System

Abstract

A system is provided.SOLUTION: The system includes a means for receiving text data input by a user, an analysis means using a generation model for analyzing the received text data and detecting inappropriate contents or contents corresponding to harassment, and a proposal means for proposing an alert message or an appropriate expression to the user on the basis of an analysis result.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] In recent years, with the spread of text-based communication tools such as email, chat, and social networking sites, inappropriate content and harassment have increased. Such behavior can cause significant psychological damage to victims and, in some cases, can lead to serious consequences. While existing mental health care services focus on victims, there are limited means to prevent the behavior of perpetrators in advance. Therefore, there is a need to prevent harassment by detecting inappropriate messages in advance and prompting users to correct them. [Means for solving the problem]

[0005] To solve the above problems, the present invention provides the following means. A system is provided that includes: means for receiving text data entered by a user; analysis means using a generative model that analyzes the received text data to detect inappropriate content or content that constitutes harassment; and suggestion means for suggesting alert messages and appropriate expressions to the user based on the analysis results. Furthermore, by providing means for analyzing the transmission time of the received text data and suggesting an appropriate transmission time, appropriate communication according to the time of day can be promoted. Furthermore, by providing means for scoring the analysis results of the text data analyzed using the generative model and issuing alerts and suggestions based on the scores, more accurate alerts and suggestions can be realized.

[0006] "User" means an individual or organization that uses the system to input and transmit text data.

[0007] "Text data" refers to character string information that a user inputs and sends via email, chat, SNS, etc.

[0008] "Reception" refers to the process in which text data is sent from the user terminal to the server and the server receives the data.

[0009] A "generative model" is an algorithm or system that uses natural language processing techniques to analyze and evaluate text data.

[0010] The "analysis means" is a function that analyzes received text data using a generative model to detect inappropriate content or harassment.

[0011] An "alert message" is a warning or caution message that is presented to the user based on the analysis results.

[0012] The "suggestion means" is a function that suggests appropriate expressions and transmission methods to the user based on the results of the analysis means.

[0013] The "transmission time" is the time when the user executes or plans to transmit the text data.

[0014] "Scoring" is the process of quantifying the degree of inappropriateness or harassment based on the analysis results.

[0015] The "appropriate transmission time" is a transmission time that is judged to be less likely to cause discomfort to the other party in communication. [Brief explanation of the drawings]

[0016] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12]FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0017] 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.

[0018] First, the terms used in the following description will be explained.

[0019] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0020] 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.

[0021] 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.

[0022] 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.

[0023] 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."

[0024] [First embodiment]

[0025] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0026] 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.

[0027] 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).

[0028] 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.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0033] 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.

[0034] 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.

[0035] 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.

[0036] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0037] This system analyzes text-based communications, detects inappropriate content and harassment in advance, and suggests appropriate expressions and sending times to users. The specific operation and functions of this system are explained below.

[0038] System Configuration

[0039] 1. Device:

[0040] A device that users use to enter text-based messages such as email, chat, and social networking sites. The terminal includes an interface for receiving user input and sending it to a server.

[0041] 2. Server:

[0042] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[0043] 3. Generative Model:

[0044] It is an algorithm that uses natural language processing technology to analyze text data, understand the context, and detect inappropriate expressions and harassment.

[0045] Program processing

[0046] Enter and send a message

[0047] User:

[0048] A user inputs a text message into an email, chat, social networking site, etc. For example, the user inputs a message such as "Please send the meeting materials immediately."

[0049] Device:

[0050] When the user presses the send button after entering a message, the device sends the text data and related meta-information (user ID, sending time, etc.) to the server.

[0051] Message Parsing

[0052] server:

[0053] The server preprocesses the received text data and prepares it for the generative model, including tokenizing and filtering the message.

[0054] Generative model:

[0055] Generative models analyze text data to understand context and detect inappropriate content or harassment. The results of this analysis are scored according to importance and identify specific issues.

[0056] Alerts and Suggestions

[0057] server:

[0058] Based on the analysis results of the generative model, appropriate alert messages and suggestions are generated. For example, if the analysis results receive a high score, a specific suggestion such as "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'" is generated.

[0059] Device:

[0060] Alert messages and suggestions sent by the server are displayed to the user in a popup format, allowing the user to review the message before sending it.

[0061] User Selection

[0062] User:

[0063] The user can review the alerts and suggestions and choose to modify the message or send it as is. For example, they can modify the message to "Could you please send the meeting materials as soon as possible?" and resend it.

[0064] Resend and final confirmation of the revised message

[0065] Device:

[0066] The corrected message is sent back to the server for final confirmation.

[0067] server:

[0068] Re-analyze the correction message and finally make sure there are no problems.

[0069] Sending a message

[0070] server:

[0071] After a final check, the resolved message is sent to the actual recipient.

[0072] Specific examples

[0073] As a concrete example, consider the case of an "commanding message sent late at night."

[0074] User:

[0075] At 2 a.m., type the message "This is an emergency and I need your help immediately!" and press the send button.

[0076] Device:

[0077] Sends text and meta information to the server.

[0078] server:

[0079] Receives messages, preprocesses them, and passes them to the generative model.

[0080] Generative model:

[0081] The message is analyzed to detect that it contains the imperative phrase "This is urgent, please respond immediately!" and that it was sent late at night.

[0082] server:

[0083] Generate an alert such as, "This message was sent late at night and contains imperative language. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'"

[0084] Device:

[0085] The alert is displayed to the user in a popup.

[0086] User:

[0087] Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[0088] Device:

[0089] The corrected message is sent back to the server for final confirmation.

[0090] server:

[0091] Re-analyze and check that there are no problems.

[0092] server:

[0093] The next morning, a clean message will be automatically sent to the recipient.

[0094] This invention aims to prevent harassment and inappropriate language in advance, and provides a concrete means for improving the quality of communication.

[0095] The processing flow will be explained below.

[0096] Step 1:

[0097] User: The user types a text message to send via email, chat, or social media.

[0098] Step 2:

[0099] User: Checks the message and clicks the send button.

[0100] Step 3:

[0101] Terminal: Sends text data and related meta-information (user ID, sending time, etc.) to the server.

[0102] Step 4:

[0103] Server: Preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message and spell checking.

[0104] Step 5:

[0105] Generative model: Analyzes tokenized text data and detects inappropriate content and harassment while understanding the context. The analysis results are scored and quantified to indicate their importance.

[0106] Step 6:

[0107] Server: Generates alert messages and suggestion messages based on the analysis results of the generative model. For example, if the score is high, it generates suggestions including specific corrections.

[0108] Step 7:

[0109] Server: Sends alert and suggestion messages to the device.

[0110] Step 8:

[0111] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[0112] Step 9:

[0113] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[0114] Step 10:

[0115] User: If you have corrected the message, click the resend button.

[0116] Step 11:

[0117] Terminal: Send the corrected text data to the server again.

[0118] Step 12:

[0119] Server: Re-parse the corrected text data and verify that the problem is resolved.

[0120] Step 13:

[0121] Server: After final confirmation, the text data is sent to the actual recipient without any problems.

[0122] Step 14:

[0123] Server: Logs the sent message and its analysis results and stores them for future reference.

[0124] Example 1

[0125] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0126] A problem with conventional text-based communication is that users can unknowingly send messages containing inappropriate content or harassment. Furthermore, sending messages at inappropriate times, such as late at night or early in the morning, can cause discomfort to the recipient. A system that addresses these issues and promotes healthy communication was needed.

[0127] 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.

[0128] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data and passing it to a generative AI model, means for analyzing the received text data using the generative AI model and detecting inappropriate content or content that constitutes harassment, means for suggesting an alert message or appropriate expression to the user based on the detected analysis results, means for displaying the alert message or suggestion to the user, means for resending and reanalyzing a message corrected by the user, and means for sending the corrected message to the actual recipient after final confirmation. This prevents the user from unknowingly sending a message containing inappropriate content and enables the message to be sent at an appropriate time.

[0129] "User" means a person who uses the system to enter and send text messages.

[0130] "Means for receiving" is a general term for hardware and software for transmitting text data entered by a user to a server and receiving it.

[0131] "Preprocessing" refers to a series of steps performed on received text data before it is passed to a generative AI model, including tokenization, removal of unnecessary characters, and grammatical analysis.

[0132] A "generative AI model" is an algorithm that uses natural language processing technology to analyze text data, understand context, and detect inappropriate content or harassment.

[0133] "Means for analysis" refers collectively to software and hardware used to analyze received text data using a generative AI model and identify inappropriate content or expressions that constitute harassment.

[0134] An "alert message" is a message that warns the user that the content contains inappropriate content or expressions based on the analysis results of the generative AI model.

[0135] "Appropriate expressions" are expressions that suggest more appropriate and desirable wording to the user based on the problems detected by the generative AI model.

[0136] "Proposed means" is a general term for software and hardware that generates alert messages and appropriate expressions based on the analysis results and presents them to the user.

[0137] "Means for retransmission" is a general term for software and hardware that has the function of sending a message corrected by the user back to the server for reanalysis.

[0138] "Means for sending after final confirmation" refers to software and hardware that has the functionality to reanalyze the modified message, confirm that there are no problems, and then send the message to the actual recipient.

[0139] This invention relates to a system that analyzes text data entered by a user, detects inappropriate content or harassment in advance, and suggests appropriate expressions and sending times. This system improves the appropriateness of messages by performing various analyses during the process of a user entering a text message via email, chat, SNS, etc. and sending it.

[0140] System Configuration and Hardware

[0141] 1. Device:

[0142] A device used by users to input text-based messages such as emails, chats, and social media. Typical devices include PCs, smartphones, and tablets.

[0143] The device includes an interface for receiving user input and sending it to a server, often using a front-end framework such as React or Vue.js for the user interface.

[0144] 2. Server:

[0145] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[0146] Natural language processing libraries used include the Python-based NLTK and spaCy. Preprocessing involves tokenizing and filtering messages.

[0147] Transformer-based models (e.g., GPT-3, BERT) are used as generative AI models.

[0148] 3. Generative AI Model:

[0149] This is an algorithm that uses natural language processing technology to analyze text data, understand the context, and then detect inappropriate language and harassment. The generative AI model is capable of highly accurate analysis and can accurately grasp the context.

[0150] System Operation

[0151] 1. Enter and send a message:

[0152] User: The user types a text message into email, chat, SNS, etc. For example, they type a message such as "Please send the meeting materials right away."

[0153] Terminal: When the user presses the send button after entering a message, the terminal sends the text data entered by the user and related meta information (user ID, sending time, etc.) to the server. The HTTP protocol is used for sending.

[0154] 2. Message analysis:

[0155] Server: The server preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message, removing unnecessary characters and tags, and simple grammatical analysis.

[0156] Generative models: Generative models analyze text data and understand context to detect inappropriate content and harassment. Specifically, transformer-based models understand context and identify things like commanding tone, unnecessary emphasis, and discriminatory language.

[0157] 3. Alert generation and suggestions:

[0158] Server: Generates appropriate alert messages and suggestions based on the analysis results obtained from the generative model. For example, it generates specific suggestions such as, "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'"

[0159] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format.

[0160] 4. User Choice:

[0161] User: The user can review the message based on the pop-up alerts and suggestions, and choose to apply the suggested changes or send it as is. For example, they can change it to "Could you please send the meeting materials as soon as possible?" and resend it.

[0162] 5. Resend and final confirmation of the corrected message:

[0163] Terminal: If the user presses the button to resend the corrected message, the terminal will resend the corrected text data to the server.

[0164] Server: Re-analyze the corrected message and finally check that there are no problems.

[0165] 6. Sending a message:

[0166] Server: Once all is well, the server sends the message to the actual recipient, typically using the SMTP protocol or an API to connect to an email or messaging system.

[0167] Specific examples

[0168] As a concrete example, consider an imperative message sent late at night.

[0169] User: At 2 a.m., type the message "This is urgent, please respond immediately!" and press the send button.

[0170] Terminal: The text data entered by the user and meta information (user ID, time of submission, etc.) are sent to the server via the HTTP protocol.

[0171] Server: Receives messages, tokenizes and filters them using NLTK, and performs parsing with spaCy.

[0172] Generative model: We use GPT-3 to analyze the message and detect the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night.

[0173] Server: Generate an alert like, "This message was sent late at night and contains an imperative. We recommend sending it the next morning and amending it to 'Please respond as soon as possible tomorrow morning.'"

[0174] Terminal: Display the alert to the user as a popup.

[0175] User: Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[0176] Device: Send the corrected message data and sending timing to the server again.

[0177] Server: Re-analyze the data and make sure there are no problems.

[0178] Server: The next morning, the corrected message is automatically sent to the recipient.

[0179] Example prompt sentence:

[0180] "A user types and attempts to send a message at 2 AM saying 'This is urgent, please respond immediately!' Analyze the appropriateness of this message and generate a prompt suggesting a correction."

[0181] The present invention is a system that prevents a user from unknowingly sending a message containing inappropriate content and supports sending messages at an appropriate time.

[0182] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0183] Step 1: User types and sends a text message

[0184] User: Enter a text message via email, chat, SNS, etc. For example, enter the message "Please send the meeting materials immediately."

[0185] Input: The text message entered by the user.

[0186] Output: Text data and meta information (user ID, sending time, etc.) ready to be sent to the device

[0187] Step 2: The device sends the text data to the server

[0188] Terminal: After receiving user input, it detects when the send button is pressed and sends the text data and meta information to the server using the HTTP protocol.

[0189] Input: Text data and meta information entered by the user

[0190] Output: Text data and meta information sent to the server

[0191] Step 3: The server preprocesses the text data

[0192] Server: Preprocesses the received text data. This preprocessing includes tokenizing the text, removing unnecessary characters and tags, and simple grammar analysis. This is done using a Python natural language processing library (NLTK, spaCy, etc.).

[0193] Input: Text data and meta information

[0194] Output: Preprocessed text data

[0195] Step 4: The server uses the generative model to analyze the text data

[0196] Server: Preprocessed text data is passed to a generative AI model (e.g., GPT-3, BERT) that understands the context and detects inappropriate content and harassment.

[0197] Input: Preprocessed text data

[0198] Output: Detection results and a score indicating their importance

[0199] Step 5: The server generates an alert message or suggestion

[0200] Server: Generates appropriate alert messages and suggestions based on the analysis results of the generative model. For example, it generates a message such as, "This message contains imperative language. We recommend revising it to 'Could you please send the meeting materials as soon as possible?'"

[0201] Input: Detection results and scoring

[0202] Output: Alert message and suggestions

[0203] Step 6: The device displays an alert message or suggestion to the user.

[0204] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format using a front-end framework (e.g., React, Vue.js).

[0205] Input: Alert message and suggestions

[0206] Output: Alerts and suggestions displayed in the user interface

[0207] Step 7: User reviews the suggested changes and edits the message

[0208] User: Based on the alerts and suggestions displayed, revise the message, for example, to "Could you please send the meeting materials as soon as possible?"

[0209] Input: The displayed alert message and suggestion

[0210] Output: The modified text message

[0211] Step 8: The device resends the corrected message

[0212] Terminal: If the user clicks the resend button after correcting the message, the corrected message data is sent to the server again.

[0213] Input: Modified text message

[0214] Output: The correction message sent to the server.

[0215] Step 9: The server reparses the modified message and performs a final check

[0216] Server: Re-parses the modified message to ensure there are no problems. This re-parse is done using the generative model, just like the initial parse.

[0217] Input: Modified text message

[0218] Output: Final confirmation result of the message after correction

[0219] Step 10: Server sends message after final confirmation

[0220] Server: After verifying that there are no problems, it sends the corrected message to the actual recipient, using the SMTP protocol or API to connect to email or messaging systems.

[0221] Input: Correction message that has been finalized

[0222] Output: The message sent to the actual recipient

[0223] (Application example 1)

[0224] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0225] The present invention aims to realize safe and secure communication by effectively detecting inappropriate content and harassment in text-based communication and suggesting appropriate expressions and sending times to users. It also provides a system that reanalyzes messages corrected based on the suggestions, ultimately transmitting problem-free text data.

[0226] 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.

