system

Abstract

Provide a system. [Solution] means for collecting profile information and behavioral data of users based on their actions; A means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer; A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile; means for transmitting the generated optimized submission document to a user's terminal; means for collecting user reaction data after said transmission, analyzing said reaction data, and feeding back the data to subsequent submitted documents; A system including:

Description

[Technical Field]

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

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

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

[0004] In modern digital marketing, providing information optimized for each individual customer is important for improving customer satisfaction and encouraging purchasing behavior. However, many companies still send uniform messages to all customers, without achieving individual optimization. This inefficient information provision can fail to attract customer attention, resulting in decreased sales and a damaged brand image. Furthermore, manual optimization tasks consume time and effort, hindering efficient marketing activities. [Means for solving the problem]

[0005] This system includes means for collecting profile information and behavioral data based on user actions, means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer, means for generating posts optimized for each customer using generative artificial intelligence based on the generated customer profile, means for sending the generated optimized posts to the user's device, and means for collecting user reaction data after sending, analyzing the reaction data, and feeding the data back into subsequent posts. This makes it possible to automatically provide information optimized for each customer, thereby realizing efficient marketing activities.

[0006] "User action" refers to operations or actions that a user performs on the system, such as sending a message or making a purchase.

[0007] "Profile information" refers to data including a user's personal information and attribute information, such as name, gender, and age.

[0008] "Behavioral data" refers to information about a user's activities on the system, such as their past operation history and purchase history.

[0009] "Means for collection" refers to devices or programs for acquiring user profile information and behavioral data and transmitting them to a server.

[0010] "Means of analysis" refers to machine learning algorithms and programs used to process collected data and analyze user interests and behavioral patterns.

[0011] A "customer profile" refers to a data set of each customer's interests and behavioral patterns generated based on analysis results.

[0012] "Generative AI" refers to an AI system that uses natural language processing technology to automatically generate posts that are optimized for users.

[0013] "Optimized Post" refers to a personalized, tailored message generated based on each customer's customer profile.

[0014] The "transmitting means" refers to a device or program that delivers the generated optimized posted document to the user's terminal.

[0015] "Reaction data" refers to data such as a user's actions after receiving a posted document, such as a reply message or purchasing behavior.

[0016] "Feedback means" refers to devices or programs that reanalyze collected response data and reflect it in the generation of subsequent submitted documents. [Brief explanation of the drawings]

[0017] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram illustrating a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0018] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0019] First, the terms used in the following description will be explained.

[0020] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0021] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0022] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0023] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0024] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0025] [First embodiment]

[0026] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0027] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0028] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0029] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0030] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0031] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0032] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0033] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0034] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0035] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0036] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0037] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0038] The present invention is a system for automatically providing information optimized for each individual user, and is realized by the following processing steps.

[0039] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and all related data to the server.

[0040] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[0041] Based on the results of the above analysis, the server creates a profile for each customer. This allows each user's detailed interests and behavioral patterns to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products."

[0042] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A.

[0043] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[0044] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0045] Finally, the server collects user response data (e.g., reply messages and purchasing behavior) after distribution and analyzes it again. Based on this response data, feedback is provided for generating subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests.

[0046] The processing flow will be explained below.

[0047] Step 1:

[0048] Detecting user actions

[0049] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[0050] Step 2:

[0051] Data collection

[0052] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[0053] Step 3:

[0054] Sending data

[0055] The device sends the collected data to the server. For example, it sends user A's profile information and behavioral data in JSON format to the server.

[0056] Step 4:

[0057] Receiving and storing data

[0058] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[0059] Step 5:

[0060] Data Preprocessing

[0061] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[0062] Step 6:

[0063] Customer Segmentation

[0064] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[0065] Step 7:

[0066] Generate customer profiles

[0067] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[0068] Step 8:

[0069] Optimized submission document generation request

[0070] The server requests the AI ​​to generate a post optimized for each customer profile, for example, by instructing it to use a template related to beauty products.

[0071] Step 9:

[0072] Generate optimized submission documents

[0073] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0074] Step 10:

[0075] Review and correct submitted documents

[0076] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0077] Step 11:

[0078] Preparing to send submitted documents

[0079] The server prepares to send the optimized post to each user's official LINE account.

[0080] Step 12:

[0081] Submitting a submission

[0082] The server sends the posted document to the user's LINE device. For example, it sends a posted document optimized for User A's LINE device.

[0083] Step 13:

[0084] Reaction data collection

[0085] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[0086] Step 14:

[0087] Reanalysis of reaction data

[0088] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[0089] Step 15:

[0090] Reflecting feedback

[0091] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[0092] Through these steps, the system automatically provides information optimized for each user, achieving continuous optimization.

[0093] Example 1

[0094] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0095] Conventional information provision systems have difficulty providing information optimized to users' interests. There is also a risk that posted documents containing inappropriate language may be distributed. Furthermore, the lack of a mechanism for effectively providing feedback on users' reactions reduces the quality and accuracy of the information provided. There is a need to solve these issues.

[0096] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0097] In this invention, the server includes: means for collecting profile information and behavioral data of a user based on the user's actions; means for receiving the collected data and cleansing it by correcting inconsistent data and filling in missing data; means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer; means for generating a post optimized for each customer using generative artificial intelligence based on the generated customer profile; means for verifying the generated optimized post, checking for inappropriate language, and manually correcting it; means for transmitting the generated optimized post to the user's device; and means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding it back into subsequent posts. This enables the provision of information optimized to the user's interests, reduces the risk of inappropriate language, and improves the quality and accuracy of information provided by effectively feeding back user reactions.

[0098] A "user action" is a series of operations that a user performs on the system, such as sending a message or purchasing a product.

[0099] "Profile information" refers to personal information such as a user's name, gender, and age.

[0100] "Behavioral data" refers to a record of a user's past behavior, such as message history and purchase history.

[0101] "Cleansing" is the process of organizing and formatting data to correct inconsistent data and fill in missing data.

[0102] "Analysis" is the process of deriving user behavior patterns and estimated interests based on collected data.

[0103] A "customer profile" is a data set generated based on analyzed data that contains detailed interests and behavioral patterns for each user.

[0104] "Generative AI" is an AI technology that generates posts optimized for each customer based on their customer profile.

[0105] A "posted document" is a message or piece of text that is sent to a user.

[0106] "Response data" refers to data such as replies and purchasing behavior of users in response to received posted documents.

[0107] "Feedback" refers to the process of reflecting collected response data in the generation of subsequent submitted documents.

[0108] MODE FOR CARRYING OUT THE INVENTION

[0109] This invention is a system that automatically provides information optimized for each individual user, and specifically, it operates in cooperation with a server, a terminal, and a user. The system of this invention operates as follows.

[0110] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (such as name, gender, and age) and behavioral data (such as message history and purchase history), consolidates this data, and sends it to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and the associated profile information and behavioral data to the server.

[0111] The server processes the data received from the device. Specifically, the server cleanses the received data, corrects inconsistent data, and fills in missing data. The server then converts the data into an analyzable format (for example, JSON format). Next, the server uses a machine learning algorithm (using Tensorflow (registered trademark) or PyTorch, etc.) to analyze the user's behavioral patterns and estimated interests. From past message history, the server learns that a particular user has a high interest in beauty products.

[0112] Based on the analysis results, the server generates a profile for each customer. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products." By generating this profile, the user's detailed interests and behavioral patterns are compiled into a single data set.

[0113] Next, the server requests a generative artificial intelligence (AI) to generate an optimized post based on the customer profile. The generative AI uses natural language processing technology such as OpenAI's GPT-4 (registered trademark) to generate a message optimized for each user. For example, a message might be generated for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out." The following is an example of a prompt sentence that can be used:

[0114] "Please create a promotional message for a new beauty product for User A, a woman in her 30s who has a high interest in beauty products."

[0115] The generated submission is validated by the server and checked for inappropriate language, which is then manually corrected.

[0116] Finally, the server sends the optimized post to each user's device, which then displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0117] After distribution, the server collects user response data (such as reply messages and purchasing behavior) and analyzes it again. This response data is used as feedback for generating subsequent posts, enabling even more optimized information to be provided. In this way, the system can continuously provide optimal information tailored to the user's interests.

[0118] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0119] Step 1:

[0120] Users can send messages to official accounts or purchase products.

[0121] Input: User action (sending a message or purchasing a product)

[0122] Output: User action detection data

[0123] Specific operation: A user sends a message to the official LINE account asking, "Do you have any new beauty products?"

[0124] Step 2:

[0125] The device detects this user's actions and collects the user's profile information and behavioral data.

[0126] Input: User action detection data

[0127] Output: Collected user profile information and behavioral data

[0128] Specific operation: The device detects the message and collects user A's name, gender, age, past message history, and purchase history.

[0129] Step 3:

[0130] The data collected by the device is consolidated and sent to the server.

[0131] Input: Collected user profile information and behavioral data

[0132] Output: Consolidated data sent from the device to the server

[0133] Specific operation: The terminal compiles the collected data of user A and sends it to the server.

[0134] Step 4:

[0135] The server cleanses the incoming data, corrects inconsistent data, and fills in missing data.

[0136] Input: Integrated data sent from the device

[0137] Output: Cleansed data

[0138] Specific behavior: The server will correct inconsistencies in the data (e.g. missing names) and complete them as needed.

[0139] Step 5:

[0140] The server converts the cleansed data into a parsable format.

[0141] Input: Cleansed data

[0142] Output: Data in a parsable format (e.g. JSON)

[0143] What happens: The server converts the data into a parsable format, such as JSON.

[0144] Step 6:

[0145] The server uses machine learning algorithms to analyze users' behavioral patterns and estimated interests.

[0146] Input: Data in a parsable format

[0147] Output: Analysis results of user behavior patterns and estimated interests

[0148] Specific operation: The server uses TensorFlow to analyze User A's interest in beauty products based on their past message history.

[0149] Step 7:

[0150] The server generates a customer profile based on the analysis results.

[0151] Input: Analysis results of user behavior patterns and estimated interests

[0152] Output: Generated customer profile

[0153] Specific operation: The server creates a profile for "a woman in her 30s with a strong interest in beauty products."

[0154] Step 8:

[0155] The server requests the generative artificial intelligence to generate an optimized submission document based on the customer profile.

[0156] Input: Generated customer profile

[0157] Output: Prompt sentence requested by the generative AI

[0158] Specific operation: The server requests a generative artificial intelligence (e.g., OpenAI GPT-4) to "generate the optimal message for User A, a woman in her 30s who has a high interest in beauty products."

[0159] Step 9:

[0160] Generative AI generates optimized submission documents.

[0161] Input: Prompt sentence requested by the generative artificial intelligence

[0162] Output: The generated optimized submission document

[0163] What it does: The generative AI generates a message saying, "The latest beauty products have arrived! Check them out at a limited-time discount."

[0164] Step 10:

[0165] The server validates the generated optimized submission document and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0166] Input: Generated optimized submission document

[0167] Output: Validated and corrected submission document

[0168] Specific behavior: The server validates the generated message, checks for errors in specific product information, etc., and manually corrects them.

[0169] Step 11:

[0170] The server transmits the optimized submitted document to the user's terminal.

[0171] Input: Validated and corrected submitted document

[0172] Output: Posted document sent to terminal

[0173] Specific operation: The server sends the optimized submission document to the terminal.

[0174] Step 12:

[0175] The terminal displays the posted document to the user.

[0176] Input: Posted document sent to terminal

[0177] Output: The submitted document as displayed to the user

[0178] Specific operation: The device displays a message on User A's LINE message saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0179] Step 13:

[0180] The server collects user response data (reply messages and purchasing behavior) after delivery and analyzes it again.

[0181] Input: User response data

[0182] Output: Parsed reaction data

[0183] Specific operation: The server collects and analyzes User A's reply messages and purchasing behavior.

[0184] Step 14:

[0185] The server provides feedback for the next submission.

[0186] Input: Parsed reaction data

[0187] Output: Feedback information reflected in the next submission generation

[0188] Specific operation: The server optimizes the process of generating the next submitted document based on the analysis results.

[0189] (Application example 1)

[0190] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0191] On modern online shopping sites, diverse users each have different interests, so providing information in a single location reduces user satisfaction. Another issue is the difficulty of effectively analyzing collected user behavior data and making personalized, appropriate product recommendations.

[0192] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0193] In this invention, the server includes a means for sending user action data to the server using a REST API, a means for generating an optimized message by making a request to the generative AI based on the analyzed data, and a means for sending the generated message again from the server to the terminal, thereby enabling personalized product recommendations based on each user's interests.

[0194] "User actions" refer to specific operations that users perform on the system, such as purchasing, browsing, sending messages, etc.

[0195] "Profile information" refers to basic personal information such as a user's name, gender, and age.

[0196] "Behavioral data" refers to data relating to specific behaviors of users, such as their message history and purchase history.

[0197] "Server" means a central management system for collecting, analyzing, generating and transmitting user profile information and behavioral data.

[0198] "Generative AI" refers to artificial intelligence techniques used to generate optimized messages based on user profiles.

[0199] An "optimized submission" is a document created by generative artificial intelligence to suit a specific user based on the user's profile.

[0200] A "user terminal" is a device, such as a smartphone or computer, that a user uses to receive information.

[0201] A "REST API" is a standard interface for sending data to a server.

[0202] "Machine learning algorithms" are data analysis techniques used to analyze user behavior patterns and inferred interests.

[0203] "Response data" refers to responses and purchasing behavior of users in response to information from the system.

[0204] The present invention is a system for a mail-order site that provides personalized information based on a user's interests. This system is realized by the following method and apparatus.

[0205] First, when a user purchases or browses a product on an online shopping site, their behavioral data (e.g., product ID, category, etc.) is automatically collected. This collection is performed by the user's device issuing a request and sending the data to the server via a REST API.

[0206] The server cleans the received data and uses machine learning algorithms to analyze user behavior patterns and inferred interests. To do this, the server uses Python programs and libraries such as TensorFlow and Scikit-learn.

[0207] Based on the analysis results, the server creates a detailed profile for each user, including the user's interests and behavioral patterns. For example, a profile might be generated that describes a woman in her 30s with a strong interest in beauty products.

[0208] Next, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3 (registered trademark)) to generate a post optimized for each user. An example of a prompt is "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message."

[0209] The generated post is verified by the server and checked for inappropriate content, after which the generated optimized post is sent back to the user's device and displayed to the user.

[0210] As described above, personalized information based on the user's interests can be provided. This system enables optimized product recommendations for each user, improving the user experience. As a specific example of use, if a woman in her 30s frequently purchases beauty products, a message will be displayed saying, "Hello, we have new beauty products that suit your tastes. They are on special discount, so please check them out."

[0211] In this way, user satisfaction at the online shopping site can be improved.

[0212] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0213] Step 1:

[0214] When a user purchases or browses products on an online shopping site, their device collects their behavioral data. This data includes information such as product ID, category, and time. The collected data is sent to the server via a REST API. This input data is output as a request to the server.

[0215] Step 2:

[0216] The server receives the behavioral data sent from the device and performs data cleansing. Specifically, it corrects inconsistent data and complements missing data. The cleansed data is output as a dataset for analysis.

[0217] Step 3:

[0218] The server then runs machine learning algorithms on the cleansed data to analyze user behavior patterns and inferred interests. This analysis uses libraries such as Scikit-learn and TensorFlow. The analysis results are output as a user profile.

[0219] Step 4:

[0220] Based on the generated user profile, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate an optimized post. For example, the prompt "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message" is used as the source data, and the generated recommendation message is output.

[0221] Step 5:

[0222] The server validates the generated optimized submission document, checking for inappropriate language, and outputs the validated document as the final message.

[0223] Step 6:

[0224] The server sends a final message to the user's terminal, which displays the message and the user receives the recommended information. The user then checks the recommended product information.

[0225] Step 7:

[0226] When a user responds to a recommended message (for example, by purchasing a product or replying to a message), the response data is collected again. The collected response data is sent to the server via a REST API. The sent data is output as feedback data.

[0227] Step 8:

[0228] The server analyzes the collected response data and provides feedback for the next submission. This feedback enables the generation of more appropriate recommendation messages in the future. The feedback results are output as the next profile analysis data.

[0229] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0230] This invention is a system that automatically provides information optimized for each user and further optimizes the information by recognizing the user's emotions, and is realized by the following processing steps.

[0231] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device sends this message and all related data to the server.

[0232] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[0233] In addition to the analysis data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. For example, when a user sends a message saying, "Do you have any new beauty products?", the emotion engine infers emotions such as "excitement" from the user's text.

[0234] Based on the analysis results and emotion estimation results, the server generates a profile for each customer. This allows each user's detailed interests and behavioral patterns, as well as their emotional state, to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, high interest in beauty products, current emotional state: excited."

[0235] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A. The tone and content of the message are also further customized based on the estimation results of the emotion engine.

