system

Abstract

We provide the system. [Solution] A means for collecting data on the user's hobbies and interests, A method for analyzing collected data to extract users' hobbies and thought patterns, A means of customizing a generative artificial intelligence based on extracted hobby and thought patterns, A means of delivering customized generative artificial intelligence to the user's device, A means of collecting user interaction data and continuously training a generative artificial intelligence, A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a persona chatbot control method performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] Conventional generation-based artificial intelligence systems can only provide general responses and have difficulty providing personalized experiences tailored to individual user hobbies and interests, resulting in problems such as decreased user satisfaction and usage frequency. There is also a problem that it is difficult to respond when a user's interests and hobbies change over time.

Means for Solving the Problems

[0005] This invention includes means for collecting and analyzing data related to a user's hobbies and interests to extract the user's hobby and thought patterns. This allows for the customization of generative artificial intelligence to suit the individual characteristics of each user. The invention also includes means for delivering the customized generative artificial intelligence to the user's terminal and continuously collecting conversational data with the user to train the generative artificial intelligence. This makes it possible to continuously adapt to changes in the user's interests and hobbies.

[0006] "User" refers to an individual or legal entity that uses the system.

[0007] "Data related to hobbies and interests" refers to information related to the user's preferences and interests, and includes survey results, browsing history, and past interaction data.

[0008] "Analysis" refers to the act of processing collected data to extract meaningful results, and in this case, machine learning algorithms are used.

[0009] "Hobbies and thought patterns" refer to tendencies and characteristics based on a user's hobbies and interests.

[0010] "Generative artificial intelligence" refers to artificial intelligence that uses natural language processing technology to generate appropriate responses to user input.

[0011] "Customization" refers to the act of adjusting or changing the characteristics and response content of a generative artificial intelligence system to suit the user's characteristics.

[0012] "Terminal" refers to electronic devices used by users to access generative artificial intelligence, such as smartphones and computers.

[0013] "Dialogue data" refers to records of conversations and interactions that take place between users and generative artificial intelligence.

[0014] "Continuous learning" refers to the process of improving the performance and response accuracy of generative artificial intelligence based on new data. [Brief explanation of the drawing]

[0015] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] This is a sequence diagram showing the processing flow of the data processing system in Example 2, which incorporates an emotion engine. [Figure 14]It is a sequence diagram showing the processing flow of a data processing system in Application Example 2 when a sentiment engine is combined.

Embodiments for Carrying Out the Invention

[0016] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0017] First, the terms used in the following description will be explained.

[0018] In the following embodiments, a numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units 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), and the like.

[0019] In the following embodiments, a numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0020] In the following embodiments, a numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0021] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0022] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0023] [First Embodiment]

[0024] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0025] As shown in Figure 1, the 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.

[0026] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0027] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0028] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0029] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0030] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0031] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0033] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0034] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0035] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0036] System Overview

[0037] This invention is a system for providing a generative artificial intelligence customized specifically to the user's hobbies and interests. This system involves a server and a user's terminal working together to optimize the generative artificial intelligence based on the user's interests and hobbies.

[0038] Program Processing Description

[0039] 1. Collection of user data

[0040] User: Create a new account and enter initial data such as profile information and hobbies. This includes answering questionnaires and selecting genres of interest (music, movies, sports, etc.).

[0041] Terminal: Sends user-entered information and survey responses to the server.

[0042] Server: Receives user data and stores it in a database for later analysis.

[0043] 2. Analysis of hobbies and thought patterns

[0044] Server: Preprocesses the collected data and filters out noise and incomplete data.

[0045] Server: Based on pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used to classify user trends into multiple categories.

[0046] Server: Saves analysis results to the database.

[0047] 3. Customization of Generative Artificial Intelligence

[0048] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0049] Server: Trains the selected template with user-specific data to create a user-specific generative artificial intelligence. This customization process includes prioritizing the training of data related to specific areas of interest.

[0050] Server: Stores customized generative artificial intelligence models in a database.

[0051] 4. Provision to users

[0052] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0053] Terminal: Notifies the user that generative artificial intelligence is available.

[0054] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0055] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0056] 5. Continuous learning and improvement

[0057] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[0058] Terminal: Sends collected conversation data to the server.

[0059] Server: Analyzes collected data to detect changes in the user's hobbies and interests.

[0060] Server: Updates generative artificial intelligence models based on new data.

[0061] Server: Redistributes the updated model to user terminals.

[0062] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[0063] Specific example

[0064] Initial setup and usage

[0065] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[0066] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[0067] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[0068] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[0069] Continuous learning and improvement

[0070] User: As I continue to interact with the generative AI, I begin to develop an interest in "jazz." I start asking more questions about jazz.

[0071] Terminal: Sends new dialogue data to the server.

[0072] Server: Analyzes collected data to detect new interests (jazz) of user A. Updates the generative artificial intelligence model based on this information.

[0073] Device: Re-download the updated AI model so that User A can engage in conversations tailored to their new interests.

[0074] This system allows users to utilize generative artificial intelligence optimized for their hobbies and interests, and because it is continuously updated, they can always receive services that cater to their latest interests.

[0075] The following describes the processing flow.

[0076] Step 1:

[0077] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[0078] Step 2:

[0079] Terminal: Sends user-entered profile information and survey responses to the server.

[0080] Step 3:

[0081] Server: Receives user information and stores it in a database for later analysis.

[0082] Step 4:

[0083] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[0084] Step 5:

[0085] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[0086] Step 6:

[0087] Server: Saves analysis results to the database.

[0088] Step 7:

[0089] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0090] Step 8:

[0091] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[0092] Step 9:

[0093] Server: Stores customized generative artificial intelligence models in a database.

[0094] Step 10:

[0095] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0096] Step 11:

[0097] Terminal: Notifies the user that generative artificial intelligence is available.

[0098] Step 12:

[0099] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0100] Step 13:

[0101] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0102] Step 14:

[0103] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[0104] Step 15:

[0105] Terminal: Sends collected conversation data to the server.

[0106] Step 16:

[0107] Server: Analyzes collected dialogue data to detect changes in the user's hobbies and interests.

[0108] Step 17:

[0109] Server: Updates generative artificial intelligence models based on new data.

[0110] Step 18:

[0111] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[0112] Step 19:

[0113] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[0114] (Example 1)

[0115] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0116] Currently, many generative artificial intelligence models are specialized in providing general information, but they have the challenge of not being able to provide services tailored to a user's individual interests and thoughts. Furthermore, they struggle to quickly adapt to changes in user interests and hobbies. In addition, there is a lack of effective methods for providing the most suitable generative artificial intelligence model for individual users.

[0117] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0118] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for filtering the collected data to remove noise and incomplete data, means for analyzing the filtered data to extract the user's hobbies and thought patterns, means for selecting a generative artificial intelligence model based on the extracted hobbies and thought patterns, means for training and customizing the selected generative artificial intelligence model with user-specific data, means for delivering the customized generative artificial intelligence model to the user's terminal, and means for collecting interaction data with the user to continuously train and update the generative artificial intelligence model. This makes it possible to provide a generative artificial intelligence model tailored to the user's hobbies and thoughts, and to respond quickly to changes in the user's interests.

[0119] A "user" refers to an individual or legal entity that utilizes the system, provides data related to their hobbies and interests, and receives services based on a customized generative artificial intelligence model.

[0120] A "server" refers to a device or system that receives and stores data sent by users, and performs data analysis, customization, distribution, and updating of generative artificial intelligence models.

[0121] A "terminal" refers to an electronic device used by a user, such as a computer, smartphone, or tablet, that communicates with a server to transmit data, download generative artificial intelligence models, and engage in dialogue.

[0122] "Data collection" refers to the process of obtaining information about users' hobbies and interests through surveys and other means.

[0123] "Filtering" refers to the process of removing noise and incomplete data from collected data, preparing it for analysis.

[0124] "Data analysis" refers to the process of analyzing pre-processed data using machine learning algorithms and other methods to extract users' hobbies and thought patterns.

[0125] A "generative artificial intelligence model" refers to an artificial intelligence template designed to interact with users and provide information using natural language processing technology.

[0126] "Customization" refers to the process of training a generative artificial intelligence model with user-specific data to optimize it for individual users.

[0127] "Distribution" refers to the process of transferring a customized generative artificial intelligence model from a server to a user's device.

[0128] "Dialogue data" refers to data that records user questions and the responses of generative artificial intelligence models, including the content of the dialogue with the model.

[0129] "Continuous learning" refers to the process of improving the performance of a generative artificial intelligence model by retraining and updating it based on user interaction data.

[0130] This invention relates to a system that provides a generative artificial intelligence model specialized according to the user's hobbies and interests. This system can provide a service optimized to the user's interests and hobbies by analyzing data provided by the user, customizing the generative artificial intelligence based on that analysis, and continuously learning and updating it.

[0131] Specifically, this system includes the following elements:

[0132] User data collection

[0133] User: Create a new account and enter basic information such as name, age, gender, and place of residence, as well as answer a questionnaire about hobbies and interests. For example, to the question "What genre of music do you like?", answer "Rock".

[0134] Terminal: This terminal sends basic information and survey responses entered by the user to the server. This communication uses the secure protocol HTTPS.

[0135] Server: Receives user data and stores it in a database management system (e.g., MySQL®) for later analysis.

[0136] Analysis of hobbies and thought patterns

[0137] Server: Preprocesses the collected data, filtering out noise and incomplete data. This preprocessing is done using data cleaning tools (e.g., OpenRefine).

[0138] Server: Based on filtered data, it applies machine learning algorithms (e.g., K-means clustering) to extract users' hobbies and thought patterns.

[0139] Server: Saves analysis results to the database.

[0140] Customization of generative artificial intelligence

[0141] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates (e.g., GPT-3®).

[0142] Server: Trains and customizes the selected generative AI template with user-specific data. For example, if the user is interested in "rock" music, a dataset in that genre is prepared and the template is fine-tuned.

[0143] Server: Stores customized generative artificial intelligence models in a database.

[0144] Provision to users

[0145] Server: Prepares the server to deliver customized generative artificial intelligence models to user terminals. This preparation includes packaging the data to be delivered.

[0146] Terminal: Notify the user that generative artificial intelligence is available. The notification will use a push notification service (e.g., Firebase Cloud Messaging).

[0147] Terminal: Downloads generative artificial intelligence models from the server. HTTPS is used as the data transfer protocol.

[0148] User: Using the downloaded generative AI, initiate conversations and information searches that match the user's interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[0149] Continuous learning and improvement

[0150] Terminal: Every time a user interacts with the generative artificial intelligence, the interaction data is collected. For example, a question like "What are some recommended new action movies?" and its answer are stored as data.

[0151] Terminal: Sends collected conversation data to the server. The HTTPS protocol is used for transmission.

[0152] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[0153] Server: Updates the generative artificial intelligence model based on new data. Prepares a new dataset and performs fine-tuning again.

[0154] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[0155] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[0156] Example of a prompt

[0157] "What rock bands do you recommend these days?"

[0158] "What are some of the latest action movies you would recommend?"

[0159] "Please tell me about some famous jazz songs."

[0160] As described above, the present invention provides a generative artificial intelligence model that is tailored to the user's hobbies and thoughts, and can continuously adapt to changes in the user's interests.

[0161] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0162] Step 1: Collecting User Data

[0163] User: Create a new account and enter basic information such as name, age, gender, and place of residence. Also, answer a questionnaire about your interests and hobbies. For example, answer "Rock" to the question, "What is your favorite music genre?"

[0164] Input: Basic information, survey responses

[0165] Output: Basic information and survey response dataset

[0166] Terminal: The terminal sends the basic information and survey responses entered by the user to the server via the HTTPS protocol. Specifically, when the submit button on the input form is pressed, the data is transferred to the server.

[0167] Input: User input data

[0168] Output: Data transfer to the server

[0169] Server: Receives user data and stores it in a database management system (e.g., MySQL). User information and survey results are stored in the database in an organized manner.

[0170] Input: Transferred user data

[0171] Output: User data stored in the database

[0172] Step 2: Analysis of hobbies and thought patterns

[0173] Server: Preprocesses collected data, filtering out noise and incomplete data. Uses data cleaning tools (e.g., OpenRefine) to remove unnecessary data and prepare it for analysis.

[0174] Input: Raw data

[0175] Output: Filtered, clean data

[0176] Server: Applies machine learning algorithms (e.g., K-means clustering) to preprocessed data to extract users' hobbies and thought patterns. This algorithm classifies users' hobbies into multiple categories.

[0177] Input: Clean data

[0178] Output: User's hobbies and thought patterns

[0179] Server: Saves analysis results to a database. Analysis results are categorized by user and used for customization at a later stage.

[0180] Input: Analysis results

[0181] Output: Analysis results saved in the database

[0182] Step 3: Customizing Generative Artificial Intelligence

[0183] Server: Based on the analysis results, it selects the most suitable template from generative artificial intelligence templates (e.g., GPT-3). It uses a template selection algorithm to choose the most appropriate template.

[0184] Input: Analysis results

[0185] Output: Selected generative AI templates

[0186] Server: Trains the selected generative AI template with user-specific data. Specifically, it prepares a dataset related to the user's areas of interest and uses that data to fine-tune the template.

[0187] Input: Generative artificial intelligence template and user data

[0188] Output: Customized generative artificial intelligence model

[0189] Server: Stores customized generative artificial intelligence models in a database. Models are stored with version control using a model management system (e.g., MLflow).

[0190] Input: Customized generative artificial intelligence model

[0191] Output: Generative artificial intelligence models stored in the database

[0192] Step 4: Provision to users

[0193] Server: Prepares to deliver customized generative artificial intelligence models to user terminals. Specifically, it packages the delivery data and configures the delivery server.

[0194] Input: Customized generative artificial intelligence model

[0195] Output: Ready for delivery

[0196] Terminal: Notifies the user that generative artificial intelligence is available. This notification uses a push notification service (e.g., Firebase Cloud Messaging).

[0197] Input: Notification content

[0198] Output: Notification to the user

[0199] Terminal: Downloads generative AI models from the server. Secure HTTPS is used as the data transfer protocol.

[0200] Input: Download link

[0201] Output: Downloaded generative artificial intelligence model

[0202] User: Using the downloaded generative AI, initiate conversations and information searches that match their interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[0203] Input: User's question (prompt)

[0204] Output: Answer from a generative artificial intelligence

[0205] Step 5: Continuous learning and improvement

[0206] Terminal: Each time a user interacts with a generative artificial intelligence, the dialogue data is collected. This dialogue data includes the content of the questions and answers.

[0207] Input: Dialogue between the user and the generative artificial intelligence.

[0208] Output: Collected dialogue data

[0209] Terminal: Sends collected conversation data to the server. Secure HTTPS protocol is used for transmission.

[0210] Input: Dialogue data

[0211] Output: Sending data to the server

[0212] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[0213] Input: Dialogue data

[0214] Output: Analysis results

[0215] Server: Updates the generative artificial intelligence model based on new data. The update includes a process of retraining the model based on newly collected data.

[0216] Input: New dataset

[0217] Output: Updated generative artificial intelligence model

[0218] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[0219] Input: Updated generative artificial intelligence model

[0220] Output: Ready for delivery

[0221] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[0222] Input: New question from the user (prompt)

[0223] Output: A new answer from generative artificial intelligence

[0224] (Application Example 1)

[0225] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0226] Traditional advertising delivery systems have not adequately personalized ads using data on users' hobbies and interests, making it difficult to improve the effectiveness and accuracy of advertisements. Furthermore, providing users with ads that resonate with them requires continuous data collection and updates, but efficient methods for doing so have been lacking. Therefore, there is a need for a system that generates and continuously updates ads optimized for users' hobbies and interests.

[0227] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0228] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for collecting conversational data with the user to continuously train the generative artificial intelligence, means for delivering the generated customized advertisements, and means for collecting and analyzing user feedback to update the advertising model. This makes it possible to generate advertisements optimized for the user's hobbies and interests and to continuously improve them.

[0229] "Data related to users' hobbies and interests" refers to information collected to clarify the hobbies and interests that users exhibit, and specifically includes genres, preferences, and past behavioral history.

[0230] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of processing collected user data and using analytical techniques and algorithms to identify users' interests and patterns.

[0231] "Methods for customizing generative artificial intelligence" refers to technologies that optimize generative artificial intelligence for each user based on their hobbies and thought patterns, thereby creating personalized generative models.

[0232] "Means for delivering generated customized advertisements" refers to the technologies and methods for sending and displaying customized advertisements on a user's device.

[0233] "Methods for collecting and analyzing user feedback to update advertising models" refers to the process of collecting user reactions and evaluations to advertisements, analyzing that data, and improving the advertising model to match the latest user tastes and interests.

[0234] "Means for collecting dialogue data and continuously training generative artificial intelligence" refers to technologies and methods for collecting data from user interactions with generative artificial intelligence and using that data to improve and update the generative model.

[0235] "Means for customizing generative artificial intelligence based on analysis results" refers to technologies and methods for optimizing the operation of generative artificial intelligence based on analyzed hobbies and thought patterns.

[0236] This invention is a system that generates and delivers advertisements customized based on the user's hobbies and interests. This system involves a server and the user's terminal working together to handle everything from collecting and analyzing user data to generating and delivering advertisements, as well as continuous learning and improvement. The embodiments for carrying out this invention will be described in detail below.

[0237] The server first collects data on hobbies and interests provided by the user. This data is obtained when the user logs into the application and enters information based on their profile and questionnaires. The server then analyzes this collected data. The analysis involves data preprocessing (such as imputing missing values ​​and normalizing the data) and clustering techniques to extract the user's hobbies and thought patterns.

[0238] Furthermore, the server customizes the generative artificial intelligence (AI) based on the extracted hobbies and thought patterns. Specifically, it trains the generative AI model to match the user's preferences and generates personalized advertisements. The generative AI used in this generation process includes TENSORFLOW® and PyTorch.

[0239] The device delivers customized, generated ads to the user. The generated ads are displayed to the user through a smartphone application. After the user views, clicks, or takes other action on an ad, feedback is sent back to the server. By analyzing this feedback data, the server updates its generative AI model to provide more accurate ads.

[0240] For example, if a user enters the prompt "Tell me about the latest recommended smartphone accessories," the generative artificial intelligence will generate and display advertisements suggesting the most suitable smartphone accessories based on the user's past interests and behavioral patterns. Through this continuous feedback loop, the system can always provide advertisements that are tailored to the user's latest hobbies and interests.

[0241] This system configuration allows users to receive advertisements that perfectly match their interests and preferences at any given time, and also enables advertisers to expect high results.

[0242] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0243] Step 1:

[0244] Users log in to the application and enter initial data, including profile information and data about their hobbies and interests from a questionnaire. This initial data includes the user's gender, age, areas of interest, and preferred product categories. The device then sends this data to the server.

