving data.

[1968] The present invention relates to an AI system for learning from data licensed by individuals or corporations. The system utilizes a generative AI model and has the ability to adjust learning based on user emotions.

[1969] System configuration

[1970] Data collection and permission acquisition

[1971] The user provides the system with data for which they have received permission to use from an individual or corporation. The terminal then sends the data provided by the user to the server. The server then checks the permission to use the received data and stores it in a database. Specifically, a database management system (e.g., MySQL or PostgreSQL) is used.

[1972] Data Preprocessing

[1973] The server preprocesses the collected data. Specifically, it cleans the data (filling in missing values ​​and removing outliers) and converts the format (tokenizing text data, etc.). The server cleans the data using Python's pandas library and converts the numeric data using NumPy.

[1974] Learning generative AI models

[1975] The server supplies the preprocessed data to a generative AI model (Generative Pretrained Transformer). The server uses an NVIDIA GPU (e.g., Tesla V100) to train the generative AI model using TensorFlow or PyTorch. The server monitors the model's performance during training and adjusts hyperparameters as needed.

[1976] Use of emotion engine

[1977] When a user inputs text or voice, the device sends the input to a server. The server uses an emotion engine to recognize the user's emotions. Specifically, it uses the NLP library Hugging Face Transformers to extract emotions from text data. The server then adjusts the learning of the generative AI based on the recognized emotions. For example, if a user inputs text that expresses joy, it prioritizes positive data in the learning process.

[1978] Output of training results

[1979] The server outputs the learning results of the generative AI and provides them to the user via a terminal. The server generates the learning results in text and graph format and displays them in a way that is easy for the user to understand. Specifically, graphs are created using Matplotlib and Seaborn.

[1980] Specific examples

[1981] For example, if the user enters the text "I'm so happy today!", the specific behavior is as follows:

[1982] 1. The user types, "I'm so happy today!"

[1983] 2. The device sends this text to the server.

[1984] 3. The server uses an emotion engine to recognize the emotion "joy" from this text.

[1985] 4. Based on this emotion of "joy," the server adjusts the training dataset for the generative AI model, giving priority to positive data.

[1986] 5. The server generates the learning results and provides the user with "text generation results that reflect positive emotions" via the terminal.

[1987] Prompt Sentence Examples

[1988] Below are some example prompts to input to a generative AI model:

[1989] If a user types "I'm so happy today!", how would the emotion engine respond and adjust the generative AI's learning?

[1990] In this way, the system effectively learns from the data using generative AI while taking into account the user's emotions.

[1991] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1992] Step 1:

[1993] Data collection and permission acquisition

[1994] A user provides the system with data for which they have received permission to use from an individual or corporation. The terminal sends the data provided by the user to the server. The server checks the permission to use the received data and stores it in a database. Specifically, a database management system (e.g., MySQL or PostgreSQL) is used. The input is the data provided by the user, and the output is the data stored in the database.

[1995] Step 2:

[1996] Data Preprocessing

[1997] The server preprocesses the collected data. Specifically, it cleans the data (filling in missing values ​​and removing outliers) and converts the format (tokenizing text data, etc.). The server cleans the data using Python's pandas library and converts numerical data using NumPy. The input is raw data obtained from the database, and the output is preprocessed data.

[1998] Step 3:

[1999] Learning generative AI models

[2000] The server supplies the preprocessed data to a generative AI model (Generative Pretrained Transformer). The server uses an NVIDIA GPU (e.g., Tesla V100) to train the generative AI model using TensorFlow or PyTorch. The server monitors the model's performance during training and adjusts hyperparameters as necessary. The input is the preprocessed data, and the output is a trained generative AI model.

[2001] Step 4:

[2002] Use of emotion engine

[2003] When a user inputs text or voice, the device sends the input to the server. The server uses an emotion engine to recognize the user's emotions. Specifically, it uses the NLP library Hugging Face Transformers to extract emotions from text data. The server then adjusts the learning of the generative AI based on the recognized emotions. For example, if a user inputs text that expresses the emotion of joy, it prioritizes learning positive data. The input is text or voice data from the user, and the output is the emotion recognition results and adjusted learning data.

[2004] Step 5:

[2005] Output of training results

[2006] The server outputs the learning results of the generative AI and provides them to the user via a terminal. The server generates the learning results in text or graph format and displays them in a way that is easy for the user to understand. Specifically, graphs are created using Matplotlib or Seaborn. The input is the trained generative AI model, and the output is the text generation results and graphs provided to the user.