[0227] In this invention, the server includes means for receiving text data entered by a user, analysis means using a generative model to analyze the received text data and detect inappropriate content or content that constitutes harassment, means for suggesting alert messages or appropriate expressions to the user based on the analysis results, means for reanalyzing the text data corrected based on the suggested content and reconfirming its appropriateness, and means for transmitting the corrected text data to the final recipient. This enables safe and secure communication by automatically checking the appropriateness of messages entered by users and suggesting corrections as necessary.

[0228] "User" means the entity that uses the system and inputs and transmits text.

[0229] "Text data" refers to character information entered and transmitted by a user.

[0230] The "receiving means" is a device or program that has the function of incorporating text data sent from a user into the system.

[0231] "Analysis means" refers to a device or program for analyzing text data to detect inappropriate expressions or harassment.

[0232] A "generative model" is an algorithm or program that uses natural language processing technology to analyze the content of text data.

[0233] The "suggestion means" is a device or program that has the function of providing the user with specific alerts and examples of corrections based on the analysis results.

[0234] The "reanalysis means" is a device or program having the function of reanalyzing text data that has been corrected by the user based on the suggestion and confirming its appropriateness.

[0235] The "transmission means" is a device or program having a function for transmitting the final confirmed text data to the recipient.

[0236] An "alert message" is a message that warns the user when the analysis means detects inappropriate content or harassment.

[0237] "Appropriate language" refers to language that is respectful to the recipient and does not contain inappropriate or harassing content.

[0238] The system for implementing this invention consists of a server connected to the Internet and a user terminal. The user inputs and sends text messages using a terminal such as a smartphone or PC. The terminal includes an interface for receiving the input text data and transmitting it to the server.

[0239] The server performs various processes to analyze the received text data. Specifically, the server preprocesses the received message and passes it to a generative model for contextual analysis. The generative model can be the Transformers library from Hugging Face. Specifically, a natural language processing model (e.g., BERT or GPT-3) is used.

[0240] The generative model analyzes text data to detect inappropriate content and harassment. At this stage, the analysis results are scored and the problem areas are identified based on the score. For example, a message such as "This is an emergency and please act immediately!" is deemed to be imperative and inappropriate.

[0241] Based on the analysis results, the server makes appropriate suggestions to the user. The suggestions are specific and are displayed in the form of, for example, "This message contains imperative language. We recommend that you correct it as follows: 'Please respond as soon as possible.'"

[0242] The user reviews the suggestions and modifies the message. The modified message is then sent back to the server for re-analysis. This re-analysis ultimately verifies the message is clean. After verification, the modified message is finally sent to the recipient.

[0243] As a concrete example, consider the case of an imperative message sent late at night. A user types the message "This is urgent, please respond immediately!" at 2 a.m. and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message and detects that it contains the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night. The server generates an alert stating, "This message was sent late at night and contains an imperative expression. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'" The device displays the alert to the user as a pop-up. The user edits the message to "Please respond as soon as possible tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message again to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The next morning, the server automatically sends the message without any problems to the recipient.

[0244] An example of a prompt is as follows:

[0245] "Analyze the following sentence and determine if it is an imperative: 'Please respond immediately!'"

[0246] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0247] Step 1:

[0248] The user enters a text message into the terminal and presses the send button. At this time, the terminal prepares the entered text data and meta information (user ID, sending time, etc.). The terminal sends the text data and meta information to the server. The input is the "text message" and "meta information," and the output is the "text data and meta information sent to the server."

[0249] Step 2:

[0250] The server preprocesses the text data it receives. Specifically, it tokenizes and filters the text data. The input is the received text data and meta information, and the output is the preprocessed text data.

[0251] Step 3:

[0252] The server passes the preprocessed text data to a generative model for contextual analysis. Specifically, it uses a natural language processing model based on the Hugging Face Transformers library (e.g., BERT or GPT-3) to analyze inappropriate language and harassment in the text. The input is the preprocessed text data, and the output is the analysis results (scores for inappropriate language and harassment).

[0253] Step 4:

[0254] The server detects inappropriate content or harassment based on the analysis results of the generative model and generates an alert message or correction suggestions. The input is the "analysis results (scores for inappropriate expressions and harassment)" and the output is the "alert message or correction suggestions." Specifically, the message generated is "This message contains imperative language. We recommend correcting it as follows: 'Please respond as soon as possible.'"

[0255] Step 5:

[0256] The server sends an alert message or correction suggestion to the terminal, which then displays it to the user in a popup format. The input is the "alert message or correction suggestion" and the output is the "displayed alert message or correction suggestion." The user then corrects the message based on the suggested suggestions.

[0257] Step 6:

[0258] The user resends the corrected text message, and the terminal sends it to the server. The input is the "corrected text message" and the output is the "corrected text data sent to the server."

[0259] Step 7:

[0260] The server re-analyzes the corrected text message and performs a final check. The input is the "corrected text data" and the output is the "final analysis result." It checks to see if there are any problems.

[0261] Step 8:

[0262] The server sends the verified modified text message to the final recipient. The input is the "verified modified text message" and the output is the "text message sent to the final recipient".

[0263] 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.

[0264] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining this system with an emotion engine, it is possible to realize more accurate alerts and suggestions based on the user's emotions.

[0265] System Configuration

[0266] 1. Device:

[0267] A device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text.

[0268] 2. Server:

[0269] It receives text data sent from the device and analyzes it using a generative model and emotion engine.

[0270] 3. Generative Model:

[0271] This algorithm uses natural language processing technology to analyze text data, understand context and meaning, and detect inappropriate content and harassment.

[0272] 4. Emotion Engine:

[0273] This is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, and neutral as analysis results.

[0274] Program processing

[0275] Enter and send a message

[0276] User:

[0277] The user types the text message "You're really useless" into a chat app.

[0278] Device:

[0279] When the send button is pressed, the terminal sends the text data and related meta information (user ID, sending time, etc.) to the server.

[0280] Message Parsing

[0281] server:

[0282] Preprocessing the received text data before passing it to the generative model includes tokenizing the message and escaping any necessary characters.

[0283] Generative model:

[0284] The system analyzes tokenized text data to detect negative content such as "unusable," and then scores it to quantify its importance.

[0285] emotion recognition

[0286] Emotion Engine:

[0287] It analyzes the user's emotions from text data and detects negative emotions, such as "anger."

[0288] Alerts and Suggestions

[0289] server:

[0290] The inappropriate content detection results from the generative model and the analysis results from the emotion engine are integrated to generate alert messages and suggestions.

[0291] For example, it generates a suggestion such as, "This message contains negative emotions and contains inappropriate language. Please try revising it to 'It would be helpful if you could provide specific feedback.'"

[0292] Device:

[0293] Displays generated alert and suggestion messages to the user in a popup format, allowing the user to review the message before sending.

[0294] User Selection

[0295] User:

[0296] Review the alert or suggestion and choose to modify the text message to "Specific feedback would be helpful" or send it as is.

[0297] Resend and final confirmation of the revised message

[0298] User:

[0299] If you have corrected the message, click the resend button.

[0300] Device:

[0301] The corrected text data is sent to the server again.

[0302] server:

[0303] Re-analyze the corrected text data to verify that the problem has been resolved.

[0304] Sending a message

[0305] server:

[0306] After final confirmation, the text data is sent to the actual recipient without any problems.

[0307] Message storage

[0308] server:

[0309] The sent message and its analysis results are recorded in a log and saved for future reference.

[0310] Specific examples

[0311] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[0312] User:

[0313] "At 2 a.m., I type a message saying, 'Deal with this right now,' and press the send button."

[0314] Device:

[0315] Sends text and meta information to the server.

[0316] server:

[0317] Receives messages, preprocesses them, and passes them to the generative model.

[0318] Generative model:

[0319] The message is analyzed to detect whether it is an "imperative expression" and is scored.

[0320] Emotion Engine:

[0321] Detects the emotion "anger" and sets a negative emotion score.

[0322] server:

[0323] Generate an alert saying, "This message was sent late at night and contains commanding language and angry sentiment. We recommend sending it the next morning and amending it to 'Please respond by tomorrow morning.'"

[0324] Device:

[0325] The alert is displayed to the user in a popup.

[0326] User:

[0327] Change the message to "Please respond by tomorrow morning" and set it to be sent the next morning.

[0328] Device:

[0329] The corrected message is sent back to the server for final confirmation.

[0330] server:

[0331] Re-analyze and check that there are no problems.

[0332] server:

[0333] The next morning, a clean message is automatically sent to the recipient. This invention provides a technological means to prevent harassment and promote better communication.

[0334] The processing flow will be explained below.

[0335] Step 1:

[0336] User: The user types a text message to send via email, chat, or social media.

[0337] Step 2:

[0338] User: Checks the message and clicks the send button.

[0339] Step 3:

[0340] Device: Sends the text message and related meta information (user ID, time sent, etc.) to the server.

[0341] Step 4:

[0342] Server: Preprocesses the received text messages and prepares them for the generative model. Preprocessing includes tokenizing the messages and normalizing the input data.

[0343] Step 5:

[0344] Generative Model: A generative model analyzes tokenized text data, understands the context, and detects inappropriate content and harassment. It then assigns a score to the inappropriate content.

[0345] Step 6:

[0346] Emotion Engine: The emotion engine analyzes the user's emotions from text data. For example, it detects emotional states such as "anger," "sadness," and "joy" and assigns a score to each.

[0347] Step 7:

[0348] Server: Integrates the analysis results of the generative model and the emotion engine to generate alert messages and suggestion messages. For example, if negative expressions and the emotion "anger" have high scores, the server will suggest corrections.

[0349] Step 8:

[0350] Server: Sends generated alert and suggestion messages to the device.

[0351] Step 9:

[0352] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[0353] Step 10:

[0354] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[0355] Step 11:

[0356] User: If you have corrected the message, click the resend button.

[0357] Step 12:

[0358] On your device: Send the modified text message back to the server.

[0359] Step 13:

[0360] Server: Re-parse the modified text message to verify that the issue has been resolved.

[0361] Step 14:

[0362] Server: After final verification, the text message is sent to the intended recipient.

[0363] Step 15:

[0364] Server: Logs the sent message and its analysis results and stores them for future reference.

[0365] Example 2

[0366] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0367] Inappropriate language and harassment are commonplace in modern communication tools, placing a heavy burden on users. Effective technology to prevent these issues has yet to be fully established. Furthermore, few systems can provide alerts or suggestions based on users' emotions, which can lead to unintended misunderstandings and conflicts. To address these issues, a system is needed that uses more advanced natural language processing technology to analyze users' message content and emotions and provide appropriate alerts and suggestions.

[0368] 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.

[0369] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data, means for analyzing the preprocessed text data using a generative model to detect inappropriate content or content that constitutes harassment, means for analyzing the received text data using an emotion engine that recognizes the user's emotions, means for suggesting an alert message or appropriate expression to the user based on the analysis results of the generative model and the emotion engine, means for sending the text data with the problem resolved to the actual recipient, and means for recording the sent message and the analysis results in a log. This makes it possible to prevent harassment and promote smooth communication by detecting in advance whether a user's message is inappropriate and what emotion it contains, and correcting it to an appropriate expression.

[0370] A "User" is an individual or entity that utilizes the system to enter and send text messages.

[0371] "Text data" is digital data of character strings including sentences and words entered by the user.

[0372] "Terminal" refers to a device that allows a user to input and send text messages, including a PC, smartphone, tablet, etc.

[0373] "Server" refers to a central management device that receives, analyzes, and stores text data sent from devices and generates alerts and suggestions.

[0374] "Preprocessing" refers to a preparatory process such as tokenizing and escaping data that is performed before the server analyzes the text data.

[0375] A "generative model" is an algorithm that uses natural language processing technology to analyze text data and detect inappropriate content or harassment.

[0376] An "emotion engine" is an algorithm that recognizes a user's emotions from text data and detects emotions such as positive, negative, or neutral.

[0377] An "alert message" is a warning message that the server sends to the user based on the analysis results of the generative model and emotion engine to inform them of the possibility of inappropriate content or harassment.

[0378] A "suggestion" is a message that the server sends to the user, based on the analysis results of the generative model and emotion engine, instructing them on more appropriate expressions and actions.

[0379] A "log" is a collection of data that records sent messages and their analysis results and is saved for later reference.

[0380] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining an emotion engine and a generative model, this system can realize highly accurate alerts and suggestions based on the user's emotions.

[0381] System Configuration

[0382] 1. Device:

[0383] A terminal is a device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text. Terminals include PCs, smartphones, and tablets.

[0384] 2. Server:

[0385] The server receives text data sent from the device and analyzes it using a generative model and emotion engine. The server also generates alert messages and suggestions based on the analysis results and sends them to the device to notify the user.

[0386] 3. Generative Model:

[0387] A generative model is an algorithm that uses natural language processing techniques to analyze text data, understand context and meaning, and detect inappropriate content and harassment. Specific techniques used include BERT and GPT-3.

[0388] 4. Emotion Engine:

[0389] An emotion engine is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, neutral, etc. Sentiment Analysis API and Emotion Detection Model are used as emotion engines.

[0390] Program processing overview

[0391] User message entry and sending

[0392] The user types the text message "You're really useless" into the chat app, and once this action is completed, the device sends the message to the server.

[0393] Message analysis and emotion recognition

[0394] The server passes the received text data to the generative model for preprocessing, which includes tokenization and escaping. The generative model then analyzes the preprocessed data and detects the inappropriate expression "useless." At the same time, the emotion engine analyzes the user's emotions and detects the emotion "anger."

[0395] Alert generation and user notification

[0396] Based on the analysis results, the server generates an alert message and a suggestion, such as "This message contains negative sentiment and contains inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The device then notifies the user of this message.

[0397] Specific examples

[0398] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[0399] At 2 a.m., the user types a message saying, "Please take care of this now," and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message, detects that it is an "imperative expression," and scores it. The emotion engine detects the emotion of "anger" and assigns a negative emotion score. The server generates an alert stating, "This message was sent late at night and contains imperative expressions and the emotion of anger. We recommend sending it the next morning and revising it to 'Please take care of this by tomorrow morning.'" The device displays the alert as a pop-up to the user. The user edits the message to "Please take care of this by tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message back to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The server automatically sends the clean message to the recipient the following morning.

[0400] This system will help prevent harassment and promote better communication.

[0401] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0402] Step 1:

[0403] User:

[0404] The user types the text message "You're really useless" into a chat app. The input is made to the device, and text data is generated.

[0405] Step 2:

[0406] Device:

[0407] When the user presses the send button, the text data and related metadata (user ID, time of sending, etc.) are sent to the server. The input here is the text data and metadata, and this is the output sent to the server.

[0408] Step 3:

[0409] server:

[0410] The server receives data sent from the terminal. It preprocesses the received data (input). Preprocessing includes tokenization (for example, breaking down "You really can't use it" into "o," "mae," "ha," "hontoku," and "tsukanenai") and character escaping (removing certain special characters). The preprocessed text data is the output.

[0411] Step 4:

[0412] server:

[0413] The preprocessed text data is passed as input to a generative model. The generative model then analyzes the text data to detect inappropriate content and harassment. The analysis involves understanding the context and extracting keywords. For example, negative expressions such as "useless" are detected. The analysis results output as inappropriate content detection results.

[0414] Step 5:

[0415] server:

[0416] Based on the analysis results from the generative model, the server passes the text data as input to the emotion engine. The emotion engine analyzes the user's emotion from the text data. For example, the emotion "anger" is detected. The emotion analysis results output the type of emotion and its intensity.

[0417] Step 6:

[0418] server:

[0419] The analysis results of the generative model and the emotion engine are integrated. Based on this integration result, the server generates an alert message and a suggestion message. For example, "This message contains negative emotions and appears to contain inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The generated alert message and suggestion message are output.

[0420] Step 7:

[0421] Device:

[0422] Receives alert messages and suggestions sent from the server and displays them to the user in a popup format. The displayed alert messages and suggestions become input information for the user to confirm. The displayed content becomes the output.