[0245] Input: User profile information, survey data

[0246] Output: Initial data sent to the server

[0247] Step 2:

[0248] The server stores the initial data received from the terminal in a database. Based on this stored data, it performs data preprocessing. Preprocessing includes imputing missing values ​​and normalizing the data.

[0249] Input: Initial data

[0250] Output: Preprocessed data

[0251] Step 3:

[0252] The server analyzes pre-processed data and applies clustering techniques to extract users' interests and thought patterns. Machine learning algorithms (such as KMeans) are used for clustering.

[0253] Input: Preprocessed data

[0254] Output: User clustering results

[0255] Step 4:

[0256] The server customizes generative artificial intelligence (AI) based on the analysis results. Specifically, it trains a generative AI model for a cluster of users and generates personalized ad templates. This process utilizes TensorFlow and PyTorch.

[0257] Input: User clustering results

[0258] Output: Customized generative AI model

[0259] Step 5:

[0260] The device receives a customized generative AI model from the server and displays advertisements generated using that model to the user. When the user takes action on an advertisement (click, purchase, etc.), the device collects the feedback and sends it to the server.

[0261] Input: Customized Generative AI Model

[0262] Output: Displayed ads, user feedback

[0263] Step 6:

[0264] The server analyzes the received feedback data. Based on this analysis, it updates the generated AI model and continuously improves the advertising model.

[0265] Input: User feedback data

[0266] Output: Updated Generative AI Model

[0267] This enables ad delivery using the latest generative AI models that are always adapted to the user's hobbies and interests.

[0268] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0269] System Overview

[0270] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[0271] Program Processing Description

[0272] 1. Collection of user data

[0273] User: Create a new account and enter profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about hobbies and interests.

[0274] Terminal: Send the profile information and questionnaire answers entered by the user to the server.

[0275] Server: Save the received user information in the database for later analysis.

[0276] 2. Analysis of Hobbies and Thinking Patterns

[0277] Server: Preprocess the user information stored in the database and filter out noise and incomplete data.

[0278] Server: Apply a machine learning algorithm using the preprocessed data to extract the user's hobbies and thinking patterns. In this process, use a clustering method to classify the user's tendencies into multiple categories.

[0279] Server: Save the analysis results in the database. [[ID=2б]]

[0280] 3. Customization of Generative AI

[0281] Server: Based on the analysis results, select the most suitable one from the generative AI templates.

[0282] Server: Let the selected template learn user-specific data to create a user-specific generative AI. At this time, preferentially learn data related to specific interest fields.

[0283] Server: Save the customized generative AI model in the database.

[0284] 4. Provision to Users

[0285] Server: Prepare to distribute a customized generative AI model to the user terminal.

[0286] Terminal: Notify the user that the generative AI is available.

[0287] Terminal: The user terminal downloads the generative AI model from the server.

[0288] User: Use the downloaded generative AI to start conversations and information searches that match their hobbies and thoughts.

[0289] 5. Emotion Recognition and Response Optimization

[0290] Terminal: During the conversation between the user and the generative AI, use an emotion engine to recognize the user's emotions. This emotion data is analyzed from voice tones, facial expressions, text content, etc.

[0291] Terminal: Send the collected emotion data to the server.

[0292] Server: Analyze the emotion data and optimize the response of the generative AI based on it. For example, if the user is feeling stressed, the generative AI generates a more calming and reassuring response.

[0293] 6. Continuous Learning and Improvement

[0294] Terminal: Every time the user interacts with the generative AI, collect the conversation data and emotion data.

[0295] Terminal: Send the collected data to the server.

[0296] Server: Analyze the collected conversation data and emotion data to detect changes in the user's hobbies, interests, and emotions.

[0297] Server: Update the generative AI model based on the new data.

[0298] Server: Redistribute the updated generative AI model to the user terminals.

[0299] User: Use the updated generative AI to conduct conversations and information searches corresponding to new interests, hobbies, and emotions.

[0300] Specific Example

[0301] Initial Setup and Usage

[0302] User: When creating an account, answer that they are particularly interested in "music" and "movies". Among music genres, they prefer "rock", and among movie genres, they select "action".

[0303] Terminal: Send this information to the server, and the server creates a profile for User A based on this.

[0304] Server: Analyze the collected data and extract User A's hobbies and thinking patterns. Based on this, customize the generative AI and create an AI model dedicated to User A.

[0305] Terminal: Download the customized AI model, and User A starts a conversation. For example, when User A asks "What are the recent recommended rock bands?", the generative AI proposes appropriate rock bands.

[0306] Emotion Recognition and Response Optimization

[0307] User: During the conversation with the generative AI, it is determined by the emotion engine that they are feeling "stress". For example, when User A says "Today has been really tough", the emotion engine recognizes stress from the voice tone and text content.

[0308] Terminal: Send this emotion data to the server.

[0309] Server: Analyzes emotional data and adjusts the generative AI's responses to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[0310] Continuous learning and improvement

[0311] User: As they continue to interact with the generative AI, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[0312] Terminal: Sends new dialogue data and emotion data to the server.

[0313] Server: Analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[0314] Terminal: Re-download the updated generative artificial intelligence model, enabling user A to engage in conversations that respond to new interests and emotions.

[0315] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[0316] The following describes the processing flow.

[0317] Step 1:

[0318] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[0319] Step 2:

[0320] Terminal: Sends user-entered profile information and survey responses to the server.

[0321] Step 3:

[0322] Server: Receives user information and stores it in a database for later analysis.

[0323] Step 4:

[0324] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[0325] Step 5:

[0326] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[0327] Step 6:

[0328] Server: Saves analysis results to the database.

[0329] Step 7:

[0330] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0331] Step 8:

[0332] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[0333] Step 9:

[0334] Server: Stores customized generative artificial intelligence models in a database.

[0335] Step 10:

[0336] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0337] Step 11:

[0338] Terminal: Notifies the user that generative artificial intelligence is available.

[0339] Step 12:

[0340] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0341] Step 13:

[0342] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0343] Step 14:

[0344] User: As you interact with the generative artificial intelligence, the emotion engine analyzes your emotions in real time. This emotion data is obtained from voice tone, facial expressions, and text content.

[0345] Step 15:

[0346] Terminal: The emotion engine analyzes the emotion data and sends it to the server.

[0347] Step 16:

[0348] Server: Analyzes emotional data and optimizes the responses of the generative artificial intelligence. For example, if the system recognizes that the user is feeling "stressed," the generative AI will generate a calmer, more comforting response.

[0349] Step 17:

[0350] User: Receive appropriate responses through interaction with generative artificial intelligence.

[0351] Step 18:

[0352] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[0353] Step 19:

[0354] Terminal: Sends collected dialogue data and emotion data to the server.

[0355] Step 20:

[0356] Server: Analyzes collected data to detect changes in users' hobbies, interests, and emotions.

[0357] Step 21:

[0358] Server: Updates generative artificial intelligence models based on new data.

[0359] Step 22:

[0360] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[0361] Step 23:

[0362] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[0363] (Example 2)

[0364] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0365] Conventional generative artificial intelligence systems struggled not only to be customized based on the user's hobbies and thoughts, but also to optimize their responses in response to changes in the user's emotions and interests. Furthermore, the lack of sufficient continuous learning and improvement for individual users resulted in a limited user experience. In addition, there was a lack of efficient means to collect initial user data and advanced methods for analyzing the collected data.

[0366] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0367] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, and means for collecting dialogue data and emotional data with the user to continuously train the generative artificial intelligence and optimize its responses. This makes it possible to provide optimal responses in accordance with changes in the user's hobbies, thoughts, and emotions, and to continuously learn and improve.

[0368] "Data on users' hobbies and interests" refers to information that users provide through their profile information and survey responses, indicating their interests and preferences in specific fields such as music, movies, literature, and sports.

[0369] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of generating specific patterns or clusters based on users' hobbies and interests using data cleansing through preprocessing and machine learning algorithms.

[0370] "Means for customizing generative artificial intelligence based on extracted hobby and thought patterns" refers to methods and techniques for optimizing a base generative artificial intelligence model to the hobbies and thoughts of individual users by utilizing the analysis results.

[0371] "Means for delivering customized generative artificial intelligence to a user's terminal" refers to a technology that transfers a generative artificial intelligence model from a server to a user's terminal, making that model available for use in the user's local environment.

[0372] "A means of collecting user interaction data and emotional data to continuously train generative artificial intelligence and optimize responses" refers to a technology that analyzes emotional data obtained from text, voice, facial expressions, etc., collected by generative artificial intelligence during interactions with users, and adjusts and improves the AI's responses to match the user's emotions and needs.

[0373] "Survey methods" refer to question-based interfaces and their implementation methods for efficiently collecting initial user data.

[0374] A "machine learning algorithm" is a mathematical model and method that automatically learns from large amounts of data to perform predictions, classifications, and clustering.

[0375] Modes for carrying out the invention

[0376] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[0377] Hardware and software

[0378] Hardware:

[0379] The user's smartphone or PC

[0380] Cloud Server

[0381] software:

[0382] Database management systems: MySQL, PostgreSQL

[0383] Machine learning libraries: TensorFlow, PyTorch

[0384] Emotion recognition software: Microsoft® Azure® Emotion API

[0385] Generative artificial intelligence: GPT-3, BERT

[0386] Program Processing Description

[0387] 1. Collection of user data:

[0388] Users create a new account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[0389] The device sends the user's entered profile information and survey responses to the server.

[0390] The server stores the received user information in a database for later analysis.

[0391] 2. Analysis of hobbies and thought patterns:

[0392] The server preprocesses the user information stored in the database, filtering out noise and incomplete data.

[0393] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[0394] The server saves the analysis results to the database.

[0395] 3. Customizing Generative Artificial Intelligence:

[0396] The server selects the most suitable template from the generative artificial intelligence templates based on the analysis results.

[0397] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[0398] The server stores customized generative artificial intelligence models in a database.

[0399] 4. Provision to users:

[0400] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[0401] The device notifies the user that generative artificial intelligence is available.

[0402] The terminal downloads the generative artificial intelligence model from the server.

[0403] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts.

[0404] 5. Emotion Recognition and Response Optimization:

[0405] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[0406] The device sends the collected emotional data to the server.

[0407] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[0408] 6. Continuous learning and improvement:

[0409] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[0410] The device sends the collected data to the server.

[0411] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[0412] The server updates the generative artificial intelligence model based on new data.

[0413] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[0414] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[0415] Specific example

[0416] 1. Initial setup and usage:

[0417] When creating an account, users indicate that they are particularly interested in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[0418] The terminal sends this information to the server, which then uses it to create a profile for user A.

[0419] The server analyzes the collected data and extracts user A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for user A.

[0420] The device downloads a customized AI model, and user A initiates a conversation. For example, if user A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[0421] 2. Emotion recognition and response optimization:

[0422] When a user interacts with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if user A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[0423] The device sends this emotion data to the server.

[0424] The server analyzes emotional data and adjusts the generative artificial intelligence's response to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[0425] 3. Continuous learning and improvement:

[0426] As the user continues to interact with the generative artificial intelligence, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[0427] The device sends new conversational and emotional data to the server.

[0428] The server analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[0429] The device will download the updated generative artificial intelligence model again, enabling user A to engage in conversations that respond to new interests and emotions.

[0430] Example of a prompt

[0431] "Could you recommend some rock bands you've been listening to lately?"

[0432] "Today was really tough."

[0433] "I've recently become interested in jazz. Do you have any recommendations?"

[0434] This allows users to utilize generative artificial intelligence optimized for their own hobbies and thoughts, and furthermore, an emotion engine enables them to receive appropriate responses based on their emotional state. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[0435] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0436] Program processing flow

[0437] User data collection

[0438] Step 1:

[0439] The user launches the application and creates a new account. They enter profile information such as their name, email address, gender, and age, and answer a questionnaire about their hobbies and interests.

[0440] Input: Name, email address, gender, age, and survey responses regarding hobbies and interests.

[0441] Output: Entered profile information and survey data.

[0442] Step 2:

[0443] The device sends the profile information and survey responses entered by the user to the server.

[0444] Input: Profile information and survey responses provided by the user.

[0445] Output: User data sent to the server.

[0446] Step 3:

[0447] The server saves the received user information to the database. It then verifies that the data was saved successfully.

[0448] Input: User data sent from the device.

[0449] Output: User data stored in the database.

[0450] Analysis of hobbies and thought patterns

[0451] Step 4:

[0452] The server preprocesses user information stored in the database, filtering out noise and incomplete data. For example, it performs data imputation and data normalization.

[0453] Input: User data stored in the database.

[0454] Output: Pre-processed, clean data.

[0455] Step 5:

[0456] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. Clustering techniques are then used to classify user trends into multiple categories.

[0457] Input: Pre-processed, clean data.

[0458] Output: A cluster that shows the user's hobbies and thought patterns.

[0459] Step 6:

[0460] The server saves the analysis results to the database. It then verifies that the analysis results were saved correctly.

[0461] Input: Analysis results (user cluster information).

[0462] Output: Analysis results stored in the database.

[0463] Customization of generative artificial intelligence

[0464] Step 7:

[0465] The server selects the optimal generative artificial intelligence template based on the analysis results.

[0466] Input: Analysis results (user cluster information).

[0467] Output: Selected AI template.

[0468] Step 8:

[0469] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. It prioritizes learning data related to the user's specific areas of interest.

[0470] Input: Selected AI template, user-specific data.

[0471] Output: A customized generative AI model.

[0472] Step 9:

[0473] The server saves the customized generative artificial intelligence model to the database. It then verifies that the saving process was successful.

[0474] Input: A customized generative AI model.

[0475] Output: Customized AI model stored in the database.

[0476] Provision to users

[0477] Step 10:

[0478] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[0479] Input: A customized AI model stored in a database.

[0480] Output: Notification that delivery is ready.

[0481] Step 11:

[0482] The device notifies the user that it is ready for distribution and downloads the generative artificial intelligence model.

[0483] Input: Server notification of preparation for delivery.

[0484] Output: Availability notification to the user and downloaded generative AI model.

[0485] Step 12:

[0486] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts. For example, they might ask, "Can you recommend some recent rock bands?"

[0487] Input: Prompt message (user question).

[0488] Output: Response from a generative artificial intelligence.

[0489] Emotion recognition and response optimization

[0490] Step 13:

[0491] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. Emotional data is analyzed from voice tone, facial expressions, and text content.

[0492] Input: User's voice tone, facial expression, and text content.

[0493] Output: Analyzed sentiment data.

[0494] Step 14:

[0495] The device sends the collected emotional data to the server.

[0496] Input: Analyzed sentiment data.

[0497] Output: Sentiment data sent to the server.

[0498] Step 15:

[0499] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the response is changed to something calmer and more comforting.

[0500] Input: Analyzed sentiment data.

[0501] Output: Optimized AI response model.

[0502] Continuous learning and improvement

[0503] Step 16:

[0504] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[0505] Input: User interaction data, sentiment data.

[0506] Output: Collected dialogue data and sentiment data.

[0507] Step 17:

[0508] The device sends the collected data to the server.

[0509] Input: Collected dialogue data and sentiment data.

[0510] Output: Dialogue data and sentiment data sent to the server.

[0511] Step 18:

[0512] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[0513] Input: Collected dialogue data and sentiment data.

[0514] Output: Analyzed data (changes in user hobbies, interests, and emotions).

[0515] Step 19:

[0516] The server updates the generative artificial intelligence model based on new data.

[0517] Input: Newly analyzed data.

[0518] Output: Updated generative AI model.

[0519] Step 20:

[0520] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[0521] Input: An updated generative AI model.

[0522] Output: The AI ​​model redistributed to the user's terminal.

[0523] Step 21:

[0524] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[0525] Input: The user's new prompt message.

[0526] Output: Response from a generative artificial intelligence.

[0527] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[0528] (Application Example 2)

[0529] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0530] Conventional generative artificial intelligence systems often fail to personalize responses based on user preferences and interests, resulting in responses that do not always meet user expectations. Furthermore, their inability to recognize user emotions makes it difficult to provide appropriate responses and services. Moreover, in environments such as autonomous vehicles, there is a growing demand for personalized services that respond to user emotions.

[0531] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing the generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for recognizing the user's emotions, and means for optimizing the response of the generative artificial intelligence based on the recognized user emotions. This enables personalized services tailored to the user's hobbies and interests, as well as appropriate responses in response to their emotions.

[0532] "Data related to hobbies and interests" refers to the preferences and interests that users have in specific fields, and specifically includes information related to preferences such as music, movies, sports, and travel.

[0533] "Methods for analyzing data to extract users' hobbies and thought patterns" refers to algorithms and methods for analyzing collected data to reveal users' hobbies and tendencies in thinking.

[0534] "Means for customizing generative artificial intelligence" refers to methods and technologies for individualizing the artificial intelligence model generated based on the extracted user's hobbies and thought patterns.

[0535] "Means for delivering customized generative artificial intelligence to a user's device" refers to systems and technologies for sending individualized artificial intelligence models to a user's device in a format that the user can use.

[0536] "Means of recognizing user emotions" refers to technologies that analyze a user's voice, facial expressions, text, etc., to understand their emotional state.

[0537] "Means for optimizing the response of generative artificial intelligence based on recognized user emotions" refers to technologies for appropriately adjusting the content and methods of responses generated by artificial intelligence based on emotion recognition.

[0538] The embodiments for carrying out the present invention are described in detail below. First, a method for collecting data on a user's hobbies and interests and customizing an artificial intelligence system based on that data is described. Furthermore, a specific method for providing the user with the most appropriate response through emotion recognition is also described.

[0539] Program Overview

[0540] 1. Collection of user data

[0541] Users create an account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[0542] The device sends the user's entered profile information and survey responses to the server.

[0543] The server stores the received user information in a database for later analysis.

[0544] 2. Analysis of hobbies and thought patterns

[0545] The server preprocesses user information stored in the database, filtering out noise and incomplete data.

[0546] The server uses pre-processed data to apply machine learning algorithms and extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[0547] The server saves the analysis results to a database.

[0548] 3. Customization of Generative Artificial Intelligence

[0549] Based on the analysis results, the server selects the most suitable template from the generative artificial intelligence templates.

[0550] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[0551] The server stores customized generative artificial intelligence models in a database.

[0552] 4. Provision to users

[0553] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[0554] The device notifies the user that generative artificial intelligence is available.

[0555] The user's terminal downloads a generative artificial intelligence model from the server.

[0556] The user then uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and preferences.

[0557] 5. Emotion Recognition and Response Optimization

[0558] During interactions between the user and generative artificial intelligence, the device uses an emotion engine to recognize the user's emotions. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[0559] The device sends the collected emotional data to the server.

[0560] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[0561] 6. Continuous learning and improvement

[0562] The device collects dialogue data and emotional data each time the user interacts with the generative artificial intelligence.

[0563] The device sends the collected data to the server.

[0564] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[0565] The server updates the generative artificial intelligence model based on new data.