[0236] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[0237] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0238] Finally, the server collects and analyzes user response data (e.g., reply messages and purchasing behavior) after the message is posted. Based on this response data, feedback is provided for the generation of subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests and emotional state.

[0239] The processing flow will be explained below.

[0240] Step 1:

[0241] Detecting user actions

[0242] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[0243] Step 2:

[0244] Data collection

[0245] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[0246] Step 3:

[0247] Sending data

[0248] The device sends the collected profile information and behavioral data to the server. For example, the device sends User A's profile information and behavioral data in JSON format to the server.

[0249] Step 4:

[0250] Receiving and storing data

[0251] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[0252] Step 5:

[0253] Data Preprocessing

[0254] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[0255] Step 6:

[0256] Customer Segmentation

[0257] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[0258] Step 7:

[0259] Generate customer profiles

[0260] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[0261] Step 8:

[0262] Emotion recognition data collection

[0263] The device sends the user's message content and behavioral data to the emotion engine. For example, a message sent by user A saying "Do you have any new beauty products?" is sent to the emotion engine.

[0264] Step 9:

[0265] Emotion estimation

[0266] The server uses the emotion engine to estimate the user's emotion from the received message. For example, the emotion engine estimates the emotion "excitement" from User A's message.

[0267] Step 10:

[0268] Optimized submission document generation request

[0269] The server requests the generative AI to generate a post optimized for each customer based on their profile and the emotion estimation results. For example, it instructs the AI ​​to use a template for users who are interested in and excited about beauty products.

[0270] Step 11:

[0271] Generate optimized submission documents

[0272] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0273] Step 12:

[0274] Review and correct submitted documents

[0275] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0276] Step 13:

[0277] Preparing to send submitted documents

[0278] The server prepares to send the optimized post to the user's device. For example, it prepares to send a message to User A's LINE device.

[0279] Step 14:

[0280] Submitting a submission

[0281] The server sends the posted document to the user's LINE device. For example, it delivers a posted document optimized for User A's LINE device.

[0282] Step 15:

[0283] Reaction data collection

[0284] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[0285] Step 16:

[0286] Reanalysis of reaction data

[0287] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[0288] Step 17:

[0289] Reflecting feedback

[0290] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[0291] Through these steps, the system not only automatically provides information optimized for each user, but also takes emotions into account to provide even more effective information.

[0292] Example 2

[0293] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0294] Conventional information delivery systems have difficulty providing optimized information that takes into account the individual interests and emotional state of each user. This has made improving user experience and engagement a challenge. Another issue is the cumbersome process of verifying and manually correcting generated messages, making it difficult to provide information efficiently.

[0295] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0296] In this invention, the server includes means for collecting profile information and behavioral data based on user actions, means for cleansing the collected data, means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer, means for estimating emotions from the content of the user's messages, means for generating a customer profile based on the analysis results and the emotion estimation results, means for generating a post optimized for each customer using generative artificial intelligence, means for verifying and manually correcting the generated optimized post, means for sending the generated optimized post to the user's device, and means for collecting user reaction data after transmission, analyzing the reaction data, and providing feedback for subsequent posts. This makes it possible to provide optimal information based on each user's detailed interests and emotional state.

[0297] "User actions" refers to a series of operations or actions performed by a user, such as sending a message to an official account or purchasing a product.

[0298] "Profile information" refers to information that can be used to personally identify a user, such as the user's name, gender, and age.

[0299] "Behavioral data" refers to data related to a user's past operations and behavior, such as message history and purchase history.

[0300] "Cleansing" refers to the process of correcting inconsistent data, filling in missing data, and converting data into an analyzable format.

[0301] "Analysis" refers to the processing of collected data to identify user behavioral patterns and interests.

[0302] "Estimated interests" refers to estimating the interests and concerns that a user is likely to have based on past behavioral data.

[0303] A "behavioral pattern" refers to a series of actions or patterns of operations that a user exhibits during a specific period of time.

[0304] "Emotion estimation" refers to the process of inferring the emotion contained in a message sent by a user from its content.

[0305] A "customer profile" refers to a data set that compiles a user's interests, behavioral patterns, and emotional state.

[0306] "Generative AI" refers to AI technology that automatically generates optimized messages and text based on given data.

[0307] "Optimized posts" refer to messages that are customized for individual users based on their customer profile.

[0308] "Response data" refers to data such as reply messages sent by users in response to messages received and subsequent purchasing behavior.

[0309] "Feedback" refers to the process of analyzing collected response data and using it to improve the quality of subsequent submissions.

[0310] This invention is a system that automatically provides information optimized for each user and further optimizes it by recognizing the user's emotions. The system is mainly composed of a server, a terminal, and a user. Details of each function and their specific implementation methods are explained below.

[0311] Basic configuration

[0312] The server plays a central role in data cleansing, data analysis, emotion estimation, and message generation. The hardware used is a high-performance data processing server, and the software used includes Hadoop, Spark, IBM Watson®, Microsoft® Azure® emotion recognition service, Scikit-learn, and TensorFlow. The generative artificial intelligence used is OpenAI GPT-4.

[0313] The terminal is responsible for detecting user actions and sending data to the server. This can be implemented on popular devices such as smartphones and PCs.

[0314] Users take action on the system and receive information optimized based on their response.

[0315] Data collection and transmission

[0316] When a user sends a message to the official account asking, "Do you have any new beauty products?", the device detects this action. The device collects the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) and sends them to the server in real time.

[0317] Data Cleansing and Analysis

[0318] The server cleanses the data it receives. This process includes correcting inconsistent data and filling in missing data, for example by filling in missing fields with appropriate default values.

[0319] The server then analyzes the cleansed data using Hadoop or Spark. Machine learning algorithms (Scikit-learn or TensorFlow) are used to analyze user behavior patterns and interests. The analysis results include information on which beauty products a particular user is frequently interested in.

[0320] Emotion estimation

[0321] The server uses IBM Watson or Microsoft Azure emotion recognition services to infer emotions from the content of a user's message. For example, it infers the emotion "excitement" from the message "Do you have any new beauty products?"

[0322] Customer profile generation

[0323] The server generates a detailed profile of the user based on the analysis results and emotion estimation results. For example, it creates a profile such as "Female in her 30s, highly interested in beauty products, current emotional state: excited."

[0324] Message Creation and Delivery

[0325] The server requests the OpenAI GPT-4 to generate an optimized message based on the generated profile. For example, the request is made with the following prompt:

[0326] Customer Profile:

[0327] Name: User A

[0328] Gender: Female

[0329] Age: 30

[0330] Interests: Beauty products

[0331] Current emotional state: Excited

[0332] message:

[0333] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[0334] The generated message is validated by the server to check for inappropriate language and, if necessary, manually corrected.

[0335] Finally, the optimized message is sent to the user's device, which displays it to the user. The user reacts, which triggers the next cycle of data collection and analysis.

[0336] Reaction data collection and feedback

[0337] The server collects user responses (reply messages and purchasing behavior) after distribution and analyzes this data. Based on the analysis results, the quality of subsequent submissions can be improved, resulting in even more optimized information provision.

[0338] This process allows the system to continually provide optimal information tailored to the user's interests and emotional state.

[0339] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0340] Step 1:

[0341] User Actions and Data Collection

[0342] A user sends a message to your official account saying, "Are there any new beauty products?"

[0343] The device detects this action and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history).

[0344] Input: User messages, profile information, behavioral data

[0345] Output: Collected dataset

[0346] Specific operation: Software in the device begins collecting data when the user sends a message.

[0347] Step 2:

[0348] Data transmission

[0349] The terminal transmits the collected data set to the server.

[0350] Input: Collected dataset

[0351] Output: The dataset sent to the server

[0352] Specific operation: The device uploads data to a server via the Internet.

[0353] Step 3:

[0354] Data Cleansing

[0355] The server cleanses the received dataset, specifically correcting inconsistent data and filling in missing data.

[0356] Input: The dataset sent to the server

[0357] Output: A cleansed dataset

[0358] Specific operations: Use data processing engines such as Hadoop and Spark to check data integrity and convert it into a unified format.

[0359] Step 4:

[0360] Data analysis

[0361] The server uses the cleansed data to analyze user behavior patterns and interests using machine learning algorithms such as Scikit-learn and TensorFlow.

[0362] Input: Cleansed dataset

[0363] Output: Analysis results (user behavior patterns, estimated interests)

[0364] Specific behavior: Apply machine learning models to training data to derive quantitative user behavior patterns and interests.

[0365] Step 5:

[0366] Emotion estimation

[0367] The server uses IBM Watson and Microsoft Azure emotion recognition services to estimate emotions from the content of the user's message.

[0368] Input: Message content, behavioral data

[0369] Output: Emotion estimation result

[0370] What it does: The emotion recognition engine analyzes text and extracts specific emotions (e.g., excitement, joy, sadness).

[0371] Step 6:

[0372] Customer profile generation

[0373] The server generates a detailed profile of the user based on the analysis results and emotion estimation results.

[0374] Input: Analysis results, emotion estimation results

[0375] Output: Customer profile

[0376] Specific operation: The analysis results are integrated with emotional data, and a detailed profile customized for each user is stored in a database.

[0377] Step 7:

[0378] Requesting Message Creation

[0379] The server requests the generative artificial intelligence (OpenAI GPT-4) to generate an optimized message based on the generated customer profile. The server instructs the generation using a prompt.

[0380] Input: Customer profile, prompt statement

[0381] Output: Generated optimization messages

[0382] Specific behavior: Provide the following prompt to the generative AI:

[0383] Customer Profile:

[0384] Name: User A

[0385] Gender: Female

[0386] Age: 30

[0387] Interests: Beauty products

[0388] Current emotional state: Excited

[0389] message:

[0390] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[0391] Step 8:

[0392] Message verification and correction

[0393] The server validates generated messages for inappropriate language and manually corrects them as necessary.

[0394] Input: The generated message

[0395] Output: Validated and corrected message

[0396] What it does: It performs text filtering and human operator review.

[0397] Step 9:

[0398] Message Delivery

[0399] The server sends the optimized message to the user's terminal.

[0400] Input: Validated and corrected message

[0401] Output: Message sent to the terminal

[0402] Specific operation: The server sends a message to the terminal via the Internet.

[0403] Step 10:

[0404] Reaction data collection and feedback

[0405] After delivery, the server collects and analyzes the user's reactions (reply messages and purchasing behavior), and reflects the results in the generation of subsequent messages.

[0406] Input: User response data

[0407] Output: Feedback data

[0408] What it does: The reaction data is stored in a database and used to retrain the machine learning model.

[0409] (Application example 2)

[0410] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0411] Current online shopping sites generally only provide product recommendations and promotional messages to users, but these messages are not optimized to reflect the interests and emotional state of individual users, making it difficult to improve the user experience. Another issue is that collected user data cannot be effectively utilized, making it difficult to provide individually optimized information in real time. These issues need to be resolved.

[0412] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0413] In this invention, the server includes means for collecting profile information and behavioral data of a user based on the user's actions, means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer, means for generating a post optimized for each customer using a generative artificial intelligence based on the generated customer profile, means for transmitting the generated optimized post to the user's terminal, means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding the data back into subsequent posts, means for providing optimized product recommendations and promotional messages in real time using a smart device, means for generating a prompt sentence for the generative artificial intelligence to generate the optimized message based on the analysis results, and means for generating the optimized message using the generative artificial intelligence using the generated prompt sentence, thereby enabling the provision of information optimized for the interests and emotional state of each individual user.

[0414] "User action" refers to the operations or inputs that a user makes to the system.

[0415] "Profile Information" refers to basic personal information such as a user's name, gender, and age.

[0416] "Behavioral data" refers to activity data such as a user's message history and purchase history.

[0417] "Analysis" refers to the process of analyzing collected data and extracting meaningful information.

[0418] "Estimated interests" refers to topics or areas in which a user is estimated to be particularly interested.

[0419] A "behavioral pattern" refers to a series of behavioral tendencies based on a user's past behavior.

[0420] A "customer profile" refers to a data set that compiles detailed interests and behavioral patterns for each user.

[0421] "Generative artificial intelligence" refers to AI that uses natural language processing technology to generate optimal messages and documents.

[0422] "Optimized posts" refer to messages that are individually customized based on a user's profile information and behavioral data.

[0423] "Terminal" refers to a device through which a user receives information, such as a smartphone or tablet.

[0424] "Response data" refers to the user's response to a posted document, such as a reply message or purchasing behavior.

[0425] "Smart device" refers to a device with advanced functionality that can connect to the Internet.

[0426] "Real-time" refers to near-instant processing and response.

[0427] A "prompt sentence" refers to an input sentence that causes a generative artificial intelligence to generate an optimized message.

[0428] This invention realizes a system for providing product recommendations and promotional messages optimized for each user on an online shopping site. This system is realized by collecting and analyzing information based on user actions, and generating and sending messages using generative artificial intelligence. Specific embodiments of this system are described below.

[0429] The user's smartphone or smart device functions as the terminal for this system. When a user searches for a product or sends a message, this action is detected by the terminal, and the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) are collected. The collected data is sent to the server in real time.

[0430] The server uses machine learning libraries like Python and TensorFlow to cleanse the incoming data and convert it into an analyzable format. Then, machine learning algorithms are used to analyze user behavior patterns and inferred interests. For example, past messaging history may indicate that a particular user is interested in beauty products.

[0431] Based on the analysis results and emotion estimation by the emotion engine, the server generates a detailed profile for each user. The generated profile is used to generate prompts for a generative AI (e.g., OpenAI's GPT-3). Based on the generated prompts, the generative AI generates optimized messages.

[0432] An example of a generated prompt is:

[0433] Ms. A (30 years old, female) is interested in a new beauty product. She is currently in an excited emotional state and has previously purchased beauty creams and face masks. Please generate the best message for her based on the following conditions:

[0434] New beauty products have arrived! Check them out for limited time discounts.

[0435] The server validates the generated optimization message and manually checks it for inappropriate content. After validation, the optimization message is sent to the user's smart device and displayed to the user in real time.

[0436] For example, if a user sends a message saying, "Do you have any new beauty products?", a customized recommendation message such as, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" will be displayed in real time on the user's device.

[0437] This invention enables the provision of information optimized for the interests and emotional state of each individual user, improving the user experience. Furthermore, by effectively utilizing collected user data, it is possible to provide individually optimized information in real time.

[0438] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0439] Step 1:

[0440] A user searches for a product or sends a message on a smartphone or smart device. The input data is the user's action, search query, and sent message. The device detects this user action as it is being performed and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history). The collected data is sent to a server in real time.

[0441] Step 2:

[0442] The server cleanses the incoming data and converts it into an analyzable format. The input data is the profile information and behavioral data collected in step 1. Inconsistent data is corrected and missing data is filled in. The cleansed data is obtained as the output.

[0443] Step 3:

[0444] The server uses machine learning algorithms to analyze the user's behavioral patterns and estimated interests. The input is cleansed data. Using machine learning libraries such as TensorFlow, the server analyzes the user's past messaging history and purchasing history to estimate specific interests. The analysis results are the user's estimated interests and behavioral patterns.

[0445] Step 4:

[0446] In addition to the analyzed data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. The input is the user's message content and behavioral data. Natural language processing technology is used to analyze emotions from the message and infer emotions such as "excitement." The emotion estimation result is obtained as the output.

[0447] Step 5:

[0448] The server generates a detailed profile for each user. The inputs are behavioral patterns, estimated interests, and estimated emotions. Each user's interests, behavioral patterns, and emotional state are compiled into a single dataset. The generated user profile is output.

[0449] Step 6:

[0450] The server generates a prompt for the generative AI to generate an optimization message. The input is the user profile. A generative AI such as OpenAI's GPT-3 is used to generate the prompt. The generated prompt is output.

[0451] Step 7:

[0452] The server uses the generated prompt sentence to generate an optimized message using a generative AI. The input is the generated prompt sentence. A generative AI model is used to generate the optimized message. The output is the optimized message.

[0453] Step 8:

[0454] The server validates the generated optimization message and checks for inappropriate expressions. The input is the generated optimization message. If any manual corrections are required, they are made. The optimized message without inappropriate expressions is output.

[0455] Step 9:

[0456] The server sends the verified optimization message to the user's terminal. The input is the verified optimization message. The message is displayed on the user's terminal in real time.

[0457] Step 10:

[0458] The server collects user reaction data after sending a message and feeds it back into subsequent messages. The input is user reaction data (reply messages, purchasing behavior, etc.). The collected reaction data is analyzed and used to generate optimized messages for subsequent messages. The results of the data analysis after feedback are output.

[0459] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0460] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0461] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0462] [Second embodiment]

[0463] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0464] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0465] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0466] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0467] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0468] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0469] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0470] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0471] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0472] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0473] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0474] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0475] The present invention is a system for automatically providing information optimized for each individual user, and is realized by the following processing steps.