[0423] Step 8:

[0424] User:

[0425] The user reviews the alerts and suggestions and chooses whether to revise the message or send it as is. For example, they might revise it to "Specific feedback would be helpful." This choice is the input, and the revised message is the output.

[0426] Step 9:

[0427] User:

[0428] If the user modifies the message, he or she presses the resend button.

[0429] Step 10:

[0430] Device:

[0431] The corrected text data is sent to the server again. The sent corrected data becomes the input and the output sent to the server.

[0432] Step 11:

[0433] server:

[0434] The server receives the corrected text data again and re-analyzes it using the generative model. As before, it performs preprocessing, analysis, and emotion recognition, and then verifies that the problem has been solved. The verification result is output.

[0435] Step 12:

[0436] server:

[0437] The text data that has been confirmed to be problem-free is sent to the actual recipient. The sent text data becomes the output.

[0438] Step 13:

[0439] server:

[0440] The sent message and its analysis results are logged for future reference. This log is the output.

[0441] In this way, the system can provide appropriate analysis and suggestions for users' messages, helping to prevent harassment and promote smooth communication.

[0442] (Application example 2)

[0443] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0444] In recent years, with the development of communication tools, there has been an increase in inappropriate language and harassment via text messages. Such problems can cause psychological stress for the recipient and, in some cases, legal issues. Furthermore, senders themselves may send misleading messages in an emotionally charged state, which can lead to future problems. There is a need to prevent these from happening in advance and ensure healthy and safe communication.

[0445] The identification process by the identification 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 receiving text data entered by a user; analysis means using a generative model for analyzing the received text data to detect inappropriate content or content that constitutes harassment; means for suggesting alert messages or appropriate expressions to the user based on the analysis results; emotion recognition means using an emotion engine for detecting the user's emotional state based on the analysis results; and means for generating a specific correction suggestion message based on the inappropriate content, harassment, and emotional state. This allows users to prevent inappropriate messages from being sent in advance and maintain healthy and smooth communication.

[0446] "User" means any person or entity that uses the System to input and send text messages.

[0447] "Text data" refers to the text information entered by the user, and is text that is subject to detection of inappropriate expressions and harassment.

[0448] The "receiving means" refers to a function or device for transmitting text data entered by a user to a server and receiving the data.

[0449] "Analysis means" refers to functions or algorithms that analyze received text data using a generative model to detect inappropriate content or harassment.

[0450] A "generative model" is a model that uses natural language processing technology to analyze text data and understand context and meaning.

[0451] An "alert message" is a message that warns the user based on the analysis results.

[0452] "Means for suggesting" refers to functions and algorithms for suggesting alert messages and appropriate expressions to users.

[0453] "Emotion recognition means" refers to functions and algorithms for detecting the user's emotional state based on the analysis results and analyzing that state.

[0454] An "emotion engine" is a system that includes an algorithm that recognizes a user's emotions from text data and detects emotional states such as positive, negative, or neutral.

[0455] A "correction suggestion message" is a message that provides the user with specific suggestions for correction based on inappropriate content, harassment, and emotional state.

[0456] The present invention aims to enable users to communicate safely and appropriately by providing a system for analyzing the content of text messages and the emotions behind them in real time. Hereinafter, an embodiment of the present invention will be described in detail.

[0457] Program processing

[0458] This system consists of a terminal where users can input and send text messages, and a server that analyzes those messages. Specifically, the system uses the following hardware and software:

[0459] Hardware

[0460] Device: Any device that allows a user to type text messages, such as a smartphone, tablet, or PC.

[0461] Server: The server for receiving and analyzing messages is run on Apache or Nginx and uses a Python-based framework (Flask or Django).

[0462] software

[0463] Generative models: Generative AI models that use natural language processing techniques, such as GPT-3 and BERT, to analyze text data and detect inappropriate content and harassment.

[0464] Emotion engine: Detects the user's emotional state from text using technologies such as IBM Watson Natural Language Understanding (NLU).

[0465] Database: A database (e.g., MySQL, PostgreSQL) for storing analysis results and transmission history.

[0466] Data processing and calculation

[0467] Message input: The user types a text message into the device, for example, "Why are you always late!"

[0468] Data transmission: The entered text data and metadata (user ID, submission time, etc.) are sent to the server.

[0469] Preprocessing: The server tokenizes the received text data and performs any necessary character escaping.

[0470] Analysis: Preprocessed text data is fed into a generative model to analyze context and meaning, detecting inappropriate content and harassment.

[0471] Emotion recognition: The emotion engine detects the user's emotional state (e.g., "anger") from text data.

[0472] Alerts and Suggestions: Based on the analysis results, appropriate alert messages and suggestions for correction are generated and notified to the user.

[0473] Save: The sent message, its analysis results, and its revision history are saved in the database.

[0474] Specific examples

[0475] As a concrete example, consider the situation where a user tries to enter the following message:

[0476] "Why are you always late?"

[0477] When attempting to send this message, the generative AI model is prompted with the following:

[0478] Example prompt:

[0479] Message: "Why are you always late?"

[0480] Emotional state: "Anger"

[0481] Purpose: "Detect profanity and provide correction suggestions"

[0482] The generated alert and suggestion will notify the user, for example, "This message contains angry emotions and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late.'" By revising the message based on this suggestion, the user can achieve more appropriate communication.

[0483] The above has described in detail the embodiments of the present invention.

[0484] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0485] Step 1:

[0486] The user inputs a text message into the terminal, for example, "Why are you always late?" This input data is passed to subsequent processing.

[0487] Step 2:

[0488] The text message is sent from the device to the server, along with metadata such as the user ID and the time of sending, allowing the server to determine who is sending what message at what time.

[0489] Step 3:

[0490] The server preprocesses the text data it receives. Specifically, it tokenizes the text data and performs necessary character escaping. This preprocessing prepares the data in a format suitable for analysis. The input is text data and metadata, and the output is the preprocessed tokenized data.

[0491] Step 4:

[0492] The preprocessed tokenized text data is passed to a generative model. The server uses the generative model to analyze this data and detect inappropriate content and harassment while understanding the context and meaning. The input is the tokenized data, and the output is the detection results for inappropriate content and harassment.

[0493] Step 5:

[0494] The server uses an emotion engine to detect the user's emotional state from the tokenized text data. For example, emotional states such as "anger" or "joy" are analyzed. The input is the tokenized data, and the output is the detected emotional state.

[0495] Step 6:

[0496] The server combines the analysis results of the generative model with the results of the emotion engine, and generates a specific correction suggestion message based on the inappropriate content and emotional state. For example, if the analysis results determine that "anger is detected and contains inappropriate language," it generates a suggestion such as "This message contains anger and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late,' and send it." The input is the detection result and emotional state, and the output is a correction suggestion message.

[0497] Step 7:

[0498] The server sends a correction suggestion message to the terminal and notifies the user. The user checks the suggestion message and makes appropriate corrections. The input is the correction suggestion message, and the output is the text message corrected by the user.

[0499] Step 8:

[0500] If the user resubmits the corrected message, it is resent from the terminal to the server. The server reanalyzes it and verifies that there are no problems. The input is the corrected message, and the output is the analysis result.

[0501] Step 9:

[0502] The server finally sends the corrected text message to the recipient and saves the sending history in a database for future reference. The input is the corrected message and the analysis result, and the output is the sending history.

[0503] This series of processing steps allows users to prevent sending inappropriate messages in advance and maintain healthy and smooth communication.

[0504] 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.

[0505] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0506] 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.

[0507] [Second embodiment]

[0508] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0509] 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.

[0510] 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).

[0511] 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.

[0512] 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.

[0513] 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).

[0514] 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.

[0515] 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.

[0516] 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.

[0517] 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.

[0518] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0519] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0520] This system analyzes text-based communications, detects inappropriate content and harassment in advance, and suggests appropriate expressions and sending times to users. The specific operation and functions of this system are explained below.

[0521] System Configuration

[0522] 1. Device:

[0523] A device that users use to enter text-based messages such as email, chat, and social networking sites. The terminal includes an interface for receiving user input and sending it to a server.

[0524] 2. Server:

[0525] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[0526] 3. Generative Model:

[0527] It is an algorithm that uses natural language processing technology to analyze text data, understand the context, and detect inappropriate expressions and harassment.

[0528] Program processing

[0529] Enter and send a message

[0530] User:

[0531] A user inputs a text message into an email, chat, social networking site, etc. For example, the user inputs a message such as "Please send the meeting materials immediately."

[0532] Device:

[0533] When the user presses the send button after entering a message, the device sends the text data and related meta-information (user ID, sending time, etc.) to the server.

[0534] Message Parsing

[0535] server:

[0536] The server preprocesses the received text data and prepares it for the generative model, including tokenizing and filtering the message.

[0537] Generative model:

[0538] Generative models analyze text data to understand context and detect inappropriate content or harassment. The results of this analysis are scored according to importance and identify specific issues.

[0539] Alerts and Suggestions

[0540] server:

[0541] Based on the analysis results of the generative model, appropriate alert messages and suggestions are generated. For example, if the analysis results receive a high score, a specific suggestion such as "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'" is generated.

[0542] Device:

[0543] Alert messages and suggestions sent by the server are displayed to the user in a popup format, allowing the user to review the message before sending it.

[0544] User Selection

[0545] User:

[0546] The user can review the alerts and suggestions and choose to modify the message or send it as is. For example, they can modify the message to "Could you please send the meeting materials as soon as possible?" and resend it.

[0547] Resend and final confirmation of the revised message

[0548] Device:

[0549] The corrected message is sent back to the server for final confirmation.

[0550] server:

[0551] Re-analyze the correction message and finally make sure there are no problems.

[0552] Sending a message

[0553] server:

[0554] After a final check, the resolved message is sent to the actual recipient.

[0555] Specific examples

[0556] As a concrete example, consider the case of an "commanding message sent late at night."

[0557] User:

[0558] At 2 a.m., type the message "This is an emergency and I need your help immediately!" and press the send button.

[0559] Device:

[0560] Sends text and meta information to the server.

[0561] server:

[0562] Receives messages, preprocesses them, and passes them to the generative model.

[0563] Generative model:

[0564] The message is analyzed to detect that it contains the imperative phrase "This is urgent, please respond immediately!" and that it was sent late at night.

[0565] server:

[0566] Generate an alert such as, "This message was sent late at night and contains imperative language. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'"

[0567] Device:

[0568] The alert is displayed to the user in a popup.

[0569] User:

[0570] Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[0571] Device:

[0572] The corrected message is sent back to the server for final confirmation.

[0573] server:

[0574] Re-analyze and check that there are no problems.

[0575] server:

[0576] The next morning, a clean message will be automatically sent to the recipient.

[0577] This invention aims to prevent harassment and inappropriate language in advance, and provides a concrete means for improving the quality of communication.

[0578] The processing flow will be explained below.

[0579] Step 1:

[0580] User: The user types a text message to send via email, chat, or social media.

[0581] Step 2:

[0582] User: Checks the message and clicks the send button.

[0583] Step 3:

[0584] Terminal: Sends text data and related meta-information (user ID, sending time, etc.) to the server.

[0585] Step 4:

[0586] Server: Preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message and spell checking.

[0587] Step 5:

[0588] Generative model: Analyzes tokenized text data and detects inappropriate content and harassment while understanding the context. The analysis results are scored and quantified to indicate their importance.

[0589] Step 6:

[0590] Server: Generates alert messages and suggestion messages based on the analysis results of the generative model. For example, if the score is high, it generates suggestions including specific corrections.

[0591] Step 7:

[0592] Server: Sends alert and suggestion messages to the device.

[0593] Step 8:

[0594] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[0595] Step 9:

[0596] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[0597] Step 10:

[0598] User: If you have corrected the message, click the resend button.

[0599] Step 11:

[0600] Terminal: Send the corrected text data to the server again.

[0601] Step 12:

[0602] Server: Re-parse the corrected text data and verify that the problem is resolved.

[0603] Step 13:

[0604] Server: After final confirmation, the text data is sent to the actual recipient without any problems.

[0605] Step 14:

[0606] Server: Logs the sent message and its analysis results and stores them for future reference.

[0607] Example 1

[0608] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0609] A problem with conventional text-based communication is that users can unknowingly send messages containing inappropriate content or harassment. Furthermore, sending messages at inappropriate times, such as late at night or early in the morning, can cause discomfort to the recipient. A system that addresses these issues and promotes healthy communication was needed.

[0610] 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.

[0611] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data and passing it to a generative AI model, means for analyzing the received text data using the generative AI model and detecting inappropriate content or content that constitutes harassment, means for suggesting an alert message or appropriate expression to the user based on the detected analysis results, means for displaying the alert message or suggestion to the user, means for resending and reanalyzing a message corrected by the user, and means for sending the corrected message to the actual recipient after final confirmation. This prevents the user from unknowingly sending a message containing inappropriate content and enables the message to be sent at an appropriate time.

[0612] "User" means a person who uses the system to enter and send text messages.

[0613] "Means for receiving" is a general term for hardware and software for transmitting text data entered by a user to a server and receiving it.

[0614] "Preprocessing" refers to a series of steps performed on received text data before it is passed to a generative AI model, including tokenization, removal of unnecessary characters, and grammatical analysis.

[0615] A "generative AI model" is an algorithm that uses natural language processing technology to analyze text data, understand context, and detect inappropriate content or harassment.

[0616] "Means for analysis" refers collectively to software and hardware used to analyze received text data using a generative AI model and identify inappropriate content or expressions that constitute harassment.

[0617] An "alert message" is a message that warns the user that the content contains inappropriate content or expressions based on the analysis results of the generative AI model.

[0618] "Appropriate expressions" are expressions that suggest more appropriate and desirable wording to the user based on the problems detected by the generative AI model.

[0619] "Proposed means" is a general term for software and hardware that generates alert messages and appropriate expressions based on the analysis results and presents them to the user.

[0620] "Means for retransmission" is a general term for software and hardware that has the function of sending a message corrected by the user back to the server for reanalysis.

[0621] "Means for sending after final confirmation" refers to software and hardware that has the functionality to reanalyze the modified message, confirm that there are no problems, and then send the message to the actual recipient.

[0622] This invention relates to a system that analyzes text data entered by a user, detects inappropriate content or harassment in advance, and suggests appropriate expressions and sending times. This system improves the appropriateness of messages by performing various analyses during the process of a user entering a text message via email, chat, SNS, etc. and sending it.

[0623] System Configuration and Hardware

[0624] 1. Device:

[0625] A device used by users to input text-based messages such as emails, chats, and social media. Typical devices include PCs, smartphones, and tablets.

[0626] The device includes an interface for receiving user input and sending it to a server, often using a front-end framework such as React or Vue.js for the user interface.

[0627] 2. Server:

[0628] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[0629] Natural language processing libraries used include the Python-based NLTK and spaCy. Preprocessing involves tokenizing and filtering messages.

[0630] Transformer-based models (e.g., GPT-3, BERT) are used as generative AI models.

[0631] 3. Generative AI Model:

[0632] This is an algorithm that uses natural language processing technology to analyze text data, understand the context, and then detect inappropriate language and harassment. The generative AI model is capable of highly accurate analysis and can accurately grasp the context.

[0633] System Operation

[0634] 1. Enter and send a message:

[0635] User: The user types a text message into email, chat, SNS, etc. For example, they type a message such as "Please send the meeting materials right away."

[0636] Terminal: When the user presses the send button after entering a message, the terminal sends the text data entered by the user and related meta information (user ID, sending time, etc.) to the server. The HTTP protocol is used for sending.

[0637] 2. Message analysis:

[0638] Server: The server preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message, removing unnecessary characters and tags, and simple grammatical analysis.

[0639] Generative models: Generative models analyze text data and understand context to detect inappropriate content and harassment. Specifically, transformer-based models understand context and identify things like commanding tone, unnecessary emphasis, and discriminatory language.

[0640] 3. Alert generation and suggestions:

[0641] Server: Generates appropriate alert messages and suggestions based on the analysis results obtained from the generative model. For example, it generates specific suggestions such as, "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'"

[0642] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format.

[0643] 4. User Choice:

[0644] User: The user can review the message based on the pop-up alerts and suggestions, and choose to apply the suggested changes or send it as is. For example, they can change it to "Could you please send the meeting materials as soon as possible?" and resend it.