[0566] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[0567] Users utilize updated generative artificial intelligence to engage in conversations and information retrieval that cater to their new interests, hobbies, and emotions.

[0568] Hardware and software to be used

[0569] hardware

[0570] Infotainment system in autonomous vehicles

[0571] Camera and microphone: For emotion recognition inside the vehicle.

[0572] Vehicle communication module: Data communication with the server

[0573] software

[0574] Server-side: Apache (registered trademark), Kafka (data streaming), TensorFlow (machine learning model), PostgreSQL (database)

[0575] Autonomous vehicle terminal: Custom Emotion Recognition Engine, in-terminal application (compatible with Android® Auto and CarPlay®)

[0576] Specific example

[0577] For example, if a user in an autonomous vehicle asks, "What are some rock bands you recommend lately?", the generative artificial intelligence and emotion recognition engine will process a prompt sentence like the following.

[0578] Example of a prompt:

[0579] "We've identified that the user is currently experiencing stress. Their hobby is rock music. Please suggest some rock bands."

[0580] Based on this prompt, the generative artificial intelligence is optimized to respond with, "Thank you for your hard work. To relax, why not listen to some songs by a recommended rock band?"

[0581] In this way, the present invention can utilize generative artificial intelligence optimized for the user's hobbies and thoughts, and furthermore, an emotion engine enables the user to receive appropriate responses according to their emotional state. This provides a more personalized and advanced service that responds to the user's needs and emotions.

[0582] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0583] Step 1:

[0584] Users create an account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests. The input data is in JSON format and includes both profile information and questionnaire responses.

[0585] Step 2:

[0586] The device sends user-entered profile information and survey responses to the server. Specifically, it transfers data to the server using an HTTP POST request, which the server receives via a REST API endpoint. Input is the transmission of user data in JSON format, and output is a notification that the data has been saved to the database.

[0587] Step 3:

[0588] The server stores the received user information in a database for later analysis. PostgreSQL is used as the database system, and user data is inserted into the appropriate tables. Input is user data in JSON format, and output is the record stored in the database.

[0589] Step 4:

[0590] The server preprocesses user information stored in the database, filtering out noise and incomplete data. Specifically, it performs tasks such as imputing missing values ​​and removing outliers. The input is the user data from the database, and the output is a preprocessed, clean dataset.

[0591] Step 5:

[0592] The server applies machine learning algorithms to preprocessed data to extract users' hobbies and thought patterns. This process uses clustering techniques (e.g., K-means clustering) to classify user tendencies into multiple categories. The input is a preprocessed dataset, and the output is clustered user data.

[0593] Step 6:

[0594] The server saves the analysis results to a database. This allows them to be used to customize subsequent generative artificial intelligence. The input is clustered data, and the output is the clustering results stored in the database.

[0595] Step 7:

[0596] The server selects the optimal template from the generative artificial intelligence templates based on the analysis results. Specifically, it selects the best template for a particular cluster. The input is the clustering result, and the output is the selected template.

[0597] Step 8:

[0598] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, data related to the user's specific areas of interest is prioritized for training. The input consists of user-specific data and templates, and the output is a customized generative AI model.

[0599] Step 9:

[0600] The server stores customized generative artificial intelligence models in a database. The input is the generative artificial intelligence model, and the output is the model stored in the database.

[0601] Step 10:

[0602] The server prepares to deliver a customized generative artificial intelligence model to the user's terminal. The input is the model stored in the database, and the output is the server's notification that it is ready to deliver.

[0603] Step 11:

[0604] The terminal notifies the user that the generative artificial intelligence is available. The notification method uses the terminal's notification system. The input is a notification that distribution is ready, and the output is a notification to the user.

[0605] Step 12:

[0606] The user's terminal downloads the generative artificial intelligence model from the server. Specifically, it uses an HTTP GET request. The input is the download link from the server, and the output is the generative artificial intelligence model stored locally.

[0607] Step 13:

[0608] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts. The input is the generative AI model, and the output is the conversation content and search results.

[0609] Step 14:

[0610] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content. The input is dialogue data, and the output is emotion data.

[0611] Step 15:

[0612] The device sends the collected emotional data to the server. The input is emotional data, and the output is the transmission of data to the server.

[0613] Step 16:

[0614] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on it. Specifically, it adjusts the response to be more comforting and relaxing. The input is emotional data, and the output is the optimized response.

[0615] Step 17:

[0616] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence. The input is the dialogue data and emotion data, and the output is the collected data.

[0617] Step 18:

[0618] The terminal sends the collected data to the server. The input is the collected data, and the output is the transmission of data to the server.

[0619] Step 19:

[0620] The server analyzes collected dialogue and emotion data to detect the user's hobbies, interests, and emotional changes. The input is the collected data, and the output is the analysis results.

[0621] Step 20:

[0622] The server updates the generative artificial intelligence model based on new data. The input is the analysis results, and the output is the updated generative artificial intelligence model.

[0623] Step 21:

[0624] The server redistributes the updated generative artificial intelligence model to the user's terminal. The input is the updated generative artificial intelligence model, and the output is a notification that the redistribution is ready.

[0625] Step 22:

[0626] Users utilize an updated generative artificial intelligence (AI) system to engage in conversations and information retrieval that address new interests, hobbies, and emotions. The input is the updated AI model, and the output is new conversation content and search results.

[0627] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0628] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0629] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0630] [Second Embodiment]

[0631] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0632] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0633] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0634] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0635] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0636] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0637] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0638] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0639] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0640] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0641] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0642] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0643] System Overview

[0644] This invention is a system for providing a generative artificial intelligence customized specifically to the user's hobbies and interests. This system involves a server and a user's terminal working together to optimize the generative artificial intelligence based on the user's interests and hobbies.

[0645] Program Processing Description

[0646] 1. Collection of user data

[0647] User: Create a new account and enter initial data such as profile information and hobbies. This includes answering questionnaires and selecting genres of interest (music, movies, sports, etc.).

[0648] Terminal: Sends user-entered information and survey responses to the server.

[0649] Server: Receives user data and stores it in a database for later analysis.

[0650] 2. Analysis of hobbies and thought patterns

[0651] Server: Preprocesses the collected data and filters out noise and incomplete data.

[0652] Server: Based on pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used to classify user trends into multiple categories.

[0653] Server: Saves analysis results to the database.

[0654] 3. Customization of Generative Artificial Intelligence

[0655] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0656] Server: Trains the selected template with user-specific data to create a user-specific generative artificial intelligence. This customization process includes prioritizing the training of data related to specific areas of interest.

[0657] Server: Stores customized generative artificial intelligence models in a database.

[0658] 4. Provision to users

[0659] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0660] Terminal: Notifies the user that generative artificial intelligence is available.

[0661] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0662] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0663] 5. Continuous learning and improvement

[0664] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[0665] Terminal: Sends collected conversation data to the server.

[0666] Server: Analyzes collected data to detect changes in the user's hobbies and interests.

[0667] Server: Updates generative artificial intelligence models based on new data.

[0668] Server: Redistributes the updated model to user terminals.

[0669] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[0670] Specific example

[0671] Initial setup and usage

[0672] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[0673] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[0674] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[0675] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[0676] Continuous learning and improvement

[0677] User: As I continue to interact with the generative AI, I begin to develop an interest in "jazz." I start asking more questions about jazz.

[0678] Terminal: Sends new dialogue data to the server.

[0679] Server: Analyzes collected data to detect new interests (jazz) of user A. Updates the generative artificial intelligence model based on this information.

[0680] Device: Re-download the updated AI model so that User A can engage in conversations tailored to their new interests.

[0681] This system allows users to utilize generative artificial intelligence optimized for their hobbies and interests, and because it is continuously updated, they can always receive services that cater to their latest interests.

[0682] The following describes the processing flow.

[0683] Step 1:

[0684] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[0685] Step 2:

[0686] Terminal: Sends user-entered profile information and survey responses to the server.

[0687] Step 3:

[0688] Server: Receives user information and stores it in a database for later analysis.

[0689] Step 4:

[0690] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[0691] Step 5:

[0692] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[0693] Step 6:

[0694] Server: Saves analysis results to the database.

[0695] Step 7:

[0696] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0697] Step 8:

[0698] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[0699] Step 9:

[0700] Server: Stores customized generative artificial intelligence models in a database.

[0701] Step 10:

[0702] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0703] Step 11:

[0704] Terminal: Notifies the user that generative artificial intelligence is available.

[0705] Step 12:

[0706] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0707] Step 13:

[0708] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0709] Step 14:

[0710] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[0711] Step 15:

[0712] Terminal: Sends collected conversation data to the server.

[0713] Step 16:

[0714] Server: Analyzes collected dialogue data to detect changes in the user's hobbies and interests.

[0715] Step 17:

[0716] Server: Updates generative artificial intelligence models based on new data.

[0717] Step 18:

[0718] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[0719] Step 19:

[0720] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[0721] (Example 1)

[0722] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0723] Currently, many generative artificial intelligence models are specialized in providing general information, but they have the challenge of not being able to provide services tailored to a user's individual interests and thoughts. Furthermore, they struggle to quickly adapt to changes in user interests and hobbies. In addition, there is a lack of effective methods for providing the most suitable generative artificial intelligence model for individual users.

[0724] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0725] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for filtering the collected data to remove noise and incomplete data, means for analyzing the filtered data to extract the user's hobbies and thought patterns, means for selecting a generative artificial intelligence model based on the extracted hobbies and thought patterns, means for training and customizing the selected generative artificial intelligence model with user-specific data, means for delivering the customized generative artificial intelligence model to the user's terminal, and means for collecting interaction data with the user to continuously train and update the generative artificial intelligence model. This makes it possible to provide a generative artificial intelligence model tailored to the user's hobbies and thoughts, and to respond quickly to changes in the user's interests.

[0726] A "user" refers to an individual or legal entity that utilizes the system, provides data related to their hobbies and interests, and receives services based on a customized generative artificial intelligence model.

[0727] A "server" refers to a device or system that receives and stores data sent by users, and performs data analysis, customization, distribution, and updating of generative artificial intelligence models.

[0728] A "terminal" refers to an electronic device used by a user, such as a computer, smartphone, or tablet, that communicates with a server to transmit data, download generative artificial intelligence models, and engage in dialogue.

[0729] "Data collection" refers to the process of obtaining information about users' hobbies and interests through surveys and other means.

[0730] "Filtering" refers to the process of removing noise and incomplete data from collected data, preparing it for analysis.

[0731] "Data analysis" refers to the process of analyzing pre-processed data using machine learning algorithms and other methods to extract users' hobbies and thought patterns.

[0732] A "generative artificial intelligence model" refers to an artificial intelligence template designed to interact with users and provide information using natural language processing technology.

[0733] "Customization" refers to the process of training a generative artificial intelligence model with user-specific data to optimize it for individual users.

[0734] "Distribution" refers to the process of transferring a customized generative artificial intelligence model from a server to a user's device.

[0735] "Dialogue data" refers to data that records user questions and the responses of generative artificial intelligence models, including the content of the dialogue with the model.

[0736] "Continuous learning" refers to the process of improving the performance of a generative artificial intelligence model by retraining and updating it based on user interaction data.

[0737] This invention relates to a system that provides a generative artificial intelligence model specialized according to the user's hobbies and interests. This system can provide a service optimized to the user's interests and hobbies by analyzing data provided by the user, customizing the generative artificial intelligence based on that analysis, and continuously learning and updating it.

[0738] Specifically, this system includes the following elements:

[0739] User data collection

[0740] User: Create a new account and enter basic information such as name, age, gender, and place of residence, as well as answer a questionnaire about hobbies and interests. For example, to the question "What genre of music do you like?", answer "Rock".

[0741] Terminal: This terminal sends basic information and survey responses entered by the user to the server. This communication uses the secure protocol HTTPS.

[0742] Server: Receives user data and stores it in a database management system (e.g., MySQL) for later analysis.

[0743] Analysis of hobbies and thought patterns

[0744] Server: Preprocesses the collected data, filtering out noise and incomplete data. This preprocessing is done using data cleaning tools (e.g., OpenRefine).

[0745] Server: Based on filtered data, it applies machine learning algorithms (e.g., K-means clustering) to extract users' hobbies and thought patterns.

[0746] Server: Saves analysis results to the database.

[0747] Customization of generative artificial intelligence

[0748] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates (e.g., GPT-3).

[0749] Server: Trains and customizes the selected generative AI template with user-specific data. For example, if the user is interested in "rock" music, a dataset in that genre is prepared and the template is fine-tuned.

[0750] Server: Stores customized generative artificial intelligence models in a database.

[0751] Provision to users

[0752] Server: Prepares the server to deliver customized generative artificial intelligence models to user terminals. This preparation includes packaging the data to be delivered.

[0753] Terminal: Notify the user that generative artificial intelligence is available. The notification will use a push notification service (e.g., Firebase Cloud Messaging).

[0754] Terminal: Downloads generative artificial intelligence models from the server. HTTPS is used as the data transfer protocol.

[0755] User: Using the downloaded generative AI, initiate conversations and information searches that match the user's interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[0756] Continuous learning and improvement

[0757] Terminal: Every time a user interacts with the generative artificial intelligence, the interaction data is collected. For example, a question like "What are some recommended new action movies?" and its answer are stored as data.

[0758] Terminal: Sends collected conversation data to the server. The HTTPS protocol is used for transmission.

[0759] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[0760] Server: Updates the generative artificial intelligence model based on new data. Prepares a new dataset and performs fine-tuning again.

[0761] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[0762] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[0763] Example of a prompt

[0764] "What rock bands do you recommend these days?"

[0765] "What are some of the latest action movies you would recommend?"

[0766] "Please tell me about some famous jazz songs."

[0767] As described above, the present invention provides a generative artificial intelligence model that is tailored to the user's hobbies and thoughts, and can continuously adapt to changes in the user's interests.

[0768] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0769] Step 1: Collecting User Data

[0770] User: Create a new account and enter basic information such as name, age, gender, and place of residence. Also, answer a questionnaire about your interests and hobbies. For example, answer "Rock" to the question, "What is your favorite music genre?"

[0771] Input: Basic information, survey responses

[0772] Output: Basic information and survey response dataset

[0773] Terminal: The terminal sends the basic information and survey responses entered by the user to the server via the HTTPS protocol. Specifically, when the submit button on the input form is pressed, the data is transferred to the server.

[0774] Input: User input data

[0775] Output: Data transfer to the server

[0776] Server: Receives user data and stores it in a database management system (e.g., MySQL). User information and survey results are stored in the database in an organized manner.

[0777] Input: Transferred user data

[0778] Output: User data stored in the database

[0779] Step 2: Analysis of hobbies and thought patterns

[0780] Server: Preprocesses collected data, filtering out noise and incomplete data. Uses data cleaning tools (e.g., OpenRefine) to remove unnecessary data and prepare it for analysis.

[0781] Input: Raw data

[0782] Output: Filtered, clean data

[0783] Server: Applies machine learning algorithms (e.g., K-means clustering) to preprocessed data to extract users' hobbies and thought patterns. This algorithm classifies users' hobbies into multiple categories.

[0784] Input: Clean data

[0785] Output: User's hobbies and thought patterns

[0786] Server: Saves analysis results to a database. Analysis results are categorized by user and used for customization at a later stage.

[0787] Input: Analysis results

[0788] Output: Analysis results saved in the database

[0789] Step 3: Customizing Generative Artificial Intelligence

[0790] Server: Based on the analysis results, it selects the most suitable template from generative artificial intelligence templates (e.g., GPT-3). It uses a template selection algorithm to choose the most appropriate template.

[0791] Input: Analysis results

[0792] Output: Selected generative AI templates

[0793] Server: Trains the selected generative AI template with user-specific data. Specifically, it prepares a dataset related to the user's areas of interest and uses that data to fine-tune the template.

[0794] Input: Generative artificial intelligence template and user data

[0795] Output: Customized generative artificial intelligence model

[0796] Server: Stores customized generative artificial intelligence models in a database. Models are stored with version control using a model management system (e.g., MLflow).

[0797] Input: Customized generative artificial intelligence model

[0798] Output: Generative artificial intelligence models stored in the database

[0799] Step 4: Provision to users

[0800] Server: Prepares to deliver customized generative artificial intelligence models to user terminals. Specifically, it packages the delivery data and configures the delivery server.

[0801] Input: Customized generative artificial intelligence model

[0802] Output: Ready for delivery

[0803] Terminal: Notifies the user that generative artificial intelligence is available. This notification uses a push notification service (e.g., Firebase Cloud Messaging).

[0804] Input: Notification content

[0805] Output: Notification to the user

[0806] Terminal: Downloads generative AI models from the server. Secure HTTPS is used as the data transfer protocol.

[0807] Input: Download link

[0808] Output: Downloaded generative artificial intelligence model

[0809] User: Using the downloaded generative AI, initiate conversations and information searches that match their interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[0810] Input: User's question (prompt)

[0811] Output: Answer from a generative artificial intelligence

[0812] Step 5: Continuous learning and improvement

[0813] Terminal: Each time a user interacts with a generative artificial intelligence, the dialogue data is collected. This dialogue data includes the content of the questions and answers.

[0814] Input: Dialogue between the user and the generative artificial intelligence.

[0815] Output: Collected dialogue data

[0816] Terminal: Sends collected conversation data to the server. Secure HTTPS protocol is used for transmission.

[0817] Input: Dialogue data

[0818] Output: Sending data to the server

[0819] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[0820] Input: Dialogue data

[0821] Output: Analysis results

[0822] Server: Updates the generative artificial intelligence model based on new data. The update includes a process of retraining the model based on newly collected data.

[0823] Input: New dataset

[0824] Output: Updated generative artificial intelligence model

[0825] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[0826] Input: Updated generative artificial intelligence model

[0827] Output: Ready for delivery

[0828] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[0829] Input: New question from the user (prompt)

[0830] Output: A new answer from generative artificial intelligence

[0831] (Application Example 1)

[0832] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0833] Traditional advertising delivery systems have not adequately personalized ads using data on users' hobbies and interests, making it difficult to improve the effectiveness and accuracy of advertisements. Furthermore, providing users with ads that resonate with them requires continuous data collection and updates, but efficient methods for doing so have been lacking. Therefore, there is a need for a system that generates and continuously updates ads optimized for users' hobbies and interests.

[0834] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0835] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for collecting conversational data with the user to continuously train the generative artificial intelligence, means for delivering the generated customized advertisements, and means for collecting and analyzing user feedback to update the advertising model. This makes it possible to generate advertisements optimized for the user's hobbies and interests and to continuously improve them.

[0836] "Data related to users' hobbies and interests" refers to information collected to clarify the hobbies and interests that users exhibit, and specifically includes genres, preferences, and past behavioral history.

[0837] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of processing collected user data and using analytical techniques and algorithms to identify users' interests and patterns.