[0476] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and all related data to the server.

[0477] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[0478] Based on the results of the above analysis, the server creates a profile for each customer. This allows each user's detailed interests and behavioral patterns to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products."

[0479] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A.

[0480] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[0481] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0482] Finally, the server collects user response data (e.g., reply messages and purchasing behavior) after distribution and analyzes it again. Based on this response data, feedback is provided for generating subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests.

[0483] The processing flow will be explained below.

[0484] Step 1:

[0485] Detecting user actions

[0486] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[0487] Step 2:

[0488] Data collection

[0489] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[0490] Step 3:

[0491] Sending data

[0492] The device sends the collected data to the server. For example, it sends user A's profile information and behavioral data in JSON format to the server.

[0493] Step 4:

[0494] Receiving and storing data

[0495] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[0496] Step 5:

[0497] Data Preprocessing

[0498] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[0499] Step 6:

[0500] Customer Segmentation

[0501] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[0502] Step 7:

[0503] Generate customer profiles

[0504] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[0505] Step 8:

[0506] Optimized submission document generation request

[0507] The server requests the AI ​​to generate a post optimized for each customer profile, for example, by instructing it to use a template related to beauty products.

[0508] Step 9:

[0509] Generate optimized submission documents

[0510] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0511] Step 10:

[0512] Review and correct submitted documents

[0513] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0514] Step 11:

[0515] Preparing to send submitted documents

[0516] The server prepares to send the optimized post to each user's official LINE account.

[0517] Step 12:

[0518] Submitting a submission

[0519] The server sends the posted document to the user's LINE device. For example, it sends a posted document optimized for User A's LINE device.

[0520] Step 13:

[0521] Reaction data collection

[0522] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[0523] Step 14:

[0524] Reanalysis of reaction data

[0525] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[0526] Step 15:

[0527] Reflecting feedback

[0528] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[0529] Through these steps, the system automatically provides information optimized for each user, achieving continuous optimization.

[0530] Example 1

[0531] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0532] Conventional information provision systems have difficulty providing information optimized to users' interests. There is also a risk that posted documents containing inappropriate language may be distributed. Furthermore, the lack of a mechanism for effectively providing feedback on users' reactions reduces the quality and accuracy of the information provided. There is a need to solve these issues.

[0533] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0534] In this invention, the server includes: means for collecting profile information and behavioral data of a user based on the user's actions; means for receiving the collected data and cleansing it by correcting inconsistent data and filling in missing data; means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer; means for generating a post optimized for each customer using generative artificial intelligence based on the generated customer profile; means for verifying the generated optimized post, checking for inappropriate language, and manually correcting it; means for transmitting the generated optimized post to the user's device; and means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding it back into subsequent posts. This enables the provision of information optimized to the user's interests, reduces the risk of inappropriate language, and improves the quality and accuracy of information provided by effectively feeding back user reactions.

[0535] A "user action" is a series of operations that a user performs on the system, such as sending a message or purchasing a product.

[0536] "Profile information" refers to personal information such as a user's name, gender, and age.

[0537] "Behavioral data" refers to a record of a user's past behavior, such as message history and purchase history.

[0538] "Cleansing" is the process of organizing and formatting data to correct inconsistent data and fill in missing data.

[0539] "Analysis" is the process of deriving user behavior patterns and estimated interests based on collected data.

[0540] A "customer profile" is a data set generated based on analyzed data that contains detailed interests and behavioral patterns for each user.

[0541] "Generative AI" is an AI technology that generates posts optimized for each customer based on their customer profile.

[0542] A "posted document" is a message or piece of text that is sent to a user.

[0543] "Response data" refers to data such as replies and purchasing behavior of users in response to received posted documents.

[0544] "Feedback" refers to the process of reflecting collected response data in the generation of subsequent submitted documents.

[0545] MODE FOR CARRYING OUT THE INVENTION

[0546] This invention is a system that automatically provides information optimized for each individual user, and specifically, it operates in cooperation with a server, a terminal, and a user. The system of this invention operates as follows.

[0547] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (such as name, gender, and age) and behavioral data (such as message history and purchase history), consolidates this data, and sends it to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and the associated profile information and behavioral data to the server.

[0548] The server processes the data received from the device. Specifically, the server cleanses the received data, corrects inconsistent data, and fills in missing data. It then converts the data into an analyzable format (for example, JSON format). Next, the server uses machine learning algorithms (using TensorFlow, PyTorch, etc.) to analyze the user's behavioral patterns and estimated interests. From past message history, the server learns that a particular user has a high interest in beauty products.

[0549] Based on the analysis results, the server generates a profile for each customer. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products." By generating this profile, the user's detailed interests and behavioral patterns are compiled into a single data set.

[0550] Next, the server requests a generative artificial intelligence (AI) to generate an optimized post based on the customer profile. The generative AI uses natural language processing technology such as OpenAI's GPT-4 to generate a message optimized for each user. For example, a message might be generated for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out." The following is an example of a prompt sentence that can be used:

[0551] "Please create a promotional message for a new beauty product for User A, a woman in her 30s who has a high interest in beauty products."

[0552] The generated submission is validated by the server and checked for inappropriate language, which is then manually corrected.

[0553] Finally, the server sends the optimized post to each user's device, which then displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0554] After distribution, the server collects user response data (such as reply messages and purchasing behavior) and analyzes it again. This response data is used as feedback for generating subsequent posts, enabling even more optimized information to be provided. In this way, the system can continuously provide optimal information tailored to the user's interests.

[0555] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0556] Step 1:

[0557] Users can send messages to official accounts or purchase products.

[0558] Input: User action (sending a message or purchasing a product)

[0559] Output: User action detection data

[0560] Specific operation: A user sends a message to the official LINE account asking, "Do you have any new beauty products?"

[0561] Step 2:

[0562] The device detects this user's actions and collects the user's profile information and behavioral data.

[0563] Input: User action detection data

[0564] Output: Collected user profile information and behavioral data

[0565] Specific operation: The device detects the message and collects user A's name, gender, age, past message history, and purchase history.

[0566] Step 3:

[0567] The data collected by the device is consolidated and sent to the server.

[0568] Input: Collected user profile information and behavioral data

[0569] Output: Consolidated data sent from the device to the server

[0570] Specific operation: The terminal compiles the collected data of user A and sends it to the server.

[0571] Step 4:

[0572] The server cleanses the incoming data, corrects inconsistent data, and fills in missing data.

[0573] Input: Integrated data sent from the device

[0574] Output: Cleansed data

[0575] Specific behavior: The server will correct inconsistencies in the data (e.g. missing names) and complete them as needed.

[0576] Step 5:

[0577] The server converts the cleansed data into a parsable format.

[0578] Input: Cleansed data

[0579] Output: Data in a parsable format (e.g. JSON)

[0580] What happens: The server converts the data into a parsable format, such as JSON.

[0581] Step 6:

[0582] The server uses machine learning algorithms to analyze users' behavioral patterns and estimated interests.

[0583] Input: Data in a parsable format

[0584] Output: Analysis results of user behavior patterns and estimated interests

[0585] Specific operation: The server uses TensorFlow to analyze User A's interest in beauty products based on their past message history.

[0586] Step 7:

[0587] The server generates a customer profile based on the analysis results.

[0588] Input: Analysis results of user behavior patterns and estimated interests

[0589] Output: Generated customer profile

[0590] Specific operation: The server creates a profile for "a woman in her 30s with a strong interest in beauty products."

[0591] Step 8:

[0592] The server requests the generative artificial intelligence to generate an optimized submission document based on the customer profile.

[0593] Input: Generated customer profile

[0594] Output: Prompt sentence requested by the generative AI

[0595] Specific operation: The server requests a generative artificial intelligence (e.g., OpenAI GPT-4) to "generate the optimal message for User A, a woman in her 30s who has a high interest in beauty products."

[0596] Step 9:

[0597] Generative AI generates optimized submission documents.

[0598] Input: Prompt sentence requested by the generative artificial intelligence

[0599] Output: The generated optimized submission document

[0600] What it does: The generative AI generates a message saying, "The latest beauty products have arrived! Check them out at a limited-time discount."

[0601] Step 10:

[0602] The server validates the generated optimized submission document and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0603] Input: Generated optimized submission document

[0604] Output: Validated and corrected submission document

[0605] Specific behavior: The server validates the generated message, checks for errors in specific product information, etc., and manually corrects them.

[0606] Step 11:

[0607] The server transmits the optimized submitted document to the user's terminal.

[0608] Input: Validated and corrected submitted document

[0609] Output: Posted document sent to terminal

[0610] Specific operation: The server sends the optimized submission document to the terminal.

[0611] Step 12:

[0612] The terminal displays the posted document to the user.

[0613] Input: Posted document sent to terminal

[0614] Output: The submitted document as displayed to the user

[0615] Specific operation: The device displays a message on User A's LINE message saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0616] Step 13:

[0617] The server collects user response data (reply messages and purchasing behavior) after delivery and analyzes it again.

[0618] Input: User response data

[0619] Output: Parsed reaction data

[0620] Specific operation: The server collects and analyzes User A's reply messages and purchasing behavior.

[0621] Step 14:

[0622] The server provides feedback for the next submission.

[0623] Input: Parsed reaction data

[0624] Output: Feedback information reflected in the next submission generation

[0625] Specific operation: The server optimizes the process of generating the next submitted document based on the analysis results.

[0626] (Application example 1)

[0627] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0628] On modern online shopping sites, diverse users each have different interests, so providing information in a single location reduces user satisfaction. Another issue is the difficulty of effectively analyzing collected user behavior data and making personalized, appropriate product recommendations.

[0629] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0630] In this invention, the server includes a means for sending user action data to the server using a REST API, a means for generating an optimized message by making a request to the generative AI based on the analyzed data, and a means for sending the generated message again from the server to the terminal, thereby enabling personalized product recommendations based on each user's interests.

[0631] "User actions" refer to specific operations that users perform on the system, such as purchasing, browsing, sending messages, etc.

[0632] "Profile information" refers to basic personal information such as a user's name, gender, and age.

[0633] "Behavioral data" refers to data relating to specific behaviors of users, such as their message history and purchase history.

[0634] "Server" means a central management system for collecting, analyzing, generating and transmitting user profile information and behavioral data.

[0635] "Generative AI" refers to artificial intelligence techniques used to generate optimized messages based on user profiles.

[0636] An "optimized submission" is a document created by generative artificial intelligence to suit a specific user based on the user's profile.

[0637] A "user terminal" is a device, such as a smartphone or computer, that a user uses to receive information.

[0638] A "REST API" is a standard interface for sending data to a server.

[0639] "Machine learning algorithms" are data analysis techniques used to analyze user behavior patterns and inferred interests.

[0640] "Response data" refers to responses and purchasing behavior of users in response to information from the system.

[0641] The present invention is a system for a mail-order site that provides personalized information based on a user's interests. This system is realized by the following method and apparatus.

[0642] First, when a user purchases or browses a product on an online shopping site, their behavioral data (e.g., product ID, category, etc.) is automatically collected. This collection is performed by the user's device issuing a request and sending the data to the server via a REST API.

[0643] The server cleans the received data and uses machine learning algorithms to analyze user behavior patterns and inferred interests. To do this, the server uses Python programs and libraries such as TensorFlow and Scikit-learn.

[0644] Based on the analysis results, the server creates a detailed profile for each user, including the user's interests and behavioral patterns. For example, a profile might be generated that describes a woman in her 30s with a strong interest in beauty products.

[0645] Next, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate a post optimized for each user. An example of a prompt is, "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message."

[0646] The generated post is verified by the server and checked for inappropriate content, after which the generated optimized post is sent back to the user's device and displayed to the user.

[0647] As described above, personalized information based on the user's interests can be provided. This system enables optimized product recommendations for each user, improving the user experience. As a specific example of use, if a woman in her 30s frequently purchases beauty products, a message will be displayed saying, "Hello, we have new beauty products that suit your tastes. They are on special discount, so please check them out."

[0648] In this way, user satisfaction at the online shopping site can be improved.

[0649] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0650] Step 1:

[0651] When a user purchases or browses products on an online shopping site, their device collects their behavioral data. This data includes information such as product ID, category, and time. The collected data is sent to the server via a REST API. This input data is output as a request to the server.

[0652] Step 2:

[0653] The server receives the behavioral data sent from the device and performs data cleansing. Specifically, it corrects inconsistent data and complements missing data. The cleansed data is output as a dataset for analysis.

[0654] Step 3:

[0655] The server then runs machine learning algorithms on the cleansed data to analyze user behavior patterns and inferred interests. This analysis uses libraries such as Scikit-learn and TensorFlow. The analysis results are output as a user profile.

[0656] Step 4:

[0657] Based on the generated user profile, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate an optimized post. For example, the prompt "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message" is used as the source data, and the generated recommendation message is output.

[0658] Step 5:

[0659] The server validates the generated optimized submission document, checking for inappropriate language, and outputs the validated document as the final message.

[0660] Step 6:

[0661] The server sends a final message to the user's terminal, which displays the message and the user receives the recommended information. The user then checks the recommended product information.

[0662] Step 7:

[0663] When a user responds to a recommended message (for example, by purchasing a product or replying to a message), the response data is collected again. The collected response data is sent to the server via a REST API. The sent data is output as feedback data.

[0664] Step 8:

[0665] The server analyzes the collected response data and provides feedback for the next submission. This feedback enables the generation of more appropriate recommendation messages in the future. The feedback results are output as the next profile analysis data.

[0666] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0667] This invention is a system that automatically provides information optimized for each user and further optimizes the information by recognizing the user's emotions, and is realized by the following processing steps.

[0668] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device sends this message and all related data to the server.

[0669] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[0670] In addition to the analysis data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. For example, when a user sends a message saying, "Do you have any new beauty products?", the emotion engine infers emotions such as "excitement" from the user's text.

[0671] Based on the analysis results and emotion estimation results, the server generates a profile for each customer. This allows each user's detailed interests and behavioral patterns, as well as their emotional state, to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, high interest in beauty products, current emotional state: excited."

[0672] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A. The tone and content of the message are also further customized based on the estimation results of the emotion engine.

[0673] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[0674] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0675] Finally, the server collects and analyzes user response data (e.g., reply messages and purchasing behavior) after the message is posted. Based on this response data, feedback is provided for the generation of subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests and emotional state.

[0676] The processing flow will be explained below.

[0677] Step 1:

[0678] Detecting user actions

[0679] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[0680] Step 2:

[0681] Data collection

[0682] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[0683] Step 3:

[0684] Sending data

[0685] The device sends the collected profile information and behavioral data to the server. For example, the device sends User A's profile information and behavioral data in JSON format to the server.

[0686] Step 4:

[0687] Receiving and storing data

[0688] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[0689] Step 5:

[0690] Data Preprocessing

[0691] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[0692] Step 6:

[0693] Customer Segmentation

[0694] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[0695] Step 7:

[0696] Generate customer profiles

[0697] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[0698] Step 8:

[0699] Emotion recognition data collection

[0700] The device sends the user's message content and behavioral data to the emotion engine. For example, a message sent by user A saying "Do you have any new beauty products?" is sent to the emotion engine.

[0701] Step 9:

[0702] Emotion estimation

[0703] The server uses the emotion engine to estimate the user's emotion from the received message. For example, the emotion engine estimates the emotion "excitement" from User A's message.

[0704] Step 10:

[0705] Optimized submission document generation request

[0706] The server requests the generative AI to generate a post optimized for each customer based on their profile and the emotion estimation results. For example, it instructs the AI ​​to use a template for users who are interested in and excited about beauty products.

[0707] Step 11:

[0708] Generate optimized submission documents

[0709] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0710] Step 12:

[0711] Review and correct submitted documents

[0712] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0713] Step 13:

[0714] Preparing to send submitted documents

[0715] The server prepares to send the optimized post to the user's device. For example, it prepares to send a message to User A's LINE device.

[0716] Step 14:

[0717] Submitting a submission

[0718] The server sends the posted document to the user's LINE device. For example, it delivers a posted document optimized for User A's LINE device.

[0719] Step 15:

[0720] Reaction data collection

[0721] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[0722] Step 16:

[0723] Reanalysis of reaction data

[0724] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[0725] Step 17:

[0726] Reflecting feedback

[0727] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[0728] Through these steps, the system not only automatically provides information optimized for each user, but also takes emotions into account to provide even more effective information.

[0729] Example 2

[0730] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0731] Conventional information delivery systems have difficulty providing optimized information that takes into account the individual interests and emotional state of each user. This has made improving user experience and engagement a challenge. Another issue is the cumbersome process of verifying and manually correcting generated messages, making it difficult to provide information efficiently.

