System

Abstract

To provide a system that semi-automatically generates messages using phrases and reply styles suited to individual users.SOLUTION: A system includes: means for acquiring a message received in a message application; a prompt sentence for instructing the system to generate a reply to the received message by imitating messages sent by a user, based on a history of the messages sent by the user in the message application; means for generating candidate replies to the received message using generative AI; means for presenting the generated candidate replies to the user, and receiving editing; and means for sending the edited result as a message of the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] Conventional messenger systems require users to manually create and send messages, which requires time and effort. They also require users to properly understand the content and urgency of received messages and generate replies accordingly. [Means for solving the problem]

[0005] The system of this invention analyzes traffic within the messenger and labels received messages to determine the recipient's request and its urgency. Furthermore, it learns from the user's sending history and semi-automatically generates text and stamps using phrases and reply styles that suit each individual user. This reduces the burden on users and enables more efficient communication. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0014] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0027] "Example 1"

[0028] The system of the present invention is integrated into a messenger application and analyzes the message traffic sent and received by users in real time. This analysis is performed using natural language processing (NLP) techniques to identify the content, tone, and urgency of the message. For example, if a message says, "Please send me the materials for tomorrow's meeting," the message is labeled as a "Request for meeting materials" and the urgency is set to "High."

[0029] "Example 2"

[0030] The system of the present invention also learns from a user's past sending history. This learning is done using deep learning techniques to understand the user's phrasing and response style. For example, if a user often responds with "Okay, I'll send it right away," the system will learn that same response style.

[0031] "Example 3"

[0032] The system of the present invention then semi-automatically generates replies to received messages. This generation is based on the learned reply style and labeled message content, and can include not only text but also non-verbal elements such as stamps. For example, in response to the above request for meeting materials, a stamp showing a nod is generated along with the sentence "Got it. I'll send it right away."

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

[0034] "Example 1"

[0035] Step 1: User opens messenger application and sends a message.

[0036] Step 2: The system of the present invention analyzes message traffic in real time using natural language processing (NLP) techniques.

[0037] Step 3: The system will understand the content, tone, and urgency of the message and label it accordingly. For example, a message saying "Please send me the materials for tomorrow's meeting" will be labeled "Request for meeting materials" and the urgency will be set to "High."

[0038] "Example 2"

[0039] Step 1: The system retrieves the user's past sending history.

[0040] Step 2: Based on the acquired sending history, the system uses deep learning technology to learn the user's phrasing and response style.

[0041] Step 3: Based on the learning results, the system will imitate the user's response style. For example, if the user often used the response "Okay, I'll send it right away" in the past, the system will learn a similar response style.

[0042] "Example 3"

[0043] Step 1: The system generates a semi-automatic response to the message received.

[0044] Step 2: Generation is based on the learned response style and labeled message content.

[0045] Step 3: The generated response can include not only text but also non-verbal elements such as stamps. For example, in response to a request for meeting materials, a stamp showing a nod is generated along with the sentence "Understood. I will send them right away."

[0046] Example 1

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

[0048] Conventional messaging applications have difficulty analyzing the content, tone, and urgency of messages sent and received by users in real time. They also lack the ability to learn from a user's sending history and semi-automatically generate text and images using phrases and reply styles that are individually tailored to each individual. As a result, users' communication efficiency declines and important messages may be overlooked.

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

[0050] In this invention, the server includes means for analyzing communications within the messenger, means for labeling received messages, means for learning from the user's sending history and semi-automatically generating text and images using phrases and reply styles suited to the individual user, means for analyzing the content, tone, and urgency of messages using natural language processing technology, and means for displaying the analysis results to the user. This makes it possible to analyze the content, tone, and urgency of messages sent and received by users in real time and semi-automatically generate replies that are suited to each individual user.

[0051] "Means for analyzing communications within a messenger" refers to a function for collecting messages and data sent and received within a messenger application in real time and analyzing their contents.

[0052] "Means for labeling received messages" refers to a function that assigns specific categories or tags to received messages based on their content.

[0053] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a function that analyzes data from messages sent by the user in the past and learns from those patterns to automatically generate replies and stamps that are appropriate for the user.

[0054] "Means for analyzing the content, tone, and urgency of a message using natural language processing technology" refers to a function that uses natural language processing (NLP) technology to analyze the text data of a message and evaluate its content, sentiment, and urgency.

[0055] "Means for displaying analysis results to the user" refers to a function that displays information such as the content and urgency of the analyzed message in a format that is easy for the user to understand.

[0056] The present invention is a system that is incorporated into a messenger application and provides a function for analyzing messages sent and received by a user in real time. A specific embodiment of this system will be described below.

[0057] Hardware and software used

[0058] Hardware:

[0059] Server: Use a server with a powerful processor and sufficient memory. The server analyzes messages and stores data.

[0060] Device: A device used by a user, such as a smartphone or computer, that runs a messenger application.

[0061] software:

[0062] Messenger application: An application that allows users to send and receive messages.

[0063] Natural Language Processing (NLP) libraries, such as Python's NLTK library or spaCy, can be used to analyze message content, tone, and urgency.

[0064] HTTP communication library: Used to send and receive data between the device and the server.

[0065] Program processing

[0066] Message capture:

[0067] The device captures messages sent and received by the user via the messenger application in real time, specifically by using the messenger application's API to retrieve the messages sent.

[0068] Sending a message:

[0069] The device sends the captured message to the server using an HTTP POST request.

[0070] Message parsing:

[0071] The server analyzes the received messages using natural language processing (NLP) techniques, specifically using Python's NLTK library and spaCy to analyze the message's content, tone, and urgency.

[0072] Generate analysis results:

[0073] The server assigns a label to the message as a result of the analysis and sets the urgency level. For example, a message saying "Please send the materials for tomorrow's meeting" is labeled as "Request for meeting materials" and the urgency level is set to "High."

[0074] Returning analysis results:

[0075] The server returns the analysis results to the terminal as an HTTP response.

[0076] Displaying the analysis results:

[0077] The device will update the UI of the messenger application and display the analysis results in a user-friendly format.

[0078] Specific examples

[0079] If a user sends a message in a messenger application saying "Please send me the materials for tomorrow's meeting," the following process will occur:

[0080] 1. The device captures the message using the messenger application's API.

[0081] 2. The device sends the captured message to the server via an HTTP POST request.

[0082] 3. The server parses the message using Python's NLTK library, specifically tokenizing the message, tagging it as a part of speech, and performing sentiment analysis.

[0083] 4. The server labels the analysis result as "Request for meeting materials" and sets the urgency to "High."

[0084] 5. The server returns the analysis results to the terminal as an HTTP response.

[0085] 6. The device updates the UI of the messenger application and displays the analysis result to the user as "Request for meeting materials (urgency: high)."

[0086] Prompt Sentence Examples

[0087] "Create a program that analyzes messages sent by users in a messenger application in real time to understand the content, tone, and urgency of the message. For example, if the message says, 'Please send me the materials for tomorrow's meeting,' label the message as a 'Request for meeting materials' and set the urgency to 'High.'"

[0088] The above is a specific embodiment for carrying out the present invention.

[0089] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0090] Step 1:

[0091] Message capture

[0092] The device captures messages sent and received by users in messenger applications in real time.

[0093] Input: A message sent by a user in a messenger application.

[0094] Specific operation: The device uses the messenger application's API to retrieve the sent message.

[0095] Output: The captured message data.

[0096] Step 2:

[0097] Sending a message

[0098] The terminal sends the captured message to the server.

[0099] Input: The captured message data.

[0100] Specific operation: The device uses an HTTP POST request to send message data to the server.

[0101] Output: The message data sent to the server.

[0102] Step 3:

[0103] Message Parsing

[0104] The server analyzes the received message using natural language processing (NLP) techniques.

[0105] Input: The message data sent to the server.

[0106] What it does: The server uses Python's NLTK library and spaCy to analyze messages for content, tone, and urgency, including tokenizing, part-of-speech tagging, and sentiment analysis.

[0107] Output: Parsed message content, tone, and urgency data.

[0108] Step 4:

[0109] Generate analysis results

[0110] The server generates the analysis results.

[0111] Input: Parsed message content, tone, and urgency data.

[0112] Specific behavior: The server assigns a label and urgency to the message based on its content. For example, a message saying "Please send me the materials for tomorrow's meeting" is labeled as "Request for meeting materials" and the urgency is set to "High."

[0113] Output: Labeled message and urgency data.

[0114] Step 5:

[0115] Returning analysis results

[0116] The server returns the analysis results to the terminal.

[0117] Input: Labeled message and urgency data.

[0118] Specific operation: The server sends the analysis results to the terminal as an HTTP response.

[0119] Output: Analysis results sent back to the device.

[0120] Step 6:

[0121] Displaying analysis results

[0122] The device displays the analysis results to the user.

[0123] Input: Analysis results sent back to the device.

[0124] What it does: The device updates the UI of the messenger application and displays the analysis results in a user-friendly format.

[0125] Output: The analysis results displayed to the user.

[0126] (Application example 1)

[0127] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0128] Conventional messenger applications have difficulty analyzing the content and urgency of messages sent and received by users in real time, making it difficult to detect security risks early. Furthermore, the automatic generation function to support user communication was limited, making it impossible to provide responses optimized for each individual user. This could delay the early detection and rapid response of security risks within a company.

[0129] 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.

[0130] In this invention, the server includes means for analyzing communications within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit the individual user, means for analyzing the content, tone, and urgency of messages in real time, means for determining security risks based on the analysis results, and means for issuing alerts according to the determined risks, thereby enabling early detection of security risks within the messenger application and rapid response.

[0131] "Means for analyzing communications within a messenger" refers to technology for analyzing the content and traffic of messages sent and received within a messenger application.

[0132] "Means for labeling received messages" refers to a technology that assigns appropriate labels to received messages based on their content and urgency.

[0133] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a technology that analyzes a user's past message sending history and automatically generates replies and stamps that are optimized for that user.

[0134] "Means for analyzing message content, tone, and urgency in real time" refers to technology that instantly analyzes the content, emotional tone, and urgency of received messages.

[0135] "Means for determining security risks based on analysis results" refers to technology that evaluates and determines potential security risks based on the analysis results of messages.

[0136] "Means for issuing alerts according to the assessed risk" refers to technology that notifies users and administrators of appropriate alerts according to the assessed level of security risk.

[0137] A system for implementing the present invention is for analyzing messages sent and received within a messenger application in real time and detecting security risks at an early stage. The system includes the following main components:

[0138] 1. Means of analyzing communications in messengers:

[0139] The server analyzes the content and traffic of messages sent and received within the messenger application, using natural language processing (NLP) libraries such as spaCy and Transformers in Python to analyze the content, tone, and urgency of messages.

[0140] 2. How to label received messages:

[0141] The server assigns appropriate labels to received messages based on their content and urgency, such as "Request for meeting materials" or "Urgent response required."

[0142] 3. A method to semi-automatically generate text and images using phrases and response styles that are personalized to each user based on their sending history:

[0143] The server analyzes the user's message history and automatically generates personalized replies and stickers, using a generative AI model to generate prompts tailored to the user's communication style.

[0144] 4. Real-time analysis of message content, tone, and urgency:

[0145] The server instantly analyzes the content, emotional tone, and urgency of the received message, allowing it to determine the importance and urgency of the message and prompt appropriate responses.

[0146] 5. How to determine security risks based on the analysis results:

[0147] The server evaluates and judges the potential security risk based on the analysis of the message. For example, if a message saying "The server is down. Urgent action is required" is received, it will be judged as a high risk.

[0148] 6. Means of issuing alerts according to the determined risk:

[0149] The server will then send appropriate alerts to users and administrators depending on the level of security risk determined, enabling prompt action.

[0150] Specific examples

[0151] Message: "Server is down. Urgent action required."

[0152] Analysis results:

[0153] Entity: [("Server", "PRODUCT")]

[0154] Emotion: [{"label": "NEGATIVE", "score": 0.99}]

[0155] Urgency: "High"

[0156] Prompt Sentence Examples

[0157] python

[0158] message = "The server is down. Urgent action required."

[0159] analysis_result = analyze_message(message)

[0160] print(analysis_result)

[0161] In this way, security risks within messenger applications can be detected early and addressed quickly.

[0162] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0163] Step 1:

[0164] The server receives messages sent and received within the messenger application. The input is the message sent by the user, and the output is the received message data. Specifically, the text content of the message is sent to the server.

[0165] Step 2:

[0166] The server analyzes the received message using a natural language processing (NLP) model. The input is the received message data, and the output is the analysis result. Specifically, it uses spaCy and Transformers to analyze the message content, tone, and urgency.

[0167] Step 3:

[0168] The server assigns a label to the message based on the analysis results. The input is the analysis results, and the output is the labeled message data. Specifically, the server assigns labels such as "Request for meeting materials" or "Urgent response required" to the message.

[0169] Step 4:

[0170] The server learns the user's sending history and semi-automatically generates sentences and images using phrases and response styles that are tailored to the individual user. The input is the user's sending history data, and the output is the generated sentences and images. Specifically, it uses a generative AI model to generate prompt sentences and provide optimized responses to the user.

[0171] Step 5:

[0172] The server analyzes the message content, tone, and urgency in real time. The input is the received message data, and the output is the analysis results. Specifically, it performs sentiment analysis and determines the urgency level to evaluate the importance of the message.

[0173] Step 6:

[0174] The server determines the security risk based on the analysis results. The input is the analysis results, and the output is the risk assessment result. Specifically, it evaluates the potential security risk from the message content and tone, and determines that the risk is high.

[0175] Step 7:

[0176] The server issues an alert based on the risk it determines. The input is the risk assessment result, and the output is an alert notification. Specifically, alerts are sent to users and administrators based on the risk level, encouraging them to take prompt action.

[0177] Example 2

[0178] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0179] Conventional messaging applications have the problem of inefficiency, requiring users to manually input messages each time. It is also difficult to automatically respond to messages that reflect the user's individual phrasing and response style, which can lead to a decline in the quality of communication. Furthermore, it is difficult to properly determine what the other party is requesting and the urgency of that request.

[0180] 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.

[0181] In this invention, the server includes a means for analyzing data in a message application, a means for classifying received messages, a means for learning from a user's sending history and semi-automatically generating text and images using phrases and response styles suited to the individual user, a means for collecting the user's past sending history and storing it in a database, a means for training a model based on the sending history data using deep learning technology, a means for receiving requests from the user and generating appropriate responses, and a means for sending the generated responses to the user. This enables automatic responses that reflect the user's individual phrases and response styles, improving the quality of communication. It also enables efficient message exchange by appropriately determining the matters the other party is requesting and their urgency.

[0182] A "messaging application" is software that allows users to send and receive text and images.

[0183] The "means for analyzing data" refers to a technique for collecting data in a message application and analyzing the content.

[0184] A "message categorization means" is a technique for categorizing received messages based on specific criteria.

[0185] "Means for learning from sending history" refers to technology that learns a user's phrasing and response style based on messages the user has sent in the past.

[0186] "Means for semi-automatic text and image generation" refers to technology in which a system automatically generates text and images based on user input and past history.

[0187] The "means for collecting sending history" is a technique for collecting messages that a user has sent in the past.

[0188] "Means of storing data in a database" refers to the technology used to organize and store collected data.

[0189] "Deep learning technology" is a technique for training models based on large amounts of data and learning patterns and features.

[0190] The "means for training a model" is a technique for training a model based on transmission history data using deep learning technology.

[0191] "Means for receiving requests" refers to the technology for receiving requests or inquiries from users.

[0192] The "means for generating an appropriate response" is a technology that automatically generates a response that matches the user's style in response to a received request.

[0193] The "means for sending a response" is a technique for sending the generated response to the user.

[0194] This invention is a system that learns a user's phrasing and reply style based on the user's past sending history, and automatically generates appropriate replies. A specific embodiment of this system will be described below.

[0195] First, the server collects text data from messages and emails sent by the user in the past. This data serves as the basis for learning the user's phrasing and response style. Common APIs such as the Gmail API and Slack API can be used to collect the data.

[0196] Next, the server saves the collected sending history in a database. This database organizes and stores the sending history for each user. A common database management system such as MySQL (registered trademark) or PostgreSQL can be used as the database.

[0197] The server then uses a deep learning framework, such as TENSORFLOW® or PyTorch, to train a deep learning model based on the historical submission data. This model is used to learn the user's specific phrasing and response style. For example, a long short-term memory (LSTM) network can be used to learn the user's phrasing and response style.

[0198] When a user sends a request, the device receives the request. For example, a user might type "Please send me the materials for the meeting" through a smartphone app. This request is then sent to the server.

[0199] The server uses a deep learning model to generate an appropriate response based on the received request. The model generates a response that matches the user's style based on information learned from past sending history. For example, it can generate a response such as "Got it. I'll send it right away."

[0200] Finally, the device displays the response generated by the server to the user. For example, a message such as "Got it. I'll send it right away" appears on the smartphone app.

[0201] As a concrete example, the following prompt sentences can be input to a generative AI model:

[0202] Prompt: In response to the request "Please send me the materials for the meeting," generate an appropriate response based on past sending history.

[0203] When this prompt sentence is input into the generative AI model, the model generates a response based on its past learning results: "Got it. I'll send it right away."

[0204] In this way, the present invention realizes automatic responses that reflect the user's individual phrasing and response style, improving the quality of communication. It also makes it possible to properly determine the matters the other party is requesting and their urgency, enabling efficient message exchange.

[0205] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0206] Step 1:

[0207] The server collects the user's past sending history. Specifically, it uses common APIs such as the Gmail API and Slack API to obtain text data of messages and emails that the user has sent in the past. The input is the raw message data obtained through the API, and the output is a list of the collected message data.

[0208] Step 2:

[0209] The server saves the collected sending history in a database. Specifically, it uses a database management system such as MySQL or PostgreSQL to organize and store the sending history for each user. The input is a list of collected message data, and the output is the organized sending history data stored in the database.

[0210] Step 3:

[0211] The server uses deep learning techniques to train a model based on the transmission history data. Specifically, it uses deep learning frameworks such as TensorFlow and PyTorch to train an LSTM (long short-term memory) network. The input is the transmission history data stored in a database, and the output is a deep learning model that has learned the user's phrasing and response style.

[0212] Step 4:

[0213] The terminal receives a request from a user. Specifically, the user inputs "Please send me the materials for the meeting" through a smartphone app. The input is the request text entered by the user, and the output is the request data to be sent to the server.

[0214] Step 5:

[0215] The server uses a deep learning model to generate an appropriate response based on the received request. Specifically, the request text is input into the model, and a response is generated based on information learned from past transmission history. The input is the user's request text and the trained deep learning model, and the output is the generated response text.

[0216] Step 6:

[0217] The device displays the response generated by the server to the user. Specifically, the smartphone app displays the message "Got it. I'll send it right away." The input is the response text sent from the server, and the output is the message displayed to the user.

[0218] (Application example 2)

[0219] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0220] Conventional messenger systems required users to create individual messages, making it difficult to respond quickly, especially in busy situations or emergencies. Furthermore, they lacked the functionality to automatically generate optimal replies for each user by utilizing the user's past comment history, which reduced the efficiency of user communication.

[0221] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from the user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the user individually, and means for learning from the user's past comment history and automatically generating a comment style that the user prefers. This enables the user to automatically generate quick and appropriate replies, improving communication efficiency.

[0222] "Means for analyzing traffic within a messenger" refers to a function for analyzing the flow of messages and data sent and received within a messenger application.

[0223] "Means for labeling received messages" refers to a function that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags or labels.

[0224] "Means for learning from the user's sending history" is a function that collects the history of messages sent by the user in the past and uses this to learn the user's wording and response style.

[0225] "Means for semi-automatically generating text and stamps" refers to a function that automatically generates appropriate text and stamps, reducing the amount of input required by the user, based on the user's past sending history and learning results.

[0226] "Means for learning a user's past comment history" is a function that collects the history of comments left by a user in the past and learns the user's commenting style based on this.

[0227] "Means for automatically generating comment styles preferred by users" is a function that automatically generates comments in a format preferred by users based on the learned comment styles of users.

[0228] The system for implementing this invention has the functions of analyzing traffic within the messenger, labeling received messages, learning from the user's sending history, and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user.It also includes a function that learns the user's past comment history and automatically generates a comment style that the user prefers.

[0229] The server first analyzes the traffic within the messenger, using software designed to analyze the flow of messages and data sent and received within the messenger application, specifically deep learning frameworks such as TensorFlow and Keras.

[0230] Next, the received messages are labeled. This labeling process identifies the content, importance, urgency, etc. of the received messages and assigns appropriate tags and labels. This makes it easier to determine the priority of the messages.

[0231] To learn from a user's sending history, the system collects messages sent by the user in the past and learns the user's phrasing and reply style based on this. Again, a deep learning model is built using TensorFlow and Keras to learn the user's characteristics.

[0232] To semi-automatically generate sentences and stamps, the trained model is used to reduce the user's input and automatically generate appropriate sentences and stamps. This process can reflect the user's past usage and style.

[0233] Furthermore, it has the ability to learn from a user's past comment history and automatically generate a comment style that the user prefers. This allows it to automatically generate comments in the format that the user prefers.

[0234] As a specific example, if a user frequently uses the comment "I understand," the system can automatically generate comments such as "I understand. I'll send it right away" or "I understand. Thank you."

[0235] An example of a prompt to input to a generative AI model is as follows:

[0236] Develop an application that learns from a user's past comment history and automatically generates a comment style that the user prefers. A user's comment history includes comments such as "Got it, I'll send it right away" and "Thank you." Use this history to build a model that generates comments that the user prefers.

[0237] In this way, users can automatically generate quick and appropriate responses, improving communication efficiency.

[0238] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0239] Step 1:

[0240] The server analyzes the traffic within the messenger. Specifically, it uses deep learning frameworks such as TensorFlow and Keras to analyze the flow of messages and data sent and received within the messenger application. The input includes all message data within the messenger, and the output extracts features from each message.

[0241] Step 2:

[0242] The server labels the received messages. It identifies the content, importance, urgency, etc. of the received messages and assigns appropriate tags and labels. The input includes the message features extracted in step 1, and the output is labeled message data.

[0243] Step 3:

[0244] The server learns from the user's sending history. It collects the user's message history and uses TensorFlow and Keras again to learn the user's phrasing and reply style based on this. The input includes the user's past message history, and the output is a model that learns the user's characteristics.

[0245] Step 4:

[0246] The server semi-automatically generates sentences and stamps. It uses the trained model to automatically generate appropriate sentences and stamps, reducing the user's input effort. The input includes the model generated in step 3 and the current message context, and the output provides the generated sentences and stamps.