[0838] "Methods for customizing generative artificial intelligence" refers to technologies that optimize generative artificial intelligence for each user based on their hobbies and thought patterns, thereby creating personalized generative models.

[0839] "Means for delivering generated customized advertisements" refers to the technologies and methods for sending and displaying customized advertisements on a user's device.

[0840] "Methods for collecting and analyzing user feedback to update advertising models" refers to the process of collecting user reactions and evaluations to advertisements, analyzing that data, and improving the advertising model to match the latest user tastes and interests.

[0841] "Means for collecting dialogue data and continuously training generative artificial intelligence" refers to technologies and methods for collecting data from user interactions with generative artificial intelligence and using that data to improve and update the generative model.

[0842] "Means for customizing generative artificial intelligence based on analysis results" refers to technologies and methods for optimizing the operation of generative artificial intelligence based on analyzed hobbies and thought patterns.

[0843] This invention is a system that generates and delivers advertisements customized based on the user's hobbies and interests. This system involves a server and the user's terminal working together to handle everything from collecting and analyzing user data to generating and delivering advertisements, as well as continuous learning and improvement. The embodiments for carrying out this invention will be described in detail below.

[0844] The server first collects data on hobbies and interests provided by the user. This data is obtained when the user logs into the application and enters information based on their profile and questionnaires. The server then analyzes this collected data. The analysis involves data preprocessing (such as imputing missing values ​​and normalizing the data) and clustering techniques to extract the user's hobbies and thought patterns.

[0845] Furthermore, the server customizes the generative artificial intelligence (AI) based on the extracted hobbies and thought patterns. Specifically, it trains the generative AI model to match the user's preferences and generates personalized advertisements. TensorFlow and PyTorch are used as the generative AI frameworks in this generation process.

[0846] The device delivers customized, generated ads to the user. The generated ads are displayed to the user through a smartphone application. After the user views, clicks, or takes other action on an ad, feedback is sent back to the server. By analyzing this feedback data, the server updates its generative AI model to provide more accurate ads.

[0847] For example, if a user enters the prompt "Tell me about the latest recommended smartphone accessories," the generative artificial intelligence will generate and display advertisements suggesting the most suitable smartphone accessories based on the user's past interests and behavioral patterns. Through this continuous feedback loop, the system can always provide advertisements that are tailored to the user's latest hobbies and interests.

[0848] This system configuration allows users to receive advertisements that perfectly match their interests and preferences at any given time, and also enables advertisers to expect high results.

[0849] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0850] Step 1:

[0851] Users log in to the application and enter initial data, including profile information and data about their hobbies and interests from a questionnaire. This initial data includes the user's gender, age, areas of interest, and preferred product categories. The device then sends this data to the server.

[0852] Input: User profile information, survey data

[0853] Output: Initial data sent to the server

[0854] Step 2:

[0855] The server stores the initial data received from the terminal in a database. Based on this stored data, it performs data preprocessing. Preprocessing includes imputing missing values ​​and normalizing the data.

[0856] Input: Initial data

[0857] Output: Preprocessed data

[0858] Step 3:

[0859] The server analyzes pre-processed data and applies clustering techniques to extract users' interests and thought patterns. Machine learning algorithms (such as KMeans) are used for clustering.

[0860] Input: Preprocessed data

[0861] Output: User clustering results

[0862] Step 4:

[0863] The server customizes generative artificial intelligence (AI) based on the analysis results. Specifically, it trains a generative AI model for a cluster of users and generates personalized ad templates. This process utilizes TensorFlow and PyTorch.

[0864] Input: User clustering results

[0865] Output: Customized generative AI model

[0866] Step 5:

[0867] The device receives a customized generative AI model from the server and displays advertisements generated using that model to the user. When the user takes action on an advertisement (click, purchase, etc.), the device collects the feedback and sends it to the server.

[0868] Input: Customized Generative AI Model

[0869] Output: Displayed ads, user feedback

[0870] Step 6:

[0871] The server analyzes the received feedback data. Based on this analysis, it updates the generated AI model and continuously improves the advertising model.

[0872] Input: User feedback data

[0873] Output: Updated Generative AI Model

[0874] This enables ad delivery using the latest generative AI models that are always adapted to the user's hobbies and interests.

[0875] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0876] System Overview

[0877] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[0878] Program Processing Description

[0879] 1. Collection of user data

[0880] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[0881] Terminal: Sends user-entered profile information and survey responses to the server.

[0882] Server: Receives user information and stores it in a database for later analysis.

[0883] 2. Analysis of hobbies and thought patterns

[0884] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[0885] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[0886] Server: Saves analysis results to the database.

[0887] 3. Customization of Generative Artificial Intelligence

[0888] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0889] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[0890] Server: Stores customized generative artificial intelligence models in a database.

[0891] 4. Provision to users

[0892] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0893] Terminal: Notifies the user that generative artificial intelligence is available.

[0894] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0895] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0896] 5. Emotion Recognition and Response Optimization

[0897] Terminal: During interaction between the user and generative artificial intelligence, the system uses an emotion engine to recognize the user's emotions. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[0898] Terminal: Sends collected emotion data to the server.

[0899] Server: Analyzes emotional data and optimizes the generative AI's response based on it. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[0900] 6. Continuous learning and improvement

[0901] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[0902] Terminal: Sends collected data to the server.

[0903] Server: Analyzes collected dialogue and emotion data to detect the user's hobbies, interests, and emotional changes.

[0904] Server: Updates generative artificial intelligence models based on new data.

[0905] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[0906] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[0907] Specific example

[0908] Initial setup and usage

[0909] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[0910] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[0911] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[0912] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[0913] Emotion recognition and response optimization

[0914] User: When interacting with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if User A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[0915] Terminal: Send this emotion data to the server.

[0916] Server: Analyzes emotional data and adjusts the generative AI's responses to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[0917] Continuous learning and improvement

[0918] User: As they continue to interact with the generative AI, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[0919] Terminal: Sends new dialogue data and emotion data to the server.

[0920] Server: Analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[0921] Terminal: Re-download the updated generative artificial intelligence model, enabling user A to engage in conversations that respond to new interests and emotions.

[0922] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[0923] The following describes the processing flow.

[0924] Step 1:

[0925] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[0926] Step 2:

[0927] Terminal: Sends user-entered profile information and survey responses to the server.

[0928] Step 3:

[0929] Server: Receives user information and stores it in a database for later analysis.

[0930] Step 4:

[0931] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[0932] Step 5:

[0933] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[0934] Step 6:

[0935] Server: Saves analysis results to the database.

[0936] Step 7:

[0937] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[0938] Step 8:

[0939] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[0940] Step 9:

[0941] Server: Stores customized generative artificial intelligence models in a database.

[0942] Step 10:

[0943] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[0944] Step 11:

[0945] Terminal: Notifies the user that generative artificial intelligence is available.

[0946] Step 12:

[0947] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[0948] Step 13:

[0949] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[0950] Step 14:

[0951] User: As you interact with the generative artificial intelligence, the emotion engine analyzes your emotions in real time. This emotion data is obtained from voice tone, facial expressions, and text content.

[0952] Step 15:

[0953] Terminal: The emotion engine analyzes the emotion data and sends it to the server.

[0954] Step 16:

[0955] Server: Analyzes emotional data and optimizes the responses of the generative artificial intelligence. For example, if the system recognizes that the user is feeling "stressed," the generative AI will generate a calmer, more comforting response.

[0956] Step 17:

[0957] User: Receive appropriate responses through interaction with generative artificial intelligence.

[0958] Step 18:

[0959] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[0960] Step 19:

[0961] Terminal: Sends collected dialogue data and emotion data to the server.

[0962] Step 20:

[0963] Server: Analyzes collected data to detect changes in users' hobbies, interests, and emotions.

[0964] Step 21:

[0965] Server: Updates generative artificial intelligence models based on new data.

[0966] Step 22:

[0967] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[0968] Step 23:

[0969] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[0970] (Example 2)

[0971] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0972] Conventional generative artificial intelligence systems struggled not only to be customized based on the user's hobbies and thoughts, but also to optimize their responses in response to changes in the user's emotions and interests. Furthermore, the lack of sufficient continuous learning and improvement for individual users resulted in a limited user experience. In addition, there was a lack of efficient means to collect initial user data and advanced methods for analyzing the collected data.

[0973] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0974] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, and means for collecting dialogue data and emotional data with the user to continuously train the generative artificial intelligence and optimize its responses. This makes it possible to provide optimal responses in accordance with changes in the user's hobbies, thoughts, and emotions, and to continuously learn and improve.

[0975] "Data on users' hobbies and interests" refers to information that users provide through their profile information and survey responses, indicating their interests and preferences in specific fields such as music, movies, literature, and sports.

[0976] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of generating specific patterns or clusters based on users' hobbies and interests using data cleansing through preprocessing and machine learning algorithms.

[0977] "Means for customizing generative artificial intelligence based on extracted hobby and thought patterns" refers to methods and techniques for optimizing a base generative artificial intelligence model to the hobbies and thoughts of individual users by utilizing the analysis results.

[0978] "Means for delivering customized generative artificial intelligence to a user's terminal" refers to a technology that transfers a generative artificial intelligence model from a server to a user's terminal, making that model available for use in the user's local environment.

[0979] "A means of collecting user interaction data and emotional data to continuously train generative artificial intelligence and optimize responses" refers to a technology that analyzes emotional data obtained from text, voice, facial expressions, etc., collected by generative artificial intelligence during interactions with users, and adjusts and improves the AI's responses to match the user's emotions and needs.

[0980] "Survey methods" refer to question-based interfaces and their implementation methods for efficiently collecting initial user data.

[0981] A "machine learning algorithm" is a mathematical model and method that automatically learns from large amounts of data to perform predictions, classifications, and clustering.

[0982] Modes for carrying out the invention

[0983] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[0984] Hardware and software

[0985] Hardware:

[0986] The user's smartphone or PC

[0987] Cloud Server

[0988] software:

[0989] Database management systems: MySQL, PostgreSQL

[0990] Machine learning libraries: TensorFlow, PyTorch

[0991] Emotion recognition software: Microsoft Azure Emotion API

[0992] Generative artificial intelligence: GPT-3, BERT

[0993] Program Processing Description

[0994] 1. Collection of user data:

[0995] Users create a new account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[0996] The device sends the user's entered profile information and survey responses to the server.

[0997] The server stores the received user information in a database for later analysis.

[0998] 2. Analysis of hobbies and thought patterns:

[0999] The server preprocesses the user information stored in the database, filtering out noise and incomplete data.

[1000] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[1001] The server saves the analysis results to the database.

[1002] 3. Customizing Generative Artificial Intelligence:

[1003] The server selects the most suitable template from the generative artificial intelligence templates based on the analysis results.

[1004] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[1005] The server stores customized generative artificial intelligence models in a database.

[1006] 4. Provision to users:

[1007] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[1008] The device notifies the user that generative artificial intelligence is available.

[1009] The terminal downloads the generative artificial intelligence model from the server.

[1010] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts.

[1011] 5. Emotion Recognition and Response Optimization:

[1012] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[1013] The device sends the collected emotional data to the server.

[1014] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[1015] 6. Continuous learning and improvement:

[1016] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[1017] The device sends the collected data to the server.

[1018] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[1019] The server updates the generative artificial intelligence model based on new data.

[1020] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[1021] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[1022] Specific example

[1023] 1. Initial setup and usage:

[1024] When creating an account, users indicate that they are particularly interested in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[1025] The terminal sends this information to the server, which then uses it to create a profile for user A.

[1026] The server analyzes the collected data and extracts user A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for user A.

[1027] The device downloads a customized AI model, and user A initiates a conversation. For example, if user A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[1028] 2. Emotion recognition and response optimization:

[1029] When a user interacts with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if user A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[1030] The device sends this emotion data to the server.

[1031] The server analyzes emotional data and adjusts the generative artificial intelligence's response to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[1032] 3. Continuous learning and improvement:

[1033] As the user continues to interact with the generative artificial intelligence, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[1034] The device sends new conversational and emotional data to the server.

[1035] The server analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[1036] The device will download the updated generative artificial intelligence model again, enabling user A to engage in conversations that respond to new interests and emotions.

[1037] Example of a prompt

[1038] "Could you recommend some rock bands you've been listening to lately?"

[1039] "Today was really tough."

[1040] "I've recently become interested in jazz. Do you have any recommendations?"

[1041] This allows users to utilize generative artificial intelligence optimized for their own hobbies and thoughts, and furthermore, an emotion engine enables them to receive appropriate responses based on their emotional state. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[1042] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1043] Program processing flow

[1044] User data collection

[1045] Step 1:

[1046] The user launches the application and creates a new account. They enter profile information such as their name, email address, gender, and age, and answer a questionnaire about their hobbies and interests.

[1047] Input: Name, email address, gender, age, and survey responses regarding hobbies and interests.

[1048] Output: Entered profile information and survey data.

[1049] Step 2:

[1050] The device sends the profile information and survey responses entered by the user to the server.

[1051] Input: Profile information and survey responses provided by the user.

[1052] Output: User data sent to the server.

[1053] Step 3:

[1054] The server saves the received user information to the database. It then verifies that the data was saved successfully.

[1055] Input: User data sent from the device.

[1056] Output: User data stored in the database.

[1057] Analysis of hobbies and thought patterns

[1058] Step 4:

[1059] The server preprocesses user information stored in the database, filtering out noise and incomplete data. For example, it performs data imputation and data normalization.

[1060] Input: User data stored in the database.

[1061] Output: Pre-processed, clean data.

[1062] Step 5:

[1063] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. Clustering techniques are then used to classify user trends into multiple categories.

[1064] Input: Pre-processed, clean data.

[1065] Output: A cluster that shows the user's hobbies and thought patterns.

[1066] Step 6:

[1067] The server saves the analysis results to the database. It then verifies that the analysis results were saved correctly.

[1068] Input: Analysis results (user cluster information).

[1069] Output: Analysis results stored in the database.

[1070] Customization of generative artificial intelligence

[1071] Step 7:

[1072] The server selects the optimal generative artificial intelligence template based on the analysis results.

[1073] Input: Analysis results (user cluster information).

[1074] Output: Selected AI template.

[1075] Step 8:

[1076] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. It prioritizes learning data related to the user's specific areas of interest.

[1077] Input: Selected AI template, user-specific data.

[1078] Output: A customized generative AI model.

[1079] Step 9:

[1080] The server saves the customized generative artificial intelligence model to the database. It then verifies that the saving process was successful.

[1081] Input: A customized generative AI model.

[1082] Output: Customized AI model stored in the database.

[1083] Provision to users

[1084] Step 10:

[1085] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[1086] Input: A customized AI model stored in a database.

[1087] Output: Notification that delivery is ready.

[1088] Step 11:

[1089] The device notifies the user that it is ready for distribution and downloads the generative artificial intelligence model.

[1090] Input: Server notification of preparation for delivery.

[1091] Output: Availability notification to the user and downloaded generative AI model.

[1092] Step 12:

[1093] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts. For example, they might ask, "Can you recommend some recent rock bands?"

[1094] Input: Prompt message (user question).

[1095] Output: Response from a generative artificial intelligence.

[1096] Emotion recognition and response optimization

[1097] Step 13:

[1098] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. Emotional data is analyzed from voice tone, facial expressions, and text content.

[1099] Input: User's voice tone, facial expression, and text content.

[1100] Output: Analyzed sentiment data.

[1101] Step 14:

[1102] The device sends the collected emotional data to the server.

[1103] Input: Analyzed sentiment data.

[1104] Output: Sentiment data sent to the server.

[1105] Step 15:

[1106] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the response is changed to something calmer and more comforting.

[1107] Input: Analyzed sentiment data.

[1108] Output: Optimized AI response model.

[1109] Continuous learning and improvement

[1110] Step 16:

[1111] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[1112] Input: User interaction data, sentiment data.

[1113] Output: Collected dialogue data and sentiment data.

[1114] Step 17:

[1115] The device sends the collected data to the server.

[1116] Input: Collected dialogue data and sentiment data.

[1117] Output: Dialogue data and sentiment data sent to the server.

[1118] Step 18:

[1119] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[1120] Input: Collected dialogue data and sentiment data.

[1121] Output: Analyzed data (changes in user hobbies, interests, and emotions).

[1122] Step 19:

[1123] The server updates the generative artificial intelligence model based on new data.

[1124] Input: Newly analyzed data.

[1125] Output: Updated generative AI model.

[1126] Step 20:

[1127] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[1128] Input: An updated generative AI model.

[1129] Output: The AI ​​model redistributed to the user's terminal.

[1130] Step 21:

[1131] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[1132] Input: The user's new prompt message.

[1133] Output: Response from a generative artificial intelligence.

[1134] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[1135] (Application Example 2)

[1136] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[1137] Conventional generative artificial intelligence systems often fail to personalize responses based on user preferences and interests, resulting in responses that do not always meet user expectations. Furthermore, their inability to recognize user emotions makes it difficult to provide appropriate responses and services. Moreover, in environments such as autonomous vehicles, there is a growing demand for personalized services that respond to user emotions.

[1138] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing the generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for recognizing the user's emotions, and means for optimizing the response of the generative artificial intelligence based on the recognized user emotions. This enables personalized services tailored to the user's hobbies and interests, as well as appropriate responses in response to their emotions.

[1139] "Data related to hobbies and interests" refers to the preferences and interests that users have in specific fields, and specifically includes information related to preferences such as music, movies, sports, and travel.

[1140] "Methods for analyzing data to extract users' hobbies and thought patterns" refers to algorithms and methods for analyzing collected data to reveal users' hobbies and tendencies in thinking.

[1141] "Means for customizing generative artificial intelligence" refers to methods and technologies for individualizing the artificial intelligence model generated based on the extracted user's hobbies and thought patterns.

[1142] "Means for delivering customized generative artificial intelligence to a user's device" refers to systems and technologies for sending individualized artificial intelligence models to a user's device in a format that the user can use.

[1143] "Means of recognizing user emotions" refers to technologies that analyze a user's voice, facial expressions, text, etc., to understand their emotional state.

[1144] "Means for optimizing the response of generative artificial intelligence based on recognized user emotions" refers to technologies for appropriately adjusting the content and methods of responses generated by artificial intelligence based on emotion recognition.

[1145] The embodiments for carrying out the present invention are described in detail below. First, a method for collecting data on a user's hobbies and interests and customizing an artificial intelligence system based on that data is described. Furthermore, a specific method for providing the user with the most appropriate response through emotion recognition is also described.

[1146] Program Overview

[1147] 1. Collection of user data

[1148] Users create an account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[1149] The device sends the user's entered profile information and survey responses to the server.

[1150] The server stores the received user information in a database for later analysis.

[1151] 2. Analysis of hobbies and thought patterns

[1152] The server preprocesses user information stored in the database, filtering out noise and incomplete data.

[1153] The server uses pre-processed data to apply machine learning algorithms and extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[1154] The server saves the analysis results to a database.