[0732] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0733] In this invention, the server includes means for collecting profile information and behavioral data based on user actions, means for cleansing the collected data, means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer, means for estimating emotions from the content of the user's messages, means for generating a customer profile based on the analysis results and the emotion estimation results, means for generating a post optimized for each customer using generative artificial intelligence, means for verifying and manually correcting the generated optimized post, means for sending the generated optimized post to the user's device, and means for collecting user reaction data after transmission, analyzing the reaction data, and providing feedback for subsequent posts. This makes it possible to provide optimal information based on each user's detailed interests and emotional state.

[0734] "User actions" refers to a series of operations or actions performed by a user, such as sending a message to an official account or purchasing a product.

[0735] "Profile information" refers to information that can be used to personally identify a user, such as the user's name, gender, and age.

[0736] "Behavioral data" refers to data related to a user's past operations and behavior, such as message history and purchase history.

[0737] "Cleansing" refers to the process of correcting inconsistent data, filling in missing data, and converting data into an analyzable format.

[0738] "Analysis" refers to the processing of collected data to identify user behavioral patterns and interests.

[0739] "Estimated interests" refers to estimating the interests and concerns that a user is likely to have based on past behavioral data.

[0740] A "behavioral pattern" refers to a series of actions or patterns of operations that a user exhibits during a specific period of time.

[0741] "Emotion estimation" refers to the process of inferring the emotion contained in a message sent by a user from its content.

[0742] A "customer profile" refers to a data set that compiles a user's interests, behavioral patterns, and emotional state.

[0743] "Generative AI" refers to AI technology that automatically generates optimized messages and text based on given data.

[0744] "Optimized posts" refer to messages that are customized for individual users based on their customer profile.

[0745] "Response data" refers to data such as reply messages sent by users in response to messages received and subsequent purchasing behavior.

[0746] "Feedback" refers to the process of analyzing collected response data and using it to improve the quality of subsequent submissions.

[0747] This invention is a system that automatically provides information optimized for each user and further optimizes it by recognizing the user's emotions. The system is mainly composed of a server, a terminal, and a user. Details of each function and their specific implementation methods are explained below.

[0748] Basic configuration

[0749] The server plays a central role in data cleansing, data analysis, emotion estimation, and message generation. The hardware used is a high-performance data processing server, and the software used includes Hadoop, Spark, IBM Watson, Microsoft Azure's emotion recognition service, Scikit-learn, and TensorFlow. The generative artificial intelligence used is OpenAI GPT-4.

[0750] The terminal is responsible for detecting user actions and sending data to the server. This can be implemented on popular devices such as smartphones and PCs.

[0751] Users take action on the system and receive information optimized based on their response.

[0752] Data collection and transmission

[0753] When a user sends a message to the official account asking, "Do you have any new beauty products?", the device detects this action. The device collects the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) and sends them to the server in real time.

[0754] Data Cleansing and Analysis

[0755] The server cleanses the data it receives. This process includes correcting inconsistent data and filling in missing data, for example by filling in missing fields with appropriate default values.

[0756] The server then analyzes the cleansed data using Hadoop or Spark. Machine learning algorithms (Scikit-learn or TensorFlow) are used to analyze user behavior patterns and interests. The analysis results include information on which beauty products a particular user is frequently interested in.

[0757] Emotion estimation

[0758] The server uses IBM Watson or Microsoft Azure emotion recognition services to infer emotions from the content of a user's message. For example, it infers the emotion "excitement" from the message "Do you have any new beauty products?"

[0759] Customer profile generation

[0760] The server generates a detailed profile of the user based on the analysis results and emotion estimation results. For example, it creates a profile such as "Female in her 30s, highly interested in beauty products, current emotional state: excited."

[0761] Message Creation and Delivery

[0762] The server requests the OpenAI GPT-4 to generate an optimized message based on the generated profile. For example, the request is made with the following prompt:

[0763] Customer Profile:

[0764] Name: User A

[0765] Gender: Female

[0766] Age: 30

[0767] Interests: Beauty products

[0768] Current emotional state: Excited

[0769] message:

[0770] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[0771] The generated message is validated by the server to check for inappropriate language and, if necessary, manually corrected.

[0772] Finally, the optimized message is sent to the user's device, which displays it to the user. The user reacts, which triggers the next cycle of data collection and analysis.

[0773] Reaction data collection and feedback

[0774] The server collects user responses (reply messages and purchasing behavior) after distribution and analyzes this data. Based on the analysis results, the quality of subsequent submissions can be improved, resulting in even more optimized information provision.

[0775] This process allows the system to continually provide optimal information tailored to the user's interests and emotional state.

[0776] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0777] Step 1:

[0778] User Actions and Data Collection

[0779] A user sends a message to your official account saying, "Are there any new beauty products?"

[0780] The device detects this action and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history).

[0781] Input: User messages, profile information, behavioral data

[0782] Output: Collected dataset

[0783] Specific operation: Software in the device begins collecting data when the user sends a message.

[0784] Step 2:

[0785] Data transmission

[0786] The terminal transmits the collected data set to the server.

[0787] Input: Collected dataset

[0788] Output: The dataset sent to the server

[0789] Specific operation: The device uploads data to a server via the Internet.

[0790] Step 3:

[0791] Data Cleansing

[0792] The server cleanses the received dataset, specifically correcting inconsistent data and filling in missing data.

[0793] Input: The dataset sent to the server

[0794] Output: A cleansed dataset

[0795] Specific operations: Use data processing engines such as Hadoop and Spark to check data integrity and convert it into a unified format.

[0796] Step 4:

[0797] Data analysis

[0798] The server uses the cleansed data to analyze user behavior patterns and interests using machine learning algorithms such as Scikit-learn and TensorFlow.

[0799] Input: Cleansed dataset

[0800] Output: Analysis results (user behavior patterns, estimated interests)

[0801] Specific behavior: Apply machine learning models to training data to derive quantitative user behavior patterns and interests.

[0802] Step 5:

[0803] Emotion estimation

[0804] The server uses IBM Watson and Microsoft Azure emotion recognition services to estimate emotions from the content of the user's message.

[0805] Input: Message content, behavioral data

[0806] Output: Emotion estimation result

[0807] What it does: The emotion recognition engine analyzes text and extracts specific emotions (e.g., excitement, joy, sadness).

[0808] Step 6:

[0809] Customer profile generation

[0810] The server generates a detailed profile of the user based on the analysis results and emotion estimation results.

[0811] Input: Analysis results, emotion estimation results

[0812] Output: Customer profile

[0813] Specific operation: The analysis results are integrated with emotional data, and a detailed profile customized for each user is stored in a database.

[0814] Step 7:

[0815] Requesting Message Creation

[0816] The server requests the generative artificial intelligence (OpenAI GPT-4) to generate an optimized message based on the generated customer profile. The server instructs the generation using a prompt.

[0817] Input: Customer profile, prompt statement

[0818] Output: Generated optimization messages

[0819] Specific behavior: Provide the following prompt to the generative AI:

[0820] Customer Profile:

[0821] Name: User A

[0822] Gender: Female

[0823] Age: 30

[0824] Interests: Beauty products

[0825] Current emotional state: Excited

[0826] message:

[0827] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[0828] Step 8:

[0829] Message verification and correction

[0830] The server validates generated messages for inappropriate language and manually corrects them as necessary.

[0831] Input: The generated message

[0832] Output: Validated and corrected message

[0833] What it does: It performs text filtering and human operator review.

[0834] Step 9:

[0835] Message Delivery

[0836] The server sends the optimized message to the user's terminal.

[0837] Input: Validated and corrected message

[0838] Output: Message sent to the terminal

[0839] Specific operation: The server sends a message to the terminal via the Internet.

[0840] Step 10:

[0841] Reaction data collection and feedback

[0842] After delivery, the server collects and analyzes the user's reactions (reply messages and purchasing behavior), and reflects the results in the generation of subsequent messages.

[0843] Input: User response data

[0844] Output: Feedback data

[0845] What it does: The reaction data is stored in a database and used to retrain the machine learning model.

[0846] (Application example 2)

[0847] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0848] Current online shopping sites generally only provide product recommendations and promotional messages to users, but these messages are not optimized to reflect the interests and emotional state of individual users, making it difficult to improve the user experience. Another issue is that collected user data cannot be effectively utilized, making it difficult to provide individually optimized information in real time. These issues need to be resolved.

[0849] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0850] In this invention, the server includes means for collecting profile information and behavioral data of a user based on the user's actions, means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer, means for generating a post optimized for each customer using a generative artificial intelligence based on the generated customer profile, means for transmitting the generated optimized post to the user's terminal, means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding the data back into subsequent posts, means for providing optimized product recommendations and promotional messages in real time using a smart device, means for generating a prompt sentence for the generative artificial intelligence to generate the optimized message based on the analysis results, and means for generating the optimized message using the generative artificial intelligence using the generated prompt sentence, thereby enabling the provision of information optimized for the interests and emotional state of each individual user.

[0851] "User action" refers to the operations or inputs that a user makes to the system.

[0852] "Profile Information" refers to basic personal information such as a user's name, gender, and age.

[0853] "Behavioral data" refers to activity data such as a user's message history and purchase history.

[0854] "Analysis" refers to the process of analyzing collected data and extracting meaningful information.

[0855] "Estimated interests" refers to topics or areas in which a user is estimated to be particularly interested.

[0856] A "behavioral pattern" refers to a series of behavioral tendencies based on a user's past behavior.

[0857] A "customer profile" refers to a data set that compiles detailed interests and behavioral patterns for each user.

[0858] "Generative artificial intelligence" refers to AI that uses natural language processing technology to generate optimal messages and documents.

[0859] "Optimized posts" refer to messages that are individually customized based on a user's profile information and behavioral data.

[0860] "Terminal" refers to a device through which a user receives information, such as a smartphone or tablet.

[0861] "Response data" refers to the user's response to a posted document, such as a reply message or purchasing behavior.

[0862] "Smart device" refers to a device with advanced functionality that can connect to the Internet.

[0863] "Real-time" refers to near-instant processing and response.

[0864] A "prompt sentence" refers to an input sentence that causes a generative artificial intelligence to generate an optimized message.

[0865] This invention realizes a system for providing product recommendations and promotional messages optimized for each user on an online shopping site. This system is realized by collecting and analyzing information based on user actions, and generating and sending messages using generative artificial intelligence. Specific embodiments of this system are described below.

[0866] The user's smartphone or smart device functions as the terminal for this system. When a user searches for a product or sends a message, this action is detected by the terminal, and the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) are collected. The collected data is sent to the server in real time.

[0867] The server uses machine learning libraries like Python and TensorFlow to cleanse the incoming data and convert it into an analyzable format. Then, machine learning algorithms are used to analyze user behavior patterns and inferred interests. For example, past messaging history may indicate that a particular user is interested in beauty products.

[0868] Based on the analysis results and emotion estimation by the emotion engine, the server generates a detailed profile for each user. The generated profile is used to generate prompts for a generative AI (e.g., OpenAI's GPT-3). Based on the generated prompts, the generative AI generates optimized messages.

[0869] An example of a generated prompt is:

[0870] Ms. A (30 years old, female) is interested in a new beauty product. She is currently in an excited emotional state and has previously purchased beauty creams and face masks. Please generate the best message for her based on the following conditions:

[0871] New beauty products have arrived! Check them out for limited time discounts.

[0872] The server validates the generated optimization message and manually checks it for inappropriate content. After validation, the optimization message is sent to the user's smart device and displayed to the user in real time.

[0873] For example, if a user sends a message saying, "Do you have any new beauty products?", a customized recommendation message such as, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" will be displayed in real time on the user's device.

[0874] This invention enables the provision of information optimized for the interests and emotional state of each individual user, improving the user experience. Furthermore, by effectively utilizing collected user data, it is possible to provide individually optimized information in real time.

[0875] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0876] Step 1:

[0877] A user searches for a product or sends a message on a smartphone or smart device. The input data is the user's action, search query, and sent message. The device detects this user action as it is being performed and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history). The collected data is sent to a server in real time.

[0878] Step 2:

[0879] The server cleanses the incoming data and converts it into an analyzable format. The input data is the profile information and behavioral data collected in step 1. Inconsistent data is corrected and missing data is filled in. The cleansed data is obtained as the output.

[0880] Step 3:

[0881] The server uses machine learning algorithms to analyze the user's behavioral patterns and estimated interests. The input is cleansed data. Using machine learning libraries such as TensorFlow, the server analyzes the user's past messaging history and purchasing history to estimate specific interests. The analysis results are the user's estimated interests and behavioral patterns.

[0882] Step 4:

[0883] In addition to the analyzed data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. The input is the user's message content and behavioral data. Natural language processing technology is used to analyze emotions from the message and infer emotions such as "excitement." The emotion estimation result is obtained as the output.

[0884] Step 5:

[0885] The server generates a detailed profile for each user. The inputs are behavioral patterns, estimated interests, and estimated emotions. Each user's interests, behavioral patterns, and emotional state are compiled into a single dataset. The generated user profile is output.

[0886] Step 6:

[0887] The server generates a prompt for the generative AI to generate an optimization message. The input is the user profile. A generative AI such as OpenAI's GPT-3 is used to generate the prompt. The generated prompt is output.

[0888] Step 7:

[0889] The server uses the generated prompt sentence to generate an optimized message using a generative AI. The input is the generated prompt sentence. A generative AI model is used to generate the optimized message. The output is the optimized message.

[0890] Step 8:

[0891] The server validates the generated optimization message and checks for inappropriate expressions. The input is the generated optimization message. If any manual corrections are required, they are made. The optimized message without inappropriate expressions is output.

[0892] Step 9:

[0893] The server sends the verified optimization message to the user's terminal. The input is the verified optimization message. The message is displayed on the user's terminal in real time.

[0894] Step 10:

[0895] The server collects user reaction data after sending a message and feeds it back into subsequent messages. The input is user reaction data (reply messages, purchasing behavior, etc.). The collected reaction data is analyzed and used to generate optimized messages for subsequent messages. The results of the data analysis after feedback are output.

[0896] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0897] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0898] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0899] [Third embodiment]

[0900] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0901] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0902] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0903] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0904] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0905] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0906] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0907] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0908] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0909] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0910] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0911] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. 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."

[0912] The present invention is a system for automatically providing information optimized for each individual user, and is realized by the following processing steps.

[0913] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and all related data to the server.

[0914] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[0915] Based on the results of the above analysis, the server creates a profile for each customer. This allows each user's detailed interests and behavioral patterns to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products."

[0916] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A.

[0917] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[0918] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0919] Finally, the server collects user response data (e.g., reply messages and purchasing behavior) after distribution and analyzes it again. Based on this response data, feedback is provided for generating subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests.

[0920] The processing flow will be explained below.

[0921] Step 1:

[0922] Detecting user actions

[0923] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[0924] Step 2:

[0925] Data collection

[0926] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[0927] Step 3:

[0928] Sending data

[0929] The device sends the collected data to the server. For example, it sends user A's profile information and behavioral data in JSON format to the server.

[0930] Step 4:

[0931] Receiving and storing data

[0932] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[0933] Step 5:

[0934] Data Preprocessing

[0935] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[0936] Step 6:

[0937] Customer Segmentation

[0938] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[0939] Step 7:

[0940] Generate customer profiles

[0941] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[0942] Step 8:

[0943] Optimized submission document generation request

[0944] The server requests the AI ​​to generate a post optimized for each customer profile, for example, by instructing it to use a template related to beauty products.

[0945] Step 9:

[0946] Generate optimized submission documents

[0947] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[0948] Step 10:

[0949] Review and correct submitted documents

[0950] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[0951] Step 11:

[0952] Preparing to send submitted documents

[0953] The server prepares to send the optimized post to each user's official LINE account.

[0954] Step 12:

[0955] Submitting a submission

[0956] The server sends the posted document to the user's LINE device. For example, it sends a posted document optimized for User A's LINE device.

[0957] Step 13:

[0958] Reaction data collection

[0959] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[0960] Step 14:

[0961] Reanalysis of reaction data

[0962] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[0963] Step 15:

[0964] Reflecting feedback

[0965] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[0966] Through these steps, the system automatically provides information optimized for each user, achieving continuous optimization.

[0967] Example 1

[0968] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0969] Conventional information provision systems have difficulty providing information optimized to users' interests. There is also a risk that posted documents containing inappropriate language may be distributed. Furthermore, the lack of a mechanism for effectively providing feedback on users' reactions reduces the quality and accuracy of the information provided. There is a need to solve these issues.

[0970] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0971] In this invention, the server includes: means for collecting profile information and behavioral data of a user based on the user's actions; means for receiving the collected data and cleansing it by correcting inconsistent data and filling in missing data; means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer; means for generating a post optimized for each customer using generative artificial intelligence based on the generated customer profile; means for verifying the generated optimized post, checking for inappropriate language, and manually correcting it; means for transmitting the generated optimized post to the user's device; and means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding it back into subsequent posts. This enables the provision of information optimized to the user's interests, reduces the risk of inappropriate language, and improves the quality and accuracy of information provided by effectively feeding back user reactions.