[0247] Step 5:

[0248] The server learns from the user's past comment history. It collects the user's past comment history and uses TensorFlow and Keras again to learn the user's commenting style based on this. The input includes the user's past comment history, and the output is a model that learns the user's commenting style.

[0249] Step 6:

[0250] The server automatically generates a comment style that the user prefers based on the trained model. The input includes the model generated in step 5 and the current comment context, and the generated comment is provided as the output.

[0251] In this way, users can automatically generate quick and appropriate responses, improving communication efficiency.

[0252] Example 3

[0253] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0254] In modern communication, users receive many messages and need to respond to them quickly and appropriately. However, responding manually takes time and effort, and efficiency is especially important in business environments. Furthermore, if the content and style of responses are inconsistent, the quality of communication may decline. To solve these issues, a system is needed that can semi-automatically generate appropriate responses to received messages.

[0255] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0256] In this invention, the server includes a means for receiving a message, a means for analyzing the received message, a means for selecting a reply style based on the analysis result, a means for generating sentences and non-verbal elements based on the selected reply style, and a means for transmitting the generated reply, thereby enabling a user to semi-automatically generate a quick and consistent reply to a received message.

[0257] The "means for receiving a message" is a function that allows the server to receive a message sent by a user.

[0258] "Means for analyzing received messages" refers to a function for analyzing the contents of received messages and extracting their intent and important information.

[0259] The "means for selecting a reply style based on the analysis results" is a function for selecting an appropriate reply style based on the contents of the analyzed message.

[0260] The "means for generating sentences and non-verbal elements based on the selected reply style" is a function for generating sentences, stamps, and other non-verbal elements according to the selected reply style.

[0261] The "means for sending the generated response" is a function for sending the generated response to the user.

[0262] "Non-verbal elements" are means of communication other than text, including stamps and emojis.

[0263] A "reply style" is a specific phrase or expression that is selected based on a user's past sending history and learning data.

[0264] The present invention relates to a system for semi-automatically generating a reply to a received message. A specific embodiment of this system will be described below.

[0265] System configuration

[0266] This system consists of three main components: a server, a terminal, and a user. The server receives messages, analyzes them, selects a reply style, generates a reply, and sends the reply. The terminal is a device that users use to send and receive messages. The user is the entity that communicates using the system.

[0267] Hardware and software used

[0268] The server is a computer system equipped with a high-performance processor and large memory capacity. Specifically, a server equipped with an Intel Xeon processor or an AMD EPYC processor is suitable. For software, generative AI models such as Google's BERT and OpenAI's GPT-3 are used to implement natural language processing (NLP) technology.

[0269] Data processing and calculation

[0270] The server processes and calculates data in the following procedure.

[0271] 1. Receiving a message:

[0272] The server receives the message sent by the user from the terminal.

[0273] 2. Message analysis:

[0274] The server analyzes the received message using generative AI models such as BERT and GPT-3, and extracts the message's intent and key information.

[0275] 3. Choose your response style:

[0276] The server then selects an appropriate response style based on the analysis results, which is extracted from a dataset of learned messages and their responses.

[0277] 4. Generate a response:

[0278] The server generates sentences and non-verbal elements (such as stamps) based on the selected response style, using a generative AI model.

[0279] 5. Sending a response:

[0280] The server generates a response and sends it to the user.

[0281] Specific examples

[0282] When a user sends a message saying "Please send me the meeting materials," the server processes the message as follows.

[0283] 1. Receiving a message:

[0284] The user sends a message from the terminal saying, "Please send me the meeting materials."

[0285] The server receives this message.

[0286] 2. Message analysis:

[0287] The server analyzes the received message.

[0288] The server extracts the keyword "conference materials" and recognizes that the user is requesting materials.

[0289] 3. Choose your response style:

[0290] The server selects the response style "Understood. I'll send it right away" based on past data.

[0291] 4. Generate a response:

[0292] The server generates a stamp showing a nod along with the text "Got it. I'll send it right away."

[0293] 5. Sending a response:

[0294] The server sends the generated text "Got it. I'll send it right away" and a stamp showing a nod to the user.

[0295] The user receives this response at the terminal.

[0296] Prompt Sentence Examples

[0297] The following is an example of a prompt sentence that generates an appropriate response to the message "Please send me the meeting materials."

[0298] Message from user: "Please send me the meeting materials"

[0299] Response style: "Okay, I'll send it right away."

[0300] Non-verbal: "Nodding Stamp"

[0301] By inputting this prompt sentence into the generative AI model, an appropriate response is generated. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0302] Step 1:

[0303] Receiving messages

[0304] The server receives messages sent by users. The input is a text message sent by the user from the terminal, and the output is the message data received by the server. In concrete terms, when a user sends a message from the terminal saying "Please send me the meeting materials," the server receives this message.

[0305] Step 2:

[0306] Message Parsing

[0307] The server analyzes the received message. The input is the received message data, and the output is the intent and important information of the analyzed message. Specifically, the server analyzes the message "Please send me the meeting materials," extracts the keyword "meeting materials," and recognizes that the user is requesting materials. This analysis uses generative AI models such as BERT and GPT-3.

[0308] Step 3:

[0309] Choose a response style

[0310] The server selects an appropriate response style based on the analysis results. The input is the intent and important information of the analyzed message, and the output is the selected response style. In concrete terms, the server selects a response style such as "Got it. I'll send it right away" from past data. This selection is made to provide an appropriate response to the user's request.

[0311] Step 4:

[0312] Generate a response

[0313] The server generates sentences and non-verbal elements based on the selected response style. The input is the selected response style, and the output is the generated sentences and non-verbal elements (such as stamps). Specifically, the server generates a stamp indicating a nod along with the sentence "Got it. I'll send it right away." A generative AI model is used to generate this.

[0314] Step 5:

[0315] Sending a response

[0316] The server sends the generated response to the user. The input is the generated text and non-verbal elements, and the output is the response message that the user receives on their device. Specifically, the server sends the generated text "Got it. I'll send it right away" and a stamp indicating a nod to the user. The user receives this response on their device.

[0317] (Application example 3)

[0318] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0319] Conventional messaging systems have the problem of making it difficult to respond quickly and appropriately to messages received by users. In particular, food delivery services are required to respond quickly to inquiries from users, but there are limitations to manual responses. In addition, because the content of responses is uniform, it is difficult to provide users with a sense of familiarity and individual attention.

[0320] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0321] In this invention, the server includes means for analyzing data in message communications, means for classifying received messages, means for learning from a user's sending history and semi-automatically generating sentences and non-verbal elements using phrases and reply styles suited to the individual user, means for adding appropriate non-verbal elements to the generated reply, and means for sending the generated reply to the user. This makes it possible to provide a quick and appropriate reply to a user's inquiry, and to create a sense of friendliness and individual attention.

[0322] "Messaging" is a means of exchanging information between users, including text and non-verbal elements.

[0323] "Means for analyzing data" refers to technology that analyzes the content and patterns of received messages and extracts information to generate an appropriate response.

[0324] "Means for classifying messages" refers to a technique for categorizing received messages based on their content and urgency.

[0325] "Means for learning from a user's sending history" refers to technology that learns a user's response style and phrasing based on their past message sending history.

[0326] "Means for semi-automatically generating sentences and non-verbal elements" refers to technology in which the system automatically generates appropriate sentences and non-verbal elements such as stamps based on the user's sending history and message content.

[0327] "Non-verbal elements" are visual expressions such as stickers and emojis that accompany text messages.

[0328] "Means for adding appropriate non-verbal elements to responses" refers to technology that automatically adds appropriate stamps and emoticons to generated text.

[0329] The "means for sending a response to a user" refers to a technique for sending the generated response to a user.

[0330] The system for implementing this invention analyzes data in message communications, classifies received messages, and learns from the user's sending history to semi-automatically generate sentences and non-verbal elements. A specific embodiment of this system is described below.

[0331] System Configuration

[0332] The system consists of the following major components:

[0333] 1. Data Analysis Module: Analyzes data in message traffic. This module analyzes the content and patterns of received messages and extracts information to generate appropriate responses.

[0334] 2. Message Classification Module: Classifies received messages based on their content and urgency. This module determines what the recipient is requesting and how urgent it is.

[0335] 3. Learning module: This module learns from the user's sending history and semi-automatically generates sentences and non-verbal elements using phrases and response styles that are appropriate for the individual user. This module learns the user's response style and phrases based on their past message sending history.

[0336] 4. Non-verbal element addition module: Adds appropriate non-verbal elements (stamps and emojis) to the generated responses.

[0337] 5. Reply sending module: Sends the generated reply to the user.

[0338] Hardware and software used

[0339] Hardware: Smartphone

[0340] Software: OpenAI GPT-3 API, Python

[0341] Data processing and calculation

[0342] The server first analyzes messages received from users using the data analysis module. Next, the message classification module classifies messages based on content and urgency. After that, the learning module learns appropriate reply styles based on the user's sending history and semi-automatically generates sentences and non-verbal elements. The non-verbal element addition module adds appropriate stamps and emojis to the generated replies. Finally, the reply sending module sends the generated replies to the user.

[0343] Specific examples

[0344] For example, if a user asks, "What is the status of my order?" the system will act as follows:

[0345] 1. Data analysis: Analyze the content of the message and extract the keyword "order status."

[0346] 2. Message Classification: Classify the message content as being about "order status."

[0347] 3. Response generation: Based on past sending history, the learning module generates a response such as, "A delivery person is currently picking up your order. Please wait a little longer."

[0348] 4. Adding non-verbal elements: Add a stamp of the delivery person holding food to the generated response.

[0349] 5. Send Response: Send the final response to the user.

[0350] Prompt Sentence Examples

[0351] Message from user: What is the status of my order?

[0352] Response style: Polite and friendly

[0353] Response:

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

[0355] Step 1:

[0356] The server inputs messages received from users into a data analysis module. The data analysis module analyzes the content and patterns of the messages and extracts information to generate an appropriate response. For example, analyzing the message "Please tell me the status of my order," the module extracts the keyword "order status."

[0357] Input: Message from user

[0358] Output: Analyzed keywords and content

[0359] Step 2:

[0360] The server inputs the analyzed keywords and content into the message classification module, which then classifies the received messages based on their content and urgency. For example, based on the keyword "order status," the message may be classified into the "order status check" category.

[0361] Input: Analyzed keywords and content

[0362] Output: Classified message categories

[0363] Step 3:

[0364] The server inputs the classified message categories into a learning module. Based on the user's sending history, the learning module semi-automatically generates sentences using phrases and response styles that are appropriate for the individual user. For example, based on past sending history, it generates a response such as, "A delivery person is currently picking up your order. Please wait a little longer."

[0365] Input: Message categories, user sending history

[0366] Output: Generated response text

[0367] Step 4:

[0368] The server inputs the generated response sentence into the non-verbal element addition module, which adds appropriate non-verbal elements (stamps and emojis) to the generated response. For example, in response to the response "A delivery person is currently picking up your order. Please wait a little longer," a stamp of a delivery person holding food is added.

[0369] Input: Generated response text

[0370] Output: Response with added non-verbal elements

[0371] Step 5:

[0372] The server inputs the response with the added non-verbal elements into the response sending module. The response sending module sends the generated response to the user. For example, it sends the final response to the user, such as "The delivery person is currently picking up your order. Please wait a little longer. (Stamp)."

[0373] Input: Response with added non-verbal elements

[0374] Output: Reply sent to the user

[0375] 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.

[0376] "Example 1"

[0377] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system extracts emotions from a user's message and tailors a response based on the emotions. Specifically, if a user sends a message such as "I'm really tired today," the emotion engine extracts emotions such as "fatigue" and "stress" from the message. Based on this emotion, the system then responds with a response such as "That was tough, please take a good rest."

[0378] In this way, a system that combines an emotion engine can provide more appropriate responses that correspond to the user's emotions.

[0379] "Example 2"

[0380] The emotion engine can also generate appropriate stamps based on the user's emotions. For example, if a user sends a message such as "Today was fun!", the emotion engine can extract emotions such as "joy" or "satisfaction" from the message. Based on this emotion, the system can generate a stamp of a smiling face, for example. In this way, a system that incorporates an emotion engine can provide more appropriate stamps that correspond to the user's emotions.

[0381] "Example 3"

[0382] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system extracts emotions from a user's message and tailors a response based on the emotions. Specifically, if a user sends a message such as "I'm really tired today," the emotion engine extracts emotions such as "fatigue" and "stress" from the message. Based on this emotion, the system then responds with a response such as "That was tough, please take a good rest."

[0383] In this way, a system that combines an emotion engine can provide more appropriate responses that correspond to the user's emotions.

[0384] The processing flow of each embodiment will be described below.

[0385] "Example 1"

[0386] Step 1: A user opens a messenger application and sends a message, for example, "I'm really tired today."

[0387] Step 2: The system analyzes the sent message in real time using natural language processing (NLP) techniques.

[0388] Step 3: The emotion engine extracts emotions such as "fatigue" or "stress" from the analyzed messages.

[0389] Step 4: The system generates a response based on the extracted sentiment, for example, "That must have been tough, take a rest."

[0390] "Example 2"

[0391] Step 1: A user opens a messenger application and sends a message, for example, "I had a great time today!"

[0392] Step 2: The system analyzes the sent message in real time using natural language processing (NLP) techniques.

[0393] Step 3: The emotion engine extracts emotions such as "joy" or "satisfaction" from the analyzed messages.

[0394] Step 4: The system generates a stamp based on the extracted emotion, such as a smiley face stamp.

[0395] Example 1

[0396] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0397] With conventional messaging applications, it was difficult to understand the content, tone, and urgency of messages sent and received by users in real time, making it impossible to generate appropriate replies. It was also difficult to provide replies based on the user's emotions. This limited the user's communication experience and hindered efficient information transmission.

[0398] 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.

[0399] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles suited to the individual user, means for analyzing the content of messages using natural language processing technology, means for extracting emotions from messages using an emotion engine, means for generating appropriate replies based on the emotions extracted using a generative AI model, and means for sending the generated replies to the user, thereby making it possible to analyze user messages in real time and provide appropriate replies.

[0400] "Means for analyzing traffic within messengers" refers to technology for monitoring and analyzing message data sent and received through messenger applications in real time.

[0401] "Means for labeling received messages" refers to a technology that analyzes the content of a received message and assigns an appropriate label (tag) based on that content.

[0402] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a technology that analyzes a user's past message sending history, learns the phrases and reply styles that are best suited to that user, and semi-automatically generates text and images.

[0403] "Means for analyzing the content of a message using natural language processing technology" refers to a technology that uses natural language processing (NLP) technology to tokenize the content of a message and extract meaning.

[0404] The "means for extracting emotions from a message using an emotion engine" is a technology that uses an emotion engine to identify and extract a user's emotions from the tone and content of a message.

[0405] "Means for generating an appropriate response based on emotions extracted using a generative AI model" is a technology that uses a generative AI model to generate an appropriate response based on emotions extracted.

[0406] The "means for sending the generated reply to the user" is a technique for sending the generated reply to the user through a messenger application.

[0407] The present invention is a system that is incorporated into a messenger application and analyzes messages sent and received by users in real time and generates appropriate responses. The system includes elements of a server, a terminal, and a user.

[0408] Hardware and software used

[0409] The server uses the following hardware and software:

[0410] Hardware: A server machine with high-performance processors, memory, and storage

[0411] Software: Natural language processing libraries (e.g., spaCy, NLTK), emotion engines (e.g., IBM Watson® Tone Analyzer), generative AI models (e.g., OpenAI GPT-3)

[0412] Data processing and calculation

[0413] The server receives the message sent by the user and performs the following actions:

[0414] 1. Message content analysis:

[0415] The server uses natural language processing libraries to tokenize the message and extract meaning.

[0416] For example, if a user sends a message saying, "Please send me the materials for tomorrow's meeting," the server will label this message as a "request for meeting materials" and set the urgency to "high."

[0417] 2. Emotion extraction:

[0418] The server uses an emotion engine to extract emotions from the messages.

[0419] For example, if a user sends a message such as "I'm really tired today," the server will extract emotions such as "fatigue" and "stress" from the message.

[0420] 3. Generate a response:

[0421] The server uses a generative AI model to generate appropriate responses based on the extracted emotions.

[0422] For example, in response to a message such as "I'm really tired today," the server generates a response such as "That must have been tough, please get some rest."

[0423] 4. Sending a response:

[0424] The server then sends the generated reply to the user, who then displays the reply to the user through the messenger application.

[0425] Specific examples

[0426] As a concrete example, consider the following scenario:

[0427] scenario:

[0428] A user sends a message through a messenger application saying, "Please send me the materials for tomorrow's meeting."

[0429] Process flow:

[0430] 1. The server receives the message and uses a natural language processing library to label it as a "request for meeting materials" and set the urgency to "high."

[0431] 2. The server uses the emotion engine to extract emotions, but since this message does not contain any specific emotion, emotion extraction is skipped.

[0432] 3. The server uses a generative AI model to generate an appropriate response, in which case it determines that no response is necessary.

[0433] 4. The server generates a response and sends it to the user, who then displays it on their device.

[0434] Prompt Sentence Examples

[0435] Below are some example prompts to input to a generative AI model:

[0436] A user sends a message saying, "I'm really tired today." Extract the sentiment from this message and generate an appropriate response.

[0437] In this way, the system can analyze the user's message in real time and provide an appropriate response.

[0438] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0439] Step 1:

[0440] Receiving messages

[0441] The server receives messages sent by users through the messenger application.

[0442] Input: The message sent by the user (e.g., "Please send me the materials for tomorrow's meeting")

[0443] Specific operation: The server receives messages through the API endpoint of the messenger application and stores them in a database.

[0444] Output: Received message data

[0445] Step 2:

[0446] Message content analysis

[0447] The server analyzes the received message using a natural language processing (NLP) library.

[0448] Input: Received message data

[0449] What it does: The server uses an NLP library (e.g., spaCy) to tokenize the message and extract meaning. For example, it parses the message "Please send me the materials for tomorrow's meeting" and labels it as "Request for meeting materials."

[0450] Output: Labeled message data (e.g., "Request for meeting materials")

[0451] Step 3:

[0452] Emotion extraction

[0453] The server uses an emotion engine to extract emotions from the messages.

[0454] Input: Labeled message data

[0455] Specific operation: The server analyzes the message using an emotion engine (e.g., IBM Watson Tone Analyzer) and extracts emotions. For example, it extracts emotions such as "fatigue" and "stress" from the message "I'm really tired today."

[0456] Output: Extracted emotion data (e.g., "fatigue," "stress")

[0457] Step 4:

[0458] Generate a response

[0459] The server uses a generative AI model to generate appropriate responses based on the extracted emotions.

[0460] Input: Extracted emotion data

[0461] Specific operation: The server inputs a prompt sentence into the generative AI model (e.g., OpenAI GPT-3). For example, in response to the message "I'm really tired today," the server inputs the prompt sentence "The user feels tired. Please generate an appropriate response." The generative AI model generates the response "That must have been hard. Please take a good rest."

[0462] Output: Generated response data (e.g. "That was tough, please take a rest").

[0463] Step 5:

[0464] Sending a response

[0465] The server generates a response and sends it to the user.

[0466] Input: Generated response data

[0467] Specific operation: The server sends the generated reply to the user through the messenger application's API.

[0468] Output: A response message that is displayed on the user's device (e.g., "That was tough, please get some rest.")

[0469] (Application example 1)

[0470] Next, a description will be given of Application Example 1 of Embodiment Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0471] Conventional messaging applications have difficulty analyzing the content, urgency, and emotions of messages sent and received by users in real time, and responding appropriately. Furthermore, they lack the functionality to instantly send appropriate alerts in emergencies, and a rapid response is required to ensure user safety.

[0472] 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.

[0473] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user, means for determining the urgency of a message, means for extracting emotions from the message using an emotion engine, and means for sending an automatic alert when the urgency is determined to be high. This makes it possible to analyze user messages in real time and take appropriate action based on the urgency and emotions.

[0474] "Means for analyzing traffic within messengers" refers to technology that analyzes the content, frequency, timing, etc. of messages sent and received by users in real time.

[0475] "Means for labeling received messages" refers to techniques for attaching appropriate tags or labels to messages based on their content.

[0476] "A means of learning from a user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit each individual user" is a technology that analyzes past message history and automatically generates appropriate replies based on the user's individual communication style.

[0477] "Means for determining the urgency of a message" refers to a technique for evaluating how urgent a message is based on its content and context.

[0478] The "means for extracting emotions from a message using an emotion engine" is a technology that uses natural language processing technology to analyze and identify emotions contained in a message.

[0479] "Means for sending automatic alerts when a message is deemed to be of high urgency" refers to technology that automatically sends a warning to pre-defined contacts (e.g., police or security companies) when a message is deemed to be of high urgency.

[0480] A system for implementing this invention includes means for analyzing traffic within a messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user, means for determining the urgency of a message, means for extracting emotions from the message using an emotion engine, and means for sending an automatic alert if the urgency is determined to be high.

[0481] Hardware and software used

[0482] Hardware: Smartphones, servers

[0483] Software: Python, NLTK library, smtplib, natural language processing (NLP) technology, emotion engine

[0484] Data processing and calculation

[0485] 1. How to analyze traffic in Messenger:

[0486] The server analyzes the content, frequency, and timing of messages sent and received by users in real time, using natural language processing (NLP) technology.

[0487] 2. How to label received messages:

[0488] The server assigns appropriate tags or labels to messages based on their content, such as "Request for meeting materials" or "High urgency."

[0489] 3. A method to learn from a user's sending history and semi-automatically generate sentences and stamps using phrases and reply styles that are personalized to that user:

[0490] The server analyzes past message history and automatically generates appropriate replies based on the user's individual communication style, using a generative AI model.

[0491] 4. How to determine the urgency of a message:

[0492] The server evaluates the urgency of a message based on its content and context, including whether it contains certain keywords (e.g., "help" or "intrusion").

[0493] 5. How to extract emotions from messages using an emotion engine:

[0494] The server uses natural language processing technology to analyze and identify emotions contained in the message, such as "fear" or "urgency."

[0495] 6. Means of sending automated alerts in case of high severity:

[0496] If the server determines that the incident is urgent, it will automatically send an alert to pre-defined contacts (e.g., police or security companies) by using smtplib to send emails.

[0497] Specific examples

[0498] If a user sends a message like "Help, someone's breaking into my house!", the server will analyze the message in real time, label it as a "break-in report," and assign a "high urgency" to it. The emotion engine will detect "fear" and "urgency" and immediately send an automatic alert to the police or security company.