[1155] 3. Customization of Generative Artificial Intelligence

[1156] Based on the analysis results, the server selects the most suitable template from the generative artificial intelligence templates.

[1157] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[1158] The server stores customized generative artificial intelligence models in a database.

[1159] 4. Provision to users

[1160] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[1161] The device notifies the user that generative artificial intelligence is available.

[1162] The user's terminal downloads a generative artificial intelligence model from the server.

[1163] The user then uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and preferences.

[1164] 5. Emotion Recognition and Response Optimization

[1165] During interactions between the user and generative artificial intelligence, the device uses an emotion engine to recognize the user's emotions. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[1166] The device sends the collected emotional data to the server.

[1167] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[1168] 6. Continuous learning and improvement

[1169] The device collects dialogue data and emotional data each time the user interacts with the generative artificial intelligence.

[1170] The device sends the collected data to the server.

[1171] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[1172] The server updates the generative artificial intelligence model based on new data.

[1173] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[1174] Users utilize updated generative artificial intelligence to engage in conversations and information retrieval that cater to their new interests, hobbies, and emotions.

[1175] Hardware and software to be used

[1176] hardware

[1177] Infotainment system in autonomous vehicles

[1178] Camera and microphone: For emotion recognition inside the vehicle.

[1179] Vehicle communication module: Data communication with the server

[1180] software

[1181] Server-side: Apache Kafka (data streaming), TensorFlow (machine learning models), PostgreSQL (database)

[1182] Autonomous vehicle terminal: Custom Emotion Recognition Engine, in-device application (Android Auto, CarPlay compatible)

[1183] Specific example

[1184] For example, if a user in an autonomous vehicle asks, "What are some rock bands you recommend lately?", the generative artificial intelligence and emotion recognition engine will process a prompt sentence like the following.

[1185] Example of a prompt:

[1186] "We've identified that the user is currently experiencing stress. Their hobby is rock music. Please suggest some rock bands."

[1187] Based on this prompt, the generative artificial intelligence is optimized to respond with, "Thank you for your hard work. To relax, why not listen to some songs by a recommended rock band?"

[1188] In this way, the present invention can utilize generative artificial intelligence optimized for the user's hobbies and thoughts, and furthermore, an emotion engine enables the user to receive appropriate responses according to their emotional state. This provides a more personalized and advanced service that responds to the user's needs and emotions.

[1189] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1190] Step 1:

[1191] Users create an account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests. The input data is in JSON format and includes both profile information and questionnaire responses.

[1192] Step 2:

[1193] The device sends user-entered profile information and survey responses to the server. Specifically, it transfers data to the server using an HTTP POST request, which the server receives via a REST API endpoint. Input is the transmission of user data in JSON format, and output is a notification that the data has been saved to the database.

[1194] Step 3:

[1195] The server stores the received user information in a database for later analysis. PostgreSQL is used as the database system, and user data is inserted into the appropriate tables. Input is user data in JSON format, and output is the record stored in the database.

[1196] Step 4:

[1197] The server preprocesses user information stored in the database, filtering out noise and incomplete data. Specifically, it performs tasks such as imputing missing values ​​and removing outliers. The input is the user data from the database, and the output is a preprocessed, clean dataset.

[1198] Step 5:

[1199] The server applies machine learning algorithms to preprocessed data to extract users' hobbies and thought patterns. This process uses clustering techniques (e.g., K-means clustering) to classify user tendencies into multiple categories. The input is a preprocessed dataset, and the output is clustered user data.

[1200] Step 6:

[1201] The server saves the analysis results to a database. This allows them to be used to customize subsequent generative artificial intelligence. The input is clustered data, and the output is the clustering results stored in the database.

[1202] Step 7:

[1203] The server selects the optimal template from the generative artificial intelligence templates based on the analysis results. Specifically, it selects the best template for a particular cluster. The input is the clustering result, and the output is the selected template.

[1204] Step 8:

[1205] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, data related to the user's specific areas of interest is prioritized for training. The input consists of user-specific data and templates, and the output is a customized generative AI model.

[1206] Step 9:

[1207] The server stores customized generative artificial intelligence models in a database. The input is the generative artificial intelligence model, and the output is the model stored in the database.

[1208] Step 10:

[1209] The server prepares to deliver a customized generative artificial intelligence model to the user's terminal. The input is the model stored in the database, and the output is the server's notification that it is ready to deliver.

[1210] Step 11:

[1211] The terminal notifies the user that the generative artificial intelligence is available. The notification method uses the terminal's notification system. The input is a notification that distribution is ready, and the output is a notification to the user.

[1212] Step 12:

[1213] The user's terminal downloads the generative artificial intelligence model from the server. Specifically, it uses an HTTP GET request. The input is the download link from the server, and the output is the generative artificial intelligence model stored locally.

[1214] Step 13:

[1215] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts. The input is the generative AI model, and the output is the conversation content and search results.

[1216] Step 14:

[1217] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content. The input is dialogue data, and the output is emotion data.

[1218] Step 15:

[1219] The device sends the collected emotional data to the server. The input is emotional data, and the output is the transmission of data to the server.

[1220] Step 16:

[1221] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on it. Specifically, it adjusts the response to be more comforting and relaxing. The input is emotional data, and the output is the optimized response.

[1222] Step 17:

[1223] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence. The input is the dialogue data and emotion data, and the output is the collected data.

[1224] Step 18:

[1225] The terminal sends the collected data to the server. The input is the collected data, and the output is the transmission of data to the server.

[1226] Step 19:

[1227] The server analyzes collected dialogue and emotion data to detect the user's hobbies, interests, and emotional changes. The input is the collected data, and the output is the analysis results.

[1228] Step 20:

[1229] The server updates the generative artificial intelligence model based on new data. The input is the analysis results, and the output is the updated generative artificial intelligence model.

[1230] Step 21:

[1231] The server redistributes the updated generative artificial intelligence model to the user's terminal. The input is the updated generative artificial intelligence model, and the output is a notification that the redistribution is ready.

[1232] Step 22:

[1233] Users utilize an updated generative artificial intelligence (AI) system to engage in conversations and information retrieval that address new interests, hobbies, and emotions. The input is the updated AI model, and the output is new conversation content and search results.

[1234] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1235] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1236] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[1237] [Third Embodiment]

[1238] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[1239] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1240] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1241] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[1242] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1243] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1244] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1245] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1246] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[1247] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1248] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1249] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[1250] System Overview

[1251] This invention is a system for providing a generative artificial intelligence customized specifically to the user's hobbies and interests. This system involves a server and a user's terminal working together to optimize the generative artificial intelligence based on the user's interests and hobbies.

[1252] Program Processing Description

[1253] 1. Collection of user data

[1254] User: Create a new account and enter initial data such as profile information and hobbies. This includes answering questionnaires and selecting genres of interest (music, movies, sports, etc.).

[1255] Terminal: Sends user-entered information and survey responses to the server.

[1256] Server: Receives user data and stores it in a database for later analysis.

[1257] 2. Analysis of hobbies and thought patterns

[1258] Server: Preprocesses the collected data and filters out noise and incomplete data.

[1259] Server: Based on pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used to classify user trends into multiple categories.

[1260] Server: Saves analysis results to the database.

[1261] 3. Customization of Generative Artificial Intelligence

[1262] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[1263] Server: Trains the selected template with user-specific data to create a user-specific generative artificial intelligence. This customization process includes prioritizing the training of data related to specific areas of interest.

[1264] Server: Stores customized generative artificial intelligence models in a database.

[1265] 4. Provision to users

[1266] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[1267] Terminal: Notifies the user that generative artificial intelligence is available.

[1268] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[1269] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[1270] 5. Continuous learning and improvement

[1271] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[1272] Terminal: Sends collected conversation data to the server.

[1273] Server: Analyzes collected data to detect changes in the user's hobbies and interests.

[1274] Server: Updates generative artificial intelligence models based on new data.

[1275] Server: Redistributes the updated model to user terminals.

[1276] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[1277] Specific example

[1278] Initial setup and usage

[1279] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[1280] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[1281] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[1282] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[1283] Continuous learning and improvement

[1284] User: As I continue to interact with the generative AI, I begin to develop an interest in "jazz." I start asking more questions about jazz.

[1285] Terminal: Sends new dialogue data to the server.

[1286] Server: Analyzes collected data to detect new interests (jazz) of user A. Updates the generative artificial intelligence model based on this information.

[1287] Device: Re-download the updated AI model so that User A can engage in conversations tailored to their new interests.

[1288] This system allows users to utilize generative artificial intelligence optimized for their hobbies and interests, and because it is continuously updated, they can always receive services that cater to their latest interests.

[1289] The following describes the processing flow.

[1290] Step 1:

[1291] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[1292] Step 2:

[1293] Terminal: Sends user-entered profile information and survey responses to the server.

[1294] Step 3:

[1295] Server: Receives user information and stores it in a database for later analysis.

[1296] Step 4:

[1297] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[1298] Step 5:

[1299] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[1300] Step 6:

[1301] Server: Saves analysis results to the database.

[1302] Step 7:

[1303] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[1304] Step 8:

[1305] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[1306] Step 9:

[1307] Server: Stores customized generative artificial intelligence models in a database.

[1308] Step 10:

[1309] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[1310] Step 11:

[1311] Terminal: Notifies the user that generative artificial intelligence is available.

[1312] Step 12:

[1313] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[1314] Step 13:

[1315] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[1316] Step 14:

[1317] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[1318] Step 15:

[1319] Terminal: Sends collected conversation data to the server.

[1320] Step 16:

[1321] Server: Analyzes collected dialogue data to detect changes in the user's hobbies and interests.

[1322] Step 17:

[1323] Server: Updates generative artificial intelligence models based on new data.

[1324] Step 18:

[1325] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[1326] Step 19:

[1327] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[1328] (Example 1)

[1329] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1330] Currently, many generative artificial intelligence models are specialized in providing general information, but they have the challenge of not being able to provide services tailored to a user's individual interests and thoughts. Furthermore, they struggle to quickly adapt to changes in user interests and hobbies. In addition, there is a lack of effective methods for providing the most suitable generative artificial intelligence model for individual users.

[1331] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1332] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for filtering the collected data to remove noise and incomplete data, means for analyzing the filtered data to extract the user's hobbies and thought patterns, means for selecting a generative artificial intelligence model based on the extracted hobbies and thought patterns, means for training and customizing the selected generative artificial intelligence model with user-specific data, means for delivering the customized generative artificial intelligence model to the user's terminal, and means for collecting interaction data with the user to continuously train and update the generative artificial intelligence model. This makes it possible to provide a generative artificial intelligence model tailored to the user's hobbies and thoughts, and to respond quickly to changes in the user's interests.

[1333] A "user" refers to an individual or legal entity that utilizes the system, provides data related to their hobbies and interests, and receives services based on a customized generative artificial intelligence model.

[1334] A "server" refers to a device or system that receives and stores data sent by users, and performs data analysis, customization, distribution, and updating of generative artificial intelligence models.

[1335] A "terminal" refers to an electronic device used by a user, such as a computer, smartphone, or tablet, that communicates with a server to transmit data, download generative artificial intelligence models, and engage in dialogue.

[1336] "Data collection" refers to the process of obtaining information about users' hobbies and interests through surveys and other means.

[1337] "Filtering" refers to the process of removing noise and incomplete data from collected data, preparing it for analysis.

[1338] "Data analysis" refers to the process of analyzing pre-processed data using machine learning algorithms and other methods to extract users' hobbies and thought patterns.

[1339] A "generative artificial intelligence model" refers to an artificial intelligence template designed to interact with users and provide information using natural language processing technology.

[1340] "Customization" refers to the process of training a generative artificial intelligence model with user-specific data to optimize it for individual users.

[1341] "Distribution" refers to the process of transferring a customized generative artificial intelligence model from a server to a user's device.

[1342] "Dialogue data" refers to data that records user questions and the responses of generative artificial intelligence models, including the content of the dialogue with the model.

[1343] "Continuous learning" refers to the process of improving the performance of a generative artificial intelligence model by retraining and updating it based on user interaction data.

[1344] This invention relates to a system that provides a generative artificial intelligence model specialized according to the user's hobbies and interests. This system can provide a service optimized to the user's interests and hobbies by analyzing data provided by the user, customizing the generative artificial intelligence based on that analysis, and continuously learning and updating it.

[1345] Specifically, this system includes the following elements:

[1346] User data collection

[1347] User: Create a new account and enter basic information such as name, age, gender, and place of residence, as well as answer a questionnaire about hobbies and interests. For example, to the question "What genre of music do you like?", answer "Rock".

[1348] Terminal: This terminal sends basic information and survey responses entered by the user to the server. This communication uses the secure protocol HTTPS.

[1349] Server: Receives user data and stores it in a database management system (e.g., MySQL) for later analysis.

[1350] Analysis of hobbies and thought patterns

[1351] Server: Preprocesses the collected data, filtering out noise and incomplete data. This preprocessing is done using data cleaning tools (e.g., OpenRefine).

[1352] Server: Based on filtered data, it applies machine learning algorithms (e.g., K-means clustering) to extract users' hobbies and thought patterns.

[1353] Server: Saves analysis results to the database.

[1354] Customization of generative artificial intelligence

[1355] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates (e.g., GPT-3).

[1356] Server: Trains and customizes the selected generative AI template with user-specific data. For example, if the user is interested in "rock" music, a dataset in that genre is prepared and the template is fine-tuned.

[1357] Server: Stores customized generative artificial intelligence models in a database.

[1358] Provision to users

[1359] Server: Prepares the server to deliver customized generative artificial intelligence models to user terminals. This preparation includes packaging the data to be delivered.

[1360] Terminal: Notify the user that generative artificial intelligence is available. The notification will use a push notification service (e.g., Firebase Cloud Messaging).

[1361] Terminal: Downloads generative artificial intelligence models from the server. HTTPS is used as the data transfer protocol.

[1362] User: Using the downloaded generative AI, initiate conversations and information searches that match the user's interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[1363] Continuous learning and improvement

[1364] Terminal: Every time a user interacts with the generative artificial intelligence, the interaction data is collected. For example, a question like "What are some recommended new action movies?" and its answer are stored as data.

[1365] Terminal: Sends collected conversation data to the server. The HTTPS protocol is used for transmission.

[1366] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[1367] Server: Updates the generative artificial intelligence model based on new data. Prepares a new dataset and performs fine-tuning again.

[1368] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[1369] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[1370] Example of a prompt

[1371] "What rock bands do you recommend these days?"

[1372] "What are some of the latest action movies you would recommend?"

[1373] "Please tell me about some famous jazz songs."

[1374] As described above, the present invention provides a generative artificial intelligence model that is tailored to the user's hobbies and thoughts, and can continuously adapt to changes in the user's interests.

[1375] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1376] Step 1: Collecting User Data

[1377] User: Create a new account and enter basic information such as name, age, gender, and place of residence. Also, answer a questionnaire about your interests and hobbies. For example, answer "Rock" to the question, "What is your favorite music genre?"

[1378] Input: Basic information, survey responses

[1379] Output: Basic information and survey response dataset

[1380] Terminal: The terminal sends the basic information and survey responses entered by the user to the server via the HTTPS protocol. Specifically, when the submit button on the input form is pressed, the data is transferred to the server.

[1381] Input: User input data

[1382] Output: Data transfer to the server

[1383] Server: Receives user data and stores it in a database management system (e.g., MySQL). User information and survey results are stored in the database in an organized manner.

[1384] Input: Transferred user data

[1385] Output: User data stored in the database

[1386] Step 2: Analysis of hobbies and thought patterns

[1387] Server: Preprocesses collected data, filtering out noise and incomplete data. Uses data cleaning tools (e.g., OpenRefine) to remove unnecessary data and prepare it for analysis.

[1388] Input: Raw data

[1389] Output: Filtered, clean data

[1390] Server: Applies machine learning algorithms (e.g., K-means clustering) to preprocessed data to extract users' hobbies and thought patterns. This algorithm classifies users' hobbies into multiple categories.

[1391] Input: Clean data

[1392] Output: User's hobbies and thought patterns

[1393] Server: Saves analysis results to a database. Analysis results are categorized by user and used for customization at a later stage.

[1394] Input: Analysis results

[1395] Output: Analysis results saved in the database

[1396] Step 3: Customizing Generative Artificial Intelligence

[1397] Server: Based on the analysis results, it selects the most suitable template from generative artificial intelligence templates (e.g., GPT-3). It uses a template selection algorithm to choose the most appropriate template.

[1398] Input: Analysis results

[1399] Output: Selected generative AI templates

[1400] Server: Trains the selected generative AI template with user-specific data. Specifically, it prepares a dataset related to the user's areas of interest and uses that data to fine-tune the template.

[1401] Input: Generative artificial intelligence template and user data

[1402] Output: Customized generative artificial intelligence model

[1403] Server: Stores customized generative artificial intelligence models in a database. Models are stored with version control using a model management system (e.g., MLflow).

[1404] Input: Customized generative artificial intelligence model

[1405] Output: Generative artificial intelligence models stored in the database

[1406] Step 4: Provision to users

[1407] Server: Prepares to deliver customized generative artificial intelligence models to user terminals. Specifically, it packages the delivery data and configures the delivery server.

[1408] Input: Customized generative artificial intelligence model

[1409] Output: Ready for delivery

[1410] Terminal: Notifies the user that generative artificial intelligence is available. This notification uses a push notification service (e.g., Firebase Cloud Messaging).

[1411] Input: Notification content

[1412] Output: Notification to the user

[1413] Terminal: Downloads generative AI models from the server. Secure HTTPS is used as the data transfer protocol.

[1414] Input: Download link

[1415] Output: Downloaded generative artificial intelligence model

[1416] User: Using the downloaded generative AI, initiate conversations and information searches that match their interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[1417] Input: User's question (prompt)

[1418] Output: Answer from a generative artificial intelligence

[1419] Step 5: Continuous learning and improvement

[1420] Terminal: Each time a user interacts with a generative artificial intelligence, the dialogue data is collected. This dialogue data includes the content of the questions and answers.

[1421] Input: Dialogue between the user and the generative artificial intelligence.

[1422] Output: Collected dialogue data

[1423] Terminal: Sends collected conversation data to the server. Secure HTTPS protocol is used for transmission.

[1424] Input: Dialogue data

[1425] Output: Sending data to the server

[1426] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[1427] Input: Dialogue data

[1428] Output: Analysis results

[1429] Server: Updates the generative artificial intelligence model based on new data. The update includes a process of retraining the model based on newly collected data.