[0645] 5. Resend and final confirmation of the corrected message:

[0646] Terminal: If the user presses the button to resend the corrected message, the terminal will resend the corrected text data to the server.

[0647] Server: Re-analyze the corrected message and finally check that there are no problems.

[0648] 6. Sending a message:

[0649] Server: Once all is well, the server sends the message to the actual recipient, typically using the SMTP protocol or an API to connect to an email or messaging system.

[0650] Specific examples

[0651] As a concrete example, consider an imperative message sent late at night.

[0652] User: At 2 a.m., type the message "This is urgent, please respond immediately!" and press the send button.

[0653] Terminal: The text data entered by the user and meta information (user ID, time of submission, etc.) are sent to the server via the HTTP protocol.

[0654] Server: Receives messages, tokenizes and filters them using NLTK, and performs parsing with spaCy.

[0655] Generative model: We use GPT-3 to analyze the message and detect the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night.

[0656] Server: Generate an alert like, "This message was sent late at night and contains an imperative. We recommend sending it the next morning and amending it to 'Please respond as soon as possible tomorrow morning.'"

[0657] Terminal: Display the alert to the user as a popup.

[0658] User: Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[0659] Device: Send the corrected message data and sending timing to the server again.

[0660] Server: Re-analyze the data and make sure there are no problems.

[0661] Server: The next morning, the corrected message is automatically sent to the recipient.

[0662] Example prompt sentence:

[0663] "A user types and attempts to send a message at 2 AM saying 'This is urgent, please respond immediately!' Analyze the appropriateness of this message and generate a prompt suggesting a correction."

[0664] The present invention is a system that prevents a user from unknowingly sending a message containing inappropriate content and supports sending messages at an appropriate time.

[0665] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0666] Step 1: User types and sends a text message

[0667] User: Enter a text message via email, chat, SNS, etc. For example, enter the message "Please send the meeting materials immediately."

[0668] Input: The text message entered by the user.

[0669] Output: Text data and meta information (user ID, sending time, etc.) ready to be sent to the device

[0670] Step 2: The device sends the text data to the server

[0671] Terminal: After receiving user input, it detects when the send button is pressed and sends the text data and meta information to the server using the HTTP protocol.

[0672] Input: Text data and meta information entered by the user

[0673] Output: Text data and meta information sent to the server

[0674] Step 3: The server preprocesses the text data

[0675] Server: Preprocesses the received text data. This preprocessing includes tokenizing the text, removing unnecessary characters and tags, and simple grammar analysis. This is done using a Python natural language processing library (NLTK, spaCy, etc.).

[0676] Input: Text data and meta information

[0677] Output: Preprocessed text data

[0678] Step 4: The server uses the generative model to analyze the text data

[0679] Server: Preprocessed text data is passed to a generative AI model (e.g., GPT-3, BERT) that understands the context and detects inappropriate content and harassment.

[0680] Input: Preprocessed text data

[0681] Output: Detection results and a score indicating their importance

[0682] Step 5: The server generates an alert message or suggestion

[0683] Server: Generates appropriate alert messages and suggestions based on the analysis results of the generative model. For example, it generates a message such as, "This message contains imperative language. We recommend revising it to 'Could you please send the meeting materials as soon as possible?'"

[0684] Input: Detection results and scoring

[0685] Output: Alert message and suggestions

[0686] Step 6: The device displays an alert message or suggestion to the user.

[0687] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format using a front-end framework (e.g., React, Vue.js).

[0688] Input: Alert message and suggestions

[0689] Output: Alerts and suggestions displayed in the user interface

[0690] Step 7: User reviews the suggested changes and edits the message

[0691] User: Based on the alerts and suggestions displayed, revise the message, for example, to "Could you please send the meeting materials as soon as possible?"

[0692] Input: The displayed alert message and suggestion

[0693] Output: The modified text message

[0694] Step 8: The device resends the corrected message

[0695] Terminal: If the user clicks the resend button after correcting the message, the corrected message data is sent to the server again.

[0696] Input: Modified text message

[0697] Output: The correction message sent to the server.

[0698] Step 9: The server reparses the modified message and performs a final check

[0699] Server: Re-parses the modified message to ensure there are no problems. This re-parse is done using the generative model, just like the initial parse.

[0700] Input: Modified text message

[0701] Output: Final confirmation result of the message after correction

[0702] Step 10: Server sends message after final confirmation

[0703] Server: After verifying that there are no problems, it sends the corrected message to the actual recipient, using the SMTP protocol or API to connect to email or messaging systems.

[0704] Input: Correction message that has been finalized

[0705] Output: The message sent to the actual recipient

[0706] (Application example 1)

[0707] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0708] The present invention aims to realize safe and secure communication by effectively detecting inappropriate content and harassment in text-based communication and suggesting appropriate expressions and sending times to users. It also provides a system that reanalyzes messages corrected based on the suggestions, ultimately transmitting problem-free text data.

[0709] 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.

[0710] In this invention, the server includes means for receiving text data entered by a user, analysis means using a generative model to analyze the received text data and detect inappropriate content or content that constitutes harassment, means for suggesting alert messages or appropriate expressions to the user based on the analysis results, means for reanalyzing the text data corrected based on the suggested content and reconfirming its appropriateness, and means for transmitting the corrected text data to the final recipient. This enables safe and secure communication by automatically checking the appropriateness of messages entered by users and suggesting corrections as necessary.

[0711] "User" means the entity that uses the system and inputs and transmits text.

[0712] "Text data" refers to character information entered and transmitted by a user.

[0713] The "receiving means" is a device or program that has the function of incorporating text data sent from a user into the system.

[0714] "Analysis means" refers to a device or program for analyzing text data to detect inappropriate expressions or harassment.

[0715] A "generative model" is an algorithm or program that uses natural language processing technology to analyze the content of text data.

[0716] The "suggestion means" is a device or program that has the function of providing the user with specific alerts and examples of corrections based on the analysis results.

[0717] The "reanalysis means" is a device or program having the function of reanalyzing text data that has been corrected by the user based on the suggestion and confirming its appropriateness.

[0718] The "transmission means" is a device or program having a function for transmitting the final confirmed text data to the recipient.

[0719] An "alert message" is a message that warns the user when the analysis means detects inappropriate content or harassment.

[0720] "Appropriate language" refers to language that is respectful to the recipient and does not contain inappropriate or harassing content.

[0721] The system for implementing this invention consists of a server connected to the Internet and a user terminal. The user inputs and sends text messages using a terminal such as a smartphone or PC. The terminal includes an interface for receiving the input text data and transmitting it to the server.

[0722] The server performs various processes to analyze the received text data. Specifically, the server preprocesses the received message and passes it to a generative model for contextual analysis. The generative model can be the Transformers library from Hugging Face. Specifically, a natural language processing model (e.g., BERT or GPT-3) is used.

[0723] The generative model analyzes text data to detect inappropriate content and harassment. At this stage, the analysis results are scored and the problem areas are identified based on the score. For example, a message such as "This is an emergency and please act immediately!" is deemed to be imperative and inappropriate.

[0724] Based on the analysis results, the server makes appropriate suggestions to the user. The suggestions are specific and are displayed in the form of, for example, "This message contains imperative language. We recommend that you correct it as follows: 'Please respond as soon as possible.'"

[0725] The user reviews the suggestions and modifies the message. The modified message is then sent back to the server for re-analysis. This re-analysis ultimately verifies the message is clean. After verification, the modified message is finally sent to the recipient.

[0726] As a concrete example, consider the case of an imperative message sent late at night. A user types the message "This is urgent, please respond immediately!" at 2 a.m. and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message and detects that it contains the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night. The server generates an alert stating, "This message was sent late at night and contains an imperative expression. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'" The device displays the alert to the user as a pop-up. The user edits the message to "Please respond as soon as possible tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message again to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The next morning, the server automatically sends the message without any problems to the recipient.

[0727] An example of a prompt is as follows:

[0728] "Analyze the following sentence and determine if it is an imperative: 'Please respond immediately!'"

[0729] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0730] Step 1:

[0731] The user enters a text message into the terminal and presses the send button. At this time, the terminal prepares the entered text data and meta information (user ID, sending time, etc.). The terminal sends the text data and meta information to the server. The input is the "text message" and "meta information," and the output is the "text data and meta information sent to the server."

[0732] Step 2:

[0733] The server preprocesses the text data it receives. Specifically, it tokenizes and filters the text data. The input is the received text data and meta information, and the output is the preprocessed text data.

[0734] Step 3:

[0735] The server passes the preprocessed text data to a generative model for contextual analysis. Specifically, it uses a natural language processing model based on the Hugging Face Transformers library (e.g., BERT or GPT-3) to analyze inappropriate language and harassment in the text. The input is the preprocessed text data, and the output is the analysis results (scores for inappropriate language and harassment).

[0736] Step 4:

[0737] The server detects inappropriate content or harassment based on the analysis results of the generative model and generates an alert message or correction suggestions. The input is the "analysis results (scores for inappropriate expressions and harassment)" and the output is the "alert message or correction suggestions." Specifically, the message generated is "This message contains imperative language. We recommend correcting it as follows: 'Please respond as soon as possible.'"

[0738] Step 5:

[0739] The server sends an alert message or correction suggestion to the terminal, which then displays it to the user in a popup format. The input is the "alert message or correction suggestion" and the output is the "displayed alert message or correction suggestion." The user then corrects the message based on the suggested suggestions.

[0740] Step 6:

[0741] The user resends the corrected text message, and the terminal sends it to the server. The input is the "corrected text message" and the output is the "corrected text data sent to the server."

[0742] Step 7:

[0743] The server re-analyzes the corrected text message and performs a final check. The input is the "corrected text data" and the output is the "final analysis result." It checks to see if there are any problems.

[0744] Step 8:

[0745] The server sends the verified modified text message to the final recipient. The input is the "verified modified text message" and the output is the "text message sent to the final recipient".

[0746] 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.

[0747] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining this system with an emotion engine, it is possible to realize more accurate alerts and suggestions based on the user's emotions.

[0748] System Configuration

[0749] 1. Device:

[0750] A device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text.

[0751] 2. Server:

[0752] It receives text data sent from the device and analyzes it using a generative model and emotion engine.

[0753] 3. Generative Model:

[0754] This algorithm uses natural language processing technology to analyze text data, understand context and meaning, and detect inappropriate content and harassment.

[0755] 4. Emotion Engine:

[0756] This is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, and neutral as analysis results.

[0757] Program processing

[0758] Enter and send a message

[0759] User:

[0760] The user types the text message "You're really useless" into a chat app.

[0761] Device:

[0762] When the send button is pressed, the terminal sends the text data and related meta information (user ID, sending time, etc.) to the server.

[0763] Message Parsing

[0764] server:

[0765] Preprocessing the received text data before passing it to the generative model includes tokenizing the message and escaping any necessary characters.

[0766] Generative model:

[0767] The system analyzes tokenized text data to detect negative content such as "unusable," and then scores it to quantify its importance.

[0768] emotion recognition

[0769] Emotion Engine:

[0770] It analyzes the user's emotions from text data and detects negative emotions, such as "anger."

[0771] Alerts and Suggestions

[0772] server:

[0773] The inappropriate content detection results from the generative model and the analysis results from the emotion engine are integrated to generate alert messages and suggestions.

[0774] For example, it generates a suggestion such as, "This message contains negative emotions and contains inappropriate language. Please try revising it to 'It would be helpful if you could provide specific feedback.'"

[0775] Device:

[0776] Displays generated alert and suggestion messages to the user in a popup format, allowing the user to review the message before sending.

[0777] User Selection

[0778] User:

[0779] Review the alert or suggestion and choose to modify the text message to "Specific feedback would be helpful" or send it as is.

[0780] Resend and final confirmation of the revised message

[0781] User:

[0782] If you have corrected the message, click the resend button.

[0783] Device:

[0784] The corrected text data is sent to the server again.

[0785] server:

[0786] Re-analyze the corrected text data to verify that the problem has been resolved.

[0787] Sending a message

[0788] server:

[0789] After final confirmation, the text data is sent to the actual recipient without any problems.

[0790] Message storage

[0791] server:

[0792] The sent message and its analysis results are recorded in a log and saved for future reference.

[0793] Specific examples

[0794] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[0795] User:

[0796] "At 2 a.m., I type a message saying, 'Deal with this right now,' and press the send button."

[0797] Device:

[0798] Sends text and meta information to the server.

[0799] server:

[0800] Receives messages, preprocesses them, and passes them to the generative model.

[0801] Generative model:

[0802] The message is analyzed to detect whether it is an "imperative expression" and is scored.

[0803] Emotion Engine:

[0804] Detects the emotion "anger" and sets a negative emotion score.

[0805] server:

[0806] Generate an alert saying, "This message was sent late at night and contains commanding language and angry sentiment. We recommend sending it the next morning and amending it to 'Please respond by tomorrow morning.'"

[0807] Device:

[0808] The alert is displayed to the user in a popup.

[0809] User:

[0810] Change the message to "Please respond by tomorrow morning" and set it to be sent the next morning.

[0811] Device:

[0812] The corrected message is sent back to the server for final confirmation.

[0813] server:

[0814] Re-analyze and check that there are no problems.

[0815] server:

[0816] The next morning, a clean message is automatically sent to the recipient. This invention provides a technological means to prevent harassment and promote better communication.

[0817] The processing flow will be explained below.

[0818] Step 1:

[0819] User: The user types a text message to send via email, chat, or social media.

[0820] Step 2:

[0821] User: Checks the message and clicks the send button.

[0822] Step 3:

[0823] Device: Sends the text message and related meta information (user ID, time sent, etc.) to the server.

[0824] Step 4:

[0825] Server: Preprocesses the received text messages and prepares them for the generative model. Preprocessing includes tokenizing the messages and normalizing the input data.

[0826] Step 5:

[0827] Generative Model: A generative model analyzes tokenized text data, understands the context, and detects inappropriate content and harassment. It then assigns a score to the inappropriate content.

[0828] Step 6:

[0829] Emotion Engine: The emotion engine analyzes the user's emotions from text data. For example, it detects emotional states such as "anger," "sadness," and "joy" and assigns a score to each.

[0830] Step 7:

[0831] Server: Integrates the analysis results of the generative model and the emotion engine to generate alert messages and suggestion messages. For example, if negative expressions and the emotion "anger" have high scores, the server will suggest corrections.

[0832] Step 8:

[0833] Server: Sends generated alert and suggestion messages to the device.

[0834] Step 9:

[0835] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[0836] Step 10:

[0837] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[0838] Step 11:

[0839] User: If you have corrected the message, click the resend button.

[0840] Step 12:

[0841] On your device: Send the modified text message back to the server.

[0842] Step 13:

[0843] Server: Re-parse the modified text message to verify that the issue has been resolved.

[0844] Step 14:

[0845] Server: After final verification, the text message is sent to the intended recipient.

[0846] Step 15:

[0847] Server: Logs the sent message and its analysis results and stores them for future reference.

[0848] Example 2

[0849] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0850] Inappropriate language and harassment are commonplace in modern communication tools, placing a heavy burden on users. Effective technology to prevent these issues has yet to be fully established. Furthermore, few systems can provide alerts or suggestions based on users' emotions, which can lead to unintended misunderstandings and conflicts. To address these issues, a system is needed that uses more advanced natural language processing technology to analyze users' message content and emotions and provide appropriate alerts and suggestions.

[0851] 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.

[0852] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data, means for analyzing the preprocessed text data using a generative model to detect inappropriate content or content that constitutes harassment, means for analyzing the received text data using an emotion engine that recognizes the user's emotions, means for suggesting an alert message or appropriate expression to the user based on the analysis results of the generative model and the emotion engine, means for sending the text data with the problem resolved to the actual recipient, and means for recording the sent message and the analysis results in a log. This makes it possible to prevent harassment and promote smooth communication by detecting in advance whether a user's message is inappropriate and what emotion it contains, and correcting it to an appropriate expression.

[0853] A "User" is an individual or entity that utilizes the system to enter and send text messages.