[0499] Prompt Sentence Examples

[0500] "Design a system that analyzes messages sent by users in real time, determines their urgency and sentiment, and sends automatic alerts as needed. For example, if the message says, 'Help, someone's broken into my house!', the system should label this message as a 'Reporting a Break-in,' set the urgency to 'Very High,' and immediately send an automatic alert to the police or security company."

[0501] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0502] Step 1:

[0503] The server receives messages sent and received by users in real time.

[0504] Input: The message sent by the user

[0505] Output: Received message data

[0506] Specific operation: The server captures messages sent from the messenger application in real time and stores them in a database for analysis.

[0507] Step 2:

[0508] The server analyzes the received message using natural language processing (NLP) techniques.

[0509] Input: Received message data

[0510] Output: Parsed message content, tone, and urgency

[0511] What it does: The server uses the NLTK library to analyze the content of the message and assess its tone and urgency. For example, it analyzes keywords and context within the message to determine its urgency.

[0512] Step 3:

[0513] The server performs labeling based on the parsed message.

[0514] Input: Parsed message content, tone, and urgency

[0515] Output: Labeled message data

[0516] Specific operation: The server assigns an appropriate label (e.g., "Request for meeting materials" or "High urgency") to each message depending on its content, making it easier to categorize messages.

[0517] Step 4:

[0518] The server learns from the user's sending history and semi-automatically generates replies using individual wording and response styles.

[0519] Input: User sending history, labeled message data

[0520] Output: The generated reply message

[0521] How it works: The server uses a generative AI model to analyze a user's past messaging history and generate appropriate replies based on their individual communication style, such as automatically selecting frequently used phrases and stickers.

[0522] Step 5:

[0523] The server reassess the urgency of the message and extracts emotions using an emotion engine if necessary.

[0524] Input: Labeled message data, generated reply message

[0525] Output: Message data with urgency and sentiment ratings

[0526] Specific operation: The server uses an emotion engine to extract emotions from messages and reassess their urgency. For example, it can detect emotions such as "fear" or "urgency."

[0527] Step 6:

[0528] The server will send an automatic alert if the emergency is deemed high.

[0529] Input: Message data with urgency and sentiment ratings

[0530] Output: Alert sent

[0531] Specific operation: If the server determines that the urgency is "high," it will automatically send a warning to pre-defined contacts (e.g., police or security companies) using smtplib, for example, by sending an email or SMS to prompt a prompt response.

[0532] Example 2

[0533] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0534] Conventional messaging systems provide uniform replies and stamps without considering the user's past sending history or emotions, making personalized communication difficult. Furthermore, they lack the functionality to automatically generate appropriate stamps based on the user's emotions, making it difficult to accurately convey the user's emotions.

[0535] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for analyzing data in message communication, means for classifying received messages, means for learning from the user's sending history and semi-automatically generating text and images using phrases and reply styles that suit the individual user, means for analyzing the user's emotions and generating appropriate images based on the emotions, and means for transmitting the generated images to the user's terminal. This enables personalized communication that takes into account the user's past sending history and emotions.

[0536] "Message communication" is a means of communication for sending and receiving information such as text and images between users.

[0537] "Means for analyzing data" refers to technology or devices for analyzing data in message communications and extracting necessary information.

[0538] "Means for classifying messages" refers to techniques or devices for classifying received messages based on their content and urgency.

[0539] "Sending history" is a record of messages a user has sent in the past.

[0540] A "means for learning" is a technique or device for learning a user's specific patterns or style based on the user's sending history.

[0541] "Means for semi-automatically generating text and images" refers to technology or devices that generate text and images in a partially automated manner based on a user's sending history and learning results.

[0542] "Means for analyzing emotions" refers to technologies and devices for extracting and analyzing emotions from users' messages.

[0543] The "means for generating an appropriate image" refers to a technology or device that generates an image to express the user's emotions based on the analyzed emotions.

[0544] "Terminal" means a device (e.g., PC, smartphone) used by a user to communicate messages.

[0545] This system learns from a user's past sending history and understands their phrasing and response style. It also uses an emotion engine to generate stamps that reflect the user's emotions. This system is implemented using a server, a terminal (PC, smartphone, etc.), and various software.

[0546] Hardware and software used

[0547] Hardware: Servers, devices (PCs, smartphones, etc.)

[0548] Software: Deep learning frameworks (TensorFlow, PyTorch, etc.), sentiment analysis engine, stamp generation engine

[0549] System Overview

[0550] 1. Collecting user sending history

[0551] The server stores messages that users have previously sent in a database.

[0552] For example, the message "Got it, I'll send it right away" is saved.

[0553] 2. Deep Learning

[0554] The server uses the stored transmission history to train a deep learning model.

[0555] The model learns your phrasing and response style.

[0556] Example: Learn that users often reply with "Okay, I'll send it right away."

[0557] 3. Sentiment analysis

[0558] When a user sends a new message, the server analyzes the sentiment of the message using a sentiment analysis engine.

[0559] Example: Extracting emotions such as "joy" and "satisfaction" from the message "Today was fun!"

[0560] 4. Stamp generation

[0561] Based on the result of the sentiment analysis, the server generates an appropriate stamp using a stamp generation engine.

[0562] Example: Generating smiley face stickers based on emotions like "happiness" or "contentment."

[0563] 5. Send stamps to your device

[0564] The server sends the generated stamp to the user's terminal.

[0565] Example: The server sends the generated smiley face sticker to the user's chat app, where it is displayed on the user's screen.

[0566] Specific examples

[0567] User sending history: "Okay, I'll send it right away."

[0568] New message: "Today was fun!"

[0569] Sentiment analysis results: "joy" and "satisfaction"

[0570] Generated stamp: Smiling stamp

[0571] Prompt Sentence Examples

[0572] "Describe a system that learns from a user's past messaging history and generates emotionally relevant stickers."

[0573] "If a user sends, 'Today was fun!' what kind of stamp would be generated?"

[0574] In this way, the system of the present invention can provide more personalized communication by learning the user's past sending history and generating stamps according to emotions.

[0575] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0576] Step 1:

[0577] A user uses a device to send a message. The input is a text message typed by the user, and the output is the message data sent to the server. Specifically, the user opens a chat app, types "Got it. I'll send it right away," and presses the send button.

[0578] Step 2:

[0579] The server stores messages sent by users in a database. The input is the message data received by the server, and the output is the message record stored in the database. Specifically, the server analyzes the received message and stores the text "Got it. I'll send it right away" in the database.

[0580] Step 3:

[0581] The server trains a deep learning model using the stored sending history. The input is past message data stored in the database, and the output is the trained deep learning model. Specifically, the server uses TensorFlow to build a neural network to learn users' message patterns and inputs past messages as training data.

[0582] Step 4:

[0583] The user sends a new message. The input is the new text message entered by the user, and the output is the new message data sent to the server. In concrete terms, the user types "Today was fun!" and presses the send button.

[0584] Step 5:

[0585] The server inputs the received new message into a sentiment analysis engine and analyzes the sentiment. The input is the new message data, and the output is the analyzed sentiment data. Specifically, the server uses a sentiment analysis engine (for example, an NLP library) to extract the sentiments of "joy" and "satisfaction" from the message "Today was fun!"

[0586] Step 6:

[0587] The server uses a stamp generation engine based on the results of the emotion analysis to generate appropriate stamps. The input is the analyzed emotion data, and the output is the generated stamp data. Specifically, the server inputs emotions such as "joy" and "satisfaction" into the stamp generation engine and generates a smiling stamp.

[0588] Step 7:

[0589] The server sends the generated stamp to the user's device. The input is the generated stamp data, and the output is the stamp displayed on the user's device. Specifically, the server sends the generated smiley face stamp to the user's chat app, where it is displayed on the user's screen.

[0590] (Application example 2)

[0591] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0592] Conventional messaging systems do not provide automatic replies or stamp generation that take into account the user's past sending history or emotions, limiting the user's communication experience. Furthermore, they lack the functionality to automatically generate appropriate stamps based on the user's emotions, which means users have to go through the trouble of choosing stamps themselves.

[0593] 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.

[0594] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stickers using phrases and reply styles that suit the individual user, means for extracting the user's emotions using an emotion engine and generating appropriate stickers, means for analyzing the user's message history using a generative AI model and generating stickers based on the emotions, and means for inputting prompt sentences into the generative AI model. This enables the generation of automatic replies and stickers that take into account the user's past sending history and emotions, improving the user's communication experience.

[0595] "Means for analyzing traffic within messengers" refers to technology that analyzes message sending and receiving data to understand users' communication patterns and frequency.

[0596] "Means for labeling received messages" refers to technology that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags to them.

[0597] "Means of learning from a user's sending history" refers to technology that collects and analyzes data on messages sent by a user in the past, and learns the user's unique phrasing and response style.

[0598] "Means for semi-automatically generating text and stamps" refers to a technology in which the system automatically suggests and generates appropriate text and stamps based on the user's past sending history and learning results.

[0599] The "emotion engine" is a technology that analyzes and extracts emotions from the content of a user's message and generates appropriate responses and stamps based on those emotions.

[0600] A "generative AI model" is a learning model that uses artificial intelligence to analyze a user's message history and generate appropriate stamps and replies.

[0601] A "prompt" is an instruction entered into a generative AI model to make it perform a specific task.

[0602] A system for carrying out this invention has the following configuration: The server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stickers using phrases and reply styles suited to the individual user, means for extracting the user's emotions using an emotion engine and generating appropriate stickers, means for analyzing the user's message history using a generative AI model and generating stickers based on the emotions, and means for inputting prompt sentences into the generative AI model.

[0603] The server analyzes the user's message history using TensorFlow and Transformers libraries. Specifically, it tokenizes the user's past messages and analyzes them with a deep learning model. This allows it to learn the user's phrasing and response style and generate appropriate auto-replies.

[0604] In addition, an emotion engine is used to extract emotions from the content of the user's message. For example, if a user sends a message saying "Today was fun!", the emotion engine will extract "joy" and generate a smiling stamp based on that emotion. This stamp is generated using the Pillow library.

[0605] Furthermore, a generative AI model is used to analyze the user's message history and generate stickers based on their emotions. At this time, instructions are given to the generative AI model using prompts. For example, by entering a prompt such as "Analyze the user's message history, extract emotions, and generate appropriate stickers," the model will generate appropriate stickers.

[0606] For example, if a user sends a message saying "Today was fun!", the server analyzes the message and extracts the emotion "joy." The emotion engine then generates a smiling sticker and suggests it to the user. This process improves the user's communication experience.

[0607] Example prompt sentence:

[0608] Analyze user's message history, extract sentiment, and generate appropriate stamps.

[0609] The above is an embodiment of the present invention.

[0610] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0611] Step 1:

[0612] The server analyzes the traffic within the messenger. Specifically, it collects data on messages sent and received and analyzes the user's communication patterns and frequency. The input is message sending and receiving data, and the output is the analysis results of the user's communication patterns and frequency.

[0613] Step 2:

[0614] The server labels the received message. Specifically, it identifies the content, importance, urgency, etc. of the received message and assigns an appropriate tag. The input is the received message, and the output is the labeled message.

[0615] Step 3:

[0616] The server learns from the user's sending history. Specifically, it collects and analyzes data on messages sent by the user in the past, and learns the user's unique phrasing and reply style. The input is the user's past message history, and the output is the learned user's phrasing and reply style.

[0617] Step 4:

[0618] The server generates sentences and stamps semi-automatically. Specifically, the system automatically suggests and generates appropriate sentences and stamps based on the user's past sending history and learning results. The inputs are the learning results and the current message, and the output is the suggested sentences and stamps.

[0619] Step 5:

[0620] The server uses an emotion engine to extract the user's emotion. Specifically, it analyzes and extracts emotions from the content of the user's message. The input is the user's message, and the output is the extracted emotion.

[0621] Step 6:

[0622] The server generates appropriate stamps based on the emotions. Specifically, it generates images using the Pillow library based on the extracted emotions. The input is the extracted emotions, and the output is the generated stamps.

[0623] Step 7:

[0624] The server uses a generative AI model to analyze the user's message history. Specifically, the generative AI model inputs a prompt sentence and analyzes the user's message history. The inputs are the prompt sentence and the user's message history, and the output is the analysis result.

[0625] Step 8:

[0626] The server uses a prompt to instruct the generative AI model. Specifically, it inputs the prompt, "Analyze the user's message history, extract emotions, and generate appropriate stamps." The input is the prompt, and the output is the generated stamps.

[0627] Example 3

[0628] Next, a description will be given of a third embodiment of the third embodiment. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0629] Conventional messaging systems have difficulty quickly generating appropriate responses to messages received by users. Furthermore, they are unable to provide responses that reflect the user's emotions, which reduces the quality of communication. Furthermore, they are unable to automatically generate responses that include non-verbal elements, making it difficult to accurately convey the user's intentions.

[0630] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0631] In this invention, the server includes means for analyzing data in message communication, means for classifying received messages, means for learning from a user's sending history and semi-automatically generating sentences and non-verbal elements using phrases and reply styles that suit the user, means for extracting emotions from the user's messages and adjusting replies based on those emotions, and means for sending the generated replies to the user. This allows users to generate quick and appropriate replies to received messages, and further improves the quality of communication by providing replies that are appropriate to the user's emotions. Furthermore, automatically generating replies that include non-verbal elements makes it possible to accurately convey the user's intentions.

[0632] "Messaging" is a means of sending and receiving information, including text and non-verbal elements, between users.

[0633] "Means for analyzing data" refers to technology for analyzing the content of received messages and understanding their intent and importance.

[0634] A "message categorization means" is a technique for organizing received messages based on specific categories or labels.

[0635] "Means for learning from sending history" refers to technology that learns a user's unique phrasing and response style based on messages the user has sent in the past.

[0636] "Means for semi-automatically generating sentences and non-verbal elements" refers to technology in which the system automatically generates appropriate sentences, stamps, and other non-verbal elements based on user input and past data.

[0637] "Means for extracting emotions" refers to technology that analyzes emotions from a user's message and identifies those emotions.

[0638] The "means for adjusting a response" is a technique for appropriately changing the content and tone of a generated response based on the extracted emotion.

[0639] A "means for sending a response" is a technique for sending the generated response to the user.

[0640] "Non-verbal elements" are elements that contain information other than text, such as stamps and emojis.

[0641] This invention is a system that generates prompt and appropriate replies to messages received by a user. The system analyzes data in message communications, classifies received messages, learns from the user's sending history, and semi-automatically generates sentences and non-verbal elements using phrases and reply styles that are personalized for the user. It also extracts emotions from the user's messages, adjusts replies based on those emotions, and sends the generated replies to the user.

[0642] Hardware and software used

[0643] The server uses the following hardware and software to run the system:

[0644] Hardware: A server machine with a powerful processor, sufficient memory, and a fast network interface.

[0645] Software: Libraries and frameworks for implementing natural language processing (NLP) techniques (e.g., TensorFlow, PyTorch), generative AI models (e.g., BERT, GPT-3), and sentiment analysis APIs (e.g., IBM Watson).

[0646] Data processing and calculation

[0647] The server processes and calculates data in the following procedure.

[0648] 1. Receiving a message:

[0649] The server receives the message sent by the user from the terminal.

[0650] 2. Message analysis:

[0651] The server analyzes the received message using natural language processing technology, specifically using generative AI models (e.g., BERT, GPT-3) to understand the intent of the message.

[0652] 3. Emotion extraction:

[0653] The server uses a sentiment analysis API (e.g., IBM Watson) to extract sentiment from the user's message.

[0654] 4. Choosing a response style:

[0655] The server selects an appropriate response style from a learned response style database, which stores labeled versions of past messages and their responses.

[0656] 5. Generate a response:

[0657] The server generates a response based on the selected response style and the extracted emotion. The generated response includes not only text but also non-verbal elements such as stamps.

[0658] 6. Sending a response:

[0659] The server generates a response and sends it to the user.

[0660] Specific examples

[0661] As a specific example, the process when a user sends a message saying "Please send me the meeting materials" is shown below.

[0662] User message: "Please send me the meeting materials."

[0663] Server response: "Got it. I'll send it right away." with a nodding stamp

[0664] An example of a prompt sentence might be:

[0665] User: I'm really tired today

[0666] Server: That was tough work, please take a rest (relaxation stamp)

[0667] In this way, the system semi-automatically generates an appropriate response to a user's message, and by combining it with the emotion engine, it is possible to provide a more appropriate response. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0668] Step 1:

[0669] Receiving messages

[0670] The server receives messages sent by users.

[0671] Input: A message sent by the user from the device (e.g., "Please send me the meeting materials").

[0672] Specific operation: The user inputs a message on the terminal and presses the send button. The server receives this message and proceeds to the next processing step.

[0673] Output: The received message data.

[0674] Step 2:

[0675] Message Parsing

[0676] The server analyzes the received message using natural language processing technology.

[0677] Input: The received message data.

[0678] What it does: The server analyzes the message using a generative AI model (e.g., BERT, GPT-3) to understand its intent.

[0679] Data processing: Message tokenization, grammatical analysis, and semantic analysis.

[0680] Output: Analysis results that represent the intent or content of the message (e.g., "Request for meeting materials").

[0681] Step 3:

[0682] Emotion extraction

[0683] The server uses a sentiment analysis API to extract sentiment from the user's message.

[0684] Input: The received message data.

[0685] Specific operation: The server sends a message to a sentiment analysis API (e.g., IBM Watson) to analyze the sentiment.

[0686] Data operations: Calculating emotion scores, identifying emotion categories.

[0687] Output: Extracted emotion data (e.g., "fatigue", "stress").

[0688] Step 4:

[0689] Choosing a response style

[0690] The server selects an appropriate response style from a learned response style database.

[0691] Input: Analysis results and extracted emotion data.

[0692] Specific operation: The server refers to a database of past messages and replies and selects the most appropriate reply style.

[0693] Data calculation: Matching and selecting response styles.

[0694] Output: Selected response style data.

[0695] Step 5:

[0696] Generate a response

[0697] The server generates a response based on the selected response style and the extracted sentiment.

[0698] Input: Selected response style data and extracted sentiment data.

[0699] Specific behavior: The server uses a generative AI model to generate appropriate sentences and non-verbal elements (e.g., stamps).

[0700] Data processing: sentence generation, selection of non-verbal elements.

[0701] Output: Generated response data (e.g., "Got it. I'll send it right away." and a stamp showing a nod).

[0702] Step 6:

[0703] Sending a response

[0704] The server generates a response and sends it to the user.

[0705] Input: The generated response data.

[0706] Specific operation: The server sends the generated response to the user's terminal.

[0707] Output: A response that is displayed on the user's device (e.g., "Got it. I'll send it right away" and a nodding stamp).

[0708] (Application example 3)

[0709] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0710] Conventional messenger systems have difficulty automatically generating appropriate responses based on the user's emotions, limiting the user experience. Furthermore, they are unable to generate responses that include non-verbal elements such as stamps and emojis, which can lead to a decline in the quality of communication. This leads to issues such as a decrease in user satisfaction and convenience.

[0711] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0712] In this invention, the server includes: means for acquiring messages received within a messaging application; a prompt sentence instructing the server to generate a reply to the received message by imitating a message sent by the user based on a history of messages sent by the user within the messaging application; means for generating reply candidates for the received message using a generative AI; means for presenting the generated reply candidates to the user and accepting editing; and means for sending the edited reply candidates as the user's message. The server also includes means for recognizing a user's emotion based on the message sent by the user, and the generating means generates reply candidates for the received message using the generative AI and a prompt sentence instructing the server to generate a reply to the received message by imitating text and stamps included in the message sent by the user based on a history of messages sent by the user within the messaging application. This allows the system to automatically generate appropriate responses based on the user's emotions, and by including stamps and emojis, allows for richer communication.

[0713] "Means for analyzing traffic within a messenger" refers to technology for analyzing the flow and content of messages sent and received within a messenger application.

[0714] "Means for labeling received messages" refers to technology that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags or labels.

[0715] "A means of semi-automatically generating text and stamps using phrases and reply styles that are suited to each individual user, based on learning from the user's sending history" refers to technology that learns the user's unique phrases and reply styles based on messages the user has sent in the past, and automatically generates replies based on that.

[0716] "Means for extracting emotions from a user's message and adjusting a response based on those emotions" refers to a technology that analyzes emotions from the content of a message sent by a user and generates an appropriate response based on those emotions.

[0717] "Means for including stamps and emojis according to emotions in generated replies" refers to a technology that automatically adds stamps and emojis appropriate to the user's emotions to generated replies.

[0718] The system for implementing this invention has the functions of analyzing traffic within the messenger, labeling received messages, learning from the user's sending history, and semi-automatically generating sentences and stamps. It can also extract emotions from the user's messages, adjust replies based on those emotions, and include stamps and emoticons corresponding to the emotions in the generated replies.

[0719] System configuration

[0720] This system uses the following hardware and software:

[0721] Hardware: Smartphone with internet connection

[0722] Software: Python, TextBlob library, OpenAI API

[0723] Program processing explanation

[0724] The server first analyzes the traffic within Messenger, identifying the content, importance, and urgency of received messages and assigning appropriate labels. Next, generative AI is used to semi-automatically generate reply candidates, including sentences and stickers, based on the user's sending history and imitating the user's unique phrasing and reply style.

[0725] Additionally, it uses the TextBlob library to extract sentiment from users' messages. After the sentiment is extracted, it uses generative AI to generate suggested replies based on the sentiment, which are automatically supplemented with stickers and emojis depending on the sentiment.

[0726] Specific examples

[0727] For example, if a user sends a message saying, "I'm really tired today," the server analyzes this message and extracts emotions such as "fatigue" and "stress." If the server then receives a message from the other party saying, "That must have been tough. Please get some rest," the server uses generative AI to mimic the user's personal message and generate candidate replies to this message (including appropriate stickers and emojis).

[0728] Prompt Sentence Examples

[0729] "User sending history: I'm really tired today User sentiment: negative Message received: That was tough, please take a rest. Based on the user's recognized sentiment and message sending history, generate a reply to the received message by imitating the text, stickers, and emojis of the message sent by the user.

[0730] In this way, appropriate responses can be automatically generated based on the user's emotions, and by including stamps and emojis, richer communication becomes possible.

[0731] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0732] Step 1:

[0733] The server analyzes the traffic within the messenger. Specifically, it analyzes the content of received messages and identifies their importance and urgency. This analysis uses natural language processing technology. The input is the received message, and the output is a labeled message.

[0734] Step 2:

[0735] The server labels the received messages. Specifically, it assigns appropriate tags or labels based on the message content. For example, it might assign a label such as "Request for meeting materials" or "Urgent inquiry." The input is the parsed message, and the output is the labeled message.