[0972] A "user action" is a series of operations that a user performs on the system, such as sending a message or purchasing a product.

[0973] "Profile information" refers to personal information such as a user's name, gender, and age.

[0974] "Behavioral data" refers to a record of a user's past behavior, such as message history and purchase history.

[0975] "Cleansing" is the process of organizing and formatting data to correct inconsistent data and fill in missing data.

[0976] "Analysis" is the process of deriving user behavior patterns and estimated interests based on collected data.

[0977] A "customer profile" is a data set generated based on analyzed data that contains detailed interests and behavioral patterns for each user.

[0978] "Generative AI" is an AI technology that generates posts optimized for each customer based on their customer profile.

[0979] A "posted document" is a message or piece of text that is sent to a user.

[0980] "Response data" refers to data such as replies and purchasing behavior of users in response to received posted documents.

[0981] "Feedback" refers to the process of reflecting collected response data in the generation of subsequent submitted documents.

[0982] MODE FOR CARRYING OUT THE INVENTION

[0983] This invention is a system that automatically provides information optimized for each individual user, and specifically, it operates in cooperation with a server, a terminal, and a user. The system of this invention operates as follows.

[0984] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (such as name, gender, and age) and behavioral data (such as message history and purchase history), consolidates this data, and sends it to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and the associated profile information and behavioral data to the server.

[0985] The server processes the data received from the device. Specifically, the server cleanses the received data, corrects inconsistent data, and fills in missing data. It then converts the data into an analyzable format (for example, JSON format). Next, the server uses machine learning algorithms (using TensorFlow, PyTorch, etc.) to analyze the user's behavioral patterns and estimated interests. From past message history, the server learns that a particular user has a high interest in beauty products.

[0986] Based on the analysis results, the server generates a profile for each customer. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products." By generating this profile, the user's detailed interests and behavioral patterns are compiled into a single data set.

[0987] Next, the server requests a generative artificial intelligence (AI) to generate an optimized post based on the customer profile. The generative AI uses natural language processing technology such as OpenAI's GPT-4 to generate a message optimized for each user. For example, a message might be generated for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out." The following is an example of a prompt sentence that can be used:

[0988] "Please create a promotional message for a new beauty product for User A, a woman in her 30s who has a high interest in beauty products."

[0989] The generated submission is validated by the server and checked for inappropriate language, which is then manually corrected.

[0990] Finally, the server sends the optimized post to each user's device, which then displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[0991] After distribution, the server collects user response data (such as reply messages and purchasing behavior) and analyzes it again. This response data is used as feedback for generating subsequent posts, enabling even more optimized information to be provided. In this way, the system can continuously provide optimal information tailored to the user's interests.

[0992] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0993] Step 1:

[0994] Users can send messages to official accounts or purchase products.

[0995] Input: User action (sending a message or purchasing a product)

[0996] Output: User action detection data

[0997] Specific operation: A user sends a message to the official LINE account asking, "Do you have any new beauty products?"

[0998] Step 2:

[0999] The device detects this user's actions and collects the user's profile information and behavioral data.

[1000] Input: User action detection data

[1001] Output: Collected user profile information and behavioral data

[1002] Specific operation: The device detects the message and collects user A's name, gender, age, past message history, and purchase history.

[1003] Step 3:

[1004] The data collected by the device is consolidated and sent to the server.

[1005] Input: Collected user profile information and behavioral data

[1006] Output: Consolidated data sent from the device to the server

[1007] Specific operation: The terminal compiles the collected data of user A and sends it to the server.

[1008] Step 4:

[1009] The server cleanses the incoming data, corrects inconsistent data, and fills in missing data.

[1010] Input: Integrated data sent from the device

[1011] Output: Cleansed data

[1012] Specific behavior: The server will correct inconsistencies in the data (e.g. missing names) and complete them as needed.

[1013] Step 5:

[1014] The server converts the cleansed data into a parsable format.

[1015] Input: Cleansed data

[1016] Output: Data in a parsable format (e.g. JSON)

[1017] What happens: The server converts the data into a parsable format, such as JSON.

[1018] Step 6:

[1019] The server uses machine learning algorithms to analyze users' behavioral patterns and estimated interests.

[1020] Input: Data in a parsable format

[1021] Output: Analysis results of user behavior patterns and estimated interests

[1022] Specific operation: The server uses TensorFlow to analyze User A's interest in beauty products based on their past message history.

[1023] Step 7:

[1024] The server generates a customer profile based on the analysis results.

[1025] Input: Analysis results of user behavior patterns and estimated interests

[1026] Output: Generated customer profile

[1027] Specific operation: The server creates a profile for "a woman in her 30s with a strong interest in beauty products."

[1028] Step 8:

[1029] The server requests the generative artificial intelligence to generate an optimized submission document based on the customer profile.

[1030] Input: Generated customer profile

[1031] Output: Prompt sentence requested by the generative AI

[1032] Specific operation: The server requests a generative artificial intelligence (e.g., OpenAI GPT-4) to "generate the optimal message for User A, a woman in her 30s who has a high interest in beauty products."

[1033] Step 9:

[1034] Generative AI generates optimized submission documents.

[1035] Input: Prompt sentence requested by the generative artificial intelligence

[1036] Output: The generated optimized submission document

[1037] What it does: The generative AI generates a message saying, "The latest beauty products have arrived! Check them out at a limited-time discount."

[1038] Step 10:

[1039] The server validates the generated optimized submission document and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[1040] Input: Generated optimized submission document

[1041] Output: Validated and corrected submission document

[1042] Specific behavior: The server validates the generated message, checks for errors in specific product information, etc., and manually corrects them.

[1043] Step 11:

[1044] The server transmits the optimized submitted document to the user's terminal.

[1045] Input: Validated and corrected submitted document

[1046] Output: Posted document sent to terminal

[1047] Specific operation: The server sends the optimized submission document to the terminal.

[1048] Step 12:

[1049] The terminal displays the posted document to the user.

[1050] Input: Posted document sent to terminal

[1051] Output: The submitted document as displayed to the user

[1052] Specific operation: The device displays a message on User A's LINE message saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[1053] Step 13:

[1054] The server collects user response data (reply messages and purchasing behavior) after delivery and analyzes it again.

[1055] Input: User response data

[1056] Output: Parsed reaction data

[1057] Specific operation: The server collects and analyzes User A's reply messages and purchasing behavior.

[1058] Step 14:

[1059] The server provides feedback for the next submission.

[1060] Input: Parsed reaction data

[1061] Output: Feedback information reflected in the next submission generation

[1062] Specific operation: The server optimizes the process of generating the next submitted document based on the analysis results.

[1063] (Application example 1)

[1064] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1065] On modern online shopping sites, diverse users each have different interests, so providing information in a single location reduces user satisfaction. Another issue is the difficulty of effectively analyzing collected user behavior data and making personalized, appropriate product recommendations.

[1066] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1067] In this invention, the server includes a means for sending user action data to the server using a REST API, a means for generating an optimized message by making a request to the generative AI based on the analyzed data, and a means for sending the generated message again from the server to the terminal, thereby enabling personalized product recommendations based on each user's interests.

[1068] "User actions" refer to specific operations that users perform on the system, such as purchasing, browsing, sending messages, etc.

[1069] "Profile information" refers to basic personal information such as a user's name, gender, and age.

[1070] "Behavioral data" refers to data relating to specific behaviors of users, such as their message history and purchase history.

[1071] "Server" means a central management system for collecting, analyzing, generating and transmitting user profile information and behavioral data.

[1072] "Generative AI" refers to artificial intelligence techniques used to generate optimized messages based on user profiles.

[1073] An "optimized submission" is a document created by generative artificial intelligence to suit a specific user based on the user's profile.

[1074] A "user terminal" is a device, such as a smartphone or computer, that a user uses to receive information.

[1075] A "REST API" is a standard interface for sending data to a server.

[1076] "Machine learning algorithms" are data analysis techniques used to analyze user behavior patterns and inferred interests.

[1077] "Response data" refers to responses and purchasing behavior of users in response to information from the system.

[1078] The present invention is a system for a mail-order site that provides personalized information based on a user's interests. This system is realized by the following method and apparatus.

[1079] First, when a user purchases or browses a product on an online shopping site, their behavioral data (e.g., product ID, category, etc.) is automatically collected. This collection is performed by the user's device issuing a request and sending the data to the server via a REST API.

[1080] The server cleans the received data and uses machine learning algorithms to analyze user behavior patterns and inferred interests. To do this, the server uses Python programs and libraries such as TensorFlow and Scikit-learn.

[1081] Based on the analysis results, the server creates a detailed profile for each user, including the user's interests and behavioral patterns. For example, a profile might be generated that describes a woman in her 30s with a strong interest in beauty products.

[1082] Next, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate a post optimized for each user. An example of a prompt is, "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message."

[1083] The generated post is verified by the server and checked for inappropriate content, after which the generated optimized post is sent back to the user's device and displayed to the user.

[1084] As described above, personalized information based on the user's interests can be provided. This system enables optimized product recommendations for each user, improving the user experience. As a specific example of use, if a woman in her 30s frequently purchases beauty products, a message will be displayed saying, "Hello, we have new beauty products that suit your tastes. They are on special discount, so please check them out."

[1085] In this way, user satisfaction at the online shopping site can be improved.

[1086] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1087] Step 1:

[1088] When a user purchases or browses products on an online shopping site, their device collects their behavioral data. This data includes information such as product ID, category, and time. The collected data is sent to the server via a REST API. This input data is output as a request to the server.

[1089] Step 2:

[1090] The server receives the behavioral data sent from the device and performs data cleansing. Specifically, it corrects inconsistent data and complements missing data. The cleansed data is output as a dataset for analysis.

[1091] Step 3:

[1092] The server then runs machine learning algorithms on the cleansed data to analyze user behavior patterns and inferred interests. This analysis uses libraries such as Scikit-learn and TensorFlow. The analysis results are output as a user profile.

[1093] Step 4:

[1094] Based on the generated user profile, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate an optimized post. For example, the prompt "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message" is used as the source data, and the generated recommendation message is output.

[1095] Step 5:

[1096] The server validates the generated optimized submission document, checking for inappropriate language, and outputs the validated document as the final message.

[1097] Step 6:

[1098] The server sends a final message to the user's terminal, which displays the message and the user receives the recommended information. The user then checks the recommended product information.

[1099] Step 7:

[1100] When a user responds to a recommended message (for example, by purchasing a product or replying to a message), the response data is collected again. The collected response data is sent to the server via a REST API. The sent data is output as feedback data.

[1101] Step 8:

[1102] The server analyzes the collected response data and provides feedback for the next submission. This feedback enables the generation of more appropriate recommendation messages in the future. The feedback results are output as the next profile analysis data.

[1103] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1104] This invention is a system that automatically provides information optimized for each user and further optimizes the information by recognizing the user's emotions, and is realized by the following processing steps.

[1105] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device sends this message and all related data to the server.

[1106] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[1107] In addition to the analysis data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. For example, when a user sends a message saying, "Do you have any new beauty products?", the emotion engine infers emotions such as "excitement" from the user's text.

[1108] Based on the analysis results and emotion estimation results, the server generates a profile for each customer. This allows each user's detailed interests and behavioral patterns, as well as their emotional state, to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, high interest in beauty products, current emotional state: excited."

[1109] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A. The tone and content of the message are also further customized based on the estimation results of the emotion engine.

[1110] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[1111] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[1112] Finally, the server collects and analyzes user response data (e.g., reply messages and purchasing behavior) after the message is posted. Based on this response data, feedback is provided for the generation of subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests and emotional state.

[1113] The processing flow will be explained below.

[1114] Step 1:

[1115] Detecting user actions

[1116] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[1117] Step 2:

[1118] Data collection

[1119] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[1120] Step 3:

[1121] Sending data

[1122] The device sends the collected profile information and behavioral data to the server. For example, the device sends User A's profile information and behavioral data in JSON format to the server.

[1123] Step 4:

[1124] Receiving and storing data

[1125] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[1126] Step 5:

[1127] Data Preprocessing

[1128] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[1129] Step 6:

[1130] Customer Segmentation

[1131] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[1132] Step 7:

[1133] Generate customer profiles

[1134] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[1135] Step 8:

[1136] Emotion recognition data collection

[1137] The device sends the user's message content and behavioral data to the emotion engine. For example, a message sent by user A saying "Do you have any new beauty products?" is sent to the emotion engine.

[1138] Step 9:

[1139] Emotion estimation

[1140] The server uses the emotion engine to estimate the user's emotion from the received message. For example, the emotion engine estimates the emotion "excitement" from User A's message.

[1141] Step 10:

[1142] Optimized submission document generation request

[1143] The server requests the generative AI to generate a post optimized for each customer based on their profile and the emotion estimation results. For example, it instructs the AI ​​to use a template for users who are interested in and excited about beauty products.

[1144] Step 11:

[1145] Generate optimized submission documents

[1146] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[1147] Step 12:

[1148] Review and correct submitted documents

[1149] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[1150] Step 13:

[1151] Preparing to send submitted documents

[1152] The server prepares to send the optimized post to the user's device. For example, it prepares to send a message to User A's LINE device.

[1153] Step 14:

[1154] Submitting a submission

[1155] The server sends the posted document to the user's LINE device. For example, it delivers a posted document optimized for User A's LINE device.

[1156] Step 15:

[1157] Reaction data collection

[1158] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[1159] Step 16:

[1160] Reanalysis of reaction data

[1161] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[1162] Step 17:

[1163] Reflecting feedback

[1164] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[1165] Through these steps, the system not only automatically provides information optimized for each user, but also takes emotions into account to provide even more effective information.

[1166] Example 2

[1167] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1168] Conventional information delivery systems have difficulty providing optimized information that takes into account the individual interests and emotional state of each user. This has made improving user experience and engagement a challenge. Another issue is the cumbersome process of verifying and manually correcting generated messages, making it difficult to provide information efficiently.

[1169] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1170] In this invention, the server includes means for collecting profile information and behavioral data based on user actions, means for cleansing the collected data, means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer, means for estimating emotions from the content of the user's messages, means for generating a customer profile based on the analysis results and the emotion estimation results, means for generating a post optimized for each customer using generative artificial intelligence, means for verifying and manually correcting the generated optimized post, means for sending the generated optimized post to the user's device, and means for collecting user reaction data after transmission, analyzing the reaction data, and providing feedback for subsequent posts. This makes it possible to provide optimal information based on each user's detailed interests and emotional state.

[1171] "User actions" refers to a series of operations or actions performed by a user, such as sending a message to an official account or purchasing a product.

[1172] "Profile information" refers to information that can be used to personally identify a user, such as the user's name, gender, and age.

[1173] "Behavioral data" refers to data related to a user's past operations and behavior, such as message history and purchase history.

[1174] "Cleansing" refers to the process of correcting inconsistent data, filling in missing data, and converting data into an analyzable format.

[1175] "Analysis" refers to the processing of collected data to identify user behavioral patterns and interests.

[1176] "Estimated interests" refers to estimating the interests and concerns that a user is likely to have based on past behavioral data.

[1177] A "behavioral pattern" refers to a series of actions or patterns of operations that a user exhibits during a specific period of time.

[1178] "Emotion estimation" refers to the process of inferring the emotion contained in a message sent by a user from its content.

[1179] A "customer profile" refers to a data set that compiles a user's interests, behavioral patterns, and emotional state.

[1180] "Generative AI" refers to AI technology that automatically generates optimized messages and text based on given data.

[1181] "Optimized posts" refer to messages that are customized for individual users based on their customer profile.

[1182] "Response data" refers to data such as reply messages sent by users in response to messages received and subsequent purchasing behavior.

[1183] "Feedback" refers to the process of analyzing collected response data and using it to improve the quality of subsequent submissions.

[1184] This invention is a system that automatically provides information optimized for each user and further optimizes it by recognizing the user's emotions. The system is mainly composed of a server, a terminal, and a user. Details of each function and their specific implementation methods are explained below.

[1185] Basic configuration

[1186] The server plays a central role in data cleansing, data analysis, emotion estimation, and message generation. The hardware used is a high-performance data processing server, and the software used includes Hadoop, Spark, IBM Watson, Microsoft Azure's emotion recognition service, Scikit-learn, and TensorFlow. The generative artificial intelligence used is OpenAI GPT-4.

[1187] The terminal is responsible for detecting user actions and sending data to the server. This can be implemented on popular devices such as smartphones and PCs.

[1188] Users take action on the system and receive information optimized based on their response.

[1189] Data collection and transmission

[1190] When a user sends a message to the official account asking, "Do you have any new beauty products?", the device detects this action. The device collects the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) and sends them to the server in real time.

[1191] Data Cleansing and Analysis

[1192] The server cleanses the data it receives. This process includes correcting inconsistent data and filling in missing data, for example by filling in missing fields with appropriate default values.