[1430] Input: New dataset

[1431] Output: Updated generative artificial intelligence model

[1432] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[1433] Input: Updated generative artificial intelligence model

[1434] Output: Ready for delivery

[1435] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[1436] Input: New question from the user (prompt)

[1437] Output: A new answer from generative artificial intelligence

[1438] (Application Example 1)

[1439] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1440] Traditional advertising delivery systems have not adequately personalized ads using data on users' hobbies and interests, making it difficult to improve the effectiveness and accuracy of advertisements. Furthermore, providing users with ads that resonate with them requires continuous data collection and updates, but efficient methods for doing so have been lacking. Therefore, there is a need for a system that generates and continuously updates ads optimized for users' hobbies and interests.

[1441] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1442] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for collecting conversational data with the user to continuously train the generative artificial intelligence, means for delivering the generated customized advertisements, and means for collecting and analyzing user feedback to update the advertising model. This makes it possible to generate advertisements optimized for the user's hobbies and interests and to continuously improve them.

[1443] "Data related to users' hobbies and interests" refers to information collected to clarify the hobbies and interests that users exhibit, and specifically includes genres, preferences, and past behavioral history.

[1444] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of processing collected user data and using analytical techniques and algorithms to identify users' interests and patterns.

[1445] "Methods for customizing generative artificial intelligence" refers to technologies that optimize generative artificial intelligence for each user based on their hobbies and thought patterns, thereby creating personalized generative models.

[1446] "Means for delivering generated customized advertisements" refers to the technologies and methods for sending and displaying customized advertisements on a user's device.

[1447] "Methods for collecting and analyzing user feedback to update advertising models" refers to the process of collecting user reactions and evaluations to advertisements, analyzing that data, and improving the advertising model to match the latest user tastes and interests.

[1448] "Means for collecting dialogue data and continuously training generative artificial intelligence" refers to technologies and methods for collecting data from user interactions with generative artificial intelligence and using that data to improve and update the generative model.

[1449] "Means for customizing generative artificial intelligence based on analysis results" refers to technologies and methods for optimizing the operation of generative artificial intelligence based on analyzed hobbies and thought patterns.

[1450] This invention is a system that generates and delivers advertisements customized based on the user's hobbies and interests. This system involves a server and the user's terminal working together to handle everything from collecting and analyzing user data to generating and delivering advertisements, as well as continuous learning and improvement. The embodiments for carrying out this invention will be described in detail below.

[1451] The server first collects data on hobbies and interests provided by the user. This data is obtained when the user logs into the application and enters information based on their profile and questionnaires. The server then analyzes this collected data. The analysis involves data preprocessing (such as imputing missing values ​​and normalizing the data) and clustering techniques to extract the user's hobbies and thought patterns.

[1452] Furthermore, the server customizes the generative artificial intelligence (AI) based on the extracted hobbies and thought patterns. Specifically, it trains the generative AI model to match the user's preferences and generates personalized advertisements. TensorFlow and PyTorch are used as the generative AI frameworks in this generation process.

[1453] The device delivers customized, generated ads to the user. The generated ads are displayed to the user through a smartphone application. After the user views, clicks, or takes other action on an ad, feedback is sent back to the server. By analyzing this feedback data, the server updates its generative AI model to provide more accurate ads.

[1454] For example, if a user enters the prompt "Tell me about the latest recommended smartphone accessories," the generative artificial intelligence will generate and display advertisements suggesting the most suitable smartphone accessories based on the user's past interests and behavioral patterns. Through this continuous feedback loop, the system can always provide advertisements that are tailored to the user's latest hobbies and interests.

[1455] This system configuration allows users to receive advertisements that perfectly match their interests and preferences at any given time, and also enables advertisers to expect high results.

[1456] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1457] Step 1:

[1458] Users log in to the application and enter initial data, including profile information and data about their hobbies and interests from a questionnaire. This initial data includes the user's gender, age, areas of interest, and preferred product categories. The device then sends this data to the server.

[1459] Input: User profile information, survey data

[1460] Output: Initial data sent to the server

[1461] Step 2:

[1462] The server stores the initial data received from the terminal in a database. Based on this stored data, it performs data preprocessing. Preprocessing includes imputing missing values ​​and normalizing the data.

[1463] Input: Initial data

[1464] Output: Preprocessed data

[1465] Step 3:

[1466] The server analyzes pre-processed data and applies clustering techniques to extract users' interests and thought patterns. Machine learning algorithms (such as KMeans) are used for clustering.

[1467] Input: Preprocessed data

[1468] Output: User clustering results

[1469] Step 4:

[1470] The server customizes generative artificial intelligence (AI) based on the analysis results. Specifically, it trains a generative AI model for a cluster of users and generates personalized ad templates. This process utilizes TensorFlow and PyTorch.

[1471] Input: User clustering results

[1472] Output: Customized generative AI model

[1473] Step 5:

[1474] The device receives a customized generative AI model from the server and displays advertisements generated using that model to the user. When the user takes action on an advertisement (click, purchase, etc.), the device collects the feedback and sends it to the server.

[1475] Input: Customized Generative AI Model

[1476] Output: Displayed ads, user feedback

[1477] Step 6:

[1478] The server analyzes the received feedback data. Based on this analysis, it updates the generated AI model and continuously improves the advertising model.

[1479] Input: User feedback data

[1480] Output: Updated Generative AI Model

[1481] This enables ad delivery using the latest generative AI models that are always adapted to the user's hobbies and interests.

[1482] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1483] System Overview

[1484] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[1485] Program Processing Description

[1486] 1. Collection of user data

[1487] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[1488] Terminal: Sends user-entered profile information and survey responses to the server.

[1489] Server: Receives user information and stores it in a database for later analysis.

[1490] 2. Analysis of hobbies and thought patterns

[1491] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[1492] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[1493] Server: Saves analysis results to the database.

[1494] 3. Customization of Generative Artificial Intelligence

[1495] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[1496] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[1497] Server: Stores customized generative artificial intelligence models in a database.

[1498] 4. Provision to users

[1499] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[1500] Terminal: Notifies the user that generative artificial intelligence is available.

[1501] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[1502] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[1503] 5. Emotion Recognition and Response Optimization

[1504] Terminal: During interaction between the user and generative artificial intelligence, the system uses an emotion engine to recognize the user's emotions. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[1505] Terminal: Sends collected emotion data to the server.

[1506] Server: Analyzes emotional data and optimizes the generative AI's response based on it. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[1507] 6. Continuous learning and improvement

[1508] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[1509] Terminal: Sends collected data to the server.

[1510] Server: Analyzes collected dialogue and emotion data to detect the user's hobbies, interests, and emotional changes.

[1511] Server: Updates generative artificial intelligence models based on new data.

[1512] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[1513] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[1514] Specific example

[1515] Initial setup and usage

[1516] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[1517] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[1518] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[1519] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[1520] Emotion recognition and response optimization

[1521] User: When interacting with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if User A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[1522] Terminal: Send this emotion data to the server.

[1523] Server: Analyzes emotional data and adjusts the generative AI's responses to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[1524] Continuous learning and improvement

[1525] User: As they continue to interact with the generative AI, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[1526] Terminal: Sends new dialogue data and emotion data to the server.

[1527] Server: Analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[1528] Terminal: Re-download the updated generative artificial intelligence model, enabling user A to engage in conversations that respond to new interests and emotions.

[1529] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[1530] The following describes the processing flow.

[1531] Step 1:

[1532] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[1533] Step 2:

[1534] Terminal: Sends user-entered profile information and survey responses to the server.

[1535] Step 3:

[1536] Server: Receives user information and stores it in a database for later analysis.

[1537] Step 4:

[1538] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[1539] Step 5:

[1540] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[1541] Step 6:

[1542] Server: Saves analysis results to the database.

[1543] Step 7:

[1544] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[1545] Step 8:

[1546] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[1547] Step 9:

[1548] Server: Stores customized generative artificial intelligence models in a database.

[1549] Step 10:

[1550] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[1551] Step 11:

[1552] Terminal: Notifies the user that generative artificial intelligence is available.

[1553] Step 12:

[1554] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[1555] Step 13:

[1556] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[1557] Step 14:

[1558] User: As you interact with the generative artificial intelligence, the emotion engine analyzes your emotions in real time. This emotion data is obtained from voice tone, facial expressions, and text content.

[1559] Step 15:

[1560] Terminal: The emotion engine analyzes the emotion data and sends it to the server.

[1561] Step 16:

[1562] Server: Analyzes emotional data and optimizes the responses of the generative artificial intelligence. For example, if the system recognizes that the user is feeling "stressed," the generative AI will generate a calmer, more comforting response.

[1563] Step 17:

[1564] User: Receive appropriate responses through interaction with generative artificial intelligence.

[1565] Step 18:

[1566] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[1567] Step 19:

[1568] Terminal: Sends collected dialogue data and emotion data to the server.

[1569] Step 20:

[1570] Server: Analyzes collected data to detect changes in users' hobbies, interests, and emotions.

[1571] Step 21:

[1572] Server: Updates generative artificial intelligence models based on new data.

[1573] Step 22:

[1574] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[1575] Step 23:

[1576] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[1577] (Example 2)

[1578] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1579] Conventional generative artificial intelligence systems struggled not only to be customized based on the user's hobbies and thoughts, but also to optimize their responses in response to changes in the user's emotions and interests. Furthermore, the lack of sufficient continuous learning and improvement for individual users resulted in a limited user experience. In addition, there was a lack of efficient means to collect initial user data and advanced methods for analyzing the collected data.

[1580] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1581] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, and means for collecting dialogue data and emotional data with the user to continuously train the generative artificial intelligence and optimize its responses. This makes it possible to provide optimal responses in accordance with changes in the user's hobbies, thoughts, and emotions, and to continuously learn and improve.

[1582] "Data on users' hobbies and interests" refers to information that users provide through their profile information and survey responses, indicating their interests and preferences in specific fields such as music, movies, literature, and sports.

[1583] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of generating specific patterns or clusters based on users' hobbies and interests using data cleansing through preprocessing and machine learning algorithms.

[1584] "Means for customizing generative artificial intelligence based on extracted hobby and thought patterns" refers to methods and techniques for optimizing a base generative artificial intelligence model to the hobbies and thoughts of individual users by utilizing the analysis results.

[1585] "Means for delivering customized generative artificial intelligence to a user's terminal" refers to a technology that transfers a generative artificial intelligence model from a server to a user's terminal, making that model available for use in the user's local environment.

[1586] "A means of collecting user interaction data and emotional data to continuously train generative artificial intelligence and optimize responses" refers to a technology that analyzes emotional data obtained from text, voice, facial expressions, etc., collected by generative artificial intelligence during interactions with users, and adjusts and improves the AI's responses to match the user's emotions and needs.

[1587] "Survey methods" refer to question-based interfaces and their implementation methods for efficiently collecting initial user data.

[1588] A "machine learning algorithm" is a mathematical model and method that automatically learns from large amounts of data to perform predictions, classifications, and clustering.

[1589] Modes for carrying out the invention

[1590] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[1591] Hardware and software

[1592] Hardware:

[1593] The user's smartphone or PC

[1594] Cloud Server

[1595] software:

[1596] Database management systems: MySQL, PostgreSQL

[1597] Machine learning libraries: TensorFlow, PyTorch

[1598] Emotion recognition software: Microsoft Azure Emotion API

[1599] Generative artificial intelligence: GPT-3, BERT

[1600] Program Processing Description

[1601] 1. Collection of user data:

[1602] Users create a new account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[1603] The device sends the user's entered profile information and survey responses to the server.

[1604] The server stores the received user information in a database for later analysis.

[1605] 2. Analysis of hobbies and thought patterns:

[1606] The server preprocesses the user information stored in the database, filtering out noise and incomplete data.

[1607] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[1608] The server saves the analysis results to the database.

[1609] 3. Customizing Generative Artificial Intelligence:

[1610] The server selects the most suitable template from the generative artificial intelligence templates based on the analysis results.

[1611] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[1612] The server stores customized generative artificial intelligence models in a database.

[1613] 4. Provision to users:

[1614] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[1615] The device notifies the user that generative artificial intelligence is available.

[1616] The terminal downloads the generative artificial intelligence model from the server.

[1617] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts.

[1618] 5. Emotion Recognition and Response Optimization:

[1619] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[1620] The device sends the collected emotional data to the server.

[1621] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[1622] 6. Continuous learning and improvement:

[1623] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[1624] The device sends the collected data to the server.

[1625] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[1626] The server updates the generative artificial intelligence model based on new data.

[1627] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[1628] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[1629] Specific example

[1630] 1. Initial setup and usage:

[1631] When creating an account, users indicate that they are particularly interested in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[1632] The terminal sends this information to the server, which then uses it to create a profile for user A.

[1633] The server analyzes the collected data and extracts user A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for user A.

[1634] The device downloads a customized AI model, and user A initiates a conversation. For example, if user A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[1635] 2. Emotion recognition and response optimization:

[1636] When a user interacts with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if user A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[1637] The device sends this emotion data to the server.

[1638] The server analyzes emotional data and adjusts the generative artificial intelligence's response to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[1639] 3. Continuous learning and improvement:

[1640] As the user continues to interact with the generative artificial intelligence, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[1641] The device sends new conversational and emotional data to the server.

[1642] The server analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[1643] The device will download the updated generative artificial intelligence model again, enabling user A to engage in conversations that respond to new interests and emotions.

[1644] Example of a prompt

[1645] "Could you recommend some rock bands you've been listening to lately?"

[1646] "Today was really tough."

[1647] "I've recently become interested in jazz. Do you have any recommendations?"

[1648] This allows users to utilize generative artificial intelligence optimized for their own hobbies and thoughts, and furthermore, an emotion engine enables them to receive appropriate responses based on their emotional state. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[1649] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1650] Program processing flow

[1651] User data collection

[1652] Step 1:

[1653] The user launches the application and creates a new account. They enter profile information such as their name, email address, gender, and age, and answer a questionnaire about their hobbies and interests.

[1654] Input: Name, email address, gender, age, and survey responses regarding hobbies and interests.

[1655] Output: Entered profile information and survey data.

[1656] Step 2:

[1657] The device sends the profile information and survey responses entered by the user to the server.

[1658] Input: Profile information and survey responses provided by the user.

[1659] Output: User data sent to the server.

[1660] Step 3:

[1661] The server saves the received user information to the database. It then verifies that the data was saved successfully.

[1662] Input: User data sent from the device.

[1663] Output: User data stored in the database.

[1664] Analysis of hobbies and thought patterns

[1665] Step 4:

[1666] The server preprocesses user information stored in the database, filtering out noise and incomplete data. For example, it performs data imputation and data normalization.

[1667] Input: User data stored in the database.

[1668] Output: Pre-processed, clean data.

[1669] Step 5:

[1670] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. Clustering techniques are then used to classify user trends into multiple categories.

[1671] Input: Pre-processed, clean data.

[1672] Output: A cluster that shows the user's hobbies and thought patterns.

[1673] Step 6:

[1674] The server saves the analysis results to the database. It then verifies that the analysis results were saved correctly.

[1675] Input: Analysis results (user cluster information).

[1676] Output: Analysis results stored in the database.

[1677] Customization of generative artificial intelligence

[1678] Step 7:

[1679] The server selects the optimal generative artificial intelligence template based on the analysis results.

[1680] Input: Analysis results (user cluster information).

[1681] Output: Selected AI template.

[1682] Step 8:

[1683] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. It prioritizes learning data related to the user's specific areas of interest.

[1684] Input: Selected AI template, user-specific data.

[1685] Output: A customized generative AI model.

[1686] Step 9:

[1687] The server saves the customized generative artificial intelligence model to the database. It then verifies that the saving process was successful.

[1688] Input: A customized generative AI model.

[1689] Output: Customized AI model stored in the database.

[1690] Provision to users

[1691] Step 10:

[1692] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[1693] Input: A customized AI model stored in a database.

[1694] Output: Notification that delivery is ready.

[1695] Step 11:

[1696] The device notifies the user that it is ready for distribution and downloads the generative artificial intelligence model.

[1697] Input: Server notification of preparation for delivery.

[1698] Output: Availability notification to the user and downloaded generative AI model.

[1699] Step 12:

[1700] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts. For example, they might ask, "Can you recommend some recent rock bands?"

[1701] Input: Prompt message (user question).

[1702] Output: Response from a generative artificial intelligence.

[1703] Emotion recognition and response optimization

[1704] Step 13:

[1705] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. Emotional data is analyzed from voice tone, facial expressions, and text content.

[1706] Input: User's voice tone, facial expression, and text content.

[1707] Output: Analyzed sentiment data.

[1708] Step 14:

[1709] The device sends the collected emotional data to the server.

[1710] Input: Analyzed sentiment data.

[1711] Output: Sentiment data sent to the server.

[1712] Step 15:

[1713] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the response is changed to something calmer and more comforting.

[1714] Input: Analyzed sentiment data.

[1715] Output: Optimized AI response model.

[1716] Continuous learning and improvement

[1717] Step 16:

[1718] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[1719] Input: User interaction data, sentiment data.

[1720] Output: Collected dialogue data and sentiment data.

[1721] Step 17:

[1722] The device sends the collected data to the server.

[1723] Input: Collected dialogue data and sentiment data.

[1724] Output: Dialogue data and sentiment data sent to the server.

[1725] Step 18:

[1726] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[1727] Input: Collected dialogue data and sentiment data.

[1728] Output: Analyzed data (changes in user hobbies, interests, and emotions).

[1729] Step 19:

[1730] The server updates the generative artificial intelligence model based on new data.

[1731] Input: Newly analyzed data.

[1732] Output: Updated generative AI model.

[1733] Step 20:

[1734] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[1735] Input: An updated generative AI model.

[1736] Output: The AI ​​model redistributed to the user's terminal.

[1737] Step 21:

[1738] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[1739] Input: The user's new prompt message.

[1740] Output: Response from a generative artificial intelligence.

[1741] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[1742] (Application Example 2)

[1743] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1744] Conventional generative artificial intelligence systems often fail to personalize responses based on user preferences and interests, resulting in responses that do not always meet user expectations. Furthermore, their inability to recognize user emotions makes it difficult to provide appropriate responses and services. Moreover, in environments such as autonomous vehicles, there is a growing demand for personalized services that respond to user emotions.

[1745] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing the generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for recognizing the user's emotions, and means for optimizing the response of the generative artificial intelligence based on the recognized user emotions. This enables personalized services tailored to the user's hobbies and interests, as well as appropriate responses in response to their emotions.

[1746] "Data related to hobbies and interests" refers to the preferences and interests that users have in specific fields, and specifically includes information related to preferences such as music, movies, sports, and travel.

[1747] "Methods for analyzing data to extract users' hobbies and thought patterns" refers to algorithms and methods for analyzing collected data to reveal users' hobbies and tendencies in thinking.

[1748] "Means for customizing generative artificial intelligence" refers to methods and technologies for individualizing the artificial intelligence model generated based on the extracted user's hobbies and thought patterns.

[1749] "Means for delivering customized generative artificial intelligence to a user's device" refers to systems and technologies for sending individualized artificial intelligence models to a user's device in a format that the user can use.