[0736] Step 3:

[0737] The server acquires the user's sending history, specifically the sending history of messages sent in the past, in order to learn the user's unique phrasing and reply style.

[0738] Step 4:

[0739] The server uses the acquired message sending history to generate sentences and stamps by using generative AI, imitating the user's messages. Specifically, it generates appropriate reply candidates for received messages. These replies include stamps and emojis. The input is the labeled message and the message sending history, and the output is the generated reply candidates.

[0740] Step 5:

[0741] The server extracts sentiment from user messages. Specifically, it uses the TextBlob library to analyze the sentiment of messages and classify them as positive, negative, or neutral. The input is the received message, and the output is the extracted sentiment.

[0742] Step 6:

[0743] The server further considers the extracted emotion and generates a response. Specifically, it uses generative AI to generate a response according to the emotion. The input is the extracted emotion and the labeled message, and the output is a candidate response.

[0744] Step 7:

[0745] The server automatically adds stickers and emojis appropriate for the content of the reply to the generated reply candidates. The input is the generated reply candidates, and the output is the final reply candidate.

[0746] Step 8:

[0747] The device displays a final reply candidate to the user. Specifically, the reply, including the generated text, stamps, and emojis, is displayed in the messenger application and accepted for editing by the user. The input is the final reply candidate, and the output is the reply edited by the user. The edited reply is sent as the user's message in the message application.

[0748] 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.

[0749] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0750] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0751] 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.

[0752] [Second embodiment]

[0753] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0754] 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.

[0755] 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).

[0756] 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.

[0757] 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.

[0758] 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).

[0759] 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.

[0760] 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.

[0761] 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.

[0762] 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.

[0763] 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.

[0764] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0765] "Example 1"

[0766] The system of the present invention is integrated into a messenger application and analyzes the message traffic sent and received by users in real time. This analysis is performed using natural language processing (NLP) techniques to identify the content, tone, and urgency of the message. For example, if a message says, "Please send me the materials for tomorrow's meeting," the message is labeled as a "Request for meeting materials" and the urgency is set to "High."

[0767] "Example 2"

[0768] The system of the present invention also learns from a user's past sending history. This learning is done using deep learning techniques to understand the user's phrasing and response style. For example, if a user often responds with "Okay, I'll send it right away," the system will learn that same response style.

[0769] "Example 3"

[0770] The system of the present invention then semi-automatically generates replies to received messages. This generation is based on the learned reply style and labeled message content, and can include not only text but also non-verbal elements such as stamps. For example, in response to the above request for meeting materials, a stamp showing a nod is generated along with the sentence "Got it. I'll send it right away."

[0771] The processing flow of each embodiment will be described below.

[0772] "Example 1"

[0773] Step 1: User opens messenger application and sends a message.

[0774] Step 2: The system of the present invention analyzes message traffic in real time using natural language processing (NLP) techniques.

[0775] Step 3: The system will understand the content, tone, and urgency of the message and label it accordingly. For example, a message saying "Please send me the materials for tomorrow's meeting" will be labeled "Request for meeting materials" and the urgency will be set to "High."

[0776] "Example 2"

[0777] Step 1: The system retrieves the user's past sending history.

[0778] Step 2: Based on the acquired sending history, the system uses deep learning technology to learn the user's phrasing and response style.

[0779] Step 3: Based on the learning results, the system will imitate the user's response style. For example, if the user often used the response "Okay, I'll send it right away" in the past, the system will learn a similar response style.

[0780] "Example 3"

[0781] Step 1: The system generates a semi-automatic response to the message received.

[0782] Step 2: Generation is based on the learned response style and labeled message content.

[0783] Step 3: The generated response can include not only text but also non-verbal elements such as stamps. For example, in response to a request for meeting materials, a stamp showing a nod is generated along with the sentence "Understood. I will send them immediately."

[0784] Example 1

[0785] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0786] Conventional messaging applications have difficulty analyzing the content, tone, and urgency of messages sent and received by users in real time. They also lack the ability to learn from a user's sending history and semi-automatically generate text and images using phrases and reply styles that are individually tailored to each individual. As a result, users' communication efficiency declines and important messages may be overlooked.

[0787] 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.

[0788] In this invention, the server includes means for analyzing communications within the messenger, means for labeling received messages, means for learning from the user's sending history and semi-automatically generating text and images using phrases and reply styles suited to the individual user, means for analyzing the content, tone, and urgency of messages using natural language processing technology, and means for displaying the analysis results to the user. This makes it possible to analyze the content, tone, and urgency of messages sent and received by users in real time and semi-automatically generate replies that are suited to each individual user.

[0789] "Means for analyzing communications within a messenger" refers to a function for collecting messages and data sent and received within a messenger application in real time and analyzing their contents.

[0790] "Means for labeling received messages" refers to a function that assigns specific categories or tags to received messages based on their content.

[0791] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a function that analyzes data from messages sent by the user in the past and learns from those patterns to automatically generate replies and stamps that are appropriate for the user.

[0792] "Means for analyzing the content, tone, and urgency of a message using natural language processing technology" refers to a function that uses natural language processing (NLP) technology to analyze the text data of a message and evaluate its content, sentiment, and urgency.

[0793] "Means for displaying analysis results to the user" refers to a function that displays information such as the content and urgency of the analyzed message in a format that is easy for the user to understand.

[0794] The present invention is a system that is incorporated into a messenger application and provides a function for analyzing messages sent and received by a user in real time. A specific embodiment of this system will be described below.

[0795] Hardware and software used

[0796] Hardware:

[0797] Server: Use a server with a powerful processor and sufficient memory. The server analyzes messages and stores data.

[0798] Device: A device used by a user, such as a smartphone or computer, that runs a messenger application.

[0799] software:

[0800] Messenger application: An application that allows users to send and receive messages.

[0801] Natural Language Processing (NLP) libraries, such as Python's NLTK library or spaCy, can be used to analyze message content, tone, and urgency.

[0802] HTTP communication library: Used to send and receive data between the device and the server.

[0803] Program processing

[0804] Message capture:

[0805] The device captures messages sent and received by the user via the messenger application in real time, specifically by using the messenger application's API to retrieve the messages sent.

[0806] Sending a message:

[0807] The device sends the captured message to the server using an HTTP POST request.

[0808] Message parsing:

[0809] The server analyzes the received messages using natural language processing (NLP) techniques, specifically using Python's NLTK library and spaCy to analyze the message's content, tone, and urgency.

[0810] Generate analysis results:

[0811] The server assigns a label to the message as a result of the analysis and sets the urgency level. For example, a message saying "Please send the materials for tomorrow's meeting" is labeled as "Request for meeting materials" and the urgency level is set to "High."

[0812] Returning analysis results:

[0813] The server returns the analysis results to the terminal as an HTTP response.

[0814] Displaying the analysis results:

[0815] The device will update the UI of the messenger application and display the analysis results in a user-friendly format.

[0816] Specific examples

[0817] If a user sends a message in a messenger application saying "Please send me the materials for tomorrow's meeting," the following process will occur:

[0818] 1. The device captures the message using the messenger application's API.

[0819] 2. The device sends the captured message to the server via an HTTP POST request.

[0820] 3. The server parses the message using Python's NLTK library, specifically tokenizing the message, tagging it as a part of speech, and performing sentiment analysis.

[0821] 4. The server labels the analysis result as "Request for meeting materials" and sets the urgency to "High."

[0822] 5. The server returns the analysis results to the terminal as an HTTP response.

[0823] 6. The device updates the UI of the messenger application and displays the analysis result to the user as "Request for meeting materials (urgency: high)."

[0824] Prompt Sentence Examples

[0825] "Create a program that analyzes messages sent by users in a messenger application in real time to understand the content, tone, and urgency of the message. For example, if the message says, 'Please send me the materials for tomorrow's meeting,' label the message as a 'Request for meeting materials' and set the urgency to 'High.'"

[0826] The above is a specific embodiment for carrying out the present invention.

[0827] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0828] Step 1:

[0829] Message capture

[0830] The device captures messages sent and received by users in messenger applications in real time.

[0831] Input: A message sent by a user in a messenger application.

[0832] Specific operation: The device uses the messenger application's API to retrieve the sent message.

[0833] Output: The captured message data.

[0834] Step 2:

[0835] Sending a message

[0836] The terminal sends the captured message to the server.

[0837] Input: The captured message data.

[0838] Specific operation: The device uses an HTTP POST request to send message data to the server.

[0839] Output: The message data sent to the server.

[0840] Step 3:

[0841] Message Parsing

[0842] The server analyzes the received message using natural language processing (NLP) techniques.

[0843] Input: The message data sent to the server.

[0844] What it does: The server uses Python's NLTK library and spaCy to analyze messages for content, tone, and urgency, including tokenizing, part-of-speech tagging, and sentiment analysis.

[0845] Output: Parsed message content, tone, and urgency data.

[0846] Step 4:

[0847] Generate analysis results

[0848] The server generates the analysis results.

[0849] Input: Parsed message content, tone, and urgency data.

[0850] Specific behavior: The server assigns a label and urgency to the message based on its content. For example, a message saying "Please send me the materials for tomorrow's meeting" is labeled as "Request for meeting materials" and the urgency is set to "High."

[0851] Output: Labeled message and urgency data.

[0852] Step 5:

[0853] Returning analysis results

[0854] The server returns the analysis results to the terminal.

[0855] Input: Labeled message and urgency data.

[0856] Specific operation: The server sends the analysis results to the terminal as an HTTP response.

[0857] Output: Analysis results sent back to the device.

[0858] Step 6:

[0859] Displaying analysis results

[0860] The device displays the analysis results to the user.

[0861] Input: Analysis results sent back to the device.

[0862] What it does: The device updates the UI of the messenger application and displays the analysis results in a user-friendly format.

[0863] Output: The analysis results displayed to the user.

[0864] (Application example 1)

[0865] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0866] Conventional messenger applications have difficulty analyzing the content and urgency of messages sent and received by users in real time, making it difficult to detect security risks early. Furthermore, the automatic generation function to support user communication was limited, making it impossible to provide responses optimized for each individual user. This could delay the early detection and rapid response of security risks within a company.

[0867] 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.

[0868] In this invention, the server includes means for analyzing communications within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit the individual user, means for analyzing the content, tone, and urgency of messages in real time, means for determining security risks based on the analysis results, and means for issuing alerts according to the determined risks, thereby enabling early detection of security risks within the messenger application and rapid response.

[0869] "Means for analyzing communications within a messenger" refers to technology for analyzing the content and traffic of messages sent and received within a messenger application.

[0870] "Means for labeling received messages" refers to a technology that assigns appropriate labels to received messages based on their content and urgency.

[0871] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a technology that analyzes a user's past message sending history and automatically generates replies and stamps that are optimized for that user.

[0872] "Means for analyzing message content, tone, and urgency in real time" refers to technology that instantly analyzes the content, emotional tone, and urgency of received messages.

[0873] "Means for determining security risks based on analysis results" refers to technology that evaluates and determines potential security risks based on the analysis results of messages.

[0874] "Means for issuing alerts according to the assessed risk" refers to technology that notifies users and administrators of appropriate alerts according to the assessed level of security risk.

[0875] A system for implementing the present invention is for analyzing messages sent and received within a messenger application in real time and detecting security risks at an early stage. The system includes the following main components:

[0876] 1. Means of analyzing communications in messengers:

[0877] The server analyzes the content and traffic of messages sent and received within the messenger application, using natural language processing (NLP) libraries such as spaCy and Transformers in Python to analyze the content, tone, and urgency of messages.

[0878] 2. How to label received messages:

[0879] The server assigns appropriate labels to received messages based on their content and urgency, such as "Request for meeting materials" or "Urgent response required."

[0880] 3. A method to semi-automatically generate text and images using phrases and response styles that are personalized to each user based on their sending history:

[0881] The server analyzes the user's message history and automatically generates personalized replies and stickers, using a generative AI model to generate prompts tailored to the user's communication style.

[0882] 4. Real-time analysis of message content, tone, and urgency:

[0883] The server instantly analyzes the content, emotional tone, and urgency of the received message, allowing it to determine the importance and urgency of the message and prompt appropriate responses.

[0884] 5. How to determine security risks based on the analysis results:

[0885] The server evaluates and judges the potential security risk based on the analysis of the message. For example, if a message saying "The server is down. Urgent action is required" is received, it will be judged as a high risk.

[0886] 6. Means of issuing alerts according to the determined risk:

[0887] The server will then send appropriate alerts to users and administrators depending on the level of security risk determined, enabling prompt action.

[0888] Specific examples

[0889] Message: "Server is down. Urgent action required."

[0890] Analysis results:

[0891] Entity: [("Server", "PRODUCT")]

[0892] Emotion: [{"label": "NEGATIVE", "score": 0.99}]

[0893] Urgency: "High"

[0894] Prompt Sentence Examples

[0895] python

[0896] message = "The server is down. Urgent action required."

[0897] analysis_result = analyze_message(message)

[0898] print(analysis_result)

[0899] In this way, security risks within messenger applications can be detected early and addressed quickly.

[0900] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0901] Step 1:

[0902] The server receives messages sent and received within the messenger application. The input is the message sent by the user, and the output is the received message data. Specifically, the text content of the message is sent to the server.

[0903] Step 2:

[0904] The server analyzes the received message using a natural language processing (NLP) model. The input is the received message data, and the output is the analysis result. Specifically, it uses spaCy and Transformers to analyze the message content, tone, and urgency.

[0905] Step 3:

[0906] The server assigns labels to messages based on the analysis results. The input is the analysis results, and the output is the labeled message data. Specifically, the server assigns labels such as "Request for meeting materials" or "Urgent response required" to messages.

[0907] Step 4:

[0908] The server learns the user's sending history and semi-automatically generates sentences and images using phrases and response styles that are tailored to the individual user. The input is the user's sending history data, and the output is the generated sentences and images. Specifically, it uses a generative AI model to generate prompt sentences and provide optimized responses to the user.

[0909] Step 5:

[0910] The server analyzes the message content, tone, and urgency in real time. The input is the received message data, and the output is the analysis results. Specifically, it performs sentiment analysis and determines the urgency level to evaluate the importance of the message.

[0911] Step 6:

[0912] The server determines the security risk based on the analysis results. The input is the analysis results, and the output is the risk assessment result. Specifically, it evaluates the potential security risk from the message content and tone, and determines that the risk is high.

[0913] Step 7:

[0914] The server issues an alert based on the risk it determines. The input is the risk assessment result, and the output is an alert notification. Specifically, alerts are sent to users and administrators based on the risk level, encouraging them to take prompt action.

[0915] Example 2

[0916] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0917] Conventional messaging applications have the problem of inefficiency, requiring users to manually input messages each time. It is also difficult to automatically respond to messages that reflect the user's individual phrasing and response style, which can lead to a decline in the quality of communication. Furthermore, it is difficult to properly determine what the other party is requesting and the urgency of that request.

[0918] 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.

[0919] In this invention, the server includes a means for analyzing data in a message application, a means for classifying received messages, a means for learning from a user's sending history and semi-automatically generating text and images using phrases and response styles suited to the individual user, a means for collecting the user's past sending history and storing it in a database, a means for training a model based on the sending history data using deep learning technology, a means for receiving requests from the user and generating appropriate responses, and a means for sending the generated responses to the user. This enables automatic responses that reflect the user's individual phrases and response styles, improving the quality of communication. It also enables efficient message exchange by appropriately determining the matters the other party is requesting and their urgency.

[0920] A "messaging application" is software that allows users to send and receive text and images.

[0921] The "means for analyzing data" refers to a technique for collecting data in a message application and analyzing the content.

[0922] A "message categorization means" is a technique for categorizing received messages based on specific criteria.

[0923] "Means for learning from sending history" refers to technology that learns a user's phrasing and response style based on messages the user has sent in the past.

[0924] "Means for semi-automatic text and image generation" refers to technology in which a system automatically generates text and images based on user input and past history.

[0925] The "means for collecting sending history" is a technique for collecting messages that a user has sent in the past.

[0926] "Means of storing data in a database" refers to the technology used to organize and store collected data.

[0927] "Deep learning technology" is a technique for training models based on large amounts of data and learning patterns and features.

[0928] The "means for training a model" is a technique for training a model based on transmission history data using deep learning technology.

[0929] "Means for receiving requests" refers to the technology for receiving requests or inquiries from users.

[0930] The "means for generating an appropriate response" is a technology that automatically generates a response that matches the user's style in response to a received request.

[0931] The "means for sending a response" is a technique for sending the generated response to the user.

[0932] This invention is a system that learns a user's phrasing and reply style based on the user's past sending history, and automatically generates appropriate replies. A specific embodiment of this system will be described below.

[0933] First, the server collects text data from messages and emails sent by the user in the past. This data serves as the basis for learning the user's phrasing and response style. Common APIs such as the Gmail API and Slack API can be used to collect the data.

[0934] Next, the server saves the collected sending history in a database. This database organizes and stores the sending history for each user. A common database management system such as MySQL or PostgreSQL can be used as the database.

[0935] The server then uses a deep learning framework such as TensorFlow or PyTorch to train a deep learning model based on the historical submission data. This model is used to learn the user's specific phrasing and response style. For example, a long short-term memory (LSTM) network can be used to learn the user's phrasing and response style.

[0936] When a user sends a request, the device receives the request. For example, a user might type "Please send me the materials for the meeting" through a smartphone app. This request is then sent to the server.

[0937] The server uses a deep learning model to generate an appropriate response based on the received request. The model generates a response that matches the user's style based on information learned from past sending history. For example, it can generate a response such as "Got it. I'll send it right away."

[0938] Finally, the device displays the response generated by the server to the user. For example, a message such as "Got it. I'll send it right away" appears on the smartphone app.

[0939] As a concrete example, the following prompt sentences can be input to a generative AI model:

[0940] Prompt: In response to the request "Please send me the materials for the meeting," generate an appropriate response based on past sending history.

[0941] When this prompt sentence is input into the generative AI model, the model generates a response based on its past learning results: "Got it. I'll send it right away."

[0942] In this way, the present invention realizes automatic responses that reflect the user's individual phrasing and response style, improving the quality of communication. It also makes it possible to properly determine the matters the other party is requesting and their urgency, enabling efficient message exchange.

[0943] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0944] Step 1:

[0945] The server collects the user's past sending history. Specifically, it uses common APIs such as the Gmail API and Slack API to obtain text data of messages and emails that the user has sent in the past. The input is the raw message data obtained through the API, and the output is a list of the collected message data.

[0946] Step 2:

[0947] The server saves the collected sending history in a database. Specifically, it uses a database management system such as MySQL or PostgreSQL to organize and store the sending history for each user. The input is a list of collected message data, and the output is the organized sending history data stored in the database.

[0948] Step 3:

[0949] The server uses deep learning techniques to train a model based on the transmission history data. Specifically, it uses deep learning frameworks such as TensorFlow and PyTorch to train an LSTM (long short-term memory) network. The input is the transmission history data stored in a database, and the output is a deep learning model that has learned the user's phrasing and response style.

[0950] Step 4:

[0951] The terminal receives a request from a user. Specifically, the user inputs "Please send me the materials for the meeting" through a smartphone app. The input is the request text entered by the user, and the output is the request data to be sent to the server.

[0952] Step 5:

[0953] The server uses a deep learning model to generate an appropriate response based on the received request. Specifically, the request text is input into the model, and a response is generated based on information learned from past transmission history. The input is the user's request text and the trained deep learning model, and the output is the generated response text.

[0954] Step 6:

[0955] The device displays the response generated by the server to the user. Specifically, the smartphone app displays the message "Got it. I'll send it right away." The input is the response text sent from the server, and the output is the message displayed to the user.

[0956] (Application example 2)

[0957] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0958] Conventional messenger systems required users to create individual messages, making it difficult to respond quickly, especially in busy situations or emergencies. Furthermore, they lacked the functionality to automatically generate optimal replies for each user by utilizing the user's past comment history, which reduced the efficiency of user communication.

[0959] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from the user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the user individually, and means for learning from the user's past comment history and automatically generating a comment style that the user prefers. This enables the user to automatically generate quick and appropriate replies, improving communication efficiency.

[0960] "Means for analyzing traffic within a messenger" refers to a function for analyzing the flow of messages and data sent and received within a messenger application.

[0961] "Means for labeling received messages" refers to a function that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags or labels.

[0962] "Means for learning from the user's sending history" is a function that collects the history of messages sent by the user in the past and uses this to learn the user's wording and response style.

[0963] "Means for semi-automatically generating text and stamps" refers to a function that automatically generates appropriate text and stamps, reducing the amount of input required by the user, based on the user's past sending history and learning results.

[0964] "Means for learning a user's past comment history" is a function that collects the history of comments left by a user in the past and learns the user's commenting style based on this.

[0965] "Means for automatically generating comment styles preferred by users" is a function that automatically generates comments in a format preferred by users based on the learned comment styles of users.

[0966] The system for implementing this invention has the functions of analyzing traffic within the messenger, labeling received messages, learning from the user's sending history, and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user.It also includes a function that learns the user's past comment history and automatically generates a comment style that the user prefers.

[0967] The server first analyzes the traffic within the messenger, using software designed to analyze the flow of messages and data sent and received within the messenger application, specifically deep learning frameworks such as TensorFlow and Keras.

[0968] Next, the received messages are labeled. This labeling process identifies the content, importance, urgency, etc. of the received messages and assigns appropriate tags and labels. This makes it easier to determine the priority of the messages.

[0969] To learn from a user's sending history, the system collects messages sent by the user in the past and learns the user's phrasing and reply style based on this. Again, a deep learning model is built using TensorFlow and Keras to learn the user's characteristics.

[0970] To semi-automatically generate sentences and stamps, the trained model is used to reduce the user's input and automatically generate appropriate sentences and stamps. This process can reflect the user's past usage and style.

[0971] Furthermore, it has the ability to learn from a user's past comment history and automatically generate a comment style that the user prefers. This allows it to automatically generate comments in the format that the user prefers.

[0972] As a specific example, if a user frequently uses the comment "I understand," the system can automatically generate comments such as "I understand. I'll send it right away" or "I understand. Thank you."

[0973] An example of a prompt to input to a generative AI model is as follows:

[0974] Develop an application that learns from a user's past comment history and automatically generates a comment style that the user prefers. A user's comment history includes comments such as "Got it, I'll send it right away" and "Thank you." Use this history to build a model that generates comments that the user prefers.

[0975] In this way, users can automatically generate quick and appropriate responses, improving communication efficiency.