[2007] (Application example 1)

[2008] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[2009] Conventional AI systems lack the ability to recognize user emotions and adjust learning based on those emotions, making it difficult to recommend products that meet individual user needs and emotions. Furthermore, they lack the ability to integrate real-time emotion analysis and product recommendations to improve the shopping experience in brick-and-mortar stores.

[2010] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[2011] In this invention, the server includes an artificial intelligence system for learning data licensed from individuals or corporations, means for managing the input and output of the data, means including an environment for providing the data, means including an emotion engine that recognizes a user's emotions and adjusts learning based on the emotions, means for recommending products based on the user's emotions and past purchase history, and means for analyzing the user's emotions when viewing or touching products in real time and providing appropriate product information and coupons. This enables personalized product recommendations based on the user's emotions and an improved real-time shopping experience in physical stores.

[2012] "Data licensed for use by an individual or corporation" means data owned by a specific individual or corporation, and for which formal permission has been given to use the data.

[2013] An "artificial intelligence system" is a collection of computer programs or hardware designed to learn from data and perform specific tasks automatically.

[2014] "Means for managing data input and output" refers to functions or devices for controlling the receipt, processing, and output of results of data.

[2015] A "data serving environment" is the physical or virtual infrastructure in which data is properly collected, stored, and accessed.

[2016] An "emotion engine" is a software or hardware component that recognizes a user's emotions and adjusts the system's behavior based on those emotions.

[2017] "Means for recommending products based on a user's emotions and past purchase history" refers to an algorithm or system that analyzes a user's emotional state and past purchase data and suggests appropriate products based on that.

[2018] "Means for analyzing in real time and providing appropriate product information and coupons" refers to a function or device that instantly evaluates a user's current behavior and emotions and instantly provides appropriate product information and discount coupons.

[2019] "Generative AI (Generative Pretrained Transformer)" is an artificial intelligence model that has the ability to learn large amounts of data in advance and generate new data.

[2020] As an embodiment of the present invention, a "smart shopping assistant" system that improves the shopping experience in brick-and-mortar stores will be described as an example.

[2021] System Configuration

[2022] The system consists of the following main components:

[2023] 1. Server:

[2024] This includes artificial intelligence systems for learning from data licensed for use by individuals or entities.

[2025] It uses generative AI (Generative Pretrained Transformer) to learn from data and recommend products based on the user's emotions and past purchase history.

[2026] It includes an emotion engine that recognizes the user's emotions and adjusts learning based on those emotions.

[2027] 2. Device (smartphone):

[2028] It uses a camera and microphone to capture the user's facial expressions and voice and recognize emotions.

[2029] It analyzes users' emotions in real time when they look at or touch a product, and provides appropriate product information and coupons.

[2030] Program processing explanation

[2031] The server processes data using the following hardware and software:

[2032] Hardware: Smartphone camera and microphone

[2033] Software: OpenCV (image processing), Transformers (generative AI model), EmotionRecognizer (emotion recognition)

[2034] Data processing and calculation

[2035] 1. Emotion recognition:

[2036] The device's camera and microphone are used to capture the user's facial expressions and voice, and emotions are recognized using EmotionRecognizer.

[2037] 2. Product recommendation:

[2038] Based on the recognized emotions and the user's past purchase history data, a generative AI model (GPT-3) is used to recommend products.

[2039] 3. Real-time feedback:

[2040] It analyzes users' emotions in real time when they look at or touch a product, and provides appropriate product information and coupons.

[2041] Specific examples

[2042] For example, when a user uses a smartphone in a physical store, the camera captures the user's facial expressions and performs emotion recognition. If the user is smiling, the system recognizes the emotion of "joy" and recommends "new fashion items" based on past product purchase data.

[2043] Prompt Sentence Examples

[2044] The user's emotion is joy. Their past purchase history is "fashion items, accessories." Please recommend appropriate products.

[2045] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[2046] Step 1:

[2047] It uses the device's camera and microphone to capture the user's facial expressions and voice.

[2048] Input: User's facial expression images and voice data

[2049] Output: Captured facial expression images and audio data

[2050] How it work...

Claims

1. A means for obtaining data that is authorized for use by a user; and A means for providing the data with the user's permission to use to a generating AI model to generate a personalized advertisement customized for the user; means for outputting the generated personalized advertisement to the user; A system including:

2. The system of claim 1 , wherein the user's authorized data includes at least one of the user's browsing history, purchasing history, and social media activity.

3. The system according to claim 1 or 2, further comprising: a sentiment engine for estimating sentiment of the user to output the personalized advertisement.

Images