[1193] The server then analyzes the cleansed data using Hadoop or Spark. Machine learning algorithms (Scikit-learn or TensorFlow) are used to analyze user behavior patterns and interests. The analysis results include information on which beauty products a particular user is frequently interested in.

[1194] Emotion estimation

[1195] The server uses IBM Watson or Microsoft Azure emotion recognition services to infer emotions from the content of a user's message. For example, it infers the emotion "excitement" from the message "Do you have any new beauty products?"

[1196] Customer profile generation

[1197] The server generates a detailed profile of the user based on the analysis results and emotion estimation results. For example, it creates a profile such as "Female in her 30s, highly interested in beauty products, current emotional state: excited."

[1198] Message Creation and Delivery

[1199] The server requests the OpenAI GPT-4 to generate an optimized message based on the generated profile. For example, the request is made with the following prompt:

[1200] Customer Profile:

[1201] Name: User A

[1202] Gender: Female

[1203] Age: 30

[1204] Interests: Beauty products

[1205] Current emotional state: Excited

[1206] message:

[1207] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[1208] The generated message is validated by the server to check for inappropriate language and, if necessary, manually corrected.

[1209] Finally, the optimized message is sent to the user's device, which displays it to the user. The user reacts, which triggers the next cycle of data collection and analysis.

[1210] Reaction data collection and feedback

[1211] The server collects user responses (reply messages and purchasing behavior) after distribution and analyzes this data. Based on the analysis results, the quality of subsequent submissions can be improved, resulting in even more optimized information provision.

[1212] This process allows the system to continually provide optimal information tailored to the user's interests and emotional state.

[1213] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1214] Step 1:

[1215] User Actions and Data Collection

[1216] A user sends a message to your official account saying, "Are there any new beauty products?"

[1217] The device detects this action and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history).

[1218] Input: User messages, profile information, behavioral data

[1219] Output: Collected dataset

[1220] Specific operation: Software in the device begins collecting data when the user sends a message.

[1221] Step 2:

[1222] Data transmission

[1223] The terminal transmits the collected data set to the server.

[1224] Input: Collected dataset

[1225] Output: The dataset sent to the server

[1226] Specific operation: The device uploads data to a server via the Internet.

[1227] Step 3:

[1228] Data Cleansing

[1229] The server cleanses the received dataset, specifically correcting inconsistent data and filling in missing data.

[1230] Input: The dataset sent to the server

[1231] Output: A cleansed dataset

[1232] Specific operations: Use data processing engines such as Hadoop and Spark to check data integrity and convert it into a unified format.

[1233] Step 4:

[1234] Data analysis

[1235] The server uses the cleansed data to analyze user behavior patterns and interests using machine learning algorithms such as Scikit-learn and TensorFlow.

[1236] Input: Cleansed dataset

[1237] Output: Analysis results (user behavior patterns, estimated interests)

[1238] Specific behavior: Apply machine learning models to training data to derive quantitative user behavior patterns and interests.

[1239] Step 5:

[1240] Emotion estimation

[1241] The server uses IBM Watson and Microsoft Azure emotion recognition services to estimate emotions from the content of the user's message.

[1242] Input: Message content, behavioral data

[1243] Output: Emotion estimation result

[1244] What it does: The emotion recognition engine analyzes text and extracts specific emotions (e.g., excitement, joy, sadness).

[1245] Step 6:

[1246] Customer profile generation

[1247] The server generates a detailed profile of the user based on the analysis results and emotion estimation results.

[1248] Input: Analysis results, emotion estimation results

[1249] Output: Customer profile

[1250] Specific operation: The analysis results are integrated with emotional data, and a detailed profile customized for each user is stored in a database.

[1251] Step 7:

[1252] Requesting Message Creation

[1253] The server requests the generative artificial intelligence (OpenAI GPT-4) to generate an optimized message based on the generated customer profile. The server instructs the generation using a prompt.

[1254] Input: Customer profile, prompt statement

[1255] Output: Generated optimization messages

[1256] Specific behavior: Provide the following prompt to the generative AI:

[1257] Customer Profile:

[1258] Name: User A

[1259] Gender: Female

[1260] Age: 30

[1261] Interests: Beauty products

[1262] Current emotional state: Excited

[1263] message:

[1264] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[1265] Step 8:

[1266] Message verification and correction

[1267] The server validates generated messages for inappropriate language and manually corrects them as necessary.

[1268] Input: The generated message

[1269] Output: Validated and corrected message

[1270] What it does: It performs text filtering and human operator review.

[1271] Step 9:

[1272] Message Delivery

[1273] The server sends the optimized message to the user's terminal.

[1274] Input: Validated and corrected message

[1275] Output: Message sent to the terminal

[1276] Specific operation: The server sends a message to the terminal via the Internet.

[1277] Step 10:

[1278] Reaction data collection and feedback

[1279] After delivery, the server collects and analyzes the user's reactions (reply messages and purchasing behavior), and reflects the results in the generation of subsequent messages.

[1280] Input: User response data

[1281] Output: Feedback data

[1282] What it does: The reaction data is stored in a database and used to retrain the machine learning model.

[1283] (Application example 2)

[1284] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1285] Current online shopping sites generally only provide product recommendations and promotional messages to users, but these messages are not optimized to reflect the interests and emotional state of individual users, making it difficult to improve the user experience. Another issue is that collected user data cannot be effectively utilized, making it difficult to provide individually optimized information in real time. These issues need to be resolved.

[1286] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1287] In this invention, the server includes means for collecting profile information and behavioral data of a user based on the user's actions, means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer, means for generating a post optimized for each customer using a generative artificial intelligence based on the generated customer profile, means for transmitting the generated optimized post to the user's terminal, means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding the data back into subsequent posts, means for providing optimized product recommendations and promotional messages in real time using a smart device, means for generating a prompt sentence for the generative artificial intelligence to generate the optimized message based on the analysis results, and means for generating the optimized message using the generative artificial intelligence using the generated prompt sentence, thereby enabling the provision of information optimized for the interests and emotional state of each individual user.

[1288] "User action" refers to the operations or inputs that a user makes to the system.

[1289] "Profile Information" refers to basic personal information such as a user's name, gender, and age.

[1290] "Behavioral data" refers to activity data such as a user's message history and purchase history.

[1291] "Analysis" refers to the process of analyzing collected data and extracting meaningful information.

[1292] "Estimated interests" refers to topics or areas in which a user is estimated to be particularly interested.

[1293] A "behavioral pattern" refers to a series of behavioral tendencies based on a user's past behavior.

[1294] A "customer profile" refers to a data set that compiles detailed interests and behavioral patterns for each user.

[1295] "Generative artificial intelligence" refers to AI that uses natural language processing technology to generate optimal messages and documents.

[1296] "Optimized posts" refer to messages that are individually customized based on a user's profile information and behavioral data.

[1297] "Terminal" refers to a device through which a user receives information, such as a smartphone or tablet.

[1298] "Response data" refers to the user's response to a posted document, such as a reply message or purchasing behavior.

[1299] "Smart device" refers to a device with advanced functionality that can connect to the Internet.

[1300] "Real-time" refers to near-instant processing and response.

[1301] A "prompt sentence" refers to an input sentence that causes a generative artificial intelligence to generate an optimized message.

[1302] This invention realizes a system for providing product recommendations and promotional messages optimized for each user on an online shopping site. This system is realized by collecting and analyzing information based on user actions, and generating and sending messages using generative artificial intelligence. Specific embodiments of this system are described below.

[1303] The user's smartphone or smart device functions as the terminal for this system. When a user searches for a product or sends a message, this action is detected by the terminal, and the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) are collected. The collected data is sent to the server in real time.

[1304] The server uses machine learning libraries like Python and TensorFlow to cleanse the incoming data and convert it into an analyzable format. Then, machine learning algorithms are used to analyze user behavior patterns and inferred interests. For example, past messaging history may indicate that a particular user is interested in beauty products.

[1305] Based on the analysis results and emotion estimation by the emotion engine, the server generates a detailed profile for each user. The generated profile is used to generate prompts for a generative AI (e.g., OpenAI's GPT-3). Based on the generated prompts, the generative AI generates optimized messages.

[1306] An example of a generated prompt is:

[1307] Ms. A (30 years old, female) is interested in a new beauty product. She is currently in an excited emotional state and has previously purchased beauty creams and face masks. Please generate the best message for her based on the following conditions:

[1308] New beauty products have arrived! Check them out for limited time discounts.

[1309] The server validates the generated optimization message and manually checks it for inappropriate content. After validation, the optimization message is sent to the user's smart device and displayed to the user in real time.

[1310] For example, if a user sends a message saying, "Do you have any new beauty products?", a customized recommendation message such as, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" will be displayed in real time on the user's device.

[1311] This invention enables the provision of information optimized for the interests and emotional state of each individual user, improving the user experience. Furthermore, by effectively utilizing collected user data, it is possible to provide individually optimized information in real time.

[1312] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1313] Step 1:

[1314] A user searches for a product or sends a message on a smartphone or smart device. The input data is the user's action, search query, and sent message. The device detects this user action as it is being performed and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history). The collected data is sent to a server in real time.

[1315] Step 2:

[1316] The server cleanses the incoming data and converts it into an analyzable format. The input data is the profile information and behavioral data collected in step 1. Inconsistent data is corrected and missing data is filled in. The cleansed data is obtained as the output.

[1317] Step 3:

[1318] The server uses machine learning algorithms to analyze the user's behavioral patterns and estimated interests. The input is cleansed data. Using machine learning libraries such as TensorFlow, the server analyzes the user's past messaging history and purchasing history to estimate specific interests. The analysis results are the user's estimated interests and behavioral patterns.

[1319] Step 4:

[1320] In addition to the analyzed data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. The input is the user's message content and behavioral data. Natural language processing technology is used to analyze emotions from the message and infer emotions such as "excitement." The emotion estimation result is obtained as the output.

[1321] Step 5:

[1322] The server generates a detailed profile for each user. The inputs are behavioral patterns, estimated interests, and estimated emotions. Each user's interests, behavioral patterns, and emotional state are compiled into a single dataset. The generated user profile is output.

[1323] Step 6:

[1324] The server generates a prompt for the generative AI to generate an optimization message. The input is the user profile. A generative AI such as OpenAI's GPT-3 is used to generate the prompt. The generated prompt is output.

[1325] Step 7:

[1326] The server uses the generated prompt sentence to generate an optimized message using a generative AI. The input is the generated prompt sentence. A generative AI model is used to generate the optimized message. The output is the optimized message.

[1327] Step 8:

[1328] The server validates the generated optimization message and checks for inappropriate expressions. The input is the generated optimization message. If any manual corrections are required, they are made. The optimized message without inappropriate expressions is output.

[1329] Step 9:

[1330] The server sends the verified optimization message to the user's terminal. The input is the verified optimization message. The message is displayed on the user's terminal in real time.

[1331] Step 10:

[1332] The server collects user reaction data after sending a message and feeds it back into subsequent messages. The input is user reaction data (reply messages, purchasing behavior, etc.). The collected reaction data is analyzed and used to generate optimized messages for subsequent messages. The results of the data analysis after feedback are output.

[1333] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1334] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1335] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1336] [Fourth embodiment]

[1337] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1338] 7, a 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.

[1339] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1340] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1341] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1342] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1343] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1344] The control object 443 includes a display device, LEDs in the eyes, and motors for driving 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 emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1345] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1346] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1347] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1348] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1349] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1350] The present invention is a system for automatically providing information optimized for each individual user, and is realized by the following processing steps.

[1351] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and all related data to the server.

[1352] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[1353] Based on the results of the above analysis, the server creates a profile for each customer. This allows each user's detailed interests and behavioral patterns to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products."

[1354] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A.

[1355] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[1356] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[1357] Finally, the server collects user response data (e.g., reply messages and purchasing behavior) after distribution and analyzes it again. Based on this response data, feedback is provided for generating subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests.

[1358] The processing flow will be explained below.

[1359] Step 1:

[1360] Detecting user actions

[1361] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[1362] Step 2:

[1363] Data collection

[1364] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[1365] Step 3:

[1366] Sending data

[1367] The device sends the collected data to the server. For example, it sends user A's profile information and behavioral data in JSON format to the server.

[1368] Step 4:

[1369] Receiving and storing data

[1370] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[1371] Step 5:

[1372] Data Preprocessing

[1373] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[1374] Step 6:

[1375] Customer Segmentation

[1376] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[1377] Step 7:

[1378] Generate customer profiles

[1379] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[1380] Step 8:

[1381] Optimized submission document generation request

[1382] The server requests the AI ​​to generate a post optimized for each customer profile, for example, by instructing it to use a template related to beauty products.

[1383] Step 9:

[1384] Generate optimized submission documents

[1385] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[1386] Step 10:

[1387] Review and correct submitted documents

[1388] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[1389] Step 11:

[1390] Preparing to send submitted documents

[1391] The server prepares to send the optimized post to each user's official LINE account.

[1392] Step 12:

[1393] Submitting a submission

[1394] The server sends the posted document to the user's LINE device. For example, it sends a posted document optimized for User A's LINE device.

[1395] Step 13:

[1396] Reaction data collection

[1397] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[1398] Step 14:

[1399] Reanalysis of reaction data

[1400] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[1401] Step 15:

[1402] Reflecting feedback

[1403] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[1404] Through these steps, the system automatically provides information optimized for each user, achieving continuous optimization.

[1405] Example 1

[1406] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1407] Conventional information provision systems have difficulty providing information optimized to users' interests. There is also a risk that posted documents containing inappropriate language may be distributed. Furthermore, the lack of a mechanism for effectively providing feedback on users' reactions reduces the quality and accuracy of the information provided. There is a need to solve these issues.

[1408] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1409] In this invention, the server includes: means for collecting profile information and behavioral data of a user based on the user's actions; means for receiving the collected data and cleansing it by correcting inconsistent data and filling in missing data; means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer; means for generating a post optimized for each customer using generative artificial intelligence based on the generated customer profile; means for verifying the generated optimized post, checking for inappropriate language, and manually correcting it; means for transmitting the generated optimized post to the user's device; and means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding it back into subsequent posts. This enables the provision of information optimized to the user's interests, reduces the risk of inappropriate language, and improves the quality and accuracy of information provided by effectively feeding back user reactions.

[1410] A "user action" is a series of operations that a user performs on the system, such as sending a message or purchasing a product.

[1411] "Profile information" refers to personal information such as a user's name, gender, and age.

[1412] "Behavioral data" refers to a record of a user's past behavior, such as message history and purchase history.

[1413] "Cleansing" is the process of organizing and formatting data to correct inconsistent data and fill in missing data.

[1414] "Analysis" is the process of deriving user behavior patterns and estimated interests based on collected data.

[1415] A "customer profile" is a data set generated based on analyzed data that contains detailed interests and behavioral patterns for each user.

[1416] "Generative AI" is an AI technology that generates posts optimized for each customer based on their customer profile.

[1417] A "posted document" is a message or piece of text that is sent to a user.

[1418] "Response data" refers to data such as replies and purchasing behavior of users in response to received posted documents.

[1419] "Feedback" refers to the process of reflecting collected response data in the generation of subsequent submitted documents.

[1420] MODE FOR CARRYING OUT THE INVENTION

[1421] This invention is a system that automatically provides information optimized for each individual user, and specifically, it operates in cooperation with a server, a terminal, and a user. The system of this invention operates as follows.

[1422] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (such as name, gender, and age) and behavioral data (such as message history and purchase history), consolidates this data, and sends it to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device will send this message and the associated profile information and behavioral data to the server.

[1423] The server processes the data received from the device. Specifically, the server cleanses the received data, corrects inconsistent data, and fills in missing data. It then converts the data into an analyzable format (for example, JSON format). Next, the server uses machine learning algorithms (using TensorFlow, PyTorch, etc.) to analyze the user's behavioral patterns and estimated interests. From past message history, the server learns that a particular user has a high interest in beauty products.

[1424] Based on the analysis results, the server generates a profile for each customer. For example, User A's profile includes information such as "female in her 30s, highly interested in beauty products." By generating this profile, the user's detailed interests and behavioral patterns are compiled into a single data set.

[1425] Next, the server requests a generative artificial intelligence (AI) to generate an optimized post based on the customer profile. The generative AI uses natural language processing technology such as OpenAI's GPT-4 to generate a message optimized for each user. For example, a message might be generated for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out." The following is an example of a prompt sentence that can be used:

[1426] "Please create a promotional message for a new beauty product for User A, a woman in her 30s who has a high interest in beauty products."

[1427] The generated submission is validated by the server and checked for inappropriate language, which is then manually corrected.

[1428] Finally, the server sends the optimized post to each user's device, which then displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[1429] After distribution, the server collects user response data (such as reply messages and purchasing behavior) and analyzes it again. This response data is used as feedback for generating subsequent posts, enabling even more optimized information to be provided. In this way, the system can continuously provide optimal information tailored to the user's interests.