[0976] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0977] Step 1:

[0978] The server analyzes the traffic within the messenger. Specifically, it uses deep learning frameworks such as TensorFlow and Keras to analyze the flow of messages and data sent and received within the messenger application. The input includes all message data within the messenger, and the output extracts features from each message.

[0979] Step 2:

[0980] The server labels the received messages. It identifies the content, importance, urgency, etc. of the received messages and assigns appropriate tags and labels. The input includes the message features extracted in step 1, and the output is labeled message data.

[0981] Step 3:

[0982] The server learns from the user's sending history. It collects the user's message history and uses TensorFlow and Keras again to learn the user's phrasing and reply style based on this. The input includes the user's past message history, and the output is a model that learns the user's characteristics.

[0983] Step 4:

[0984] The server semi-automatically generates sentences and stamps. It uses the trained model to automatically generate appropriate sentences and stamps, reducing the user's input effort. The input includes the model generated in step 3 and the current message context, and the output provides the generated sentences and stamps.

[0985] Step 5:

[0986] The server learns from the user's past comment history. It collects the user's past comment history and uses TensorFlow and Keras again to learn the user's commenting style based on this. The input includes the user's past comment history, and the output is a model that learns the user's commenting style.

[0987] Step 6:

[0988] The server automatically generates a comment style that the user prefers based on the trained model. The input includes the model generated in step 5 and the current comment context, and the generated comment is provided as the output.

[0989] In this way, users can automatically generate quick and appropriate responses, improving communication efficiency.

[0990] Example 3

[0991] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0992] In modern communication, users receive many messages and need to respond to them quickly and appropriately. However, responding manually takes time and effort, and efficiency is especially important in business environments. Furthermore, if the content and style of responses are inconsistent, the quality of communication may decline. To solve these issues, a system is needed that can semi-automatically generate appropriate responses to received messages.

[0993] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0994] In this invention, the server includes a means for receiving a message, a means for analyzing the received message, a means for selecting a reply style based on the analysis result, a means for generating sentences and non-verbal elements based on the selected reply style, and a means for transmitting the generated reply, thereby enabling a user to semi-automatically generate a quick and consistent reply to a received message.

[0995] The "means for receiving a message" is a function that allows the server to receive a message sent by a user.

[0996] "Means for analyzing received messages" refers to a function for analyzing the contents of received messages and extracting their intent and important information.

[0997] The "means for selecting a reply style based on the analysis results" is a function for selecting an appropriate reply style based on the contents of the analyzed message.

[0998] The "means for generating sentences and non-verbal elements based on the selected reply style" is a function for generating sentences, stamps, and other non-verbal elements according to the selected reply style.

[0999] The "means for sending the generated response" is a function for sending the generated response to the user.

[1000] "Non-verbal elements" are means of communication other than text, including stamps and emojis.

[1001] A "reply style" is a specific phrase or expression that is selected based on a user's past sending history and learning data.

[1002] The present invention relates to a system for semi-automatically generating a reply to a received message. A specific embodiment of this system will be described below.

[1003] System configuration

[1004] This system consists of three main components: a server, a terminal, and a user. The server receives messages, analyzes them, selects a reply style, generates a reply, and sends the reply. The terminal is a device that users use to send and receive messages. The user is the entity that communicates using the system.

[1005] Hardware and software used

[1006] The server is a computer system equipped with a high-performance processor and large memory capacity. Specifically, a server equipped with an Intel Xeon processor or an AMD EPYC processor is suitable. For software, generative AI models such as Google's BERT and OpenAI's GPT-3 are used to implement natural language processing (NLP) technology.

[1007] Data processing and calculation

[1008] The server processes and calculates data in the following procedure.

[1009] 1. Receiving a message:

[1010] The server receives the message sent by the user from the terminal.

[1011] 2. Message analysis:

[1012] The server analyzes the received message using generative AI models such as BERT and GPT-3, and extracts the message's intent and key information.

[1013] 3. Choose your response style:

[1014] The server then selects an appropriate response style based on the analysis results, which is extracted from a dataset of learned messages and their responses.

[1015] 4. Generate a response:

[1016] The server generates sentences and non-verbal elements (such as stamps) based on the selected response style, using a generative AI model.

[1017] 5. Sending a response:

[1018] The server generates a response and sends it to the user.

[1019] Specific examples

[1020] When a user sends a message saying "Please send me the meeting materials," the server processes the message as follows.

[1021] 1. Receiving a message:

[1022] The user sends a message from the terminal saying, "Please send me the meeting materials."

[1023] The server receives this message.

[1024] 2. Message analysis:

[1025] The server analyzes the received message.

[1026] The server extracts the keyword "conference materials" and recognizes that the user is requesting materials.

[1027] 3. Choose your response style:

[1028] The server selects the response style "Understood. I'll send it right away" based on past data.

[1029] 4. Generate a response:

[1030] The server generates a stamp showing a nod along with the text "Got it. I'll send it right away."

[1031] 5. Sending a response:

[1032] The server sends the generated text "Got it. I'll send it right away" and a stamp showing a nod to the user.

[1033] The user receives this response at the terminal.

[1034] Prompt Sentence Examples

[1035] The following is an example of a prompt sentence that generates an appropriate response to the message "Please send me the meeting materials."

[1036] Message from user: "Please send me the meeting materials"

[1037] Response style: "Okay, I'll send it right away."

[1038] Non-verbal: "Nodding Stamp"

[1039] By inputting this prompt sentence into the generative AI model, an appropriate response is generated. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1040] Step 1:

[1041] Receiving messages

[1042] The server receives messages sent by users. The input is a text message sent by the user from the terminal, and the output is the message data received by the server. In concrete terms, when a user sends a message from the terminal saying "Please send me the meeting materials," the server receives this message.

[1043] Step 2:

[1044] Message Parsing

[1045] The server analyzes the received message. The input is the received message data, and the output is the intent and important information of the analyzed message. Specifically, the server analyzes the message "Please send me the meeting materials," extracts the keyword "meeting materials," and recognizes that the user is requesting materials. This analysis uses generative AI models such as BERT and GPT-3.

[1046] Step 3:

[1047] Choose a response style

[1048] The server selects an appropriate response style based on the analysis results. The input is the intent and important information of the analyzed message, and the output is the selected response style. In concrete terms, the server selects a response style such as "Got it. I'll send it right away" from past data. This selection is made to provide an appropriate response to the user's request.

[1049] Step 4:

[1050] Generate a response

[1051] The server generates sentences and non-verbal elements based on the selected response style. The input is the selected response style, and the output is the generated sentences and non-verbal elements (such as stamps). Specifically, the server generates a stamp indicating a nod along with the sentence "Got it. I'll send it right away." A generative AI model is used to generate this.

[1052] Step 5:

[1053] Sending a response

[1054] The server sends the generated response to the user. The input is the generated text and non-verbal elements, and the output is the response message that the user receives on their device. Specifically, the server sends the generated text "Got it. I'll send it right away" and a stamp indicating a nod to the user. The user receives this response on their device.

[1055] (Application example 3)

[1056] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1057] Conventional messaging systems have the problem of making it difficult to respond quickly and appropriately to messages received by users. In particular, food delivery services are required to respond quickly to inquiries from users, but there are limitations to manual responses. In addition, because the content of responses is uniform, it is difficult to provide users with a sense of familiarity and individual attention.

[1058] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1059] In this invention, the server includes means for analyzing data in message communications, means for classifying received messages, means for learning from a user's sending history and semi-automatically generating sentences and non-verbal elements using phrases and reply styles suited to the individual user, means for adding appropriate non-verbal elements to the generated reply, and means for sending the generated reply to the user. This makes it possible to provide a quick and appropriate reply to a user's inquiry, and to create a sense of friendliness and individual attention.

[1060] "Messaging" is a means of exchanging information between users, including text and non-verbal elements.

[1061] "Means for analyzing data" refers to technology that analyzes the content and patterns of received messages and extracts information to generate an appropriate response.

[1062] "Means for classifying messages" refers to a technique for categorizing received messages based on their content and urgency.

[1063] "Means for learning from a user's sending history" refers to technology that learns a user's response style and phrasing based on their past message sending history.

[1064] "Means for semi-automatically generating sentences and non-verbal elements" refers to technology in which the system automatically generates appropriate sentences and non-verbal elements such as stamps based on the user's sending history and message content.

[1065] "Non-verbal elements" are visual expressions such as stickers and emojis that accompany text messages.

[1066] "Means for adding appropriate non-verbal elements to responses" refers to technology that automatically adds appropriate stamps and emoticons to generated text.

[1067] The "means for sending a response to a user" refers to a technique for sending the generated response to a user.

[1068] The system for implementing this invention analyzes data in message communications, classifies received messages, and learns from the user's sending history to semi-automatically generate sentences and non-verbal elements. A specific embodiment of this system is described below.

[1069] System Configuration

[1070] The system consists of the following major components:

[1071] 1. Data Analysis Module: Analyzes data in message traffic. This module analyzes the content and patterns of received messages and extracts information to generate appropriate responses.

[1072] 2. Message Classification Module: Classifies received messages based on their content and urgency. This module determines what the recipient is requesting and how urgent it is.

[1073] 3. Learning module: This module learns from the user's sending history and semi-automatically generates sentences and non-verbal elements using phrases and response styles that are appropriate for the individual user. This module learns the user's response style and phrases based on their past message sending history.

[1074] 4. Non-verbal element addition module: Adds appropriate non-verbal elements (stamps and emojis) to the generated responses.

[1075] 5. Reply sending module: Sends the generated reply to the user.

[1076] Hardware and software used

[1077] Hardware: Smartphone

[1078] Software: OpenAI GPT-3 API, Python

[1079] Data processing and calculation

[1080] The server first analyzes messages received from users using the data analysis module. Next, the message classification module classifies messages based on content and urgency. After that, the learning module learns appropriate reply styles based on the user's sending history and semi-automatically generates sentences and non-verbal elements. The non-verbal element addition module adds appropriate stamps and emojis to the generated replies. Finally, the reply sending module sends the generated replies to the user.

[1081] Specific examples

[1082] For example, if a user asks, "What is the status of my order?" the system will act as follows:

[1083] 1. Data analysis: Analyze the content of the message and extract the keyword "order status."

[1084] 2. Message Classification: Classify the message content as being about "order status."

[1085] 3. Response generation: Based on past sending history, the learning module generates a response such as, "A delivery person is currently picking up your order. Please wait a little longer."

[1086] 4. Adding non-verbal elements: Add a stamp of the delivery person holding food to the generated response.

[1087] 5. Send Response: Send the final response to the user.

[1088] Prompt Sentence Examples

[1089] Message from user: What is the status of my order?

[1090] Response style: Polite and friendly

[1091] Response:

[1092] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1093] Step 1:

[1094] The server inputs messages received from users into a data analysis module. The data analysis module analyzes the content and patterns of the messages and extracts information to generate an appropriate response. For example, analyzing the message "Please tell me the status of my order," the module extracts the keyword "order status."

[1095] Input: Message from user

[1096] Output: Analyzed keywords and content

[1097] Step 2:

[1098] The server inputs the analyzed keywords and content into the message classification module, which then classifies the received messages based on their content and urgency. For example, based on the keyword "order status," the message may be classified into the "order status check" category.

[1099] Input: Analyzed keywords and content

[1100] Output: Classified message categories

[1101] Step 3:

[1102] The server inputs the classified message categories into a learning module. Based on the user's sending history, the learning module semi-automatically generates sentences using phrases and response styles that are appropriate for the individual user. For example, based on past sending history, it generates a response such as, "A delivery person is currently picking up your order. Please wait a little longer."

[1103] Input: Message categories, user sending history

[1104] Output: Generated response text

[1105] Step 4:

[1106] The server inputs the generated response sentence into the non-verbal element addition module, which adds appropriate non-verbal elements (stamps and emojis) to the generated response. For example, in response to the response "A delivery person is currently picking up your order. Please wait a little longer," a stamp of a delivery person holding food is added.

[1107] Input: Generated response text

[1108] Output: Response with added non-verbal elements

[1109] Step 5:

[1110] The server inputs the response with the added non-verbal elements into the response sending module. The response sending module sends the generated response to the user. For example, it sends the final response to the user, such as "The delivery person is currently picking up your order. Please wait a little longer. (Stamp)."

[1111] Input: Response with added non-verbal elements

[1112] Output: Reply sent to the user

[1113] 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.

[1114] "Example 1"

[1115] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system extracts emotions from a user's message and adjusts a response based on the emotions. Specifically, if a user sends a message such as "I'm really tired today," the emotion engine extracts emotions such as "fatigue" or "stress" from the message. Then, based on the emotions, the system generates a response such as "That must have been hard. Please take a good rest." In this way, the system incorporating an emotion engine can provide a more appropriate response that corresponds to the user's emotions.

[1116] "Example 2"

[1117] The emotion engine can also generate appropriate stamps based on the user's emotions. For example, if a user sends a message such as "Today was fun!", the emotion engine can extract emotions such as "joy" or "satisfaction" from the message. Based on this emotion, the system can generate a stamp of a smiling face, for example. In this way, a system that incorporates an emotion engine can provide more appropriate stamps that correspond to the user's emotions.

[1118] "Example 3"

[1119] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system extracts emotions from a user's message and tailors a response based on the emotions. Specifically, if a user sends a message such as "I'm really tired today," the emotion engine extracts emotions such as "fatigue" and "stress" from the message. Based on this emotion, the system then responds with a response such as "That was tough, please take a good rest."

[1120] In this way, a system that combines an emotion engine can provide more appropriate responses that correspond to the user's emotions.

[1121] The processing flow of each embodiment will be described below.

[1122] "Example 1"

[1123] Step 1: A user opens a messenger application and sends a message, for example, "I'm really tired today."

[1124] Step 2: The system analyzes the sent message in real time using natural language processing (NLP) techniques.

[1125] Step 3: The emotion engine extracts emotions such as "fatigue" or "stress" from the analyzed messages.

[1126] Step 4: The system generates a response based on the extracted sentiment, for example, "That must have been tough, take a rest."

[1127] "Example 2"

[1128] Step 1: A user opens a messenger application and sends a message, for example, "I had a great time today!"

[1129] Step 2: The system analyzes the sent message in real time using natural language processing (NLP) techniques.

[1130] Step 3: The emotion engine extracts emotions such as "joy" or "satisfaction" from the analyzed messages.

[1131] Step 4: The system generates a stamp based on the extracted emotion, such as a smiley face stamp.

[1132] Example 1

[1133] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1134] With conventional messaging applications, it was difficult to understand the content, tone, and urgency of messages sent and received by users in real time, making it impossible to generate appropriate replies. It was also difficult to provide replies based on the user's emotions. This limited the user's communication experience and hindered efficient information transmission.

[1135] 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.

[1136] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles suited to the individual user, means for analyzing the content of messages using natural language processing technology, means for extracting emotions from messages using an emotion engine, means for generating appropriate replies based on the emotions extracted using a generative AI model, and means for sending the generated replies to the user, thereby making it possible to analyze user messages in real time and provide appropriate replies.

[1137] "Means for analyzing traffic within messengers" refers to technology for monitoring and analyzing message data sent and received through messenger applications in real time.

[1138] "Means for labeling received messages" refers to a technology that analyzes the content of a received message and assigns an appropriate label (tag) based on that content.

[1139] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a technology that analyzes a user's past message sending history, learns the phrases and reply styles that are best suited to that user, and semi-automatically generates text and images.

[1140] "Means for analyzing the content of a message using natural language processing technology" refers to a technology that uses natural language processing (NLP) technology to tokenize the content of a message and extract meaning.

[1141] The "means for extracting emotions from a message using an emotion engine" is a technology that uses an emotion engine to identify and extract a user's emotions from the tone and content of a message.

[1142] "Means for generating an appropriate response based on emotions extracted using a generative AI model" is a technology that uses a generative AI model to generate an appropriate response based on emotions extracted.

[1143] The "means for sending the generated reply to the user" is a technique for sending the generated reply to the user through a messenger application.

[1144] The present invention is a system that is incorporated into a messenger application and analyzes messages sent and received by users in real time and generates appropriate responses. The system includes elements of a server, a terminal, and a user.

[1145] Hardware and software used

[1146] The server uses the following hardware and software:

[1147] Hardware: A server machine with high-performance processors, memory, and storage

[1148] Software: Natural language processing libraries (e.g., spaCy, NLTK), emotion engines (e.g., IBM Watson Tone Analyzer), generative AI models (e.g., OpenAI GPT-3)

[1149] Data processing and calculation

[1150] The server receives the message sent by the user and performs the following actions:

[1151] 1. Message content analysis:

[1152] The server uses natural language processing libraries to tokenize the message and extract meaning.

[1153] For example, if a user sends a message saying, "Please send me the materials for tomorrow's meeting," the server will label this message as a "request for meeting materials" and set the urgency to "high."

[1154] 2. Emotion extraction:

[1155] The server uses an emotion engine to extract emotions from the messages.

[1156] For example, if a user sends a message such as "I'm really tired today," the server will extract emotions such as "fatigue" and "stress" from the message.

[1157] 3. Generate a response:

[1158] The server uses a generative AI model to generate appropriate responses based on the extracted emotions.

[1159] For example, in response to a message such as "I'm really tired today," the server generates a response such as "That must have been tough, please get some rest."

[1160] 4. Sending a response:

[1161] The server then sends the generated reply to the user, who then displays the reply to the user through the messenger application.

[1162] Specific examples

[1163] As a concrete example, consider the following scenario:

[1164] scenario:

[1165] A user sends a message through a messenger application saying, "Please send me the materials for tomorrow's meeting."

[1166] Process flow:

[1167] 1. The server receives the message and uses a natural language processing library to label it as a "request for meeting materials" and set the urgency to "high."

[1168] 2. The server uses the emotion engine to extract emotions, but since this message does not contain any specific emotion, emotion extraction is skipped.

[1169] 3. The server uses a generative AI model to generate an appropriate response, in which case it determines that no response is necessary.

[1170] 4. The server generates a response and sends it to the user, who then displays it on their device.

[1171] Prompt Sentence Examples

[1172] Below are some example prompts to input to a generative AI model:

[1173] A user sends a message saying, "I'm really tired today." Extract the sentiment from this message and generate an appropriate response.

[1174] In this way, the system can analyze the user's message in real time and provide an appropriate response.

[1175] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1176] Step 1:

[1177] Receiving messages

[1178] The server receives messages sent by users through the messenger application.

[1179] Input: The message sent by the user (e.g., "Please send me the materials for tomorrow's meeting")

[1180] Specific operation: The server receives messages through the API endpoint of the messenger application and stores them in a database.

[1181] Output: Received message data

[1182] Step 2:

[1183] Message content analysis

[1184] The server analyzes the received message using a natural language processing (NLP) library.

[1185] Input: Received message data

[1186] What happens: The server uses an NLP library (e.g., spaCy) to tokenize the message and extract meaning. For example, it parses the message "Please send me the materials for tomorrow's meeting" and labels it as "Request for meeting materials."

[1187] Output: Labeled message data (e.g., "Request for meeting materials")

[1188] Step 3:

[1189] Emotion extraction

[1190] The server uses an emotion engine to extract emotions from the messages.

[1191] Input: Labeled message data

[1192] Specific operation: The server analyzes the message using an emotion engine (e.g., IBM Watson Tone Analyzer) and extracts emotions. For example, it extracts emotions such as "fatigue" and "stress" from the message "I'm really tired today."

[1193] Output: Extracted emotion data (e.g., "fatigue," "stress")

[1194] Step 4:

[1195] Generate a response

[1196] The server uses a generative AI model to generate appropriate responses based on the extracted emotions.

[1197] Input: Extracted emotion data

[1198] Specific operation: The server inputs a prompt sentence into the generative AI model (e.g., OpenAI GPT-3). For example, in response to the message "I'm really tired today," the server inputs the prompt sentence "The user feels tired. Please generate an appropriate response." The generative AI model generates the response "That must have been hard. Please take a good rest."

[1199] Output: Generated response data (e.g. "That was tough, please take a rest").

[1200] Step 5:

[1201] Sending a response

[1202] The server generates a response and sends it to the user.

[1203] Input: Generated response data

[1204] Specific operation: The server sends the generated reply to the user through the messenger application's API.

[1205] Output: A response message that is displayed on the user's device (e.g., "That was tough, please get some rest.")

[1206] (Application example 1)

[1207] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1208] Conventional messaging applications have difficulty analyzing the content, urgency, and emotions of messages sent and received by users in real time, and responding appropriately. Furthermore, they lack the functionality to instantly send appropriate alerts in emergencies, and a rapid response is required to ensure user safety.

[1209] 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.

[1210] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user, means for determining the urgency of a message, means for extracting emotions from the message using an emotion engine, and means for sending an automatic alert when the urgency is determined to be high. This makes it possible to analyze user messages in real time and take appropriate action based on the urgency and emotions.

[1211] "Means for analyzing traffic within messengers" refers to technology that analyzes the content, frequency, timing, etc. of messages sent and received by users in real time.

[1212] "Means for labeling received messages" refers to techniques for attaching appropriate tags or labels to messages based on their content.

[1213] "A means of learning from a user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit each individual user" is a technology that analyzes past message history and automatically generates appropriate replies based on the user's individual communication style.

[1214] "Means for determining the urgency of a message" refers to a technique for evaluating how urgent a message is based on its content and context.

[1215] The "means for extracting emotions from a message using an emotion engine" is a technology that uses natural language processing technology to analyze and identify emotions contained in a message.

[1216] "Means for sending automatic alerts when a message is deemed to be of high urgency" refers to technology that automatically sends a warning to pre-defined contacts (e.g., police or security companies) when a message is deemed to be of high urgency.

[1217] A system for implementing this invention includes means for analyzing traffic within a messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user, means for determining the urgency of a message, means for extracting emotions from the message using an emotion engine, and means for sending an automatic alert if the urgency is determined to be high.

[1218] Hardware and software used

[1219] Hardware: Smartphones, servers

[1220] Software: Python, NLTK library, smtplib, natural language processing (NLP) technology, emotion engine

[1221] Data processing and calculation

[1222] 1. How to analyze traffic in Messenger:

[1223] The server analyzes the content, frequency, and timing of messages sent and received by users in real time, using natural language processing (NLP) technology.

[1224] 2. How to label received messages:

[1225] The server assigns appropriate tags or labels to messages based on their content, such as "Request for meeting materials" or "High urgency."

[1226] 3. A method to learn from a user's sending history and semi-automatically generate sentences and stamps using phrases and reply styles that are personalized to that user:

[1227] The server analyzes past message history and automatically generates appropriate replies based on the user's individual communication style, using a generative AI model.