[0854] "Text data" is digital data of character strings including sentences and words entered by the user.

[0855] "Terminal" refers to a device that allows a user to input and send text messages, including a PC, smartphone, tablet, etc.

[0856] "Server" refers to a central management device that receives, analyzes, and stores text data sent from devices and generates alerts and suggestions.

[0857] "Preprocessing" refers to a preparatory process such as tokenizing and escaping data that is performed before the server analyzes the text data.

[0858] A "generative model" is an algorithm that uses natural language processing technology to analyze text data and detect inappropriate content or harassment.

[0859] An "emotion engine" is an algorithm that recognizes a user's emotions from text data and detects emotions such as positive, negative, or neutral.

[0860] An "alert message" is a warning message that the server sends to the user based on the analysis results of the generative model and emotion engine to inform them of the possibility of inappropriate content or harassment.

[0861] A "suggestion" is a message that the server sends to the user, based on the analysis results of the generative model and emotion engine, instructing them on more appropriate expressions and actions.

[0862] A "log" is a collection of data that records sent messages and their analysis results and is saved for later reference.

[0863] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining an emotion engine and a generative model, this system can realize highly accurate alerts and suggestions based on the user's emotions.

[0864] System Configuration

[0865] 1. Device:

[0866] A terminal is a device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text. Terminals include PCs, smartphones, and tablets.

[0867] 2. Server:

[0868] The server receives text data sent from the device and analyzes it using a generative model and emotion engine. The server also generates alert messages and suggestions based on the analysis results and sends them to the device to notify the user.

[0869] 3. Generative Model:

[0870] A generative model is an algorithm that uses natural language processing techniques to analyze text data, understand context and meaning, and detect inappropriate content and harassment. Specific techniques used include BERT and GPT-3.

[0871] 4. Emotion Engine:

[0872] An emotion engine is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, neutral, etc. Sentiment Analysis API and Emotion Detection Model are used as emotion engines.

[0873] Program processing overview

[0874] User message entry and sending

[0875] The user types the text message "You're really useless" into the chat app, and once this action is completed, the device sends the message to the server.

[0876] Message analysis and emotion recognition

[0877] The server passes the received text data to the generative model for preprocessing, which includes tokenization and escaping. The generative model then analyzes the preprocessed data and detects the inappropriate expression "useless." At the same time, the emotion engine analyzes the user's emotions and detects the emotion "anger."

[0878] Alert generation and user notification

[0879] Based on the analysis results, the server generates an alert message and a suggestion, such as "This message contains negative sentiment and contains inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The device then notifies the user of this message.

[0880] Specific examples

[0881] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[0882] At 2 a.m., the user types a message saying, "Please take care of this now," and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message, detects that it is an "imperative expression," and scores it. The emotion engine detects the emotion of "anger" and assigns a negative emotion score. The server generates an alert stating, "This message was sent late at night and contains imperative expressions and the emotion of anger. We recommend sending it the next morning and revising it to 'Please take care of this by tomorrow morning.'" The device displays the alert as a pop-up to the user. The user edits the message to "Please take care of this by tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message back to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The server automatically sends the clean message to the recipient the following morning.

[0883] This system will help prevent harassment and promote better communication.

[0884] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0885] Step 1:

[0886] User:

[0887] The user types the text message "You're really useless" into a chat app. The input is made to the device, and text data is generated.

[0888] Step 2:

[0889] Device:

[0890] When the user presses the send button, the text data and related metadata (user ID, time of sending, etc.) are sent to the server. The input here is the text data and metadata, and this is the output sent to the server.

[0891] Step 3:

[0892] server:

[0893] The server receives data sent from the terminal. It preprocesses the received data (input). Preprocessing includes tokenization (for example, breaking down "You really can't use it" into "o," "mae," "ha," "hontoku," and "tsukanenai") and character escaping (removing certain special characters). The preprocessed text data is the output.

[0894] Step 4:

[0895] server:

[0896] The preprocessed text data is passed as input to a generative model. The generative model then analyzes the text data to detect inappropriate content and harassment. The analysis involves understanding the context and extracting keywords. For example, negative expressions such as "useless" are detected. The analysis results output as inappropriate content detection results.

[0897] Step 5:

[0898] server:

[0899] Based on the analysis results from the generative model, the server passes the text data as input to the emotion engine. The emotion engine analyzes the user's emotion from the text data. For example, the emotion "anger" is detected. The emotion analysis results output the type of emotion and its intensity.

[0900] Step 6:

[0901] server:

[0902] The analysis results of the generative model and the emotion engine are integrated. Based on this integration result, the server generates an alert message and a suggestion message. For example, "This message contains negative emotions and appears to contain inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The generated alert message and suggestion message are output.

[0903] Step 7:

[0904] Device:

[0905] Receives alert messages and suggestions sent from the server and displays them to the user in a popup format. The displayed alert messages and suggestions become input information for the user to confirm. The displayed content becomes the output.

[0906] Step 8:

[0907] User:

[0908] The user reviews the alerts and suggestions and chooses whether to revise the message or send it as is. For example, they might revise it to "Specific feedback would be helpful." This choice is the input, and the revised message is the output.

[0909] Step 9:

[0910] User:

[0911] If the user modifies the message, he or she presses the resend button.

[0912] Step 10:

[0913] Device:

[0914] The corrected text data is sent to the server again. The sent corrected data becomes the input and the output sent to the server.

[0915] Step 11:

[0916] server:

[0917] The server receives the corrected text data again and re-analyzes it using the generative model. As before, it performs preprocessing, analysis, and emotion recognition, and then verifies that the problem has been solved. The verification result is output.

[0918] Step 12:

[0919] server:

[0920] The text data that has been confirmed to be problem-free is sent to the actual recipient. The sent text data becomes the output.

[0921] Step 13:

[0922] server:

[0923] The sent message and its analysis results are logged for future reference. This log is the output.

[0924] In this way, the system can provide appropriate analysis and suggestions for users' messages, helping to prevent harassment and promote smooth communication.

[0925] (Application example 2)

[0926] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0927] In recent years, with the development of communication tools, there has been an increase in inappropriate language and harassment via text messages. Such problems can cause psychological stress for the recipient and, in some cases, legal issues. Furthermore, senders themselves may send misleading messages in an emotionally charged state, which can lead to future problems. There is a need to prevent these from happening in advance and ensure healthy and safe communication.

[0928] The identification process by the identification 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 receiving text data entered by a user; analysis means using a generative model for analyzing the received text data to detect inappropriate content or content that constitutes harassment; means for suggesting alert messages or appropriate expressions to the user based on the analysis results; emotion recognition means using an emotion engine for detecting the user's emotional state based on the analysis results; and means for generating a specific correction suggestion message based on the inappropriate content, harassment, and emotional state. This allows users to prevent inappropriate messages from being sent in advance and maintain healthy and smooth communication.

[0929] "User" means any person or entity that uses the System to input and send text messages.

[0930] "Text data" refers to the text information entered by the user, and is text that is subject to detection of inappropriate expressions and harassment.

[0931] The "receiving means" refers to a function or device for transmitting text data entered by a user to a server and receiving the data.

[0932] "Analysis means" refers to functions or algorithms that analyze received text data using a generative model to detect inappropriate content or harassment.

[0933] A "generative model" is a model that uses natural language processing technology to analyze text data and understand context and meaning.

[0934] An "alert message" is a message that warns the user based on the analysis results.

[0935] "Means for suggesting" refers to functions and algorithms for suggesting alert messages and appropriate expressions to users.

[0936] "Emotion recognition means" refers to functions and algorithms for detecting the user's emotional state based on the analysis results and analyzing that state.

[0937] An "emotion engine" is a system that includes an algorithm that recognizes a user's emotions from text data and detects emotional states such as positive, negative, or neutral.

[0938] A "correction suggestion message" is a message that provides the user with specific suggestions for correction based on inappropriate content, harassment, and emotional state.

[0939] The present invention aims to enable users to communicate safely and appropriately by providing a system for analyzing the content of text messages and the emotions behind them in real time. Hereinafter, an embodiment of the present invention will be described in detail.

[0940] Program processing

[0941] This system consists of a terminal where users can input and send text messages, and a server that analyzes those messages. Specifically, the system uses the following hardware and software:

[0942] Hardware

[0943] Device: Any device that allows a user to type text messages, such as a smartphone, tablet, or PC.

[0944] Server: The server for receiving and analyzing messages is run on Apache or Nginx and uses a Python-based framework (Flask or Django).

[0945] software

[0946] Generative models: Generative AI models that use natural language processing techniques, such as GPT-3 and BERT, to analyze text data and detect inappropriate content and harassment.

[0947] Emotion engine: Detects the user's emotional state from text using technologies such as IBM Watson Natural Language Understanding (NLU).

[0948] Database: A database (e.g., MySQL, PostgreSQL) for storing analysis results and transmission history.

[0949] Data processing and calculation

[0950] Message input: The user types a text message into the device, for example, "Why are you always late!"

[0951] Data transmission: The entered text data and metadata (user ID, submission time, etc.) are sent to the server.

[0952] Preprocessing: The server tokenizes the received text data and performs any necessary character escaping.

[0953] Analysis: Preprocessed text data is fed into a generative model to analyze context and meaning, detecting inappropriate content and harassment.

[0954] Emotion recognition: The emotion engine detects the user's emotional state (e.g., "anger") from text data.

[0955] Alerts and Suggestions: Based on the analysis results, appropriate alert messages and suggestions for correction are generated and notified to the user.

[0956] Save: The sent message, its analysis results, and its revision history are saved in the database.

[0957] Specific examples

[0958] As a concrete example, consider the situation where a user tries to enter the following message:

[0959] "Why are you always late?"

[0960] When attempting to send this message, the generative AI model is prompted with the following:

[0961] Example prompt:

[0962] Message: "Why are you always late?"

[0963] Emotional state: "Anger"

[0964] Purpose: "Detect profanity and provide correction suggestions"

[0965] The generated alert and suggestion will notify the user, for example, "This message contains angry emotions and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late.'" By revising the message based on this suggestion, the user can achieve more appropriate communication.

[0966] The above has described in detail the embodiments of the present invention.

[0967] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0968] Step 1:

[0969] The user inputs a text message into the terminal, for example, "Why are you always late?" This input data is passed to subsequent processing.

[0970] Step 2:

[0971] The text message is sent from the device to the server, along with metadata such as the user ID and the time of sending, allowing the server to determine who is sending what message at what time.

[0972] Step 3:

[0973] The server preprocesses the text data it receives. Specifically, it tokenizes the text data and performs necessary character escaping. This preprocessing prepares the data in a format suitable for analysis. The input is text data and metadata, and the output is the preprocessed tokenized data.

[0974] Step 4:

[0975] The preprocessed tokenized text data is passed to a generative model. The server uses the generative model to analyze this data and detect inappropriate content and harassment while understanding the context and meaning. The input is the tokenized data, and the output is the detection results for inappropriate content and harassment.

[0976] Step 5:

[0977] The server uses an emotion engine to detect the user's emotional state from the tokenized text data. For example, emotional states such as "anger" or "joy" are analyzed. The input is the tokenized data, and the output is the detected emotional state.

[0978] Step 6:

[0979] The server combines the analysis results of the generative model with the results of the emotion engine, and generates a specific correction suggestion message based on the inappropriate content and emotional state. For example, if the analysis results determine that "anger is detected and contains inappropriate language," it generates a suggestion such as "This message contains anger and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late,' and send it." The input is the detection result and emotional state, and the output is a correction suggestion message.

[0980] Step 7:

[0981] The server sends a correction suggestion message to the terminal and notifies the user. The user checks the suggestion message and makes appropriate corrections. The input is the correction suggestion message, and the output is the text message corrected by the user.

[0982] Step 8:

[0983] If the user resubmits the corrected message, it is resent from the terminal to the server. The server reanalyzes it and verifies that there are no problems. The input is the corrected message, and the output is the analysis result.

[0984] Step 9:

[0985] The server finally sends the corrected text message to the recipient and saves the sending history in a database for future reference. The input is the corrected message and the analysis result, and the output is the sending history.

[0986] This series of processing steps allows users to prevent sending inappropriate messages in advance and maintain healthy and smooth communication.

[0987] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0988] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0989] 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.

[0990] [Third embodiment]

[0991] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0992] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0993] 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).

[0994] 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.

[0995] 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.

[0996] 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).

[0997] 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.

[0998] 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.

[0999] 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.

[1000] 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.

[1001] 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.

[1002] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1003] This system analyzes text-based communications, detects inappropriate content and harassment in advance, and suggests appropriate expressions and sending times to users. The specific operation and functions of this system are explained below.

[1004] System Configuration

[1005] 1. Device:

[1006] A device that users use to enter text-based messages such as email, chat, and social networking sites. The terminal includes an interface for receiving user input and sending it to a server.

[1007] 2. Server:

[1008] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[1009] 3. Generative Model:

[1010] It is an algorithm that uses natural language processing technology to analyze text data, understand the context, and detect inappropriate expressions and harassment.

[1011] Program processing

[1012] Enter and send a message

[1013] User:

[1014] A user inputs a text message into an email, chat, social networking site, etc. For example, the user inputs a message such as "Please send the meeting materials immediately."

[1015] Device:

[1016] When the user presses the send button after entering a message, the device sends the text data and related meta-information (user ID, sending time, etc.) to the server.

[1017] Message Parsing

[1018] server:

[1019] The server preprocesses the received text data and prepares it for the generative model, including tokenizing and filtering the message.

[1020] Generative model:

[1021] Generative models analyze text data to understand context and detect inappropriate content or harassment. The results of this analysis are scored according to importance and identify specific issues.

[1022] Alerts and Suggestions

[1023] server:

[1024] Based on the analysis results of the generative model, appropriate alert messages and suggestions are generated. For example, if the analysis results receive a high score, a specific suggestion such as "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'" is generated.

[1025] Device:

[1026] Alert messages and suggestions sent by the server are displayed to the user in a popup format, allowing the user to review the message before sending it.

[1027] User Selection

[1028] User:

[1029] The user can review the alerts and suggestions and choose to modify the message or send it as is. For example, they can modify the message to "Could you please send the meeting materials as soon as possible?" and resend it.

[1030] Resend and final confirmation of the revised message

[1031] Device:

[1032] The corrected message is sent back to the server for final confirmation.

[1033] server:

[1034] Re-analyze the correction message and finally make sure there are no problems.

[1035] Sending a message

[1036] server:

[1037] After a final check, the resolved message is sent to the actual recipient.

[1038] Specific examples

[1039] As a concrete example, consider the case of an "commanding message sent late at night."

[1040] User:

[1041] At 2 a.m., type the message "This is an emergency and I need your help immediately!" and press the send button.

[1042] Device:

[1043] Sends text and meta information to the server.

[1044] server:

[1045] Receives messages, preprocesses them, and passes them to the generative model.

[1046] Generative model:

[1047] The message is analyzed to detect that it contains the imperative phrase "This is urgent, please respond immediately!" and that it was sent late at night.

[1048] server:

[1049] Generate an alert such as, "This message was sent late at night and contains imperative language. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'"

[1050] Device:

[1051] The alert is displayed to the user in a popup.

[1052] User:

[1053] Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[1054] Device:

[1055] The corrected message is sent back to the server for final confirmation.

[1056] server:

[1057] Re-analyze and check that there are no problems.

[1058] server:

[1059] The next morning, a clean message will be automatically sent to the recipient.

[1060] This invention aims to prevent harassment and inappropriate language in advance, and provides a concrete means for improving the quality of communication.

[1061] The processing flow will be explained below.

[1062] Step 1:

[1063] User: The user types a text message to send via email, chat, or social media.

[1064] Step 2:

[1065] User: Checks the message and clicks the send button.

[1066] Step 3:

[1067] Terminal: Sends text data and related meta-information (user ID, sending time, etc.) to the server.

[1068] Step 4:

[1069] Server: Preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message and spell checking.

[1070] Step 5:

[1071] Generative model: Analyzes tokenized text data and detects inappropriate content and harassment while understanding the context. The analysis results are scored and quantified to indicate their importance.