[1430] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1431] Step 1:

[1432] Users can send messages to official accounts or purchase products.

[1433] Input: User action (sending a message or purchasing a product)

[1434] Output: User action detection data

[1435] Specific operation: A user sends a message to the official LINE account asking, "Do you have any new beauty products?"

[1436] Step 2:

[1437] The device detects this user's actions and collects the user's profile information and behavioral data.

[1438] Input: User action detection data

[1439] Output: Collected user profile information and behavioral data

[1440] Specific operation: The device detects the message and collects user A's name, gender, age, past message history, and purchase history.

[1441] Step 3:

[1442] The data collected by the device is consolidated and sent to the server.

[1443] Input: Collected user profile information and behavioral data

[1444] Output: Consolidated data sent from the device to the server

[1445] Specific operation: The terminal compiles the collected data of user A and sends it to the server.

[1446] Step 4:

[1447] The server cleanses the incoming data, corrects inconsistent data, and fills in missing data.

[1448] Input: Integrated data sent from the device

[1449] Output: Cleansed data

[1450] Specific behavior: The server will correct inconsistencies in the data (e.g. missing names) and complete them as needed.

[1451] Step 5:

[1452] The server converts the cleansed data into a parsable format.

[1453] Input: Cleansed data

[1454] Output: Data in a parsable format (e.g. JSON)

[1455] What happens: The server converts the data into a parsable format, such as JSON.

[1456] Step 6:

[1457] The server uses machine learning algorithms to analyze users' behavioral patterns and estimated interests.

[1458] Input: Data in a parsable format

[1459] Output: Analysis results of user behavior patterns and estimated interests

[1460] Specific operation: The server uses TensorFlow to analyze User A's interest in beauty products based on their past message history.

[1461] Step 7:

[1462] The server generates a customer profile based on the analysis results.

[1463] Input: Analysis results of user behavior patterns and estimated interests

[1464] Output: Generated customer profile

[1465] Specific operation: The server creates a profile for "a woman in her 30s with a strong interest in beauty products."

[1466] Step 8:

[1467] The server requests the generative artificial intelligence to generate an optimized submission document based on the customer profile.

[1468] Input: Generated customer profile

[1469] Output: Prompt sentence requested by the generative AI

[1470] Specific operation: The server requests a generative artificial intelligence (e.g., OpenAI GPT-4) to "generate the optimal message for User A, a woman in her 30s who has a high interest in beauty products."

[1471] Step 9:

[1472] Generative AI generates optimized submission documents.

[1473] Input: Prompt sentence requested by the generative artificial intelligence

[1474] Output: The generated optimized submission document

[1475] What it does: The generative AI generates a message saying, "The latest beauty products have arrived! Check them out at a limited-time discount."

[1476] Step 10:

[1477] The server validates the generated optimized submission document and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[1478] Input: Generated optimized submission document

[1479] Output: Validated and corrected submission document

[1480] Specific behavior: The server validates the generated message, checks for errors in specific product information, etc., and manually corrects them.

[1481] Step 11:

[1482] The server transmits the optimized submitted document to the user's terminal.

[1483] Input: Validated and corrected submitted document

[1484] Output: Posted document sent to terminal

[1485] Specific operation: The server sends the optimized submission document to the terminal.

[1486] Step 12:

[1487] The terminal displays the posted document to the user.

[1488] Input: Posted document sent to terminal

[1489] Output: The submitted document as displayed to the user

[1490] Specific operation: The device displays a message on User A's LINE message saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[1491] Step 13:

[1492] The server collects user response data (reply messages and purchasing behavior) after delivery and analyzes it again.

[1493] Input: User response data

[1494] Output: Parsed reaction data

[1495] Specific operation: The server collects and analyzes User A's reply messages and purchasing behavior.

[1496] Step 14:

[1497] The server provides feedback for the next submission.

[1498] Input: Parsed reaction data

[1499] Output: Feedback information reflected in the next submission generation

[1500] Specific operation: The server optimizes the process of generating the next submitted document based on the analysis results.

[1501] (Application example 1)

[1502] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1503] On modern online shopping sites, diverse users each have different interests, so providing information in a single location reduces user satisfaction. Another issue is the difficulty of effectively analyzing collected user behavior data and making personalized, appropriate product recommendations.

[1504] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1505] In this invention, the server includes a means for sending user action data to the server using a REST API, a means for generating an optimized message by making a request to the generative AI based on the analyzed data, and a means for sending the generated message again from the server to the terminal, thereby enabling personalized product recommendations based on each user's interests.

[1506] "User actions" refer to specific operations that users perform on the system, such as purchasing, browsing, sending messages, etc.

[1507] "Profile information" refers to basic personal information such as a user's name, gender, and age.

[1508] "Behavioral data" refers to data relating to specific behaviors of users, such as their message history and purchase history.

[1509] "Server" means a central management system for collecting, analyzing, generating and transmitting user profile information and behavioral data.

[1510] "Generative AI" refers to artificial intelligence techniques used to generate optimized messages based on user profiles.

[1511] An "optimized submission" is a document created by generative artificial intelligence to suit a specific user based on the user's profile.

[1512] A "user terminal" is a device, such as a smartphone or computer, that a user uses to receive information.

[1513] A "REST API" is a standard interface for sending data to a server.

[1514] "Machine learning algorithms" are data analysis techniques used to analyze user behavior patterns and inferred interests.

[1515] "Response data" refers to responses and purchasing behavior of users in response to information from the system.

[1516] The present invention is a system for a mail-order site that provides personalized information based on a user's interests. This system is realized by the following method and apparatus.

[1517] First, when a user purchases or browses a product on an online shopping site, their behavioral data (e.g., product ID, category, etc.) is automatically collected. This collection is performed by the user's device issuing a request and sending the data to the server via a REST API.

[1518] The server cleans the received data and uses machine learning algorithms to analyze user behavior patterns and inferred interests. To do this, the server uses Python programs and libraries such as TensorFlow and Scikit-learn.

[1519] Based on the analysis results, the server creates a detailed profile for each user, including the user's interests and behavioral patterns. For example, a profile might be generated that describes a woman in her 30s with a strong interest in beauty products.

[1520] Next, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate a post optimized for each user. An example of a prompt is, "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message."

[1521] The generated post is verified by the server and checked for inappropriate content, after which the generated optimized post is sent back to the user's device and displayed to the user.

[1522] As described above, personalized information based on the user's interests can be provided. This system enables optimized product recommendations for each user, improving the user experience. As a specific example of use, if a woman in her 30s frequently purchases beauty products, a message will be displayed saying, "Hello, we have new beauty products that suit your tastes. They are on special discount, so please check them out."

[1523] In this way, user satisfaction at the online shopping site can be improved.

[1524] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1525] Step 1:

[1526] When a user purchases or browses products on an online shopping site, their device collects their behavioral data. This data includes information such as product ID, category, and time. The collected data is sent to the server via a REST API. This input data is output as a request to the server.

[1527] Step 2:

[1528] The server receives the behavioral data sent from the device and performs data cleansing. Specifically, it corrects inconsistent data and complements missing data. The cleansed data is output as a dataset for analysis.

[1529] Step 3:

[1530] The server then runs machine learning algorithms on the cleansed data to analyze user behavior patterns and inferred interests. This analysis uses libraries such as Scikit-learn and TensorFlow. The analysis results are output as a user profile.

[1531] Step 4:

[1532] Based on the generated user profile, the server sends a prompt to a generative artificial intelligence (e.g., OpenAI GPT-3) to generate an optimized post. For example, the prompt "User profile: Female in her 30s, highly interested in beauty products. Please create a recommendation message" is used as the source data, and the generated recommendation message is output.

[1533] Step 5:

[1534] The server validates the generated optimized submission document, checking for inappropriate language, and outputs the validated document as the final message.

[1535] Step 6:

[1536] The server sends a final message to the user's terminal, which displays the message and the user receives the recommended information. The user then checks the recommended product information.

[1537] Step 7:

[1538] When a user responds to a recommended message (for example, by purchasing a product or replying to a message), the response data is collected again. The collected response data is sent to the server via a REST API. The sent data is output as feedback data.

[1539] Step 8:

[1540] The server analyzes the collected response data and provides feedback for the next submission. This feedback enables the generation of more appropriate recommendation messages in the future. The feedback results are output as the next profile analysis data.

[1541] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1542] This invention is a system that automatically provides information optimized for each user and further optimizes the information by recognizing the user's emotions, and is realized by the following processing steps.

[1543] When a user sends a message to an official account or purchases a product, the device detects the user's action. The device collects the user's profile information (e.g., name, gender, age) and behavioral data (e.g., message history and purchase history) and sends them to the server. For example, if a user sends a message saying, "Do you have any new beauty products?", the device sends this message and all related data to the server.

[1544] The server processes the data received from the device. First, the server cleanses the data (for example, correcting inconsistent data and filling in missing data) and converts it into a format that can be analyzed. Next, the server uses machine learning algorithms to analyze the user's behavioral patterns and inferred interests. For example, the server may learn from past messaging history that a particular user is frequently interested in beauty products.

[1545] In addition to the analysis data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. For example, when a user sends a message saying, "Do you have any new beauty products?", the emotion engine infers emotions such as "excitement" from the user's text.

[1546] Based on the analysis results and emotion estimation results, the server generates a profile for each customer. This allows each user's detailed interests and behavioral patterns, as well as their emotional state, to be compiled into a single data set. For example, User A's profile includes information such as "female in her 30s, high interest in beauty products, current emotional state: excited."

[1547] Using the generated customer profile, the server requests a generative artificial intelligence (AI) to generate an optimized post document. The generative AI uses natural language processing technology to generate an optimized message based on each individual customer profile. For example, a message such as "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" is generated for User A. The tone and content of the message are also further customized based on the estimation results of the emotion engine.

[1548] The server then validates the generated optimized submission document, checking for inappropriate language, and manually corrects any inappropriate language found.

[1549] The server sends the optimized post to each user's device. The device displays the post to the user. For example, information about the beauty product mentioned above is sent to User A's LINE.

[1550] Finally, the server collects and analyzes user response data (e.g., reply messages and purchasing behavior) after the message is posted. Based on this response data, feedback is provided for the generation of subsequent posts, enabling even more optimized information provision. In this way, the system can continuously provide optimal information tailored to the user's interests and emotional state.

[1551] The processing flow will be explained below.

[1552] Step 1:

[1553] Detecting user actions

[1554] The device detects user actions (such as sending messages or purchasing behavior). When a user sends a message saying, "Are there any new beauty products?", the device recognizes this action.

[1555] Step 2:

[1556] Data collection

[1557] The device collects user profile information (such as name, gender, and age) and behavioral data (such as past message history and purchase history). For example, it retrieves user A's profile information and most recent purchase history from a database.

[1558] Step 3:

[1559] Sending data

[1560] The device sends the collected profile information and behavioral data to the server. For example, the device sends User A's profile information and behavioral data in JSON format to the server.

[1561] Step 4:

[1562] Receiving and storing data

[1563] The server receives the data sent from the terminal and stores it in the database. For example, the information of User A is added to the database as new data.

[1564] Step 5:

[1565] Data Preprocessing

[1566] The server cleanses the data it receives and converts it into an analyzable format, for example by correcting inconsistent data or filling in missing data.

[1567] Step 6:

[1568] Customer Segmentation

[1569] The server uses a machine learning algorithm to segment (cluster) each customer. For example, User A is classified into the "beauty products category."

[1570] Step 7:

[1571] Generate customer profiles

[1572] The server generates individual customer profiles based on the segmentation results. For example, it adds information such as "female in her 30s, highly interested in beauty products" to User A's profile.

[1573] Step 8:

[1574] Emotion recognition data collection

[1575] The device sends the user's message content and behavioral data to the emotion engine. For example, a message sent by user A saying "Do you have any new beauty products?" is sent to the emotion engine.

[1576] Step 9:

[1577] Emotion estimation

[1578] The server uses the emotion engine to estimate the user's emotion from the received message. For example, the emotion engine estimates the emotion "excitement" from User A's message.

[1579] Step 10:

[1580] Optimized submission document generation request

[1581] The server requests the generative AI to generate a post optimized for each customer based on their profile and the emotion estimation results. For example, it instructs the AI ​​to use a template for users who are interested in and excited about beauty products.

[1582] Step 11:

[1583] Generate optimized submission documents

[1584] Generative AI uses natural language processing technology to generate posts optimized for each customer. For example, it might generate a message for User A saying, "The latest beauty products have arrived! They're on limited-time discounts, so be sure to check them out."

[1585] Step 12:

[1586] Review and correct submitted documents

[1587] The server validates the generated submission and checks for inappropriate language. If inappropriate language is found, it is manually corrected.

[1588] Step 13:

[1589] Preparing to send submitted documents

[1590] The server prepares to send the optimized post to the user's device. For example, it prepares to send a message to User A's LINE device.

[1591] Step 14:

[1592] Submitting a submission

[1593] The server sends the posted document to the user's LINE device. For example, it delivers a posted document optimized for User A's LINE device.

[1594] Step 15:

[1595] Reaction data collection

[1596] The device collects user responses (such as reply messages and purchasing behavior) and sends them to the server. For example, data is collected when User A replies to a message saying, "I want to buy it."

[1597] Step 16:

[1598] Reanalysis of reaction data

[1599] The server receives the response data and performs analysis, for example, to assess how effective a particular message was.

[1600] Step 17:

[1601] Reflecting feedback

[1602] The server will use the analysis results to provide feedback for subsequent submissions. For example, it may provide more targeted reward information to User A.

[1603] Through these steps, the system not only automatically provides information optimized for each user, but also takes emotions into account to provide even more effective information.

[1604] Example 2

[1605] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1606] Conventional information delivery systems have difficulty providing optimized information that takes into account the individual interests and emotional state of each user. This has made improving user experience and engagement a challenge. Another issue is the cumbersome process of verifying and manually correcting generated messages, making it difficult to provide information efficiently.

[1607] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1608] In this invention, the server includes means for collecting profile information and behavioral data based on user actions, means for cleansing the collected data, means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer, means for estimating emotions from the content of the user's messages, means for generating a customer profile based on the analysis results and the emotion estimation results, means for generating a post optimized for each customer using generative artificial intelligence, means for verifying and manually correcting the generated optimized post, means for sending the generated optimized post to the user's device, and means for collecting user reaction data after transmission, analyzing the reaction data, and providing feedback for subsequent posts. This makes it possible to provide optimal information based on each user's detailed interests and emotional state.

[1609] "User actions" refers to a series of operations or actions performed by a user, such as sending a message to an official account or purchasing a product.

[1610] "Profile information" refers to information that can be used to personally identify a user, such as the user's name, gender, and age.

[1611] "Behavioral data" refers to data related to a user's past operations and behavior, such as message history and purchase history.

[1612] "Cleansing" refers to the process of correcting inconsistent data, filling in missing data, and converting data into an analyzable format.

[1613] "Analysis" refers to the processing of collected data to identify user behavioral patterns and interests.

[1614] "Estimated interests" refers to estimating the interests and concerns that a user is likely to have based on past behavioral data.

[1615] A "behavioral pattern" refers to a series of actions or patterns of operations that a user exhibits during a specific period of time.

[1616] "Emotion estimation" refers to the process of inferring the emotion contained in a message sent by a user from its content.

[1617] A "customer profile" refers to a data set that compiles a user's interests, behavioral patterns, and emotional state.

[1618] "Generative AI" refers to AI technology that automatically generates optimized messages and text based on given data.

[1619] "Optimized posts" refer to messages that are customized for individual users based on their customer profile.

[1620] "Response data" refers to data such as reply messages sent by users in response to messages received and subsequent purchasing behavior.

[1621] "Feedback" refers to the process of analyzing collected response data and using it to improve the quality of subsequent submissions.

[1622] This invention is a system that automatically provides information optimized for each user and further optimizes it by recognizing the user's emotions. The system is mainly composed of a server, a terminal, and a user. Details of each function and their specific implementation methods are explained below.

[1623] Basic configuration

[1624] The server plays a central role in data cleansing, data analysis, emotion estimation, and message generation. The hardware used is a high-performance data processing server, and the software used includes Hadoop, Spark, IBM Watson, Microsoft Azure's emotion recognition service, Scikit-learn, and TensorFlow. The generative artificial intelligence used is OpenAI GPT-4.

[1625] The terminal is responsible for detecting user actions and sending data to the server. This can be implemented on popular devices such as smartphones and PCs.

[1626] Users take action on the system and receive information optimized based on their response.

[1627] Data collection and transmission

[1628] When a user sends a message to the official account asking, "Do you have any new beauty products?", the device detects this action. The device collects the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) and sends them to the server in real time.

[1629] Data Cleansing and Analysis

[1630] The server cleanses the data it receives. This process includes correcting inconsistent data and filling in missing data, for example by filling in missing fields with appropriate default values.