[1228] 4. How to determine the urgency of a message:

[1229] The server evaluates the urgency of a message based on its content and context, including whether it contains certain keywords (e.g., "help" or "intrusion").

[1230] 5. How to extract emotions from messages using an emotion engine:

[1231] The server uses natural language processing technology to analyze and identify emotions contained in the message, such as "fear" or "urgency."

[1232] 6. Means of sending automated alerts in case of high severity:

[1233] If the server determines that the incident is urgent, it will automatically send an alert to pre-defined contacts (e.g., police or security companies) by using smtplib to send emails.

[1234] Specific examples

[1235] If a user sends a message like "Help, someone's breaking into my house!", the server will analyze the message in real time, label it as a "break-in report," and assign a "high urgency" to it. The emotion engine will detect "fear" and "urgency" and immediately send an automatic alert to the police or security company.

[1236] Prompt Sentence Examples

[1237] "Design a system that analyzes messages sent by users in real time, determines their urgency and sentiment, and sends automatic alerts as needed. For example, if the message says, 'Help, someone's broken into my house!', the system should label this message as a 'Reporting a Break-in,' set the urgency to 'Very High,' and immediately send an automatic alert to the police or security company."

[1238] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1239] Step 1:

[1240] The server receives messages sent and received by users in real time.

[1241] Input: The message sent by the user

[1242] Output: Received message data

[1243] Specific operation: The server captures messages sent from the messenger application in real time and stores them in a database for analysis.

[1244] Step 2:

[1245] The server analyzes the received message using natural language processing (NLP) techniques.

[1246] Input: Received message data

[1247] Output: Parsed message content, tone, and urgency

[1248] What it does: The server uses the NLTK library to analyze the content of the message and assess its tone and urgency. For example, it analyzes keywords and context within the message to determine its urgency.

[1249] Step 3:

[1250] The server performs labeling based on the parsed message.

[1251] Input: Parsed message content, tone, and urgency

[1252] Output: Labeled message data

[1253] Specific operation: The server assigns an appropriate label (e.g., "Request for meeting materials" or "High urgency") to each message depending on its content, making it easier to categorize messages.

[1254] Step 4:

[1255] The server learns from the user's sending history and semi-automatically generates replies using individual wording and response styles.

[1256] Input: User sending history, labeled message data

[1257] Output: The generated reply message

[1258] How it works: The server uses a generative AI model to analyze a user's past messaging history and generate appropriate replies based on their individual communication style, such as automatically selecting frequently used phrases and stickers.

[1259] Step 5:

[1260] The server reassess the urgency of the message and extracts emotions using an emotion engine if necessary.

[1261] Input: Labeled message data, generated reply message

[1262] Output: Message data with urgency and sentiment ratings

[1263] Specific operation: The server uses an emotion engine to extract emotions from messages and reassess their urgency. For example, it can detect emotions such as "fear" or "urgency."

[1264] Step 6:

[1265] The server will send an automatic alert if the emergency is deemed high.

[1266] Input: Message data with urgency and sentiment ratings

[1267] Output: Alert sent

[1268] Specific operation: If the server determines that the urgency is "high," it will automatically send a warning to pre-defined contacts (e.g., police or security companies) using smtplib, for example, by sending an email or SMS to prompt a prompt response.

[1269] Example 2

[1270] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1271] Conventional messaging systems provide uniform replies and stamps without considering the user's past sending history or emotions, making personalized communication difficult. Furthermore, they lack the functionality to automatically generate appropriate stamps based on the user's emotions, making it difficult to accurately convey the user's emotions.

[1272] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for analyzing data in message communication, means for classifying received messages, means for learning from the user's sending history and semi-automatically generating text and images using phrases and reply styles that suit the individual user, means for analyzing the user's emotions and generating appropriate images based on the emotions, and means for transmitting the generated images to the user's terminal. This enables personalized communication that takes into account the user's past sending history and emotions.

[1273] "Message communication" is a means of communication for sending and receiving information such as text and images between users.

[1274] "Means for analyzing data" refers to technology or devices for analyzing data in message communications and extracting necessary information.

[1275] "Means for classifying messages" refers to techniques or devices for classifying received messages based on their content and urgency.

[1276] "Sending history" is a record of messages a user has sent in the past.

[1277] A "means for learning" is a technique or device for learning a user's specific patterns or style based on the user's sending history.

[1278] "Means for semi-automatically generating text and images" refers to technology or devices that generate text and images in a partially automated manner based on a user's sending history and learning results.

[1279] "Means for analyzing emotions" refers to technologies and devices for extracting and analyzing emotions from users' messages.

[1280] The "means for generating an appropriate image" refers to a technology or device that generates an image to express the user's emotions based on the analyzed emotions.

[1281] "Terminal" means a device (e.g., PC, smartphone) used by a user to communicate messages.

[1282] This system learns from a user's past sending history and understands their phrasing and response style. It also uses an emotion engine to generate stamps that reflect the user's emotions. This system is implemented using a server, a terminal (PC, smartphone, etc.), and various software.

[1283] Hardware and software used

[1284] Hardware: Servers, devices (PCs, smartphones, etc.)

[1285] Software: Deep learning frameworks (TensorFlow, PyTorch, etc.), sentiment analysis engine, stamp generation engine

[1286] System Overview

[1287] 1. Collecting user sending history

[1288] The server stores messages that users have previously sent in a database.

[1289] For example, the message "Got it, I'll send it right away" is saved.

[1290] 2. Deep Learning

[1291] The server uses the stored transmission history to train a deep learning model.

[1292] The model learns your phrasing and response style.

[1293] Example: Learn that users often reply with "Okay, I'll send it right away."

[1294] 3. Sentiment analysis

[1295] When a user sends a new message, the server analyzes the sentiment of the message using a sentiment analysis engine.

[1296] Example: Extracting emotions such as "joy" and "satisfaction" from the message "Today was fun!"

[1297] 4. Stamp generation

[1298] Based on the result of the sentiment analysis, the server generates an appropriate stamp using a stamp generation engine.

[1299] Example: Generating smiley face stickers based on emotions like "happiness" or "contentment."

[1300] 5. Send stamps to your device

[1301] The server sends the generated stamp to the user's terminal.

[1302] Example: The server sends the generated smiley face sticker to the user's chat app, where it is displayed on the user's screen.

[1303] Specific examples

[1304] User sending history: "Okay, I'll send it right away."

[1305] New message: "Today was fun!"

[1306] Sentiment analysis results: "joy" and "satisfaction"

[1307] Generated stamp: Smiling stamp

[1308] Prompt Sentence Examples

[1309] "Describe a system that learns from a user's past messaging history and generates emotionally relevant stickers."

[1310] "If a user sends, 'Today was fun!' what kind of stamp would be generated?"

[1311] In this way, the system of the present invention can provide more personalized communication by learning the user's past sending history and generating stamps according to emotions.

[1312] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1313] Step 1:

[1314] A user uses a device to send a message. The input is a text message typed by the user, and the output is the message data sent to the server. Specifically, the user opens a chat app, types "Got it. I'll send it right away," and presses the send button.

[1315] Step 2:

[1316] The server stores messages sent by users in a database. The input is the message data received by the server, and the output is the message record stored in the database. Specifically, the server analyzes the received message and stores the text "Got it. I'll send it right away" in the database.

[1317] Step 3:

[1318] The server trains a deep learning model using the stored sending history. The input is past message data stored in the database, and the output is the trained deep learning model. Specifically, the server uses TensorFlow to build a neural network to learn users' message patterns and inputs past messages as training data.

[1319] Step 4:

[1320] The user sends a new message. The input is the new text message entered by the user, and the output is the new message data sent to the server. In concrete terms, the user types "Today was fun!" and presses the send button.

[1321] Step 5:

[1322] The server inputs the received new message into a sentiment analysis engine and analyzes the sentiment. The input is the new message data, and the output is the analyzed sentiment data. Specifically, the server uses a sentiment analysis engine (for example, an NLP library) to extract the sentiments of "joy" and "satisfaction" from the message "Today was fun!"

[1323] Step 6:

[1324] The server uses a stamp generation engine based on the results of the emotion analysis to generate appropriate stamps. The input is the analyzed emotion data, and the output is the generated stamp data. Specifically, the server inputs emotions such as "joy" and "satisfaction" into the stamp generation engine and generates a smiling stamp.

[1325] Step 7:

[1326] The server sends the generated stamp to the user's device. The input is the generated stamp data, and the output is the stamp displayed on the user's device. Specifically, the server sends the generated smiley face stamp to the user's chat app, where it is displayed on the user's screen.

[1327] (Application example 2)

[1328] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1329] Conventional messaging systems do not provide automatic replies or stamp generation that take into account the user's past sending history or emotions, limiting the user's communication experience. Furthermore, they lack the functionality to automatically generate appropriate stamps based on the user's emotions, which means users have to go through the trouble of choosing stamps themselves.

[1330] 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.

[1331] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stickers using phrases and reply styles that suit the individual user, means for extracting the user's emotions using an emotion engine and generating appropriate stickers, means for analyzing the user's message history using a generative AI model and generating stickers based on the emotions, and means for inputting prompt sentences into the generative AI model. This enables the generation of automatic replies and stickers that take into account the user's past sending history and emotions, improving the user's communication experience.

[1332] "Means for analyzing traffic within messengers" refers to technology that analyzes message sending and receiving data to understand users' communication patterns and frequency.

[1333] "Means for labeling received messages" refers to technology that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags to them.

[1334] "Means of learning from a user's sending history" refers to technology that collects and analyzes data on messages sent by a user in the past, and learns the user's unique phrasing and response style.

[1335] "Means for semi-automatically generating text and stamps" refers to a technology in which the system automatically suggests and generates appropriate text and stamps based on the user's past sending history and learning results.

[1336] The "emotion engine" is a technology that analyzes and extracts emotions from the content of a user's message and generates appropriate responses and stamps based on those emotions.

[1337] A "generative AI model" is a learning model that uses artificial intelligence to analyze a user's message history and generate appropriate stamps and replies.

[1338] A "prompt" is an instruction entered into a generative AI model to make it perform a specific task.

[1339] A system for carrying out this invention has the following configuration: The server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating sentences and stickers using phrases and reply styles suited to the individual user, means for extracting the user's emotions using an emotion engine and generating appropriate stickers, means for analyzing the user's message history using a generative AI model and generating stickers based on the emotions, and means for inputting prompt sentences into the generative AI model.

[1340] The server analyzes the user's message history using TensorFlow and Transformers libraries. Specifically, it tokenizes the user's past messages and analyzes them with a deep learning model. This allows it to learn the user's phrasing and response style and generate appropriate auto-replies.

[1341] In addition, an emotion engine is used to extract emotions from the content of the user's message. For example, if a user sends a message saying "Today was fun!", the emotion engine will extract "joy" and generate a smiling stamp based on that emotion. This stamp is generated using the Pillow library.

[1342] Furthermore, a generative AI model is used to analyze the user's message history and generate stickers based on their emotions. At this time, instructions are given to the generative AI model using prompts. For example, by entering a prompt such as "Analyze the user's message history, extract emotions, and generate appropriate stickers," the model will generate appropriate stickers.

[1343] For example, if a user sends a message saying "Today was fun!", the server analyzes the message and extracts the emotion "joy." The emotion engine then generates a smiling sticker and suggests it to the user. This process improves the user's communication experience.

[1344] Example prompt sentence:

[1345] Analyze user's message history, extract sentiment, and generate appropriate stamps.

[1346] The above is an embodiment of the present invention.

[1347] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1348] Step 1:

[1349] The server analyzes the traffic within the messenger. Specifically, it collects data on messages sent and received and analyzes the user's communication patterns and frequency. The input is message sending and receiving data, and the output is the analysis results of the user's communication patterns and frequency.

[1350] Step 2:

[1351] The server labels the received message. Specifically, it identifies the content, importance, urgency, etc. of the received message and assigns an appropriate tag. The input is the received message, and the output is the labeled message.

[1352] Step 3:

[1353] The server learns from the user's sending history. Specifically, it collects and analyzes data on messages sent by the user in the past, and learns the user's unique phrasing and reply style. The input is the user's past message history, and the output is the learned user's phrasing and reply style.

[1354] Step 4:

[1355] The server generates sentences and stamps semi-automatically. Specifically, the system automatically suggests and generates appropriate sentences and stamps based on the user's past sending history and learning results. The inputs are the learning results and the current message, and the output is the suggested sentences and stamps.

[1356] Step 5:

[1357] The server uses an emotion engine to extract the user's emotion. Specifically, it analyzes and extracts emotions from the content of the user's message. The input is the user's message, and the output is the extracted emotion.

[1358] Step 6:

[1359] The server generates appropriate stamps based on the emotions. Specifically, it generates images using the Pillow library based on the extracted emotions. The input is the extracted emotions, and the output is the generated stamps.

[1360] Step 7:

[1361] The server uses a generative AI model to analyze the user's message history. Specifically, the generative AI model inputs a prompt sentence and analyzes the user's message history. The inputs are the prompt sentence and the user's message history, and the output is the analysis result.

[1362] Step 8:

[1363] The server uses a prompt to instruct the generative AI model. Specifically, it inputs the prompt, "Analyze the user's message history, extract emotions, and generate appropriate stamps." The input is the prompt, and the output is the generated stamps.

[1364] Example 3

[1365] Next, a description will be given of Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1366] Conventional messaging systems have difficulty quickly generating appropriate responses to messages received by users. Furthermore, they are unable to provide responses that reflect the user's emotions, which reduces the quality of communication. Furthermore, they are unable to automatically generate responses that include non-verbal elements, making it difficult to accurately convey the user's intentions.

[1367] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1368] In this invention, the server includes means for analyzing data in message communication, means for classifying received messages, means for learning from a user's sending history and semi-automatically generating sentences and non-verbal elements using phrases and reply styles that suit the user, means for extracting emotions from the user's messages and adjusting replies based on those emotions, and means for sending the generated replies to the user. This allows users to generate quick and appropriate replies to received messages, and further improves the quality of communication by providing replies that are appropriate to the user's emotions. Furthermore, automatically generating replies that include non-verbal elements makes it possible to accurately convey the user's intentions.

[1369] "Messaging" is a means of sending and receiving information, including text and non-verbal elements, between users.

[1370] "Means for analyzing data" refers to technology for analyzing the content of received messages and understanding their intent and importance.

[1371] A "message categorization means" is a technique for organizing received messages based on specific categories or labels.

[1372] "Means for learning from sending history" refers to technology that learns a user's unique phrasing and response style based on messages the user has sent in the past.

[1373] "Means for semi-automatically generating sentences and non-verbal elements" refers to technology in which the system automatically generates appropriate sentences, stamps, and other non-verbal elements based on user input and past data.

[1374] "Means for extracting emotions" refers to technology that analyzes emotions from a user's message and identifies those emotions.

[1375] The "means for adjusting a response" is a technique for appropriately changing the content and tone of a generated response based on the extracted emotion.

[1376] A "means for sending a response" is a technique for sending the generated response to the user.

[1377] "Non-verbal elements" are elements that contain information other than text, such as stamps and emojis.

[1378] This invention is a system that generates prompt and appropriate replies to messages received by a user. The system analyzes data in message communications, classifies received messages, learns from the user's sending history, and semi-automatically generates sentences and non-verbal elements using phrases and reply styles that are personalized for the user. It also extracts emotions from the user's messages, adjusts replies based on those emotions, and sends the generated replies to the user.

[1379] Hardware and software used

[1380] The server uses the following hardware and software to run the system:

[1381] Hardware: A server machine with a powerful processor, sufficient memory, and a fast network interface.

[1382] Software: Libraries and frameworks for implementing natural language processing (NLP) techniques (e.g., TensorFlow, PyTorch), generative AI models (e.g., BERT, GPT-3), and sentiment analysis APIs (e.g., IBM Watson).

[1383] Data processing and calculation

[1384] The server processes and calculates data in the following procedure.

[1385] 1. Receiving a message:

[1386] The server receives the message sent by the user from the terminal.

[1387] 2. Message analysis:

[1388] The server analyzes the received message using natural language processing technology, specifically using generative AI models (e.g., BERT, GPT-3) to understand the intent of the message.

[1389] 3. Emotion extraction:

[1390] The server uses a sentiment analysis API (e.g., IBM Watson) to extract sentiment from the user's message.

[1391] 4. Choosing a response style:

[1392] The server selects an appropriate response style from a learned response style database, which stores labeled versions of past messages and their responses.

[1393] 5. Generate a response:

[1394] The server generates a response based on the selected response style and the extracted emotion. The generated response includes not only text but also non-verbal elements such as stamps.

[1395] 6. Sending a response:

[1396] The server generates a response and sends it to the user.

[1397] Specific examples

[1398] As a specific example, the process when a user sends a message saying "Please send me the meeting materials" is shown below.

[1399] User message: "Please send me the meeting materials."

[1400] Server response: "Got it. I'll send it right away." with a nodding stamp

[1401] An example of a prompt sentence might be:

[1402] User: I'm really tired today

[1403] Server: That was tough work, please take a rest (relaxation stamp)

[1404] In this way, the system semi-automatically generates an appropriate response to a user's message, and by combining it with the emotion engine, it is possible to provide a more appropriate response. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1405] Step 1:

[1406] Receiving messages

[1407] The server receives messages sent by users.

[1408] Input: A message sent by the user from the device (e.g., "Please send me the meeting materials").

[1409] Specific operation: The user inputs a message on the terminal and presses the send button. The server receives this message and proceeds to the next processing step.

[1410] Output: The received message data.

[1411] Step 2:

[1412] Message Parsing

[1413] The server analyzes the received message using natural language processing technology.

[1414] Input: Received message data.

[1415] What it does: The server analyzes the message using a generative AI model (e.g., BERT, GPT-3) to understand its intent.

[1416] Data processing: Message tokenization, grammatical analysis, and semantic analysis.

[1417] Output: Analysis results that represent the intent or content of the message (e.g., "Request for meeting materials").

[1418] Step 3:

[1419] Emotion extraction

[1420] The server uses a sentiment analysis API to extract sentiment from the user's message.

[1421] Input: Received message data.

[1422] Specific operation: The server sends a message to a sentiment analysis API (e.g., IBM Watson) to analyze the sentiment.

[1423] Data operations: Calculating emotion scores, identifying emotion categories.

[1424] Output: Extracted emotion data (e.g., "fatigue", "stress").

[1425] Step 4:

[1426] Choosing a response style

[1427] The server selects an appropriate response style from a learned response style database.

[1428] Input: Analysis results and extracted emotion data.

[1429] Specific behavior: The server refers to a database of past messages and replies and selects the most appropriate reply style.

[1430] Data calculation: Matching and selecting response styles.

[1431] Output: Selected response style data.

[1432] Step 5:

[1433] Generate a response

[1434] The server generates a response based on the selected response style and the extracted sentiment.

[1435] Input: Selected response style data and extracted sentiment data.

[1436] Specific behavior: The server uses a generative AI model to generate appropriate sentences and non-verbal elements (e.g., stamps).

[1437] Data processing: sentence generation, selection of non-verbal elements.

[1438] Output: Generated response data (e.g., "Got it. I'll send it right away." and a stamp showing a nod).

[1439] Step 6:

[1440] Sending a response

[1441] The server generates a response and sends it to the user.

[1442] Input: The generated response data.

[1443] Specific operation: The server sends the generated response to the user's terminal.

[1444] Output: A response that is displayed on the user's device (e.g., "Got it. I'll send it right away" and a nodding stamp).

[1445] (Application example 3)

[1446] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1447] Conventional messenger systems have difficulty automatically generating appropriate responses based on the user's emotions, limiting the user experience. Furthermore, they are unable to generate responses that include non-verbal elements such as stamps and emojis, which can lead to a decline in the quality of communication. This leads to issues such as a decrease in user satisfaction and convenience.

[1448] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1449] In this invention, the server includes: means for acquiring messages received within a messaging application; a prompt sentence instructing the server to generate a reply to the received message by imitating a message sent by the user based on a history of messages sent by the user within the messaging application; means for generating reply candidates for the received message using a generative AI; means for presenting the generated reply candidates to the user and accepting editing; and means for sending the edited reply candidates as the user's message. The server also includes means for recognizing a user's emotion based on the message sent by the user, and the generating means generates reply candidates for the received message using the generative AI and a prompt sentence instructing the server to generate a reply to the received message by imitating text and stamps included in the message sent by the user based on a history of messages sent by the user within the messaging application. This allows the system to automatically generate appropriate responses based on the user's emotions, and by including stamps and emojis, allows for richer communication.

[1450] "Means for analyzing traffic within a messenger" refers to technology for analyzing the flow and content of messages sent and received within a messenger application.

[1451] "Means for labeling received messages" refers to technology that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags or labels.

[1452] "A means of semi-automatically generating text and stamps using phrases and reply styles that are suited to each individual user, based on learning from the user's sending history" refers to technology that learns the user's unique phrases and reply styles based on messages the user has sent in the past, and automatically generates replies based on that.

[1453] "Means for extracting emotions from a user's message and adjusting a response based on those emotions" refers to a technology that analyzes emotions from the content of a message sent by a user and generates an appropriate response based on those emotions.

[1454] "Means for including stamps and emojis according to emotions in generated replies" refers to a technology that automatically adds stamps and emojis appropriate to the user's emotions to generated replies.

[1455] The system for implementing this invention has the functions of analyzing traffic within the messenger, labeling received messages, learning from the user's sending history, and semi-automatically generating sentences and stamps. It can also extract emotions from the user's messages, adjust replies based on those emotions, and include stamps and emoticons corresponding to the emotions in the generated replies.

[1456] System configuration

[1457] This system uses the following hardware and software:

[1458] Hardware: Smartphone with internet connection

[1459] Software: Python, TextBlob library, OpenAI API

[1460] Program processing explanation

[1461] The server first analyzes the traffic within Messenger, identifying the content, importance, and urgency of received messages and assigning appropriate labels. Next, generative AI is used to semi-automatically generate reply candidates, including sentences and stickers, based on the user's sending history and imitating the user's unique phrasing and reply style.

[1462] Additionally, it uses the TextBlob library to extract sentiment from users' messages. After the sentiment is extracted, it uses generative AI to generate suggested replies based on the sentiment, which are automatically supplemented with stickers and emojis depending on the sentiment.

[1463] Specific examples

[1464] For example, if a user sends a message saying, "I'm really tired today," the server analyzes this message and extracts emotions such as "fatigue" and "stress." If the server then receives a message from the other party saying, "That must have been tough. Please get some rest," the server uses generative AI to mimic the user's personal message and generate candidate replies to this message (including appropriate stickers and emojis).