[1072] Step 6:

[1073] Server: Generates alert messages and suggestion messages based on the analysis results of the generative model. For example, if the score is high, it generates suggestions including specific corrections.

[1074] Step 7:

[1075] Server: Sends alert and suggestion messages to the device.

[1076] Step 8:

[1077] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[1078] Step 9:

[1079] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[1080] Step 10:

[1081] User: If you have corrected the message, click the resend button.

[1082] Step 11:

[1083] Terminal: Send the corrected text data to the server again.

[1084] Step 12:

[1085] Server: Re-parse the corrected text data and verify that the problem is resolved.

[1086] Step 13:

[1087] Server: After final confirmation, the text data is sent to the actual recipient without any problems.

[1088] Step 14:

[1089] Server: Logs the sent message and its analysis results and stores them for future reference.

[1090] Example 1

[1091] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1092] A problem with conventional text-based communication is that users can unknowingly send messages containing inappropriate content or harassment. Furthermore, sending messages at inappropriate times, such as late at night or early in the morning, can cause discomfort to the recipient. A system that addresses these issues and promotes healthy communication was needed.

[1093] 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.

[1094] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data and passing it to a generative AI model, means for analyzing the received text data using the generative AI model and detecting inappropriate content or content that constitutes harassment, means for suggesting an alert message or appropriate expression to the user based on the detected analysis results, means for displaying the alert message or suggestion to the user, means for resending and reanalyzing a message corrected by the user, and means for sending the corrected message to the actual recipient after final confirmation. This prevents the user from unknowingly sending a message containing inappropriate content and enables the message to be sent at an appropriate time.

[1095] "User" means a person who uses the system to enter and send text messages.

[1096] "Means for receiving" is a general term for hardware and software for transmitting text data entered by a user to a server and receiving it.

[1097] "Preprocessing" refers to a series of steps performed on received text data before it is passed to a generative AI model, including tokenization, removal of unnecessary characters, and grammatical analysis.

[1098] A "generative AI model" is an algorithm that uses natural language processing technology to analyze text data, understand context, and detect inappropriate content or harassment.

[1099] "Means for analysis" refers collectively to software and hardware used to analyze received text data using a generative AI model and identify inappropriate content or expressions that constitute harassment.

[1100] An "alert message" is a message that warns the user that the content contains inappropriate content or expressions based on the analysis results of the generative AI model.

[1101] "Appropriate expressions" are expressions that suggest more appropriate and desirable wording to the user based on the problems detected by the generative AI model.

[1102] "Proposed means" is a general term for software and hardware that generates alert messages and appropriate expressions based on the analysis results and presents them to the user.

[1103] "Means for retransmission" is a general term for software and hardware that has the function of sending a message corrected by the user back to the server for reanalysis.

[1104] "Means for sending after final confirmation" refers to software and hardware that has the functionality to reanalyze the modified message, confirm that there are no problems, and then send the message to the actual recipient.

[1105] This invention relates to a system that analyzes text data entered by a user, detects inappropriate content or harassment in advance, and suggests appropriate expressions and sending times. This system improves the appropriateness of messages by performing various analyses during the process of a user entering a text message via email, chat, SNS, etc. and sending it.

[1106] System Configuration and Hardware

[1107] 1. Device:

[1108] A device used by users to input text-based messages such as emails, chats, and social media. Typical devices include PCs, smartphones, and tablets.

[1109] The device includes an interface for receiving user input and sending it to a server, often using a front-end framework such as React or Vue.js for the user interface.

[1110] 2. Server:

[1111] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[1112] Natural language processing libraries used include the Python-based NLTK and spaCy. Preprocessing involves tokenizing and filtering messages.

[1113] Transformer-based models (e.g., GPT-3, BERT) are used as generative AI models.

[1114] 3. Generative AI Model:

[1115] This is an algorithm that uses natural language processing technology to analyze text data, understand the context, and then detect inappropriate language and harassment. The generative AI model is capable of highly accurate analysis and can accurately grasp the context.

[1116] System Operation

[1117] 1. Enter and send a message:

[1118] User: The user types a text message into email, chat, SNS, etc. For example, they type a message such as "Please send the meeting materials right away."

[1119] Terminal: When the user presses the send button after entering a message, the terminal sends the text data entered by the user and related meta information (user ID, sending time, etc.) to the server. The HTTP protocol is used for sending.

[1120] 2. Message analysis:

[1121] Server: The server preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message, removing unnecessary characters and tags, and simple grammatical analysis.

[1122] Generative models: Generative models analyze text data and understand context to detect inappropriate content and harassment. Specifically, transformer-based models understand context and identify things like commanding tone, unnecessary emphasis, and discriminatory language.

[1123] 3. Alert generation and suggestions:

[1124] Server: Generates appropriate alert messages and suggestions based on the analysis results obtained from the generative model. For example, it generates specific suggestions such as, "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'"

[1125] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format.

[1126] 4. User Choice:

[1127] User: The user can review the message based on the pop-up alerts and suggestions, and choose to apply the suggested changes or send it as is. For example, they can change it to "Could you please send the meeting materials as soon as possible?" and resend it.

[1128] 5. Resend and final confirmation of the corrected message:

[1129] Terminal: If the user presses the button to resend the corrected message, the terminal will resend the corrected text data to the server.

[1130] Server: Re-analyze the corrected message and finally check that there are no problems.

[1131] 6. Sending a message:

[1132] Server: Once all is well, the server sends the message to the actual recipient, typically using the SMTP protocol or an API to connect to an email or messaging system.

[1133] Specific examples

[1134] As a concrete example, consider an imperative message sent late at night.

[1135] User: At 2 a.m., type the message "This is urgent, please respond immediately!" and press the send button.

[1136] Terminal: The text data entered by the user and meta information (user ID, time of submission, etc.) are sent to the server via the HTTP protocol.

[1137] Server: Receives messages, tokenizes and filters them using NLTK, and performs parsing with spaCy.

[1138] Generative model: We use GPT-3 to analyze the message and detect the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night.

[1139] Server: Generate an alert like, "This message was sent late at night and contains an imperative. We recommend sending it the next morning and amending it to 'Please respond as soon as possible tomorrow morning.'"

[1140] Terminal: Display the alert to the user as a popup.

[1141] User: Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[1142] Device: Send the corrected message data and sending timing to the server again.

[1143] Server: Re-analyze the data and make sure there are no problems.

[1144] Server: The next morning, the corrected message is automatically sent to the recipient.

[1145] Example prompt sentence:

[1146] "A user types and attempts to send a message at 2 AM saying 'This is urgent, please respond immediately!' Analyze the appropriateness of this message and generate a prompt suggesting a correction."

[1147] The present invention is a system that prevents a user from unknowingly sending a message containing inappropriate content and supports sending messages at an appropriate time.

[1148] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1149] Step 1: User types and sends a text message

[1150] User: Enter a text message via email, chat, SNS, etc. For example, enter the message "Please send the meeting materials immediately."

[1151] Input: The text message entered by the user.

[1152] Output: Text data and meta information (user ID, sending time, etc.) ready to be sent to the device

[1153] Step 2: The device sends the text data to the server

[1154] Terminal: After receiving user input, it detects when the send button is pressed and sends the text data and meta information to the server using the HTTP protocol.

[1155] Input: Text data and meta information entered by the user

[1156] Output: Text data and meta information sent to the server

[1157] Step 3: The server preprocesses the text data

[1158] Server: Preprocesses the received text data. This preprocessing includes tokenizing the text, removing unnecessary characters and tags, and simple grammar analysis. This is done using a Python natural language processing library (NLTK, spaCy, etc.).

[1159] Input: Text data and meta information

[1160] Output: Preprocessed text data

[1161] Step 4: The server uses the generative model to analyze the text data

[1162] Server: Preprocessed text data is passed to a generative AI model (e.g., GPT-3, BERT) that understands the context and detects inappropriate content and harassment.

[1163] Input: Preprocessed text data

[1164] Output: Detection results and a score indicating their importance

[1165] Step 5: The server generates an alert message or suggestion

[1166] Server: Generates appropriate alert messages and suggestions based on the analysis results of the generative model. For example, it generates a message such as, "This message contains imperative language. We recommend revising it to 'Could you please send the meeting materials as soon as possible?'"

[1167] Input: Detection results and scoring

[1168] Output: Alert message and suggestions

[1169] Step 6: The device displays an alert message or suggestion to the user.

[1170] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format using a front-end framework (e.g., React, Vue.js).

[1171] Input: Alert message and suggestions

[1172] Output: Alerts and suggestions displayed in the user interface

[1173] Step 7: User reviews the suggested changes and edits the message

[1174] User: Based on the alerts and suggestions displayed, revise the message, for example, to "Could you please send the meeting materials as soon as possible?"

[1175] Input: The displayed alert message and suggestion

[1176] Output: The modified text message

[1177] Step 8: The device resends the corrected message

[1178] Terminal: If the user clicks the resend button after correcting the message, the corrected message data is sent to the server again.

[1179] Input: Modified text message

[1180] Output: The correction message sent to the server.

[1181] Step 9: The server reparses the modified message and performs a final check

[1182] Server: Re-parses the modified message to ensure there are no problems. This re-parse is done using the generative model, just like the initial parse.

[1183] Input: Modified text message

[1184] Output: Final confirmation result of the message after correction

[1185] Step 10: Server sends message after final confirmation

[1186] Server: After verifying that there are no problems, it sends the corrected message to the actual recipient, using the SMTP protocol or API to connect to email or messaging systems.

[1187] Input: Correction message that has been finalized

[1188] Output: The message sent to the actual recipient

[1189] (Application example 1)

[1190] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1191] The present invention aims to realize safe and secure communication by effectively detecting inappropriate content and harassment in text-based communication and suggesting appropriate expressions and sending times to users. It also provides a system that reanalyzes messages corrected based on the suggestions, ultimately transmitting problem-free text data.

[1192] 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.

[1193] In this invention, the server includes means for receiving text data entered by a user, analysis means using a generative model to analyze the received text data and detect inappropriate content or content that constitutes harassment, means for suggesting alert messages or appropriate expressions to the user based on the analysis results, means for reanalyzing the text data corrected based on the suggested content and reconfirming its appropriateness, and means for transmitting the corrected text data to the final recipient. This enables safe and secure communication by automatically checking the appropriateness of messages entered by users and suggesting corrections as necessary.

[1194] "User" means the entity that uses the system and inputs and transmits text.

[1195] "Text data" refers to character information entered and transmitted by a user.

[1196] The "receiving means" is a device or program that has the function of incorporating text data sent from a user into the system.

[1197] "Analysis means" refers to a device or program for analyzing text data to detect inappropriate expressions or harassment.

[1198] A "generative model" is an algorithm or program that uses natural language processing technology to analyze the content of text data.

[1199] The "suggestion means" is a device or program that has the function of providing the user with specific alerts and examples of corrections based on the analysis results.

[1200] The "reanalysis means" is a device or program having the function of reanalyzing text data that has been corrected by the user based on the suggestion and confirming its appropriateness.

[1201] The "transmission means" is a device or program having a function for transmitting the final confirmed text data to the recipient.

[1202] An "alert message" is a message that warns the user when the analysis means detects inappropriate content or harassment.

[1203] "Appropriate language" refers to language that is respectful to the recipient and does not contain inappropriate or harassing content.

[1204] The system for implementing this invention consists of a server connected to the Internet and a user terminal. The user inputs and sends text messages using a terminal such as a smartphone or PC. The terminal includes an interface for receiving the input text data and transmitting it to the server.

[1205] The server performs various processes to analyze the received text data. Specifically, the server preprocesses the received message and passes it to a generative model for contextual analysis. The generative model can be the Transformers library from Hugging Face. Specifically, a natural language processing model (e.g., BERT or GPT-3) is used.

[1206] The generative model analyzes text data to detect inappropriate content and harassment. At this stage, the analysis results are scored and the problem areas are identified based on the score. For example, a message such as "This is an emergency and please act immediately!" is deemed to be imperative and inappropriate.

[1207] Based on the analysis results, the server makes appropriate suggestions to the user. The suggestions are specific and are displayed in the form of, for example, "This message contains imperative language. We recommend that you correct it as follows: 'Please respond as soon as possible.'"

[1208] The user reviews the suggestions and modifies the message. The modified message is then sent back to the server for re-analysis. This re-analysis ultimately verifies the message is clean. After verification, the modified message is finally sent to the recipient.

[1209] As a concrete example, consider the case of an imperative message sent late at night. A user types the message "This is urgent, please respond immediately!" at 2 a.m. and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message and detects that it contains the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night. The server generates an alert stating, "This message was sent late at night and contains an imperative expression. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'" The device displays the alert to the user as a pop-up. The user edits the message to "Please respond as soon as possible tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message again to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The next morning, the server automatically sends the message without any problems to the recipient.

[1210] An example of a prompt is as follows:

[1211] "Analyze the following sentence and determine if it is an imperative: 'Please respond immediately!'"

[1212] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1213] Step 1:

[1214] The user enters a text message into the terminal and presses the send button. At this time, the terminal prepares the entered text data and meta information (user ID, sending time, etc.). The terminal sends the text data and meta information to the server. The input is the "text message" and "meta information," and the output is the "text data and meta information sent to the server."

[1215] Step 2:

[1216] The server preprocesses the text data it receives. Specifically, it tokenizes and filters the text data. The input is the received text data and meta information, and the output is the preprocessed text data.

[1217] Step 3:

[1218] The server passes the preprocessed text data to a generative model for contextual analysis. Specifically, it uses a natural language processing model based on the Hugging Face Transformers library (e.g., BERT or GPT-3) to analyze inappropriate language and harassment in the text. The input is the preprocessed text data, and the output is the analysis results (scores for inappropriate language and harassment).

[1219] Step 4:

[1220] The server detects inappropriate content or harassment based on the analysis results of the generative model and generates an alert message or correction suggestions. The input is the "analysis results (scores for inappropriate expressions and harassment)" and the output is the "alert message or correction suggestions." Specifically, the message generated is "This message contains imperative language. We recommend correcting it as follows: 'Please respond as soon as possible.'"

[1221] Step 5:

[1222] The server sends an alert message or correction suggestion to the terminal, which then displays it to the user in a popup format. The input is the "alert message or correction suggestion" and the output is the "displayed alert message or correction suggestion." The user then corrects the message based on the suggested suggestions.

[1223] Step 6:

[1224] The user resends the corrected text message, and the terminal sends it to the server. The input is the "corrected text message" and the output is the "corrected text data sent to the server."

[1225] Step 7:

[1226] The server re-analyzes the corrected text message and performs a final check. The input is the "corrected text data" and the output is the "final analysis result." It checks to see if there are any problems.

[1227] Step 8:

[1228] The server sends the verified modified text message to the final recipient. The input is the "verified modified text message" and the output is the "text message sent to the final recipient".

[1229] 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.

[1230] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining this system with an emotion engine, it is possible to realize more accurate alerts and suggestions based on the user's emotions.

[1231] System Configuration

[1232] 1. Device:

[1233] A device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text.

[1234] 2. Server:

[1235] It receives text data sent from the device and analyzes it using a generative model and emotion engine.

[1236] 3. Generative Model:

[1237] This algorithm uses natural language processing technology to analyze text data, understand context and meaning, and detect inappropriate content and harassment.

[1238] 4. Emotion Engine:

[1239] This is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, and neutral as analysis results.

[1240] Program processing

[1241] Enter and send a message

[1242] User:

[1243] The user types the text message "You're really useless" into a chat app.

[1244] Device:

[1245] When the send button is pressed, the terminal sends the text data and related meta information (user ID, sending time, etc.) to the server.

[1246] Message Parsing

[1247] server:

[1248] Preprocessing the received text data before passing it to the generative model includes tokenizing the message and escaping any necessary characters.

[1249] Generative model:

[1250] The system analyzes tokenized text data to detect negative content such as "unusable," and then scores it to quantify its importance.

[1251] emotion recognition

[1252] Emotion Engine:

[1253] It analyzes the user's emotions from text data and detects negative emotions, such as "anger."

[1254] Alerts and Suggestions

[1255] server:

[1256] The inappropriate content detection results from the generative model and the analysis results from the emotion engine are integrated to generate alert messages and suggestions.