[1750] "Means of recognizing user emotions" refers to technologies that analyze a user's voice, facial expressions, text, etc., to understand their emotional state.

[1751] "Means for optimizing the response of generative artificial intelligence based on recognized user emotions" refers to technologies for appropriately adjusting the content and methods of responses generated by artificial intelligence based on emotion recognition.

[1752] The embodiments for carrying out the present invention are described in detail below. First, a method for collecting data on a user's hobbies and interests and customizing an artificial intelligence system based on that data is described. Furthermore, a specific method for providing the user with the most appropriate response through emotion recognition is also described.

[1753] Program Overview

[1754] 1. Collection of user data

[1755] Users create an account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[1756] The device sends the user's entered profile information and survey responses to the server.

[1757] The server stores the received user information in a database for later analysis.

[1758] 2. Analysis of hobbies and thought patterns

[1759] The server preprocesses user information stored in the database, filtering out noise and incomplete data.

[1760] The server uses pre-processed data to apply machine learning algorithms and extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[1761] The server saves the analysis results to a database.

[1762] 3. Customization of Generative Artificial Intelligence

[1763] Based on the analysis results, the server selects the most suitable template from the generative artificial intelligence templates.

[1764] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[1765] The server stores customized generative artificial intelligence models in a database.

[1766] 4. Provision to users

[1767] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[1768] The device notifies the user that generative artificial intelligence is available.

[1769] The user's terminal downloads a generative artificial intelligence model from the server.

[1770] The user then uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and preferences.

[1771] 5. Emotion Recognition and Response Optimization

[1772] During interactions between the user and generative artificial intelligence, the device uses an emotion engine to recognize the user's emotions. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[1773] The device sends the collected emotional data to the server.

[1774] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[1775] 6. Continuous learning and improvement

[1776] The device collects dialogue data and emotional data each time the user interacts with the generative artificial intelligence.

[1777] The device sends the collected data to the server.

[1778] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[1779] The server updates the generative artificial intelligence model based on new data.

[1780] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[1781] Users utilize updated generative artificial intelligence to engage in conversations and information retrieval that cater to their new interests, hobbies, and emotions.

[1782] Hardware and software to be used

[1783] hardware

[1784] Infotainment system in autonomous vehicles

[1785] Camera and microphone: For emotion recognition inside the vehicle.

[1786] Vehicle communication module: Data communication with the server

[1787] software

[1788] Server-side: Apache Kafka (data streaming), TensorFlow (machine learning models), PostgreSQL (database)

[1789] Autonomous vehicle terminal: Custom Emotion Recognition Engine, in-device application (Android Auto, CarPlay compatible)

[1790] Specific example

[1791] For example, if a user in an autonomous vehicle asks, "What are some rock bands you recommend lately?", the generative artificial intelligence and emotion recognition engine will process a prompt sentence like the following.

[1792] Example of a prompt:

[1793] "We've identified that the user is currently experiencing stress. Their hobby is rock music. Please suggest some rock bands."

[1794] Based on this prompt, the generative artificial intelligence is optimized to respond with, "Thank you for your hard work. To relax, why not listen to some songs by a recommended rock band?"

[1795] In this way, the present invention can utilize generative artificial intelligence optimized for the user's hobbies and thoughts, and furthermore, an emotion engine enables the user to receive appropriate responses according to their emotional state. This provides a more personalized and advanced service that responds to the user's needs and emotions.

[1796] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1797] Step 1:

[1798] Users create an account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests. The input data is in JSON format and includes both profile information and questionnaire responses.

[1799] Step 2:

[1800] The device sends user-entered profile information and survey responses to the server. Specifically, it transfers data to the server using an HTTP POST request, which the server receives via a REST API endpoint. Input is the transmission of user data in JSON format, and output is a notification that the data has been saved to the database.

[1801] Step 3:

[1802] The server stores the received user information in a database for later analysis. PostgreSQL is used as the database system, and user data is inserted into the appropriate tables. Input is user data in JSON format, and output is the record stored in the database.

[1803] Step 4:

[1804] The server preprocesses user information stored in the database, filtering out noise and incomplete data. Specifically, it performs tasks such as imputing missing values ​​and removing outliers. The input is the user data from the database, and the output is a preprocessed, clean dataset.

[1805] Step 5:

[1806] The server applies machine learning algorithms to preprocessed data to extract users' hobbies and thought patterns. This process uses clustering techniques (e.g., K-means clustering) to classify user tendencies into multiple categories. The input is a preprocessed dataset, and the output is clustered user data.

[1807] Step 6:

[1808] The server saves the analysis results to a database. This allows them to be used to customize subsequent generative artificial intelligence. The input is clustered data, and the output is the clustering results stored in the database.

[1809] Step 7:

[1810] The server selects the optimal template from the generative artificial intelligence templates based on the analysis results. Specifically, it selects the best template for a particular cluster. The input is the clustering result, and the output is the selected template.

[1811] Step 8:

[1812] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, data related to the user's specific areas of interest is prioritized for training. The input consists of user-specific data and templates, and the output is a customized generative AI model.

[1813] Step 9:

[1814] The server stores customized generative artificial intelligence models in a database. The input is the generative artificial intelligence model, and the output is the model stored in the database.

[1815] Step 10:

[1816] The server prepares to deliver a customized generative artificial intelligence model to the user's terminal. The input is the model stored in the database, and the output is the server's notification that it is ready to deliver.

[1817] Step 11:

[1818] The terminal notifies the user that the generative artificial intelligence is available. The notification method uses the terminal's notification system. The input is a notification that distribution is ready, and the output is a notification to the user.

[1819] Step 12:

[1820] The user's terminal downloads the generative artificial intelligence model from the server. Specifically, it uses an HTTP GET request. The input is the download link from the server, and the output is the generative artificial intelligence model stored locally.

[1821] Step 13:

[1822] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts. The input is the generative AI model, and the output is the conversation content and search results.

[1823] Step 14:

[1824] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content. The input is dialogue data, and the output is emotion data.

[1825] Step 15:

[1826] The device sends the collected emotional data to the server. The input is emotional data, and the output is the transmission of data to the server.

[1827] Step 16:

[1828] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on it. Specifically, it adjusts the response to be more comforting and relaxing. The input is emotional data, and the output is the optimized response.

[1829] Step 17:

[1830] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence. The input is the dialogue data and emotion data, and the output is the collected data.

[1831] Step 18:

[1832] The terminal sends the collected data to the server. The input is the collected data, and the output is the transmission of data to the server.

[1833] Step 19:

[1834] The server analyzes collected dialogue and emotion data to detect the user's hobbies, interests, and emotional changes. The input is the collected data, and the output is the analysis results.

[1835] Step 20:

[1836] The server updates the generative artificial intelligence model based on new data. The input is the analysis results, and the output is the updated generative artificial intelligence model.

[1837] Step 21:

[1838] The server redistributes the updated generative artificial intelligence model to the user's terminal. The input is the updated generative artificial intelligence model, and the output is a notification that the redistribution is ready.

[1839] Step 22:

[1840] Users utilize an updated generative artificial intelligence (AI) system to engage in conversations and information retrieval that address new interests, hobbies, and emotions. The input is the updated AI model, and the output is new conversation content and search results.

[1841] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1842] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1843] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1844] [Fourth Embodiment]

[1845] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1846] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1847] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1848] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1849] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1850] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1851] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1852] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1853] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1854] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[1855] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1856] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1857] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1858] System Overview

[1859] This invention is a system for providing a generative artificial intelligence customized specifically to the user's hobbies and interests. This system involves a server and a user's terminal working together to optimize the generative artificial intelligence based on the user's interests and hobbies.

[1860] Program Processing Description

[1861] 1. Collection of user data

[1862] User: Create a new account and enter initial data such as profile information and hobbies. This includes answering questionnaires and selecting genres of interest (music, movies, sports, etc.).

[1863] Terminal: Sends user-entered information and survey responses to the server.

[1864] Server: Receives user data and stores it in a database for later analysis.

[1865] 2. Analysis of hobbies and thought patterns

[1866] Server: Preprocesses the collected data and filters out noise and incomplete data.

[1867] Server: Based on pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used to classify user trends into multiple categories.

[1868] Server: Saves analysis results to the database.

[1869] 3. Customization of Generative Artificial Intelligence

[1870] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[1871] Server: Trains the selected template with user-specific data to create a user-specific generative artificial intelligence. This customization process includes prioritizing the training of data related to specific areas of interest.

[1872] Server: Stores customized generative artificial intelligence models in a database.

[1873] 4. Provision to users

[1874] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[1875] Terminal: Notifies the user that generative artificial intelligence is available.

[1876] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[1877] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[1878] 5. Continuous learning and improvement

[1879] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[1880] Terminal: Sends collected conversation data to the server.

[1881] Server: Analyzes collected data to detect changes in the user's hobbies and interests.

[1882] Server: Updates generative artificial intelligence models based on new data.

[1883] Server: Redistributes the updated model to user terminals.

[1884] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[1885] Specific example

[1886] Initial setup and usage

[1887] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[1888] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[1889] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[1890] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[1891] Continuous learning and improvement

[1892] User: As I continue to interact with the generative AI, I begin to develop an interest in "jazz." I start asking more questions about jazz.

[1893] Terminal: Sends new dialogue data to the server.

[1894] Server: Analyzes collected data to detect new interests (jazz) of user A. Updates the generative artificial intelligence model based on this information.

[1895] Device: Re-download the updated AI model so that User A can engage in conversations tailored to their new interests.

[1896] This system allows users to utilize generative artificial intelligence optimized for their hobbies and interests, and because it is continuously updated, they can always receive services that cater to their latest interests.

[1897] The following describes the processing flow.

[1898] Step 1:

[1899] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[1900] Step 2:

[1901] Terminal: Sends user-entered profile information and survey responses to the server.

[1902] Step 3:

[1903] Server: Receives user information and stores it in a database for later analysis.

[1904] Step 4:

[1905] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[1906] Step 5:

[1907] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[1908] Step 6:

[1909] Server: Saves analysis results to the database.

[1910] Step 7:

[1911] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[1912] Step 8:

[1913] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[1914] Step 9:

[1915] Server: Stores customized generative artificial intelligence models in a database.

[1916] Step 10:

[1917] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[1918] Step 11:

[1919] Terminal: Notifies the user that generative artificial intelligence is available.

[1920] Step 12:

[1921] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[1922] Step 13:

[1923] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[1924] Step 14:

[1925] Terminal: Collects dialogue data each time the user interacts with the generative artificial intelligence.

[1926] Step 15:

[1927] Terminal: Sends collected conversation data to the server.

[1928] Step 16:

[1929] Server: Analyzes collected dialogue data to detect changes in the user's hobbies and interests.

[1930] Step 17:

[1931] Server: Updates generative artificial intelligence models based on new data.

[1932] Step 18:

[1933] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[1934] Step 19:

[1935] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that cater to new interests and hobbies.

[1936] (Example 1)

[1937] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1938] Currently, many generative artificial intelligence models are specialized in providing general information, but they have the challenge of not being able to provide services tailored to a user's individual interests and thoughts. Furthermore, they struggle to quickly adapt to changes in user interests and hobbies. In addition, there is a lack of effective methods for providing the most suitable generative artificial intelligence model for individual users.

[1939] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1940] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for filtering the collected data to remove noise and incomplete data, means for analyzing the filtered data to extract the user's hobbies and thought patterns, means for selecting a generative artificial intelligence model based on the extracted hobbies and thought patterns, means for training and customizing the selected generative artificial intelligence model with user-specific data, means for delivering the customized generative artificial intelligence model to the user's terminal, and means for collecting interaction data with the user to continuously train and update the generative artificial intelligence model. This makes it possible to provide a generative artificial intelligence model tailored to the user's hobbies and thoughts, and to respond quickly to changes in the user's interests.

[1941] A "user" refers to an individual or legal entity that utilizes the system, provides data related to their hobbies and interests, and receives services based on a customized generative artificial intelligence model.

[1942] A "server" refers to a device or system that receives and stores data sent by users, and performs data analysis, customization, distribution, and updating of generative artificial intelligence models.

[1943] A "terminal" refers to an electronic device used by a user, such as a computer, smartphone, or tablet, that communicates with a server to transmit data, download generative artificial intelligence models, and engage in dialogue.

[1944] "Data collection" refers to the process of obtaining information about users' hobbies and interests through surveys and other means.

[1945] "Filtering" refers to the process of removing noise and incomplete data from collected data, preparing it for analysis.

[1946] "Data analysis" refers to the process of analyzing pre-processed data using machine learning algorithms and other methods to extract users' hobbies and thought patterns.

[1947] A "generative artificial intelligence model" refers to an artificial intelligence template designed to interact with users and provide information using natural language processing technology.

[1948] "Customization" refers to the process of training a generative artificial intelligence model with user-specific data to optimize it for individual users.

[1949] "Distribution" refers to the process of transferring a customized generative artificial intelligence model from a server to a user's device.

[1950] "Dialogue data" refers to data that records user questions and the responses of generative artificial intelligence models, including the content of the dialogue with the model.

[1951] "Continuous learning" refers to the process of improving the performance of a generative artificial intelligence model by retraining and updating it based on user interaction data.

[1952] This invention relates to a system that provides a generative artificial intelligence model specialized according to the user's hobbies and interests. This system can provide a service optimized to the user's interests and hobbies by analyzing data provided by the user, customizing the generative artificial intelligence based on that analysis, and continuously learning and updating it.

[1953] Specifically, this system includes the following elements:

[1954] User data collection

[1955] User: Create a new account and enter basic information such as name, age, gender, and place of residence, as well as answer a questionnaire about hobbies and interests. For example, to the question "What genre of music do you like?", answer "Rock".

[1956] Terminal: This terminal sends basic information and survey responses entered by the user to the server. This communication uses the secure protocol HTTPS.

[1957] Server: Receives user data and stores it in a database management system (e.g., MySQL) for later analysis.

[1958] Analysis of hobbies and thought patterns

[1959] Server: Preprocesses the collected data, filtering out noise and incomplete data. This preprocessing is done using data cleaning tools (e.g., OpenRefine).

[1960] Server: Based on filtered data, it applies machine learning algorithms (e.g., K-means clustering) to extract users' hobbies and thought patterns.

[1961] Server: Saves analysis results to the database.

[1962] Customization of generative artificial intelligence

[1963] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates (e.g., GPT-3).

[1964] Server: Trains and customizes the selected generative AI template with user-specific data. For example, if the user is interested in "rock" music, a dataset in that genre is prepared and the template is fine-tuned.

[1965] Server: Stores customized generative artificial intelligence models in a database.

[1966] Provision to users

[1967] Server: Prepares the server to deliver customized generative artificial intelligence models to user terminals. This preparation includes packaging the data to be delivered.

[1968] Terminal: Notify the user that generative artificial intelligence is available. The notification will use a push notification service (e.g., Firebase Cloud Messaging).

[1969] Terminal: Downloads generative artificial intelligence models from the server. HTTPS is used as the data transfer protocol.

[1970] User: Using the downloaded generative AI, initiate conversations and information searches that match the user's interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[1971] Continuous learning and improvement

[1972] Terminal: Every time a user interacts with the generative artificial intelligence, the interaction data is collected. For example, a question like "What are some recommended new action movies?" and its answer are stored as data.

[1973] Terminal: Sends collected conversation data to the server. The HTTPS protocol is used for transmission.

[1974] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[1975] Server: Updates the generative artificial intelligence model based on new data. Prepares a new dataset and performs fine-tuning again.

[1976] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[1977] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[1978] Example of a prompt

[1979] "What rock bands do you recommend these days?"

[1980] "What are some of the latest action movies you would recommend?"

[1981] "Please tell me about some famous jazz songs."

[1982] As described above, the present invention provides a generative artificial intelligence model that is tailored to the user's hobbies and thoughts, and can continuously adapt to changes in the user's interests.

[1983] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1984] Step 1: Collecting User Data

[1985] User: Create a new account and enter basic information such as name, age, gender, and place of residence. Also, answer a questionnaire about your interests and hobbies. For example, answer "Rock" to the question, "What is your favorite music genre?"

[1986] Input: Basic information, survey responses

[1987] Output: Basic information and survey response dataset

[1988] Terminal: The terminal sends the basic information and survey responses entered by the user to the server via the HTTPS protocol. Specifically, when the submit button on the input form is pressed, the data is transferred to the server.

[1989] Input: User input data

[1990] Output: Data transfer to the server

[1991] Server: Receives user data and stores it in a database management system (e.g., MySQL). User information and survey results are stored in the database in an organized manner.

[1992] Input: Transferred user data

[1993] Output: User data stored in the database

[1994] Step 2: Analysis of hobbies and thought patterns

[1995] Server: Preprocesses collected data, filtering out noise and incomplete data. Uses data cleaning tools (e.g., OpenRefine) to remove unnecessary data and prepare it for analysis.

[1996] Input: Raw data

[1997] Output: Filtered, clean data

[1998] Server: Applies machine learning algorithms (e.g., K-means clustering) to preprocessed data to extract users' hobbies and thought patterns. This algorithm classifies users' hobbies into multiple categories.

[1999] Input: Clean data

[2000] Output: User's hobbies and thought patterns

[2001] Server: Saves analysis results to a database. Analysis results are categorized by user and used for customization at a later stage.

[2002] Input: Analysis results

[2003] Output: Analysis results saved in the database

[2004] Step 3: Customizing Generative Artificial Intelligence

[2005] Server: Based on the analysis results, it selects the most suitable template from generative artificial intelligence templates (e.g., GPT-3). It uses a template selection algorithm to choose the most appropriate template.

[2006] Input: Analysis results

[2007] Output: Selected generative AI templates

[2008] Server: Trains the selected generative AI template with user-specific data. Specifically, it prepares a dataset related to the user's areas of interest and uses that data to fine-tune the template.

[2009] Input: Generative artificial intelligence template and user data

[2010] Output: Customized generative artificial intelligence model

[2011] Server: Stores customized generative artificial intelligence models in a database. Models are stored with version control using a model management system (e.g., MLflow).

[2012] Input: Customized generative artificial intelligence model

[2013] Output: Generative artificial intelligence models stored in the database

[2014] Step 4: Provision to users

[2015] Server: Prepares to deliver customized generative artificial intelligence models to user terminals. Specifically, it packages the delivery data and configures the delivery server.

[2016] Input: Customized generative artificial intelligence model

[2017] Output: Ready for delivery

[2018] Terminal: Notifies the user that generative artificial intelligence is available. This notification uses a push notification service (e.g., Firebase Cloud Messaging).

[2019] Input: Notification content

[2020] Output: Notification to the user

[2021] Terminal: Downloads generative AI models from the server. Secure HTTPS is used as the data transfer protocol.

[2022] Input: Download link

[2023] Output: Downloaded generative artificial intelligence model

[2024] User: Using the downloaded generative AI, initiate conversations and information searches that match their interests and thoughts. For example, by asking "What are some recommended rock bands lately?", the generative AI will provide an appropriate answer.