[1465] Prompt Sentence Examples

[1466] "User sending history: I'm really tired today User sentiment: negative Message received: That was tough, please take a rest. Based on the user's recognized sentiment and message sending history, generate a reply to the received message by imitating the text, stickers, and emojis of the message sent by the user.

[1467] In this way, appropriate responses can be automatically generated based on the user's emotions, and by including stamps and emojis, richer communication becomes possible.

[1468] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1469] Step 1:

[1470] The server analyzes the traffic within the messenger. Specifically, it analyzes the content of received messages and identifies their importance and urgency. This analysis uses natural language processing technology. The input is the received message, and the output is a labeled message.

[1471] Step 2:

[1472] The server labels the received messages. Specifically, it assigns appropriate tags or labels based on the message content. For example, it might assign a label such as "Request for meeting materials" or "Urgent inquiry." The input is the parsed message, and the output is the labeled message.

[1473] Step 3:

[1474] The server acquires the user's sending history, specifically the sending history of messages sent in the past, in order to learn the user's unique phrasing and reply style.

[1475] Step 4:

[1476] The server uses the acquired message sending history to generate sentences and stamps by using generative AI, imitating the user's messages. Specifically, it generates appropriate reply candidates for received messages. These replies include stamps and emojis. The input is the labeled message and the message sending history, and the output is the generated reply candidates.

[1477] Step 5:

[1478] The server extracts sentiment from user messages. Specifically, it uses the TextBlob library to analyze the sentiment of messages and classify them as positive, negative, or neutral. The input is the received message, and the output is the extracted sentiment.

[1479] Step 6:

[1480] The server further considers the extracted emotion and generates a response. Specifically, it uses generative AI to generate a response according to the emotion. The input is the extracted emotion and the labeled message, and the output is a candidate response.

[1481] Step 7:

[1482] The server automatically adds stickers and emojis appropriate for the content of the reply to the generated reply candidates. The input is the generated reply candidates, and the output is the final reply candidate.

[1483] Step 8:

[1484] The device displays a final reply candidate to the user. Specifically, the reply, including the generated text, stamps, and emojis, is displayed in the messenger application and accepted for editing by the user. The input is the final reply candidate, and the output is the reply edited by the user. The edited reply is sent as the user's message in the message application.

[1485] 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 a sound indicating a user input regarding the result of the specific processing.

[1486] The unit 46A transmits voice data representing 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.

[1487] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1488] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1489] 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.

[1490] [Third embodiment]

[1491] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1492] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1493] 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).

[1494] 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.

[1495] 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.

[1496] The camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[1497] For example, an imaging range defined by an angle of view equivalent to the width of the field of view of a typical healthy person is captured.

[1498] 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.

[1499] 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.

[1500] 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.

[1501] 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.

[1502] 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.

[1503] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1504] "Example 1"

[1505] The system of the present invention is integrated into a messenger application and analyzes the message traffic sent and received by users in real time. This analysis is performed using natural language processing (NLP) techniques to identify the content, tone, and urgency of the message. For example, if a message says, "Please send me the materials for tomorrow's meeting," the message is labeled as a "Request for meeting materials" and the urgency is set to "High."

[1506] "Example 2"

[1507] The system of the present invention also learns from a user's past sending history. This learning is done using deep learning techniques to understand the user's phrasing and response style. For example, if a user often responds with "Okay, I'll send it right away," the system will learn that same response style.

[1508] "Example 3"

[1509] The system of the present invention then semi-automatically generates replies to received messages. This generation is based on the learned reply style and labeled message content, and can include not only text but also non-verbal elements such as stamps. For example, in response to the above request for meeting materials, a stamp showing a nod is generated along with the sentence "Got it. I'll send it right away."

[1510] The processing flow of each embodiment will be described below.

[1511] "Example 1"

[1512] Step 1: The user opens the messenger application and sends a message.

[1513] Step 2: The system of the present invention analyzes message traffic in real time using natural language processing (NLP) techniques.

[1514] Step 3: The system will understand the content, tone, and urgency of the message and label it accordingly. For example, a message saying "Please send me the materials for tomorrow's meeting" will be labeled "Request for meeting materials" and the urgency will be set to "High."

[1515] "Example 2"

[1516] Step 1: The system retrieves the user's past sending history.

[1517] Step 2: Based on the acquired sending history, the system uses deep learning technology to learn the user's phrasing and response style.

[1518] Step 3: Based on the learning results, the system will imitate the user's response style. For example, if the user often used the response "Okay, I'll send it right away" in the past, the system will learn a similar response style.

[1519] "Example 3"

[1520] Step 1: The system generates a semi-automatic response to the message received.

[1521] Step 2: Generation is based on the learned response style and labeled message content.

[1522] Step 3: The generated response can include not only text but also non-verbal elements such as stamps. For example, in response to a request for meeting materials, a stamp showing a nod is generated along with the sentence "Understood. I will send them right away."

[1523] Example 1

[1524] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1525] Conventional messaging applications have difficulty analyzing the content, tone, and urgency of messages sent and received by users in real time. They also lack the ability to learn from a user's sending history and semi-automatically generate text and images using phrases and reply styles that are individually tailored to each individual. As a result, users' communication efficiency declines and important messages may be overlooked.

[1526] 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.

[1527] In this invention, the server includes means for analyzing communications within the messenger, means for labeling received messages, means for learning from the user's sending history and semi-automatically generating text and images using phrases and reply styles suited to the individual user, means for analyzing the content, tone, and urgency of messages using natural language processing technology, and means for displaying the analysis results to the user. This makes it possible to analyze the content, tone, and urgency of messages sent and received by users in real time and semi-automatically generate replies that are suited to each individual user.

[1528] "Means for analyzing communications within a messenger" refers to a function for collecting messages and data sent and received within a messenger application in real time and analyzing their contents.

[1529] "Means for labeling received messages" refers to a function that assigns specific categories or tags to received messages based on their content.

[1530] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a function that analyzes data from messages sent by the user in the past and learns from those patterns to automatically generate replies and stamps that are appropriate for the user.

[1531] "Means for analyzing the content, tone, and urgency of a message using natural language processing technology" refers to a function that uses natural language processing (NLP) technology to analyze the text data of a message and evaluate its content, sentiment, and urgency.

[1532] "Means for displaying analysis results to the user" refers to a function that displays information such as the content and urgency of the analyzed message in a format that is easy for the user to understand.

[1533] The present invention is a system that is incorporated into a messenger application and provides a function for analyzing messages sent and received by a user in real time. A specific embodiment of this system will be described below.

[1534] Hardware and software used

[1535] Hardware:

[1536] Server: Use a server with a powerful processor and sufficient memory. The server analyzes messages and stores data.

[1537] Device: A device used by a user, such as a smartphone or computer, that runs a messenger application.

[1538] software:

[1539] Messenger application: An application that allows users to send and receive messages.

[1540] Natural Language Processing (NLP) libraries, such as Python's NLTK library or spaCy, can be used to analyze message content, tone, and urgency.

[1541] HTTP communication library: Used to send and receive data between the device and the server.

[1542] Program processing

[1543] Message capture:

[1544] The device captures messages sent and received by the user via the messenger application in real time, specifically by using the messenger application's API to retrieve the messages sent.

[1545] Sending a message:

[1546] The device sends the captured message to the server using an HTTP POST request.

[1547] Message parsing:

[1548] The server analyzes the received messages using natural language processing (NLP) techniques, specifically using Python's NLTK library and spaCy to analyze the message's content, tone, and urgency.

[1549] Generate analysis results:

[1550] The server assigns a label to the message as a result of the analysis and sets the urgency level. For example, a message saying "Please send the materials for tomorrow's meeting" is labeled as "Request for meeting materials" and the urgency level is set to "High."

[1551] Returning analysis results:

[1552] The server returns the analysis results to the terminal as an HTTP response.

[1553] Displaying the analysis results:

[1554] The device will update the UI of the messenger application and display the analysis results in a user-friendly format.

[1555] Specific examples

[1556] If a user sends a message in a messenger application saying "Please send me the materials for tomorrow's meeting," the following process will occur:

[1557] 1. The device captures the message using the messenger application's API.

[1558] 2. The device sends the captured message to the server via an HTTP POST request.

[1559] 3. The server parses the message using Python's NLTK library, specifically tokenizing the message, tagging it as a part of speech, and performing sentiment analysis.

[1560] 4. The server labels the analysis result as "Request for meeting materials" and sets the urgency to "High."

[1561] 5. The server returns the analysis results to the terminal as an HTTP response.

[1562] 6. The device updates the UI of the messenger application and displays the analysis result to the user as "Request for meeting materials (urgency: high)."

[1563] Prompt Sentence Examples

[1564] "Create a program that analyzes messages sent by users in a messenger application in real time to understand the content, tone, and urgency of the message. For example, if the message says, 'Please send me the materials for tomorrow's meeting,' label the message as a 'Request for meeting materials' and set the urgency to 'High.'"

[1565] The above is a specific embodiment for carrying out the present invention.

[1566] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1567] Step 1:

[1568] Message capture

[1569] The device captures messages sent and received by users in messenger applications in real time.

[1570] Input: A message sent by a user in a messenger application.

[1571] Specific operation: The device uses the messenger application's API to retrieve the sent message.

[1572] Output: The captured message data.

[1573] Step 2:

[1574] Sending a message

[1575] The terminal sends the captured message to the server.

[1576] Input: The captured message data.

[1577] Specific operation: The device uses an HTTP POST request to send message data to the server.

[1578] Output: The message data sent to the server.

[1579] Step 3:

[1580] Message Parsing

[1581] The server analyzes the received message using natural language processing (NLP) techniques.

[1582] Input: The message data sent to the server.

[1583] What it does: The server uses Python's NLTK library and spaCy to analyze messages for content, tone, and urgency, including tokenizing, part-of-speech tagging, and sentiment analysis.

[1584] Output: Parsed message content, tone, and urgency data.

[1585] Step 4:

[1586] Generate analysis results

[1587] The server generates the analysis results.

[1588] Input: Parsed message content, tone, and urgency data.

[1589] Specific behavior: The server assigns a label and urgency to the message based on its content. For example, a message saying "Please send me the materials for tomorrow's meeting" is labeled as "Request for meeting materials" and the urgency is set to "High."

[1590] Output: Labeled message and urgency data.

[1591] Step 5:

[1592] Returning analysis results

[1593] The server returns the analysis results to the terminal.

[1594] Input: Labeled message and urgency data.

[1595] Specific operation: The server sends the analysis results to the terminal as an HTTP response.

[1596] Output: Analysis results sent back to the device.

[1597] Step 6:

[1598] Displaying analysis results

[1599] The device displays the analysis results to the user.

[1600] Input: Analysis results sent back to the device.

[1601] What it does: The device updates the UI of the messenger application and displays the analysis results in a user-friendly format.

[1602] Output: The analysis results displayed to the user.

[1603] (Application example 1)

[1604] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1605] Conventional messenger applications have difficulty analyzing the content and urgency of messages sent and received by users in real time, making it difficult to detect security risks early. Furthermore, the automatic generation function to support user communication was limited, making it impossible to provide responses optimized for each individual user. This could delay the early detection and rapid response of security risks within a company.

[1606] 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.

[1607] In this invention, the server includes means for analyzing communications within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit the individual user, means for analyzing the content, tone, and urgency of messages in real time, means for determining security risks based on the analysis results, and means for issuing alerts according to the determined risks, thereby enabling early detection of security risks within the messenger application and rapid response.

[1608] "Means for analyzing communications within a messenger" refers to technology for analyzing the content and traffic of messages sent and received within a messenger application.

[1609] "Means for labeling received messages" refers to a technology that assigns appropriate labels to received messages based on their content and urgency.

[1610] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a technology that analyzes a user's past message sending history and automatically generates replies and stamps that are optimized for that user.

[1611] "Means for analyzing message content, tone, and urgency in real time" refers to technology that instantly analyzes the content, emotional tone, and urgency of received messages.

[1612] "Means for determining security risks based on analysis results" refers to technology that evaluates and determines potential security risks based on the analysis results of messages.

[1613] "Means for issuing alerts according to the assessed risk" refers to technology that notifies users and administrators of appropriate alerts according to the assessed level of security risk.

[1614] A system for implementing the present invention is for analyzing messages sent and received within a messenger application in real time and detecting security risks at an early stage. The system includes the following main components:

[1615] 1. Means of analyzing communications in messengers:

[1616] The server analyzes the content and traffic of messages sent and received within the messenger application, using natural language processing (NLP) libraries such as spaCy and Transformers in Python to analyze the content, tone, and urgency of messages.

[1617] 2. Means of labeling received messages:

[1618] The server assigns appropriate labels to received messages based on their content and urgency, such as "Request for meeting materials" or "Urgent response required."

[1619] 3. A method to semi-automatically generate text and images using phrases and response styles that are personalized to each user based on their sending history:

[1620] The server analyzes the user's message history and automatically generates personalized replies and stickers, using a generative AI model to generate prompts tailored to the user's communication style.

[1621] 4. Real-time analysis of message content, tone, and urgency:

[1622] The server instantly analyzes the content, emotional tone, and urgency of the received message, allowing it to determine the importance and urgency of the message and prompt appropriate responses.

[1623] 5. How to determine security risks based on the analysis results:

[1624] The server evaluates and judges the potential security risk based on the analysis of the message. For example, if a message saying "The server is down. Urgent action is required" is received, it will be judged as a high risk.

[1625] 6. Means of issuing alerts according to the determined risk:

[1626] The server will then send appropriate alerts to users and administrators depending on the level of security risk determined, enabling prompt action.

[1627] Specific examples

[1628] Message: "Server is down. Urgent action required."

[1629] Analysis results:

[1630] Entity: [("Server", "PRODUCT")]

[1631] Emotion: [{"label": "NEGATIVE", "score": 0.99}]

[1632] Urgency: "High"

[1633] Prompt Sentence Examples

[1634] python

[1635] message = "The server is down. Urgent action required."

[1636] analysis_result = analyze_message(message)

[1637] print(analysis_result)

[1638] In this way, security risks within messenger applications can be detected early and addressed quickly.

[1639] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1640] Step 1:

[1641] The server receives messages sent and received within the messenger application. The input is the message sent by the user, and the output is the received message data. Specifically, the text content of the message is sent to the server.

[1642] Step 2:

[1643] The server analyzes the received message using a natural language processing (NLP) model. The input is the received message data, and the output is the analysis result. Specifically, it uses spaCy and Transformers to analyze the message content, tone, and urgency.

[1644] Step 3:

[1645] The server assigns labels to messages based on the analysis results. The input is the analysis results, and the output is the labeled message data. Specifically, the server assigns labels such as "Request for meeting materials" or "Urgent response required" to messages.

[1646] Step 4:

[1647] The server learns the user's sending history and semi-automatically generates sentences and images using phrases and response styles that are tailored to the individual user. The input is the user's sending history data, and the output is the generated sentences and images. Specifically, it uses a generative AI model to generate prompt sentences and provide optimized responses to the user.

[1648] Step 5:

[1649] The server analyzes the message content, tone, and urgency in real time. The input is the received message data, and the output is the analysis results. Specifically, it performs sentiment analysis and determines the urgency level to evaluate the importance of the message.

[1650] Step 6:

[1651] The server determines the security risk based on the analysis results. The input is the analysis results, and the output is the risk assessment result. Specifically, it evaluates the potential security risk from the message content and tone, and determines that the risk is high.

[1652] Step 7:

[1653] The server issues an alert based on the risk it determines. The input is the risk assessment result, and the output is an alert notification. Specifically, alerts are sent to users and administrators based on the risk level, encouraging them to take prompt action.

[1654] Example 2

[1655] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1656] Conventional messaging applications have the problem of inefficiency, requiring users to manually input messages each time. It is also difficult to automatically respond to messages that reflect the user's individual phrasing and response style, which can lead to a decline in the quality of communication. Furthermore, it is difficult to properly determine what the other party is requesting and the urgency of that request.

[1657] 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.

[1658] In this invention, the server includes a means for analyzing data in a message application, a means for classifying received messages, a means for learning from a user's sending history and semi-automatically generating text and images using phrases and response styles suited to the individual user, a means for collecting the user's past sending history and storing it in a database, a means for training a model based on the sending history data using deep learning technology, a means for receiving requests from the user and generating appropriate responses, and a means for sending the generated responses to the user. This enables automatic responses that reflect the user's individual phrases and response styles, improving the quality of communication. It also enables efficient message exchange by appropriately determining the matters the other party is requesting and their urgency.

[1659] A "messaging application" is software that allows users to send and receive text and images.

[1660] The "means for analyzing data" refers to a technique for collecting data in a message application and analyzing the content.

[1661] A "message categorization means" is a technique for categorizing received messages based on specific criteria.

[1662] "Means for learning from sending history" refers to technology that learns a user's phrasing and response style based on messages the user has sent in the past.

[1663] "Means for semi-automatic text and image generation" refers to technology in which a system automatically generates text and images based on user input and past history.

[1664] The "means for collecting sending history" is a technique for collecting messages that a user has sent in the past.

[1665] "Means of storing data in a database" refers to the technology used to organize and store collected data.

[1666] "Deep learning technology" is a technique for training models based on large amounts of data and learning patterns and features.

[1667] The "means for training a model" is a technique for training a model based on transmission history data using deep learning technology.

[1668] "Means for receiving requests" refers to the technology for receiving requests or inquiries from users.

[1669] The "means for generating an appropriate response" is a technology that automatically generates a response that matches the user's style in response to a received request.

[1670] The "means for sending a response" is a technique for sending the generated response to the user.

[1671] This invention is a system that learns a user's phrasing and reply style based on the user's past sending history, and automatically generates appropriate replies. A specific embodiment of this system will be described below.

[1672] First, the server collects text data from messages and emails sent by the user in the past. This data serves as the basis for learning the user's phrasing and response style. Common APIs such as the Gmail API and Slack API can be used to collect the data.

[1673] Next, the server saves the collected sending history in a database. This database organizes and stores the sending history for each user. A common database management system such as MySQL or PostgreSQL can be used as the database.

[1674] The server then uses a deep learning framework such as TensorFlow or PyTorch to train a deep learning model based on the historical submission data. This model is used to learn the user's specific phrasing and response style. For example, a long short-term memory (LSTM) network can be used to learn the user's phrasing and response style.

[1675] When a user sends a request, the device receives the request. For example, a user might type "Please send me the materials for the meeting" through a smartphone app. This request is then sent to the server.

[1676] The server uses a deep learning model to generate an appropriate response based on the received request. The model generates a response that matches the user's style based on information learned from past sending history. For example, it can generate a response such as "Got it. I'll send it right away."

[1677] Finally, the device displays the response generated by the server to the user. For example, a message such as "Got it. I'll send it right away" appears on the smartphone app.

[1678] As a concrete example, the following prompt sentences can be input to a generative AI model:

[1679] Prompt: In response to the request "Please send me the materials for the meeting," generate an appropriate response based on past sending history.

[1680] When this prompt sentence is input into the generative AI model, the model generates a response based on its past learning results: "Got it. I'll send it right away."

[1681] In this way, the present invention realizes automatic responses that reflect the user's individual phrasing and response style, improving the quality of communication. It also makes it possible to properly determine the matters the other party is requesting and their urgency, enabling efficient message exchange.

[1682] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1683] Step 1:

[1684] The server collects the user's past sending history. Specifically, it uses common APIs such as the Gmail API and Slack API to obtain text data of messages and emails that the user has sent in the past. The input is the raw message data obtained through the API, and the output is a list of the collected message data.

[1685] Step 2:

[1686] The server saves the collected sending history in a database. Specifically, it uses a database management system such as MySQL or PostgreSQL to organize and store the sending history for each user. The input is a list of collected message data, and the output is the organized sending history data stored in the database.

[1687] Step 3:

[1688] The server uses deep learning techniques to train a model based on the transmission history data. Specifically, it uses deep learning frameworks such as TensorFlow and PyTorch to train an LSTM (long short-term memory) network. The input is the transmission history data stored in a database, and the output is a deep learning model that has learned the user's phrasing and response style.

[1689] Step 4:

[1690] The terminal receives a request from a user. Specifically, the user inputs "Please send me the materials for the meeting" through a smartphone app. The input is the request text entered by the user, and the output is the request data to be sent to the server.

[1691] Step 5:

[1692] The server uses a deep learning model to generate an appropriate response based on the received request. Specifically, the request text is input into the model, and a response is generated based on information learned from past transmission history. The input is the user's request text and the trained deep learning model, and the output is the generated response text.

[1693] Step 6:

[1694] The device displays the response generated by the server to the user. Specifically, the smartphone app displays the message "Got it. I'll send it right away." The input is the response text sent from the server, and the output is the message displayed to the user.

[1695] (Application example 2)

[1696] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1697] Conventional messenger systems required users to create individual messages, making it difficult to respond quickly, especially in busy situations or emergencies. Furthermore, they lacked the functionality to automatically generate optimal replies for each user by utilizing the user's past comment history, which reduced the efficiency of user communication.

[1698] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from the user's sending history and semi-automatically generating sentences and stamps using phrases and reply styles that suit the user individually, and means for learning from the user's past comment history and automatically generating a comment style that the user prefers. This enables the user to automatically generate quick and appropriate replies, improving communication efficiency.

[1699] "Means for analyzing traffic within a messenger" refers to a function for analyzing the flow of messages and data sent and received within a messenger application.

[1700] "Means for labeling received messages" refers to a function that identifies the content, importance, urgency, etc. of received messages and assigns appropriate tags or labels.

[1701] "Means for learning from the user's sending history" is a function that collects the history of messages sent by the user in the past and uses this to learn the user's wording and response style.

[1702] "Means for semi-automatically generating text and stamps" refers to a function that automatically generates appropriate text and stamps, reducing the amount of input required by the user, based on the user's past sending history and learning results.

[1703] "Means for learning a user's past comment history" is a function that collects the history of comments left by a user in the past and learns the user's commenting style based on this.

[1704] "Means for automatically generating comment styles preferred by users" is a function that automatically generates comments in a format preferred by users based on the learned comment styles of users.

[1705] The system for implementing this invention has the functions of analyzing traffic within the messenger, labeling received messages, learning from the user's sending history, and semi-automatically generating sentences and stamps using phrases and reply styles that suit the individual user.It also includes a function that learns the user's past comment history and automatically generates a comment style that the user prefers.

[1706] The server first analyzes the traffic within the messenger, using software designed to analyze the flow of messages and data sent and received within the messenger application, specifically deep learning frameworks such as TensorFlow and Keras.

[1707] Next, the received messages are labeled. This labeling process identifies the content, importance, urgency, etc. of the received messages and assigns appropriate tags and labels. This makes it easier to determine the priority of the messages.