[1257] For example, it generates a suggestion such as, "This message contains negative emotions and contains inappropriate language. Please try revising it to 'It would be helpful if you could provide specific feedback.'"

[1258] Device:

[1259] Displays generated alert and suggestion messages to the user in a popup format, allowing the user to review the message before sending.

[1260] User Selection

[1261] User:

[1262] Review the alert or suggestion and choose to modify the text message to "Specific feedback would be helpful" or send it as is.

[1263] Resend and final confirmation of the revised message

[1264] User:

[1265] If you have corrected the message, click the resend button.

[1266] Device:

[1267] The corrected text data is sent to the server again.

[1268] server:

[1269] Re-analyze the corrected text data to verify that the problem has been resolved.

[1270] Sending a message

[1271] server:

[1272] After final confirmation, the text data is sent to the actual recipient without any problems.

[1273] Message storage

[1274] server:

[1275] The sent message and its analysis results are recorded in a log and saved for future reference.

[1276] Specific examples

[1277] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[1278] User:

[1279] "At 2 a.m., I type a message saying, 'Deal with this right now,' and press the send button."

[1280] Device:

[1281] Sends text and meta information to the server.

[1282] server:

[1283] Receives messages, preprocesses them, and passes them to the generative model.

[1284] Generative model:

[1285] The message is analyzed to detect whether it is an "imperative expression" and is scored.

[1286] Emotion Engine:

[1287] Detects the emotion "anger" and sets a negative emotion score.

[1288] server:

[1289] Generate an alert saying, "This message was sent late at night and contains commanding language and angry sentiment. We recommend sending it the next morning and amending it to 'Please respond by tomorrow morning.'"

[1290] Device:

[1291] The alert is displayed to the user in a popup.

[1292] User:

[1293] Change the message to "Please respond by tomorrow morning" and set it to be sent the next morning.

[1294] Device:

[1295] The corrected message is sent back to the server for final confirmation.

[1296] server:

[1297] Re-analyze and check that there are no problems.

[1298] server:

[1299] The next morning, a clean message is automatically sent to the recipient. This invention provides a technological means to prevent harassment and promote better communication.

[1300] The processing flow will be explained below.

[1301] Step 1:

[1302] User: The user types a text message to send via email, chat, or social media.

[1303] Step 2:

[1304] User: Checks the message and clicks the send button.

[1305] Step 3:

[1306] Device: Sends the text message and related meta information (user ID, time sent, etc.) to the server.

[1307] Step 4:

[1308] Server: Preprocesses the received text messages and prepares them for the generative model. Preprocessing includes tokenizing the messages and normalizing the input data.

[1309] Step 5:

[1310] Generative Model: A generative model analyzes tokenized text data, understands the context, and detects inappropriate content and harassment. It then assigns a score to the inappropriate content.

[1311] Step 6:

[1312] Emotion Engine: The emotion engine analyzes the user's emotions from text data. For example, it detects emotional states such as "anger," "sadness," and "joy" and assigns a score to each.

[1313] Step 7:

[1314] Server: Integrates the analysis results of the generative model and the emotion engine to generate alert messages and suggestion messages. For example, if negative expressions and the emotion "anger" have high scores, the server will suggest corrections.

[1315] Step 8:

[1316] Server: Sends generated alert and suggestion messages to the device.

[1317] Step 9:

[1318] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[1319] Step 10:

[1320] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[1321] Step 11:

[1322] User: If you have corrected the message, click the resend button.

[1323] Step 12:

[1324] On your device: Send the modified text message back to the server.

[1325] Step 13:

[1326] Server: Re-parse the modified text message to verify that the issue has been resolved.

[1327] Step 14:

[1328] Server: After final verification, the text message is sent to the intended recipient.

[1329] Step 15:

[1330] Server: Logs the sent message and its analysis results and stores them for future reference.

[1331] Example 2

[1332] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1333] Inappropriate language and harassment are commonplace in modern communication tools, placing a heavy burden on users. Effective technology to prevent these issues has yet to be fully established. Furthermore, few systems can provide alerts or suggestions based on users' emotions, which can lead to unintended misunderstandings and conflicts. To address these issues, a system is needed that uses more advanced natural language processing technology to analyze users' message content and emotions and provide appropriate alerts and suggestions.

[1334] 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.

[1335] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data, means for analyzing the preprocessed text data using a generative model to detect inappropriate content or content that constitutes harassment, means for analyzing the received text data using an emotion engine that recognizes the user's emotions, means for suggesting an alert message or appropriate expression to the user based on the analysis results of the generative model and the emotion engine, means for sending the text data with the problem resolved to the actual recipient, and means for recording the sent message and the analysis results in a log. This makes it possible to prevent harassment and promote smooth communication by detecting in advance whether a user's message is inappropriate and what emotion it contains, and correcting it to an appropriate expression.

[1336] A "User" is an individual or entity that utilizes the system to enter and send text messages.

[1337] "Text data" is digital data of character strings including sentences and words entered by the user.

[1338] "Terminal" refers to a device that allows a user to input and send text messages, including a PC, smartphone, tablet, etc.

[1339] "Server" refers to a central management device that receives, analyzes, and stores text data sent from devices and generates alerts and suggestions.

[1340] "Preprocessing" refers to a preparatory process such as tokenizing and escaping data that is performed before the server analyzes the text data.

[1341] A "generative model" is an algorithm that uses natural language processing technology to analyze text data and detect inappropriate content or harassment.

[1342] An "emotion engine" is an algorithm that recognizes a user's emotions from text data and detects emotions such as positive, negative, or neutral.

[1343] An "alert message" is a warning message that the server sends to the user based on the analysis results of the generative model and emotion engine to inform them of the possibility of inappropriate content or harassment.

[1344] A "suggestion" is a message that the server sends to the user, based on the analysis results of the generative model and emotion engine, instructing them on more appropriate expressions and actions.

[1345] A "log" is a collection of data that records sent messages and their analysis results and is saved for later reference.

[1346] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining an emotion engine and a generative model, this system can realize highly accurate alerts and suggestions based on the user's emotions.

[1347] System Configuration

[1348] 1. Device:

[1349] A terminal is a device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text. Terminals include PCs, smartphones, and tablets.

[1350] 2. Server:

[1351] The server receives text data sent from the device and analyzes it using a generative model and emotion engine. The server also generates alert messages and suggestions based on the analysis results and sends them to the device to notify the user.

[1352] 3. Generative Model:

[1353] A generative model is an algorithm that uses natural language processing techniques to analyze text data, understand context and meaning, and detect inappropriate content and harassment. Specific techniques used include BERT and GPT-3.

[1354] 4. Emotion Engine:

[1355] An emotion engine is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, neutral, etc. Sentiment Analysis API and Emotion Detection Model are used as emotion engines.

[1356] Program processing overview

[1357] User message entry and sending

[1358] The user types the text message "You're really useless" into the chat app, and once this action is completed, the device sends the message to the server.

[1359] Message analysis and emotion recognition

[1360] The server passes the received text data to the generative model for preprocessing, which includes tokenization and escaping. The generative model then analyzes the preprocessed data and detects the inappropriate expression "useless." At the same time, the emotion engine analyzes the user's emotions and detects the emotion "anger."

[1361] Alert generation and user notification

[1362] Based on the analysis results, the server generates an alert message and a suggestion, such as "This message contains negative sentiment and contains inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The device then notifies the user of this message.

[1363] Specific examples

[1364] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[1365] At 2 a.m., the user types a message saying, "Please take care of this now," and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message, detects that it is an "imperative expression," and scores it. The emotion engine detects the emotion of "anger" and assigns a negative emotion score. The server generates an alert stating, "This message was sent late at night and contains imperative expressions and the emotion of anger. We recommend sending it the next morning and revising it to 'Please take care of this by tomorrow morning.'" The device displays the alert as a pop-up to the user. The user edits the message to "Please take care of this by tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message back to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The server automatically sends the clean message to the recipient the following morning.

[1366] This system will help prevent harassment and promote better communication.

[1367] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1368] Step 1:

[1369] User:

[1370] The user types the text message "You're really useless" into a chat app. The input is made to the device, and text data is generated.

[1371] Step 2:

[1372] Device:

[1373] When the user presses the send button, the text data and related metadata (user ID, time of sending, etc.) are sent to the server. The input here is the text data and metadata, and this is the output sent to the server.

[1374] Step 3:

[1375] server:

[1376] The server receives data sent from the terminal. It preprocesses the received data (input). Preprocessing includes tokenization (for example, breaking down "You really can't use it" into "o," "mae," "ha," "hontoku," and "tsukanenai") and character escaping (removing certain special characters). The preprocessed text data is the output.

[1377] Step 4:

[1378] server:

[1379] The preprocessed text data is passed as input to a generative model. The generative model then analyzes the text data to detect inappropriate content and harassment. The analysis involves understanding the context and extracting keywords. For example, negative expressions such as "useless" are detected. The analysis results output as inappropriate content detection results.

[1380] Step 5:

[1381] server:

[1382] Based on the analysis results from the generative model, the server passes the text data as input to the emotion engine. The emotion engine analyzes the user's emotion from the text data. For example, the emotion "anger" is detected. The emotion analysis results output the type of emotion and its intensity.

[1383] Step 6:

[1384] server:

[1385] The analysis results of the generative model and the emotion engine are integrated. Based on this integration result, the server generates an alert message and a suggestion message. For example, "This message contains negative emotions and appears to contain inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The generated alert message and suggestion message are output.

[1386] Step 7:

[1387] Device:

[1388] Receives alert messages and suggestions sent from the server and displays them to the user in a popup format. The displayed alert messages and suggestions become input information for the user to confirm. The displayed content becomes the output.

[1389] Step 8:

[1390] User:

[1391] The user reviews the alerts and suggestions and chooses whether to revise the message or send it as is. For example, they might revise it to "Specific feedback would be helpful." This choice is the input, and the revised message is the output.

[1392] Step 9:

[1393] User:

[1394] If the user modifies the message, he or she presses the resend button.

[1395] Step 10:

[1396] Device:

[1397] The corrected text data is sent to the server again. The sent corrected data becomes the input and the output sent to the server.

[1398] Step 11:

[1399] server:

[1400] The server receives the corrected text data again and re-analyzes it using the generative model. As before, it performs preprocessing, analysis, and emotion recognition, and then verifies that the problem has been solved. The verification result is output.

[1401] Step 12:

[1402] server:

[1403] The text data that has been confirmed to be problem-free is sent to the actual recipient. The sent text data becomes the output.

[1404] Step 13:

[1405] server:

[1406] The sent message and its analysis results are logged for future reference. This log is the output.

[1407] In this way, the system can provide appropriate analysis and suggestions for users' messages, helping to prevent harassment and promote smooth communication.

[1408] (Application example 2)

[1409] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1410] In recent years, with the development of communication tools, there has been an increase in inappropriate language and harassment via text messages. Such problems can cause psychological stress for the recipient and, in some cases, legal issues. Furthermore, senders themselves may send misleading messages in an emotionally charged state, which can lead to future problems. There is a need to prevent these from happening in advance and ensure healthy and safe communication.

[1411] The identification process by the identification 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 receiving text data entered by a user; analysis means using a generative model for analyzing the received text data to detect inappropriate content or content that constitutes harassment; means for suggesting alert messages or appropriate expressions to the user based on the analysis results; emotion recognition means using an emotion engine for detecting the user's emotional state based on the analysis results; and means for generating a specific correction suggestion message based on the inappropriate content, harassment, and emotional state. This allows users to prevent inappropriate messages from being sent in advance and maintain healthy and smooth communication.

[1412] "User" means any person or entity that uses the System to input and send text messages.

[1413] "Text data" refers to the text information entered by the user, and is text that is subject to detection of inappropriate expressions and harassment.

[1414] The "receiving means" refers to a function or device for transmitting text data entered by a user to a server and receiving the data.

[1415] "Analysis means" refers to functions or algorithms that analyze received text data using a generative model to detect inappropriate content or harassment.

[1416] A "generative model" is a model that uses natural language processing technology to analyze text data and understand context and meaning.

[1417] An "alert message" is a message that warns the user based on the analysis results.

[1418] "Means for suggesting" refers to functions and algorithms for suggesting alert messages and appropriate expressions to users.

[1419] "Emotion recognition means" refers to functions and algorithms for detecting the user's emotional state based on the analysis results and analyzing that state.

[1420] An "emotion engine" is a system that includes an algorithm that recognizes a user's emotions from text data and detects emotional states such as positive, negative, or neutral.

[1421] A "correction suggestion message" is a message that provides the user with specific suggestions for correction based on inappropriate content, harassment, and emotional state.

[1422] The present invention aims to enable users to communicate safely and appropriately by providing a system for analyzing the content of text messages and the emotions behind them in real time. Hereinafter, an embodiment of the present invention will be described in detail.

[1423] Program processing

[1424] This system consists of a terminal where users can input and send text messages, and a server that analyzes those messages. Specifically, the system uses the following hardware and software:

[1425] Hardware

[1426] Device: Any device that allows a user to type text messages, such as a smartphone, tablet, or PC.

[1427] Server: The server for receiving and analyzing messages is run on Apache or Nginx and uses a Python-based framework (Flask or Django).

[1428] software

[1429] Generative models: Generative AI models that use natural language processing techniques, such as GPT-3 and BERT, to analyze text data and detect inappropriate content and harassment.

[1430] Emotion engine: Detects the user's emotional state from text using technologies such as IBM Watson Natural Language Understanding (NLU).

[1431] Database: A database (e.g., MySQL, PostgreSQL) for storing analysis results and transmission history.

[1432] Data processing and calculation

[1433] Message input: The user types a text message into the device, for example, "Why are you always late!"

[1434] Data transmission: The entered text data and metadata (user ID, submission time, etc.) are sent to the server.

[1435] Preprocessing: The server tokenizes the received text data and performs any necessary character escaping.

[1436] Analysis: Preprocessed text data is fed into a generative model to analyze context and meaning, detecting inappropriate content and harassment.

[1437] Emotion recognition: The emotion engine detects the user's emotional state (e.g., "anger") from text data.

[1438] Alerts and Suggestions: Based on the analysis results, appropriate alert messages and suggestions for correction are generated and notified to the user.

[1439] Save: The sent message, its analysis results, and its revision history are saved in the database.

[1440] Specific examples

[1441] As a concrete example, consider the situation where a user tries to enter the following message:

[1442] "Why are you always late?"

[1443] When attempting to send this message, the generative AI model is prompted with the following:

[1444] Example prompt:

[1445] Message: "Why are you always late?"

[1446] Emotional state: "Anger"

[1447] Purpose: "Detect profanity and provide correction suggestions"

[1448] The generated alert and suggestion will notify the user, for example, "This message contains angry emotions and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late.'" By revising the message based on this suggestion, the user can achieve more appropriate communication.

[1449] The above has described in detail the embodiments of the present invention.

[1450] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1451] Step 1:

[1452] The user inputs a text message into the terminal, for example, "Why are you always late?" This input data is passed to subsequent processing.

[1453] Step 2:

[1454] The text message is sent from the device to the server, along with metadata such as the user ID and the time of sending, allowing the server to determine who is sending what message at what time.

[1455] Step 3:

[1456] The server preprocesses the text data it receives. Specifically, it tokenizes the text data and performs necessary character escaping. This preprocessing prepares the data in a format suitable for analysis. The input is text data and metadata, and the output is the preprocessed tokenized data.

[1457] Step 4:

[1458] The preprocessed tokenized text data is passed to a generative model. The server uses the generative model to analyze this data and detect inappropriate content and harassment while understanding the context and meaning. The input is the tokenized data, and the output is the detection results for inappropriate content and harassment.

[1459] Step 5:

[1460] The server uses an emotion engine to detect the user's emotional state from the tokenized text data. For example, emotional states such as "anger" or "joy" are analyzed. The input is the tokenized data, and the output is the detected emotional state.