[1631] The server then analyzes the cleansed data using Hadoop or Spark. Machine learning algorithms (Scikit-learn or TensorFlow) are used to analyze user behavior patterns and interests. The analysis results include information on which beauty products a particular user is frequently interested in.

[1632] Emotion estimation

[1633] The server uses IBM Watson or Microsoft Azure emotion recognition services to infer emotions from the content of a user's message. For example, it infers the emotion "excitement" from the message "Do you have any new beauty products?"

[1634] Customer profile generation

[1635] The server generates a detailed profile of the user based on the analysis results and emotion estimation results. For example, it creates a profile such as "Female in her 30s, highly interested in beauty products, current emotional state: excited."

[1636] Message Creation and Delivery

[1637] The server requests the OpenAI GPT-4 to generate an optimized message based on the generated profile. For example, the request is made with the following prompt:

[1638] Customer Profile:

[1639] Name: User A

[1640] Gender: Female

[1641] Age: 30

[1642] Interests: Beauty products

[1643] Current emotional state: Excited

[1644] message:

[1645] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[1646] The generated message is validated by the server to check for inappropriate language and, if necessary, manually corrected.

[1647] Finally, the optimized message is sent to the user's device, which displays it to the user. The user reacts, which triggers the next cycle of data collection and analysis.

[1648] Reaction data collection and feedback

[1649] The server collects user responses (reply messages and purchasing behavior) after distribution and analyzes this data. Based on the analysis results, the quality of subsequent submissions can be improved, resulting in even more optimized information provision.

[1650] This process allows the system to continually provide optimal information tailored to the user's interests and emotional state.

[1651] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1652] Step 1:

[1653] User Actions and Data Collection

[1654] A user sends a message to your official account saying, "Are there any new beauty products?"

[1655] The device detects this action and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history).

[1656] Input: User messages, profile information, behavioral data

[1657] Output: Collected dataset

[1658] Specific operation: Software in the device begins collecting data when the user sends a message.

[1659] Step 2:

[1660] Data transmission

[1661] The terminal transmits the collected data set to the server.

[1662] Input: Collected dataset

[1663] Output: The dataset sent to the server

[1664] Specific operation: The device uploads data to a server via the Internet.

[1665] Step 3:

[1666] Data Cleansing

[1667] The server cleanses the received dataset, specifically correcting inconsistent data and filling in missing data.

[1668] Input: The dataset sent to the server

[1669] Output: A cleansed dataset

[1670] Specific operations: Use data processing engines such as Hadoop and Spark to check data integrity and convert it into a unified format.

[1671] Step 4:

[1672] Data analysis

[1673] The server uses the cleansed data to analyze user behavior patterns and interests using machine learning algorithms such as Scikit-learn and TensorFlow.

[1674] Input: Cleansed dataset

[1675] Output: Analysis results (user behavior patterns, estimated interests)

[1676] Specific behavior: Apply machine learning models to training data to derive quantitative user behavior patterns and interests.

[1677] Step 5:

[1678] Emotion estimation

[1679] The server uses IBM Watson and Microsoft Azure emotion recognition services to estimate emotions from the content of the user's message.

[1680] Input: Message content, behavioral data

[1681] Output: Emotion estimation result

[1682] What it does: The emotion recognition engine analyzes text and extracts specific emotions (e.g., excitement, joy, sadness).

[1683] Step 6:

[1684] Customer profile generation

[1685] The server generates a detailed profile of the user based on the analysis results and emotion estimation results.

[1686] Input: Analysis results, emotion estimation results

[1687] Output: Customer profile

[1688] Specific operation: The analysis results are integrated with emotional data, and a detailed profile customized for each user is stored in a database.

[1689] Step 7:

[1690] Requesting Message Creation

[1691] The server requests the generative artificial intelligence (OpenAI GPT-4) to generate an optimized message based on the generated customer profile. The server instructs the generation using a prompt.

[1692] Input: Customer profile, prompt statement

[1693] Output: Generated optimization messages

[1694] Specific behavior: Provide the following prompt to the generative AI:

[1695] Customer Profile:

[1696] Name: User A

[1697] Gender: Female

[1698] Age: 30

[1699] Interests: Beauty products

[1700] Current emotional state: Excited

[1701] message:

[1702] New beauty products have arrived and are on limited-time discounts, so be sure to check them out!

[1703] Step 8:

[1704] Message verification and correction

[1705] The server validates generated messages for inappropriate language and manually corrects them as necessary.

[1706] Input: The generated message

[1707] Output: Validated and corrected message

[1708] What it does: It performs text filtering and human operator review.

[1709] Step 9:

[1710] Message Delivery

[1711] The server sends the optimized message to the user's terminal.

[1712] Input: Validated and corrected message

[1713] Output: Message sent to the terminal

[1714] Specific operation: The server sends a message to the terminal via the Internet.

[1715] Step 10:

[1716] Reaction data collection and feedback

[1717] After delivery, the server collects and analyzes the user's reactions (reply messages and purchasing behavior), and reflects the results in the generation of subsequent messages.

[1718] Input: User response data

[1719] Output: Feedback data

[1720] What it does: The reaction data is stored in a database and used to retrain the machine learning model.

[1721] (Application example 2)

[1722] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1723] Current online shopping sites generally only provide product recommendations and promotional messages to users, but these messages are not optimized to reflect the interests and emotional state of individual users, making it difficult to improve the user experience. Another issue is that collected user data cannot be effectively utilized, making it difficult to provide individually optimized information in real time. These issues need to be resolved.

[1724] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1725] In this invention, the server includes means for collecting profile information and behavioral data of a user based on the user's actions, means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer, means for generating a post optimized for each customer using a generative artificial intelligence based on the generated customer profile, means for transmitting the generated optimized post to the user's terminal, means for collecting user reaction data after the transmission, analyzing the reaction data, and feeding the data back into subsequent posts, means for providing optimized product recommendations and promotional messages in real time using a smart device, means for generating a prompt sentence for the generative artificial intelligence to generate the optimized message based on the analysis results, and means for generating the optimized message using the generative artificial intelligence using the generated prompt sentence, thereby enabling the provision of information optimized for the interests and emotional state of each individual user.

[1726] "User action" refers to the operations or inputs that a user makes to the system.

[1727] "Profile Information" refers to basic personal information such as a user's name, gender, and age.

[1728] "Behavioral data" refers to activity data such as a user's message history and purchase history.

[1729] "Analysis" refers to the process of analyzing collected data and extracting meaningful information.

[1730] "Estimated interests" refers to topics or areas in which a user is estimated to be particularly interested.

[1731] A "behavioral pattern" refers to a series of behavioral tendencies based on a user's past behavior.

[1732] A "customer profile" refers to a data set that compiles detailed interests and behavioral patterns for each user.

[1733] "Generative artificial intelligence" refers to AI that uses natural language processing technology to generate optimal messages and documents.

[1734] "Optimized posts" refer to messages that are individually customized based on a user's profile information and behavioral data.

[1735] "Terminal" refers to a device through which a user receives information, such as a smartphone or tablet.

[1736] "Response data" refers to the user's response to a posted document, such as a reply message or purchasing behavior.

[1737] "Smart device" refers to a device with advanced functionality that can connect to the Internet.

[1738] "Real-time" refers to near-instant processing and response.

[1739] A "prompt sentence" refers to an input sentence that causes a generative artificial intelligence to generate an optimized message.

[1740] This invention realizes a system for providing product recommendations and promotional messages optimized for each user on an online shopping site. This system is realized by collecting and analyzing information based on user actions, and generating and sending messages using generative artificial intelligence. Specific embodiments of this system are described below.

[1741] The user's smartphone or smart device functions as the terminal for this system. When a user searches for a product or sends a message, this action is detected by the terminal, and the user's profile information (name, gender, age, etc.) and behavioral data (message history, purchase history, etc.) are collected. The collected data is sent to the server in real time.

[1742] The server uses machine learning libraries like Python and TensorFlow to cleanse the incoming data and convert it into an analyzable format. Then, machine learning algorithms are used to analyze user behavior patterns and inferred interests. For example, past messaging history may indicate that a particular user is interested in beauty products.

[1743] Based on the analysis results and emotion estimation by the emotion engine, the server generates a detailed profile for each user. The generated profile is used to generate prompts for a generative AI (e.g., OpenAI's GPT-3). Based on the generated prompts, the generative AI generates optimized messages.

[1744] An example of a generated prompt is:

[1745] Ms. A (30 years old, female) is interested in a new beauty product. She is currently in an excited emotional state and has previously purchased beauty creams and face masks. Please generate the best message for her based on the following conditions:

[1746] New beauty products have arrived! Check them out for limited time discounts.

[1747] The server validates the generated optimization message and manually checks it for inappropriate content. After validation, the optimization message is sent to the user's smart device and displayed to the user in real time.

[1748] For example, if a user sends a message saying, "Do you have any new beauty products?", a customized recommendation message such as, "New beauty products have arrived! They're on limited-time discounts, so be sure to check them out" will be displayed in real time on the user's device.

[1749] This invention enables the provision of information optimized for the interests and emotional state of each individual user, improving the user experience. Furthermore, by effectively utilizing collected user data, it is possible to provide individually optimized information in real time.

[1750] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1751] Step 1:

[1752] A user searches for a product or sends a message on a smartphone or smart device. The input data is the user's action, search query, and sent message. The device detects this user action as it is being performed and collects the user's profile information (name, gender, age) and behavioral data (message history, purchase history). The collected data is sent to a server in real time.

[1753] Step 2:

[1754] The server cleanses the incoming data and converts it into an analyzable format. The input data is the profile information and behavioral data collected in step 1. Inconsistent data is corrected and missing data is filled in. The cleansed data is obtained as the output.

[1755] Step 3:

[1756] The server uses machine learning algorithms to analyze the user's behavioral patterns and estimated interests. The input is cleansed data. Using machine learning libraries such as TensorFlow, the server analyzes the user's past messaging history and purchasing history to estimate specific interests. The analysis results are the user's estimated interests and behavioral patterns.

[1757] Step 4:

[1758] In addition to the analyzed data, the server uses an emotion engine to infer emotions from the user's message content and behavioral data. The input is the user's message content and behavioral data. Natural language processing technology is used to analyze emotions from the message and infer emotions such as "excitement." The emotion estimation result is obtained as the output.

[1759] Step 5:

[1760] The server generates a detailed profile for each user. The inputs are behavioral patterns, estimated interests, and estimated emotions. Each user's interests, behavioral patterns, and emotional state are compiled into a single dataset. The generated user profile is output.

[1761] Step 6:

[1762] The server generates a prompt for the generative AI to generate an optimization message. The input is the user profile. A generative AI such as OpenAI's GPT-3 is used to generate the prompt. The generated prompt is output.

[1763] Step 7:

[1764] The server uses the generated prompt sentence to generate an optimized message using a generative AI. The input is the generated prompt sentence. A generative AI model is used to generate the optimized message. The output is the optimized message.

[1765] Step 8:

[1766] The server validates the generated optimization message and checks for inappropriate expressions. The input is the generated optimization message. If any manual corrections are required, they are made. The optimized message without inappropriate expressions is output.

[1767] Step 9:

[1768] The server sends the verified optimization message to the user's terminal. The input is the verified optimization message. The message is displayed on the user's terminal in real time.

[1769] Step 10:

[1770] The server collects user reaction data after sending a message and feeds it back into subsequent messages. The input is user reaction data (reply messages, purchasing behavior, etc.). The collected reaction data is analyzed and used to generate optimized messages for subsequent messages. The results of the data analysis after feedback are output.

[1771] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1772] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1773] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1774] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1775] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1776] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1777] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1778] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1779] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1780] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1781] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1782] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1783] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1784] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1785] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1786] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1787] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1788] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1789] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1790] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1791] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1792] The following is further disclosed regarding the above embodiment.

[1793] (Claim 1)

[1794] means for collecting profile information and behavioral data of users based on their actions;

[1795] A means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer;

[1796] A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile;

[1797] means for transmitting the generated optimized submission document to a user's terminal;

[1798] means for collecting user reaction data after said transmission, analyzing said reaction data, and feeding back the data to subsequent submitted documents;

[1799] A system including:

[1800] (Claim 2)

[1801] 2. The system according to claim 1, wherein the means for generating estimated interests and behavioral patterns for each customer uses a machine learning algorithm.

[1802] (Claim 3)

[1803] 2. The system of claim 1, further comprising means for verifying the generated submission and manually correcting it.

[1804] "Example 1"

[1805] (Claim 1)

[1806] means for collecting profile information and behavioral data of users based on their actions;

[1807] a means for receiving the collected data and cleansing it by correcting inconsistent data and completing missing data;

[1808] A means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer;

[1809] A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile;

[1810] A means to review the generated optimized submission document, check for inappropriate language, and manually correct it; and

[1811] means for transmitting the generated optimized submission document to a user's terminal;

[1812] means for collecting user reaction data after said transmission, analyzing said reaction data, and feeding back the data to subsequent submitted documents;

[1813] A system including:

[1814] (Claim 2)

[1815] 2. The system according to claim 1, wherein the means for generating estimated interests and behavioral patterns for each customer uses a machine learning algorithm.

[1816] (Claim 3)

[1817] 2. The system of claim 1, further comprising means for verifying the generated submission and manually correcting it.

[1818] "Application Example 1"

[1819] (Claim 1)

[1820] means for collecting profile information and behavioral data of users based on their actions;

[1821] A means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer;

[1822] A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile;

[1823] means for displaying the generated optimized submission on a user's terminal;

[1824] A means for transmitting user action data to a server using a REST API;

[1825] A means for requesting a generative artificial intelligence to generate an optimization message based on the data analyzed by the machine learning algorithm;

[1826] means for transmitting the generated message from the server to the terminal again;

[1827] means for collecting user reaction data after said transmission, analyzing said reaction data, and feeding back the data to subsequent submitted documents;

[1828] A system including:

[1829] (Claim 2)

[1830] 2. The system according to claim 1, wherein the means for generating estimated interests and behavioral patterns for each customer uses a machine learning algorithm.

[1831] (Claim 3)

[1832] 2. The system of claim 1, further comprising means for verifying the generated submission and manually correcting it.

[1833] "Example 2: Combining Emotion Engines"

[1834] (Claim 1)

[1835] means for collecting user profile information and behavioral data based on user actions;

[1836] a means for cleansing the collected data;

[1837] A means for analyzing the cleansed data to generate estimated interests and behavioral patterns for each customer;

[1838] means for estimating a user's emotion from the content of the user's message;

[1839] A means for generating a customer profile based on the analysis result and the emotion estimation result;

[1840] A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile;

[1841] A means for verifying and manually correcting the generated optimized submission documentation; and

[1842] means for transmitting the generated optimized submission document to a user's terminal;

[1843] A means for collecting user reaction data after submission, analyzing the reaction data, and providing feedback for subsequent submissions;

[1844] A system including:

[1845] (Claim 2)

[1846] 2. The system according to claim 1, wherein the means for generating estimated interests and behavioral patterns for each customer uses a machine learning algorithm.

[1847] (Claim 3)

[1848] 2. The system according to claim 1, further comprising means for verifying the generated posted document and checking for inappropriate language.

[1849] "Application example 2 when combining emotion engines"

[1850] (Claim 1)

[1851] means for collecting profile information and behavioral data of users based on their actions;

[1852] A means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer;

[1853] A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile;

[1854] means for transmitting the generated optimized submission document to a user's terminal;

[1855] means for collecting user reaction data after said transmission, analyzing said reaction data, and feeding back the data to subsequent submitted documents;

[1856] A means to provide real-time optimized product recommendations and promotional messages using smart devices,

[1857] a means for generating a prompt sentence for the generative artificial intelligence to generate an optimization message based on the analysis result;

[1858] A means for generating an optimization message using a generative artificial intelligence using the generated prompt sentence;

[1859] A system including:

[1860] (Claim 2)

[1861] 2. The system according to claim 1, wherein the means for generating estimated interests and behavioral patterns for each customer uses a machine learning algorithm.

[1862] (Claim 3)

[1863] 2. The system of claim 1, further comprising means for verifying the generated submission and manually correcting it. [Explanation of symbols]

[1864] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for collecting profile information and behavioral data of users based on their actions; A means for analyzing the collected data to generate estimated interests and behavioral patterns for each customer; A means for generating a posting document optimized for each customer using generative artificial intelligence based on the generated customer profile; means for transmitting the generated optimized submission document to a user's terminal; means for collecting user reaction data after said transmission, analyzing said reaction data, and feeding back the data to subsequent submitted documents; A system including:

2. 2. The system according to claim 1, wherein the means for generating estimated interests and behavioral patterns for each customer uses a machine learning algorithm.

3. 2. The system of claim 1, further comprising means for verifying the generated submission and for manually correcting it.

Images