[1708] To learn from a user's sending history, the system collects messages sent by the user in the past and learns the user's phrasing and reply style based on this. Again, a deep learning model is built using TensorFlow and Keras to learn the user's characteristics.

[1709] To semi-automatically generate sentences and stamps, the trained model is used to reduce the user's input and automatically generate appropriate sentences and stamps. This process can reflect the user's past usage and style.

[1710] Furthermore, it has the ability to learn from a user's past comment history and automatically generate a comment style that the user prefers. This allows it to automatically generate comments in the format that the user prefers.

[1711] As a specific example, if a user frequently uses the comment "I understand," the system can automatically generate comments such as "I understand. I'll send it right away" or "I understand. Thank you."

[1712] An example of a prompt to input to a generative AI model is as follows:

[1713] Develop an application that learns from a user's past comment history and automatically generates a comment style that the user prefers. A user's comment history includes comments such as "Got it, I'll send it right away" and "Thank you." Use this history to build a model that generates comments that the user prefers.

[1714] In this way, users can automatically generate quick and appropriate responses, improving communication efficiency.

[1715] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1716] Step 1:

[1717] The server analyzes the traffic within the messenger. Specifically, it uses deep learning frameworks such as TensorFlow and Keras to analyze the flow of messages and data sent and received within the messenger application. The input includes all message data within the messenger, and the output extracts features from each message.

[1718] Step 2:

[1719] The server labels the received messages. It identifies the content, importance, urgency, etc. of the received messages and assigns appropriate tags and labels. The input includes the message features extracted in step 1, and the output is labeled message data.

[1720] Step 3:

[1721] The server learns from the user's sending history. It collects the user's message history and uses TensorFlow and Keras again to learn the user's phrasing and reply style based on this. The input includes the user's past message history, and the output is a model that learns the user's characteristics.

[1722] Step 4:

[1723] The server semi-automatically generates sentences and stamps. It uses the trained model to automatically generate appropriate sentences and stamps, reducing the user's input effort. The input includes the model generated in step 3 and the current message context, and the output provides the generated sentences and stamps.

[1724] Step 5:

[1725] The server learns from the user's past comment history. It collects the user's past comment history and uses TensorFlow and Keras again to learn the user's commenting style based on this. The input includes the user's past comment history, and the output is a model that learns the user's commenting style.

[1726] Step 6:

[1727] The server automatically generates a comment style that the user prefers based on the trained model. The input includes the model generated in step 5 and the current comment context, and the generated comment is provided as the output.

[1728] In this way, users can automatically generate quick and appropriate responses, improving communication efficiency.

[1729] Example 3

[1730] Next, a third embodiment of the third embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1731] In modern communication, users receive many messages and need to respond to them quickly and appropriately. However, responding manually takes time and effort, and efficiency is especially important in business environments. Furthermore, if the content and style of responses are inconsistent, the quality of communication may decline. To solve these issues, a system is needed that can semi-automatically generate appropriate responses to received messages.

[1732] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1733] In this invention, the server includes a means for receiving a message, a means for analyzing the received message, a means for selecting a reply style based on the analysis result, a means for generating sentences and non-verbal elements based on the selected reply style, and a means for transmitting the generated reply, thereby enabling a user to semi-automatically generate a quick and consistent reply to a received message.

[1734] The "means for receiving a message" is a function that allows the server to receive a message sent by a user.

[1735] "Means for analyzing received messages" refers to a function for analyzing the contents of received messages and extracting their intent and important information.

[1736] The "means for selecting a reply style based on the analysis results" is a function for selecting an appropriate reply style based on the contents of the analyzed message.

[1737] The "means for generating sentences and non-verbal elements based on the selected reply style" is a function for generating sentences, stamps, and other non-verbal elements according to the selected reply style.

[1738] The "means for sending the generated response" is a function for sending the generated response to the user.

[1739] "Non-verbal elements" are means of communication other than text, including stamps and emojis.

[1740] A "reply style" is a specific phrase or expression that is selected based on a user's past sending history and learning data.

[1741] The present invention relates to a system for semi-automatically generating a reply to a received message. A specific embodiment of this system will be described below.

[1742] System configuration

[1743] This system consists of three main components: a server, a terminal, and a user. The server receives messages, analyzes them, selects a reply style, generates a reply, and sends the reply. The terminal is a device that users use to send and receive messages. The user is the entity that communicates using the system.

[1744] Hardware and software used

[1745] The server is a computer system equipped with a high-performance processor and large memory capacity. Specifically, a server equipped with an Intel Xeon processor or an AMD EPYC processor is suitable. For software, generative AI models such as Google's BERT and OpenAI's GPT-3 are used to implement natural language processing (NLP) technology.

[1746] Data processing and calculation

[1747] The server processes and calculates data in the following procedure.

[1748] 1. Receiving a message:

[1749] The server receives the message sent by the user from the terminal.

[1750] 2. Message analysis:

[1751] The server analyzes the received message using generative AI models such as BERT and GPT-3, and extracts the message's intent and key information.

[1752] 3. Choose your response style:

[1753] The server then selects an appropriate response style based on the analysis results, which is extracted from a dataset of learned messages and their responses.

[1754] 4. Generate a response:

[1755] The server generates sentences and non-verbal elements (such as stamps) based on the selected response style, using a generative AI model.

[1756] 5. Sending a response:

[1757] The server generates a response and sends it to the user.

[1758] Specific examples

[1759] When a user sends a message saying "Please send me the meeting materials," the server processes the message as follows.

[1760] 1. Receiving a message:

[1761] The user sends a message from the terminal saying, "Please send me the meeting materials."

[1762] The server receives this message.

[1763] 2. Message analysis:

[1764] The server analyzes the received message.

[1765] The server extracts the keyword "conference materials" and recognizes that the user is requesting materials.

[1766] 3. Choose your response style:

[1767] The server selects the response style "Understood. I'll send it right away" based on past data.

[1768] 4. Generate a response:

[1769] The server generates a stamp showing a nod along with the text "Got it. I'll send it right away."

[1770] 5. Sending a response:

[1771] The server sends the generated text "Got it. I'll send it right away" and a stamp showing a nod to the user.

[1772] The user receives this response at the terminal.

[1773] Prompt Sentence Examples

[1774] The following is an example of a prompt sentence that generates an appropriate response to the message "Please send me the meeting materials."

[1775] Message from user: "Please send me the meeting materials"

[1776] Response style: "Okay, I'll send it right away."

[1777] Non-verbal: "Nodding Stamp"

[1778] By inputting this prompt sentence into the generative AI model, an appropriate response is generated. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1779] Step 1:

[1780] Receiving messages

[1781] The server receives messages sent by users. The input is a text message sent by the user from the terminal, and the output is the message data received by the server. In concrete terms, when a user sends a message from the terminal saying "Please send me the meeting materials," the server receives this message.

[1782] Step 2:

[1783] Message Parsing

[1784] The server analyzes the received message. The input is the received message data, and the output is the intent and important information of the analyzed message. Specifically, the server analyzes the message "Please send me the meeting materials," extracts the keyword "meeting materials," and recognizes that the user is requesting materials. This analysis uses generative AI models such as BERT and GPT-3.

[1785] Step 3:

[1786] Choose a response style

[1787] The server selects an appropriate response style based on the analysis results. The input is the intent and important information of the analyzed message, and the output is the selected response style. In concrete terms, the server selects a response style such as "Got it. I'll send it right away" from past data. This selection is made to provide an appropriate response to the user's request.

[1788] Step 4:

[1789] Generate a response

[1790] The server generates sentences and non-verbal elements based on the selected response style. The input is the selected response style, and the output is the generated sentences and non-verbal elements (such as stamps). Specifically, the server generates a stamp indicating a nod along with the sentence "Got it. I'll send it right away." A generative AI model is used to generate this.

[1791] Step 5:

[1792] Sending a response

[1793] The server sends the generated response to the user. The input is the generated text and non-verbal elements, and the output is the response message that the user receives on their device. Specifically, the server sends the generated text "Got it. I'll send it right away" and a stamp indicating a nod to the user. The user receives this response on their device.

[1794] (Application example 3)

[1795] Next, a description will be given of Application Example 3 of Form Example 3. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1796] Conventional messaging systems have the problem of making it difficult to respond quickly and appropriately to messages received by users. In particular, food delivery services are required to respond quickly to inquiries from users, but there are limitations to manual responses. In addition, because the content of responses is uniform, it is difficult to provide users with a sense of familiarity and individual attention.

[1797] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1798] In this invention, the server includes means for analyzing data in message communications, means for classifying received messages, means for learning from a user's sending history and semi-automatically generating sentences and non-verbal elements using phrases and reply styles suited to the individual user, means for adding appropriate non-verbal elements to the generated reply, and means for sending the generated reply to the user. This makes it possible to provide a quick and appropriate reply to a user's inquiry, and to create a sense of friendliness and individual attention.

[1799] "Messaging" is a means of exchanging information between users, including text and non-verbal elements.

[1800] "Means for analyzing data" refers to technology that analyzes the content and patterns of received messages and extracts information to generate an appropriate response.

[1801] "Means for classifying messages" refers to a technique for categorizing received messages based on their content and urgency.

[1802] "Means for learning from a user's sending history" refers to technology that learns a user's response style and phrasing based on their past message sending history.

[1803] "Means for semi-automatically generating sentences and non-verbal elements" refers to technology in which the system automatically generates appropriate sentences and non-verbal elements such as stamps based on the user's sending history and message content.

[1804] "Non-verbal elements" are visual expressions such as stickers and emojis that accompany text messages.

[1805] "Means for adding appropriate non-verbal elements to responses" refers to technology that automatically adds appropriate stamps and emoticons to generated text.

[1806] The "means for sending a response to a user" refers to a technique for sending the generated response to a user.

[1807] The system for implementing this invention analyzes data in message communications, classifies received messages, and learns from the user's sending history to semi-automatically generate sentences and non-verbal elements. A specific embodiment of this system is described below.

[1808] System Configuration

[1809] The system consists of the following major components:

[1810] 1. Data Analysis Module: Analyzes data in message traffic. This module analyzes the content and patterns of received messages and extracts information to generate appropriate responses.

[1811] 2. Message Classification Module: Classifies received messages based on their content and urgency. This module determines what the recipient is requesting and how urgent it is.

[1812] 3. Learning module: This module learns from the user's sending history and semi-automatically generates sentences and non-verbal elements using phrases and response styles that are appropriate for the individual user. This module learns the user's response style and phrases based on their past message sending history.

[1813] 4. Non-verbal element addition module: Adds appropriate non-verbal elements (stamps and emojis) to the generated responses.

[1814] 5. Reply sending module: Sends the generated reply to the user.

[1815] Hardware and software used

[1816] Hardware: Smartphone

[1817] Software: OpenAI GPT-3 API, Python

[1818] Data processing and calculation

[1819] The server first analyzes messages received from users using the data analysis module. Next, the message classification module classifies messages based on content and urgency. After that, the learning module learns appropriate reply styles based on the user's sending history and semi-automatically generates sentences and non-verbal elements. The non-verbal element addition module adds appropriate stamps and emojis to the generated replies. Finally, the reply sending module sends the generated replies to the user.

[1820] Specific examples

[1821] For example, if a user asks, "What is the status of my order?" the system will act as follows:

[1822] 1. Data analysis: Analyze the content of the message and extract the keyword "order status."

[1823] 2. Message Classification: Classify the message content as being about "order status."

[1824] 3. Response generation: Based on past sending history, the learning module generates a response such as, "A delivery person is currently picking up your order. Please wait a little longer."

[1825] 4. Adding non-verbal elements: Add a stamp of the delivery person holding food to the generated response.

[1826] 5. Send Response: Send the final response to the user.

[1827] Prompt Sentence Examples

[1828] Message from user: What is the status of my order?

[1829] Response style: Polite and friendly

[1830] Response:

[1831] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1832] Step 1:

[1833] The server inputs messages received from users into a data analysis module. The data analysis module analyzes the content and patterns of the messages and extracts information to generate an appropriate response. For example, analyzing the message "Please tell me the status of my order," the module extracts the keyword "order status."

[1834] Input: Message from user

[1835] Output: Analyzed keywords and content

[1836] Step 2:

[1837] The server inputs the analyzed keywords and content into the message classification module, which then classifies the received messages based on their content and urgency. For example, based on the keyword "order status," the message may be classified into the "order status check" category.

[1838] Input: Analyzed keywords and content

[1839] Output: Classified message categories

[1840] Step 3:

[1841] The server inputs the classified message categories into a learning module. Based on the user's sending history, the learning module semi-automatically generates sentences using phrases and response styles that are appropriate for the individual user. For example, based on past sending history, it generates a response such as, "A delivery person is currently picking up your order. Please wait a little longer."

[1842] Input: Message categories, user sending history

[1843] Output: Generated response text

[1844] Step 4:

[1845] The server inputs the generated response sentence into the non-verbal element addition module, which adds appropriate non-verbal elements (stamps and emojis) to the generated response. For example, in response to the response "A delivery person is currently picking up your order. Please wait a little longer," a stamp of a delivery person holding food is added.

[1846] Input: Generated response text

[1847] Output: Response with added non-verbal elements

[1848] Step 5:

[1849] The server inputs the response with the added non-verbal elements into the response sending module. The response sending module sends the generated response to the user. For example, it sends the final response to the user, such as "The delivery person is currently picking up your order. Please wait a little longer. (Stamp)."

[1850] Input: Response with added non-verbal elements

[1851] Output: Reply sent to the user

[1852] 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.

[1853] "Example 1"

[1854] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system extracts emotions from a user's message and adjusts a response based on the emotions. Specifically, if a user sends a message such as "I'm really tired today," the emotion engine extracts emotions such as "fatigue" or "stress" from the message. Then, based on the emotions, the system generates a response such as "That must have been hard. Please take a good rest." In this way, the system incorporating an emotion engine can provide a more appropriate response that corresponds to the user's emotions.

[1855] "Example 2"

[1856] The emotion engine can also generate appropriate stamps based on the user's emotions. For example, if a user sends a message such as "Today was fun!", the emotion engine can extract emotions such as "joy" or "satisfaction" from the message. Based on this emotion, the system can generate a stamp of a smiling face, for example. In this way, a system that incorporates an emotion engine can provide more appropriate stamps that correspond to the user's emotions.

[1857] "Example 3"

[1858] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system extracts emotions from a user's message and tailors a response based on the emotions. Specifically, if a user sends a message such as "I'm really tired today," the emotion engine extracts emotions such as "fatigue" and "stress" from the message. Based on this emotion, the system then responds with a response such as "That was tough, please take a good rest."

[1859] In this way, a system that combines an emotion engine can provide more appropriate responses that correspond to the user's emotions.

[1860] The processing flow of each embodiment will be described below.

[1861] "Example 1"

[1862] Step 1: A user opens a messenger application and sends a message, for example, "I'm really tired today."

[1863] Step 2: The system analyzes the sent message in real time using natural language processing (NLP) techniques.

[1864] Step 3: The emotion engine extracts emotions such as "fatigue" or "stress" from the analyzed messages.

[1865] Step 4: The system generates a response based on the extracted sentiment, for example, "That must have been tough, take a rest."

[1866] "Example 2"

[1867] Step 1: A user opens a messenger application and sends a message, for example, "Today was fun!"

[1868] Step 2: The system analyzes the sent message in real time using natural language processing (NLP) techniques.

[1869] Step 3: The emotion engine extracts emotions such as "joy" or "satisfaction" from the analyzed messages.

[1870] Step 4: The system generates a stamp based on the extracted emotion, such as a smiley face stamp.

[1871] Example 1

[1872] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1873] With conventional messaging applications, it was difficult to understand the content, tone, and urgency of messages sent and received by users in real time, making it impossible to generate appropriate replies. It was also difficult to provide replies based on the user's emotions. This limited the user's communication experience and hindered efficient information transmission.

[1874] 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.

[1875] In this invention, the server includes means for analyzing traffic within the messenger, means for labeling received messages, means for learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles suited to the individual user, means for analyzing the content of messages using natural language processing technology, means for extracting emotions from messages using an emotion engine, means for generating appropriate replies based on the emotions extracted using a generative AI model, and means for sending the generated replies to the user, thereby making it possible to analyze user messages in real time and provide appropriate replies.

[1876] "Means for analyzing traffic within messengers" refers to technology for monitoring and analyzing message data sent and received through messenger applications in real time.

[1877] "Means for labeling received messages" refers to a technology that analyzes the content of a received message and assigns an appropriate label (tag) based on that content.

[1878] "A means of learning from a user's sending history and semi-automatically generating text and images using phrases and reply styles that suit each individual user" is a technology that analyzes a user's past message sending history, learns the phrases and reply styles that are best suited to that user, and semi-automatically generates text and images.

[1879] "Means for analyzing the content of a message using natural language processing technology" refers to a technology that uses natural language processing (NLP) technology to tokenize the content of a message and extract meaning.

[1880] The "means for extracting emotions from a message using an emotion engine" is a technology that uses an emotion engine to identify and extract a user's emotions from the tone and content of a message.

[1881] "Means for generating an appropriate response based on emotions extracted using a generative AI model" is a technology that uses a generative AI model to generate an appropriate response based on emotions extracted.

[1882] The "means for sending the generated reply to the user" is a technique for sending the generated reply to the user through a messenger application.

[1883] The present invention is a system that is incorporated into a messenger application and analyzes messages sent and received by users in real time and generates appropriate responses. The system includes elements of a server, a terminal, and a user.

[1884] Hardware and software used

[1885] The server uses the following hardware and software:

[1886] Hardware: A server machine with high-performance processors, memory, and storage

[1887] Software: Natural language processing libraries (e.g., spaCy, NLTK), emotion engines (e.g., IBM Watson Tone Analyzer), generative AI models (e.g., OpenAI GPT-3)

[1888] Data processing and calculation

[1889] The server receives the message sent by the user and performs the following actions:

[1890] 1. Message content analysis:

[1891] The server uses natural language processing libraries to tokenize the message and extract meaning.

[1892] For example, if a user sends a message saying, "Please send me the materials for tomorrow's meeting," the server will label this message as a "request for meeting materials" and set the urgency to "high."

[1893] 2. Emotion extraction:

[1894] The server uses an emotion engine to extract emotions from the messages.

[1895] For example, if a user sends a message such as "I'm really tired today," the server will extract emotions such as "fatigue" and "stress" from the message.

[1896] 3. Generate a response:

[1897] The server uses a generative AI model to generate appropriate responses based on the extracted emotions.

[1898] For example, in response to a message such as "I'm really tired today," the server generates a response such as "That must have been tough, please get some rest."

[1899] 4. Sending a response:

[1900] The server then sends the generated reply to the user, who then displays the reply to the user through the messenger application.

[1901] Specific examples

[1902] As a concrete example, consider the following scenario:

[1903] scenario:

[1904] A user sends a message through a messenger application saying, "Please send me the materials for tomorrow's meeting."

[1905] Process flow:

[1906] 1. The server receives the message and uses a natural language processing library to label it as a "request for meeting materials" and set the urgency to "high."

[1907] 2. The server uses the emotion engine to extract emotions, but since this message does not contain any specific emotion, emotion extraction is skipped.

[1908] 3. The server uses a generative AI model to generate an appropriate response, in which case it determines that no response is necessary.

[1909] 4. The server generates a response and sends it to the user, who then displays it on their device.

[1910] Prompt Sentence Examples

[1911] Below are some example prompts to input to a generative AI model:

[1912] A user sends a message saying, "I'm really tired today." Extract the sentiment from this message and generate an appropriate response.

[1913] In this way, the system can analyze the user's message in real time and provide an appropriate response.

[1914] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1915] Step 1:

[1916] Receiving messages

[1917] The server receives messages sent by users through the messenger application.

[1918] Input: The message sent by the user (e.g., "Please send me the materials for tomorrow's meeting")

[1919] Specific operation: The server receives messages through the API endpoint of the messenger application and stores them in a database.

[1920] Output: Received message data

[1921] Step 2:

[1922] Message content analysis

[1923] The server analyzes the received message using a natural language processing (NLP) library.

[1924] Input: Received message data

[1925] What happens: The server uses an NLP library (e.g., spaCy) to tokenize the message and extract meaning. For example, it parses the message "Please send me the materials for tomorrow's meeting" and labels it as "Request for meeting materials."

[1926] Output: Labeled message data (e.g., "Request for meeting materials")

[1927] Step 3:

[1928] Emotion extraction

[1929] The server uses an emotion engine to extract emotions from the messages.

[1930] Input: Labeled message data

[1931] Specific operation: The server analyzes the message using an emotion engine (e.g., IBM Watson Tone Analyzer) and extracts emotions. For example, it extracts emotions such as "fatigue" and "stress" from the message "I'm really tired today."

[1932] Output: Extracted emotion data (e.g., "fatigue," "stress")

[1933] Step 4:

[1934] Generate a response

[1935] The server uses a generative AI model to generate appropriate responses based on the extracted emotions.

[1936] Input: Extracted emotion data

[1937] Specific operation: The server inputs a prompt sentence into the generative AI model (e.g., OpenAI GPT-3). For example, in response to the message "I'm really tired today," the server inputs the prompt sentence "The user feels tired. Please generate an appropriate response." The generative AI model generates the response "That must have been hard. Please take a good rest."

[1938] Output: Generated response data (e.g. "That was tough, please take a rest").

[1939] Step 5:

[1940] Sending a response

[1941] The server generates a response and sends it to the user.

[1942] Input: Generated response data

[1943] Specific operation: The server sends the generated reply to the user through the messenger application's API.

[1944] Output: A response message that is displayed on the user's device (e.g., "That was tough, please get some rest.")

[1945] (Application example 1)

[1946] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referre...

Claims

1. means for retrieving received messages within a messaging application; A means for generating a candidate reply to the received message using a prompt sentence that instructs the user to generate a reply to the received message by imitating a message sent by the user based on a history of messages sent by the user in a message application, and a generative AI; means for presenting the generated reply candidates to the user and accepting editing of the candidates; means for transmitting the edited result as a message to the user; A system including:

2. means for recognizing an emotion of a user based on a message sent by the user; The system described in claim 1, wherein the generating means generates candidate replies to the received message based on the recognized user's emotions and the message history, using a prompt sentence that instructs the user to generate a reply to the received message, and the generative AI.

3. The system of claim 1, wherein the generating means generates candidate replies to the received message using a prompt sentence that instructs the user to generate a reply to the received message by imitating text and stamps contained in messages sent by the user based on a history of messages sent by the user in a message application, and the generation AI.

Images