[2025] Input: User's question (prompt)

[2026] Output: Answer from a generative artificial intelligence

[2027] Step 5: Continuous learning and improvement

[2028] Terminal: Each time a user interacts with a generative artificial intelligence, the dialogue data is collected. This dialogue data includes the content of the questions and answers.

[2029] Input: Dialogue between the user and the generative artificial intelligence.

[2030] Output: Collected dialogue data

[2031] Terminal: Sends collected conversation data to the server. Secure HTTPS protocol is used for transmission.

[2032] Input: Dialogue data

[2033] Output: Sending data to the server

[2034] Server: Analyzes collected dialogue data to detect changes in the user's interests and hobbies. This analysis uses natural language processing algorithms (e.g., TF-IDF).

[2035] Input: Dialogue data

[2036] Output: Analysis results

[2037] Server: Updates the generative artificial intelligence model based on new data. The update includes a process of retraining the model based on newly collected data.

[2038] Input: New dataset

[2039] Output: Updated generative artificial intelligence model

[2040] Server: Redistributes the updated generative artificial intelligence model to the user's terminal. The redistribution procedure is the same as the initial one.

[2041] Input: Updated generative artificial intelligence model

[2042] Output: Ready for delivery

[2043] User: Utilizing the updated generative AI, the user engages in conversations and information searches that cater to new hobbies and interests. For example, by asking "Please tell me about famous jazz songs," the new generative AI will respond.

[2044] Input: New question from the user (prompt)

[2045] Output: A new answer from generative artificial intelligence

[2046] (Application Example 1)

[2047] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[2048] Traditional advertising delivery systems have not adequately personalized ads using data on users' hobbies and interests, making it difficult to improve the effectiveness and accuracy of advertisements. Furthermore, providing users with ads that resonate with them requires continuous data collection and updates, but efficient methods for doing so have been lacking. Therefore, there is a need for a system that generates and continuously updates ads optimized for users' hobbies and interests.

[2049] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[2050] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, means for collecting conversational data with the user to continuously train the generative artificial intelligence, means for delivering the generated customized advertisements, and means for collecting and analyzing user feedback to update the advertising model. This makes it possible to generate advertisements optimized for the user's hobbies and interests and to continuously improve them.

[2051] "Data related to users' hobbies and interests" refers to information collected to clarify the hobbies and interests that users exhibit, and specifically includes genres, preferences, and past behavioral history.

[2052] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of processing collected user data and using analytical techniques and algorithms to identify users' interests and patterns.

[2053] "Methods for customizing generative artificial intelligence" refers to technologies that optimize generative artificial intelligence for each user based on their hobbies and thought patterns, thereby creating personalized generative models.

[2054] "Means for delivering generated customized advertisements" refers to the technologies and methods for sending and displaying customized advertisements on a user's device.

[2055] "Methods for collecting and analyzing user feedback to update advertising models" refers to the process of collecting user reactions and evaluations to advertisements, analyzing that data, and improving the advertising model to match the latest user tastes and interests.

[2056] "Means for collecting dialogue data and continuously training generative artificial intelligence" refers to technologies and methods for collecting data from user interactions with generative artificial intelligence and using that data to improve and update the generative model.

[2057] "Means for customizing generative artificial intelligence based on analysis results" refers to technologies and methods for optimizing the operation of generative artificial intelligence based on analyzed hobbies and thought patterns.

[2058] This invention is a system that generates and delivers advertisements customized based on the user's hobbies and interests. This system involves a server and the user's terminal working together to handle everything from collecting and analyzing user data to generating and delivering advertisements, as well as continuous learning and improvement. The embodiments for carrying out this invention will be described in detail below.

[2059] The server first collects data on hobbies and interests provided by the user. This data is obtained when the user logs into the application and enters information based on their profile and questionnaires. The server then analyzes this collected data. The analysis involves data preprocessing (such as imputing missing values ​​and normalizing the data) and clustering techniques to extract the user's hobbies and thought patterns.

[2060] Furthermore, the server customizes the generative artificial intelligence (AI) based on the extracted hobbies and thought patterns. Specifically, it trains the generative AI model to match the user's preferences and generates personalized advertisements. TensorFlow and PyTorch are used as the generative AI frameworks in this generation process.

[2061] The device delivers customized, generated ads to the user. The generated ads are displayed to the user through a smartphone application. After the user views, clicks, or takes other action on an ad, feedback is sent back to the server. By analyzing this feedback data, the server updates its generative AI model to provide more accurate ads.

[2062] For example, if a user enters the prompt "Tell me about the latest recommended smartphone accessories," the generative artificial intelligence will generate and display advertisements suggesting the most suitable smartphone accessories based on the user's past interests and behavioral patterns. Through this continuous feedback loop, the system can always provide advertisements that are tailored to the user's latest hobbies and interests.

[2063] This system configuration allows users to receive advertisements that perfectly match their interests and preferences at any given time, and also enables advertisers to expect high results.

[2064] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[2065] Step 1:

[2066] Users log in to the application and enter initial data, including profile information and data about their hobbies and interests from a questionnaire. This initial data includes the user's gender, age, areas of interest, and preferred product categories. The device then sends this data to the server.

[2067] Input: User profile information, survey data

[2068] Output: Initial data sent to the server

[2069] Step 2:

[2070] The server stores the initial data received from the terminal in a database. Based on this stored data, it performs data preprocessing. Preprocessing includes imputing missing values ​​and normalizing the data.

[2071] Input: Initial data

[2072] Output: Preprocessed data

[2073] Step 3:

[2074] The server analyzes pre-processed data and applies clustering techniques to extract users' interests and thought patterns. Machine learning algorithms (such as KMeans) are used for clustering.

[2075] Input: Preprocessed data

[2076] Output: User clustering results

[2077] Step 4:

[2078] The server customizes generative artificial intelligence (AI) based on the analysis results. Specifically, it trains a generative AI model for a cluster of users and generates personalized ad templates. This process utilizes TensorFlow and PyTorch.

[2079] Input: User clustering results

[2080] Output: Customized generative AI model

[2081] Step 5:

[2082] The device receives a customized generative AI model from the server and displays advertisements generated using that model to the user. When the user takes action on an advertisement (click, purchase, etc.), the device collects the feedback and sends it to the server.

[2083] Input: Customized Generative AI Model

[2084] Output: Displayed ads, user feedback

[2085] Step 6:

[2086] The server analyzes the received feedback data. Based on this analysis, it updates the generated AI model and continuously improves the advertising model.

[2087] Input: User feedback data

[2088] Output: Updated Generative AI Model

[2089] This enables ad delivery using the latest generative AI models that are always adapted to the user's hobbies and interests.

[2090] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[2091] System Overview

[2092] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[2093] Program Processing Description

[2094] 1. Collection of user data

[2095] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[2096] Terminal: Sends user-entered profile information and survey responses to the server.

[2097] Server: Receives user information and stores it in a database for later analysis.

[2098] 2. Analysis of hobbies and thought patterns

[2099] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[2100] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[2101] Server: Saves analysis results to the database.

[2102] 3. Customization of Generative Artificial Intelligence

[2103] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[2104] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[2105] Server: Stores customized generative artificial intelligence models in a database.

[2106] 4. Provision to users

[2107] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[2108] Terminal: Notifies the user that generative artificial intelligence is available.

[2109] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[2110] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[2111] 5. Emotion Recognition and Response Optimization

[2112] Terminal: During interaction between the user and generative artificial intelligence, the system uses an emotion engine to recognize the user's emotions. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[2113] Terminal: Sends collected emotion data to the server.

[2114] Server: Analyzes emotional data and optimizes the generative AI's response based on it. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[2115] 6. Continuous learning and improvement

[2116] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[2117] Terminal: Sends collected data to the server.

[2118] Server: Analyzes collected dialogue and emotion data to detect the user's hobbies, interests, and emotional changes.

[2119] Server: Updates generative artificial intelligence models based on new data.

[2120] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[2121] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[2122] Specific example

[2123] Initial setup and usage

[2124] User: When creating an account, they indicated a particular interest in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[2125] Terminal: This information is sent to the server, which then uses it to create a profile for user A.

[2126] Server: Analyzes collected data to extract User A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for User A.

[2127] Terminal: Download a customized AI model, and User A begins a conversation. For example, if User A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[2128] Emotion recognition and response optimization

[2129] User: When interacting with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if User A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[2130] Terminal: Send this emotion data to the server.

[2131] Server: Analyzes emotional data and adjusts the generative AI's responses to be gentle and comforting. For example, it might generate a response like, "You must be tired. To help you relax, why not listen to some music by a rock band I recommend?"

[2132] Continuous learning and improvement

[2133] User: As they continue to interact with the generative AI, they begin to develop an interest in "jazz." Furthermore, emotional data is accumulated as their stress levels decrease.

[2134] Terminal: Sends new dialogue data and emotion data to the server.

[2135] Server: Analyzes the collected data and reflects user A's new interests and emotional changes in the model.

[2136] Terminal: Re-download the updated generative artificial intelligence model, enabling user A to engage in conversations that respond to new interests and emotions.

[2137] This system allows users to utilize generative artificial intelligence optimized for their hobbies and thoughts, and further enables them to receive appropriate responses based on their emotional state through an emotion engine. This results in a more personalized and advanced service that responds to the user's needs and emotions.

[2138] The following describes the processing flow.

[2139] Step 1:

[2140] User: Create a new account and enter your profile information (name, email address, gender, age, etc.). Also, answer a questionnaire about your hobbies and interests.

[2141] Step 2:

[2142] Terminal: Sends user-entered profile information and survey responses to the server.

[2143] Step 3:

[2144] Server: Receives user information and stores it in a database for later analysis.

[2145] Step 4:

[2146] Server: Preprocesses user information stored in the database, filtering out noise and incomplete data.

[2147] Step 5:

[2148] Server: Using pre-processed data, machine learning algorithms are applied to extract users' hobbies and thought patterns. Clustering techniques are used during this process to classify user tendencies into multiple categories.

[2149] Step 6:

[2150] Server: Saves analysis results to the database.

[2151] Step 7:

[2152] Server: Based on the analysis results, it selects the most suitable template from the generative artificial intelligence templates.

[2153] Step 8:

[2154] Server: The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes learning data related to the user's specific areas of interest.

[2155] Step 9:

[2156] Server: Stores customized generative artificial intelligence models in a database.

[2157] Step 10:

[2158] Server: Prepares to deliver customized generative artificial intelligence models to user terminals.

[2159] Step 11:

[2160] Terminal: Notifies the user that generative artificial intelligence is available.

[2161] Step 12:

[2162] Terminal: The user's terminal downloads the generative artificial intelligence model from the server.

[2163] Step 13:

[2164] User: Use the downloaded generative artificial intelligence to initiate conversations and information retrieval that match your interests and thoughts.

[2165] Step 14:

[2166] User: As you interact with the generative artificial intelligence, the emotion engine analyzes your emotions in real time. This emotion data is obtained from voice tone, facial expressions, and text content.

[2167] Step 15:

[2168] Terminal: The emotion engine analyzes the emotion data and sends it to the server.

[2169] Step 16:

[2170] Server: Analyzes emotional data and optimizes the responses of the generative artificial intelligence. For example, if the system recognizes that the user is feeling "stressed," the generative AI will generate a calmer, more comforting response.

[2171] Step 17:

[2172] User: Receive appropriate responses through interaction with generative artificial intelligence.

[2173] Step 18:

[2174] Terminal: Each time a user interacts with a generative artificial intelligence, it collects dialogue data and emotional data.

[2175] Step 19:

[2176] Terminal: Sends collected dialogue data and emotion data to the server.

[2177] Step 20:

[2178] Server: Analyzes collected data to detect changes in users' hobbies, interests, and emotions.

[2179] Step 21:

[2180] Server: Updates generative artificial intelligence models based on new data.

[2181] Step 22:

[2182] Server: Redistributes the updated generative artificial intelligence model to the user's terminal.

[2183] Step 23:

[2184] User: Use updated generative artificial intelligence to engage in conversations and information retrieval that respond to new interests, hobbies, and emotions.

[2185] (Example 2)

[2186] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[2187] Conventional generative artificial intelligence systems struggled not only to be customized based on the user's hobbies and thoughts, but also to optimize their responses in response to changes in the user's emotions and interests. Furthermore, the lack of sufficient continuous learning and improvement for individual users resulted in a limited user experience. In addition, there was a lack of efficient means to collect initial user data and advanced methods for analyzing the collected data.

[2188] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[2189] In this invention, the server includes means for collecting data on the user's hobbies and interests, means for analyzing the collected data to extract the user's hobbies and thought patterns, means for customizing a generative artificial intelligence based on the extracted hobbies and thought patterns, means for delivering the customized generative artificial intelligence to the user's terminal, and means for collecting dialogue data and emotional data with the user to continuously train the generative artificial intelligence and optimize its responses. This makes it possible to provide optimal responses in accordance with changes in the user's hobbies, thoughts, and emotions, and to continuously learn and improve.

[2190] "Data on users' hobbies and interests" refers to information that users provide through their profile information and survey responses, indicating their interests and preferences in specific fields such as music, movies, literature, and sports.

[2191] "Methods for analyzing collected data to extract users' hobbies and thought patterns" refers to the process of generating specific patterns or clusters based on users' hobbies and interests using data cleansing through preprocessing and machine learning algorithms.

[2192] "Means for customizing generative artificial intelligence based on extracted hobby and thought patterns" refers to methods and techniques for optimizing a base generative artificial intelligence model to the hobbies and thoughts of individual users by utilizing the analysis results.

[2193] "Means for delivering customized generative artificial intelligence to a user's terminal" refers to a technology that transfers a generative artificial intelligence model from a server to a user's terminal, making that model available for use in the user's local environment.

[2194] "A means of collecting user interaction data and emotional data to continuously train generative artificial intelligence and optimize responses" refers to a technology that analyzes emotional data obtained from text, voice, facial expressions, etc., collected by generative artificial intelligence during interactions with users, and adjusts and improves the AI's responses to match the user's emotions and needs.

[2195] "Survey methods" refer to question-based interfaces and their implementation methods for efficiently collecting initial user data.

[2196] A "machine learning algorithm" is a mathematical model and method that automatically learns from large amounts of data to perform predictions, classifications, and clustering.

[2197] Modes for carrying out the invention

[2198] This invention is a system that combines a generative artificial intelligence customized specifically for the user's hobbies and thoughts with an emotion engine that recognizes the user's emotions. In this system, the server and the user's terminal work together to optimize the generative artificial intelligence based on the user's interests, hobbies, and emotions.

[2199] Hardware and software

[2200] Hardware:

[2201] The user's smartphone or PC

[2202] Cloud Server

[2203] software:

[2204] Database management systems: MySQL, PostgreSQL

[2205] Machine learning libraries: TensorFlow, PyTorch

[2206] Emotion recognition software: Microsoft Azure Emotion API

[2207] Generative artificial intelligence: GPT-3, BERT

[2208] Program Processing Description

[2209] 1. Collection of user data:

[2210] Users create a new account and enter profile information such as their name, email address, gender, and age. They also answer a questionnaire about their hobbies and interests.

[2211] The device sends the user's entered profile information and survey responses to the server.

[2212] The server stores the received user information in a database for later analysis.

[2213] 2. Analysis of hobbies and thought patterns:

[2214] The server preprocesses the user information stored in the database, filtering out noise and incomplete data.

[2215] The server applies machine learning algorithms to pre-processed data to extract users' hobbies and thought patterns. During this process, clustering techniques are used to classify user tendencies into multiple categories.

[2216] The server saves the analysis results to the database.

[2217] 3. Customizing Generative Artificial Intelligence:

[2218] The server selects the most suitable template from the generative artificial intelligence templates based on the analysis results.

[2219] The server trains the selected template with user-specific data to create a user-specific generative artificial intelligence. During this process, it prioritizes training the AI ​​with data related to the user's specific areas of interest.

[2220] The server stores customized generative artificial intelligence models in a database.

[2221] 4. Provision to users:

[2222] The server prepares to deliver the customized generative artificial intelligence model to the user's terminal.

[2223] The device notifies the user that generative artificial intelligence is available.

[2224] The terminal downloads the generative artificial intelligence model from the server.

[2225] The user uses the downloaded generative artificial intelligence to initiate conversations and information searches that match their interests and thoughts.

[2226] 5. Emotion Recognition and Response Optimization:

[2227] The device uses an emotion engine to recognize the user's emotions during interactions with the generative artificial intelligence. This emotion data is analyzed from factors such as voice tone, facial expressions, and text content.

[2228] The device sends the collected emotional data to the server.

[2229] The server analyzes emotional data and optimizes the generative artificial intelligence's response based on that analysis. For example, if the user is feeling stressed, the generative AI will generate a calmer, more comforting response.

[2230] 6. Continuous learning and improvement:

[2231] The device collects dialogue data and emotion data each time the user interacts with the generative artificial intelligence.

[2232] The device sends the collected data to the server.

[2233] The server analyzes the collected dialogue and emotion data to detect the user's hobbies, interests, and changes in their emotions.

[2234] The server updates the generative artificial intelligence model based on new data.

[2235] The server redistributes the updated generative artificial intelligence model to the user's terminal.

[2236] Users will use the updated generative artificial intelligence to engage in conversations and information retrieval that respond to their new interests, hobbies, and emotions.

[2237] Specific example

[2238] 1. Initial setup and usage:

[2239] When creating an account, users indicate that they are particularly interested in "music" and "movies." They prefer "rock" as a music genre and "action" as a movie genre.

[2240] The terminal sends this information to the server, which then uses it to create a profile for user A.

[2241] The server analyzes the collected data and extracts user A's hobbies and thought patterns. Based on this, it customizes a generative artificial intelligence to create an AI model specifically for user A.

[2242] The device downloads a customized AI model, and user A initiates a conversation. For example, if user A asks, "What are some recommended rock bands lately?", the generative artificial intelligence will suggest appropriate rock bands.

[2243] 2. Emotion recognition and response optimization:

[2244] When a user interacts with a generative artificial intelligence, the emotion engine determines that the user is experiencing "stress." For example, if user A says, "Today was tough," the emotion engine recognizes stress based on the tone of voice and text content.

[2245] The device sends this emotion data to the server.

[2246] ...

Claims

1. A means of collecting data about users' hobbies and interests, A method for analyzing collected data to extract users' hobbies and thought patterns, A means of customizing a generative artificial intelligence based on extracted hobbies and thought patterns, A means of delivering customized generative artificial intelligence to the user's device, A means of collecting user interaction data and continuously training a generative artificial intelligence, A system that includes this.

2. The system according to claim 1, including a means for obtaining initial user data through a questionnaire.

3. The system according to claim 1, comprising means for clustering users' hobbies and interests using a machine learning algorithm.

Images