[1461] Step 6:

[1462] The server combines the analysis results of the generative model with the results of the emotion engine, and generates a specific correction suggestion message based on the inappropriate content and emotional state. For example, if the analysis results determine that "anger is detected and contains inappropriate language," it generates a suggestion such as "This message contains anger and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late,' and send it." The input is the detection result and emotional state, and the output is a correction suggestion message.

[1463] Step 7:

[1464] The server sends a correction suggestion message to the terminal and notifies the user. The user checks the suggestion message and makes appropriate corrections. The input is the correction suggestion message, and the output is the text message corrected by the user.

[1465] Step 8:

[1466] If the user resubmits the corrected message, it is resent from the terminal to the server. The server reanalyzes it and verifies that there are no problems. The input is the corrected message, and the output is the analysis result.

[1467] Step 9:

[1468] The server finally sends the corrected text message to the recipient and saves the sending history in a database for future reference. The input is the corrected message and the analysis result, and the output is the sending history.

[1469] This series of processing steps allows users to prevent sending inappropriate messages in advance and maintain healthy and smooth communication.

[1470] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1471] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1472] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1473] [Fourth embodiment]

[1474] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1475] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1476] 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).

[1477] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1478] 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.

[1479] 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).

[1480] 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.

[1481] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1482] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1483] 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.

[1484] 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.

[1485] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1486] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1487] This system analyzes text-based communications, detects inappropriate content and harassment in advance, and suggests appropriate expressions and sending times to users. The specific operation and functions of this system are explained below.

[1488] System Configuration

[1489] 1. Device:

[1490] A device that users use to enter text-based messages such as email, chat, and social networking sites. The terminal includes an interface for receiving user input and sending it to a server.

[1491] 2. Server:

[1492] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[1493] 3. Generative Model:

[1494] It is an algorithm that uses natural language processing technology to analyze text data, understand the context, and detect inappropriate expressions and harassment.

[1495] Program processing

[1496] Enter and send a message

[1497] User:

[1498] A user inputs a text message into an email, chat, social networking site, etc. For example, the user inputs a message such as "Please send the meeting materials immediately."

[1499] Device:

[1500] When the user presses the send button after entering a message, the device sends the text data and related meta-information (user ID, sending time, etc.) to the server.

[1501] Message Parsing

[1502] server:

[1503] The server preprocesses the received text data and prepares it for the generative model, including tokenizing and filtering the message.

[1504] Generative model:

[1505] Generative models analyze text data to understand context and detect inappropriate content or harassment. The results of this analysis are scored according to importance and identify specific issues.

[1506] Alerts and Suggestions

[1507] server:

[1508] Based on the analysis results of the generative model, appropriate alert messages and suggestions are generated. For example, if the analysis results receive a high score, a specific suggestion such as "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'" is generated.

[1509] Device:

[1510] Alert messages and suggestions sent by the server are displayed to the user in a popup format, allowing the user to review the message before sending it.

[1511] User Selection

[1512] User:

[1513] The user can review the alerts and suggestions and choose to modify the message or send it as is. For example, they can modify the message to "Could you please send the meeting materials as soon as possible?" and resend it.

[1514] Resend and final confirmation of the revised message

[1515] Device:

[1516] The corrected message is sent back to the server for final confirmation.

[1517] server:

[1518] Re-analyze the correction message and finally make sure there are no problems.

[1519] Sending a message

[1520] server:

[1521] After a final check, the resolved message is sent to the actual recipient.

[1522] Specific examples

[1523] As a concrete example, consider the case of an "commanding message sent late at night."

[1524] User:

[1525] At 2 a.m., type the message "This is an emergency and I need your help immediately!" and press the send button.

[1526] Device:

[1527] Sends text and meta information to the server.

[1528] server:

[1529] Receives messages, preprocesses them, and passes them to the generative model.

[1530] Generative model:

[1531] The message is analyzed to detect that it contains the imperative phrase "This is urgent, please respond immediately!" and that it was sent late at night.

[1532] server:

[1533] Generate an alert such as, "This message was sent late at night and contains imperative language. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'"

[1534] Device:

[1535] The alert is displayed to the user in a popup.

[1536] User:

[1537] Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[1538] Device:

[1539] The corrected message is sent back to the server for final confirmation.

[1540] server:

[1541] Re-analyze and check that there are no problems.

[1542] server:

[1543] The next morning, a clean message will be automatically sent to the recipient.

[1544] This invention aims to prevent harassment and inappropriate language in advance, and provides a concrete means for improving the quality of communication.

[1545] The processing flow will be explained below.

[1546] Step 1:

[1547] User: The user types a text message to send via email, chat, or social media.

[1548] Step 2:

[1549] User: Checks the message and clicks the send button.

[1550] Step 3:

[1551] Terminal: Sends text data and related meta-information (user ID, sending time, etc.) to the server.

[1552] Step 4:

[1553] Server: Preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message and spell checking.

[1554] Step 5:

[1555] Generative model: Analyzes tokenized text data and detects inappropriate content and harassment while understanding the context. The analysis results are scored and quantified to indicate their importance.

[1556] Step 6:

[1557] Server: Generates alert messages and suggestion messages based on the analysis results of the generative model. For example, if the score is high, it generates suggestions including specific corrections.

[1558] Step 7:

[1559] Server: Sends alert and suggestion messages to the device.

[1560] Step 8:

[1561] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[1562] Step 9:

[1563] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[1564] Step 10:

[1565] User: If you have corrected the message, click the resend button.

[1566] Step 11:

[1567] Terminal: Send the corrected text data to the server again.

[1568] Step 12:

[1569] Server: Re-parse the corrected text data and verify that the problem is resolved.

[1570] Step 13:

[1571] Server: After final confirmation, the text data is sent to the actual recipient without any problems.

[1572] Step 14:

[1573] Server: Logs the sent message and its analysis results and stores them for future reference.

[1574] Example 1

[1575] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1576] A problem with conventional text-based communication is that users can unknowingly send messages containing inappropriate content or harassment. Furthermore, sending messages at inappropriate times, such as late at night or early in the morning, can cause discomfort to the recipient. A system that addresses these issues and promotes healthy communication was needed.

[1577] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1578] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data and passing it to a generative AI model, means for analyzing the received text data using the generative AI model and detecting inappropriate content or content that constitutes harassment, means for suggesting an alert message or appropriate expression to the user based on the detected analysis results, means for displaying the alert message or suggestion to the user, means for resending and reanalyzing a message corrected by the user, and means for sending the corrected message to the actual recipient after final confirmation. This prevents the user from unknowingly sending a message containing inappropriate content and enables the message to be sent at an appropriate time.

[1579] "User" means a person who uses the system to enter and send text messages.

[1580] "Means for receiving" is a general term for hardware and software for transmitting text data entered by a user to a server and receiving it.

[1581] "Preprocessing" refers to a series of steps performed on received text data before it is passed to a generative AI model, including tokenization, removal of unnecessary characters, and grammatical analysis.

[1582] A "generative AI model" is an algorithm that uses natural language processing technology to analyze text data, understand context, and detect inappropriate content or harassment.

[1583] "Means for analysis" refers collectively to software and hardware used to analyze received text data using a generative AI model and identify inappropriate content or expressions that constitute harassment.

[1584] An "alert message" is a message that warns the user that the content contains inappropriate content or expressions based on the analysis results of the generative AI model.

[1585] "Appropriate expressions" are expressions that suggest more appropriate and desirable wording to the user based on the problems detected by the generative AI model.

[1586] "Proposed means" is a general term for software and hardware that generates alert messages and appropriate expressions based on the analysis results and presents them to the user.

[1587] "Means for retransmission" is a general term for software and hardware that has the function of sending a message corrected by the user back to the server for reanalysis.

[1588] "Means for sending after final confirmation" refers to software and hardware that has the functionality to reanalyze the modified message, confirm that there are no problems, and then send the message to the actual recipient.

[1589] This invention relates to a system that analyzes text data entered by a user, detects inappropriate content or harassment in advance, and suggests appropriate expressions and sending times. This system improves the appropriateness of messages by performing various analyses during the process of a user entering a text message via email, chat, SNS, etc. and sending it.

[1590] System Configuration and Hardware

[1591] 1. Device:

[1592] A device used by users to input text-based messages such as emails, chats, and social media. Typical devices include PCs, smartphones, and tablets.

[1593] The device includes an interface for receiving user input and sending it to a server, often using a front-end framework such as React or Vue.js for the user interface.

[1594] 2. Server:

[1595] It analyzes the received text data and uses a generative model to detect inappropriate content and harassment.

[1596] Natural language processing libraries used include the Python-based NLTK and spaCy. Preprocessing involves tokenizing and filtering messages.

[1597] Transformer-based models (e.g., GPT-3, BERT) are used as generative AI models.

[1598] 3. Generative AI Model:

[1599] This is an algorithm that uses natural language processing technology to analyze text data, understand the context, and then detect inappropriate language and harassment. The generative AI model is capable of highly accurate analysis and can accurately grasp the context.

[1600] System Operation

[1601] 1. Enter and send a message:

[1602] User: The user types a text message into email, chat, SNS, etc. For example, they type a message such as "Please send the meeting materials right away."

[1603] Terminal: When the user presses the send button after entering a message, the terminal sends the text data entered by the user and related meta information (user ID, sending time, etc.) to the server. The HTTP protocol is used for sending.

[1604] 2. Message analysis:

[1605] Server: The server preprocesses the received text data and passes it to the generative model. Preprocessing includes tokenizing the message, removing unnecessary characters and tags, and simple grammatical analysis.

[1606] Generative models: Generative models analyze text data and understand context to detect inappropriate content and harassment. Specifically, transformer-based models understand context and identify things like commanding tone, unnecessary emphasis, and discriminatory language.

[1607] 3. Alert generation and suggestions:

[1608] Server: Generates appropriate alert messages and suggestions based on the analysis results obtained from the generative model. For example, it generates specific suggestions such as, "This message contains imperative language. We recommend revising it as follows: 'Could you please send the meeting materials as soon as possible?'"

[1609] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format.

[1610] 4. User Choice:

[1611] User: The user can review the message based on the pop-up alerts and suggestions, and choose to apply the suggested changes or send it as is. For example, they can change it to "Could you please send the meeting materials as soon as possible?" and resend it.

[1612] 5. Resend and final confirmation of the corrected message:

[1613] Terminal: If the user presses the button to resend the corrected message, the terminal will resend the corrected text data to the server.

[1614] Server: Re-analyze the corrected message and finally check that there are no problems.

[1615] 6. Sending a message:

[1616] Server: Once all is well, the server sends the message to the actual recipient, typically using the SMTP protocol or an API to connect to an email or messaging system.

[1617] Specific examples

[1618] As a concrete example, consider an imperative message sent late at night.

[1619] User: At 2 a.m., type the message "This is urgent, please respond immediately!" and press the send button.

[1620] Terminal: The text data entered by the user and meta information (user ID, time of submission, etc.) are sent to the server via the HTTP protocol.

[1621] Server: Receives messages, tokenizes and filters them using NLTK, and performs parsing with spaCy.

[1622] Generative model: We use GPT-3 to analyze the message and detect the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night.

[1623] Server: Generate an alert like, "This message was sent late at night and contains an imperative. We recommend sending it the next morning and amending it to 'Please respond as soon as possible tomorrow morning.'"

[1624] Terminal: Display the alert to the user as a popup.

[1625] User: Change the message to "Please respond as soon as possible tomorrow morning" and set it to be sent the next morning.

[1626] Device: Send the corrected message data and sending timing to the server again.

[1627] Server: Re-analyze the data and make sure there are no problems.

[1628] Server: The next morning, the corrected message is automatically sent to the recipient.

[1629] Example prompt sentence:

[1630] "A user types and attempts to send a message at 2 AM saying 'This is urgent, please respond immediately!' Analyze the appropriateness of this message and generate a prompt suggesting a correction."

[1631] The present invention is a system that prevents a user from unknowingly sending a message containing inappropriate content and supports sending messages at an appropriate time.

[1632] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1633] Step 1: User types and sends a text message

[1634] User: Enter a text message via email, chat, SNS, etc. For example, enter the message "Please send the meeting materials immediately."

[1635] Input: The text message entered by the user.

[1636] Output: Text data and meta information (user ID, sending time, etc.) ready to be sent to the device

[1637] Step 2: The device sends the text data to the server

[1638] Terminal: After receiving user input, it detects when the send button is pressed and sends the text data and meta information to the server using the HTTP protocol.

[1639] Input: Text data and meta information entered by the user

[1640] Output: Text data and meta information sent to the server

[1641] Step 3: The server preprocesses the text data

[1642] Server: Preprocesses the received text data. This preprocessing includes tokenizing the text, removing unnecessary characters and tags, and simple grammar analysis. This is done using a Python natural language processing library (NLTK, spaCy, etc.).

[1643] Input: Text data and meta information

[1644] Output: Preprocessed text data

[1645] Step 4: The server uses the generative model to analyze the text data

[1646] Server: Preprocessed text data is passed to a generative AI model (e.g., GPT-3, BERT) that understands the context and detects inappropriate content and harassment.

[1647] Input: Preprocessed text data

[1648] Output: Detection results and a score indicating their importance

[1649] Step 5: The server generates an alert message or suggestion

[1650] Server: Generates appropriate alert messages and suggestions based on the analysis results of the generative model. For example, it generates a message such as, "This message contains imperative language. We recommend revising it to 'Could you please send the meeting materials as soon as possible?'"

[1651] Input: Detection results and scoring

[1652] Output: Alert message and suggestions

[1653] Step 6: The device displays an alert message or suggestion to the user.

[1654] Terminal: Alert messages and suggestions sent from the server are displayed to the user in a popup format using a front-end framework (e.g., React, Vue.js).

[1655] Input: Alert message and suggestions

[1656] Output: Alerts and suggestions displayed in the user interface

[1657] Step 7: User reviews the suggested changes and edits the message

[1658] User: Based on the alerts and suggestions displayed, revise the message, for example, to "Could you please send the meeting materials as soon as possible?"

[1659] Input: The displayed alert message and suggestion

[1660] Output: The modified text message

[1661] Step 8: The device resends the corrected message

[1662] Terminal: If the user clicks the resend button after correcting the message, the corrected message data is sent to the server again.

[1663] Input: Modified text message

[1664] Output: The correction message sent to the server.

[1665] Step 9: The server reparses the modified message and performs a final check

[1666] Server: Re-parses the modified message to ensure there are no problems. This re-parse is done using the generative model, just like the initial parse.

[1667] Input: Modified text message

[1668] Output: Final confirmation result of the message after correction

[1669] Step 10: Server sends message after final confirmation

[1670] Server: After verifying that there are no problems, it sends the corrected message to the actual recipient, using the SMTP protocol or API to connect to email or messaging systems.

[1671] Input: Correction message that has been finalized

[1672] Output: The message sent to the actual recipient

[1673] (Application example 1)

[1674] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1675] The present invention aims to realize safe and secure communication by effectively detecting inappropriate content and harassment in text-based communication and suggesting appropriate expressions and sending times to users. It also provides a system that reanalyzes messages corrected based on the suggestions, ultimately transmitting problem-free text data.

[1676] 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.

[1677] In this invention, the server includes means for receiving text data entered by a user, analysis means using a generative model to analyze the received text data and detect inappropriate content or content that constitutes harassment, means for suggesting alert messages or appropriate expressions to the user based on the analysis results, means for reanalyzing the text data corrected based on the suggested content and reconfirming its appropriateness, and means for transmitting the corrected text data to the final recipient. This enables safe and secure communication by automatically checking the appropriateness of messages entered by users and suggesting corrections as necessary.

[1678] "User" means the entity that uses the system and inputs and transmits text.

[1679] "Text data" refers to character information entered and transmitted by a user.

[1680] The "receiving means" is a device or program that has the function of incorporating text data sent from a user into the system.

[1681] "Analysis means" refers to a device or program for analyzing text data to detect inappropriate expressions or harassment.

[1682] A "generative model" is an algorithm or program that uses natural language processing technology to analyze the content of text data.

[1683] The "suggestion means" is a device or program that has the function of providing the user with specific alerts and examples of corrections based on the analysis results.

[1684] The "reanalysis means" is a device or program having the function of reanalyzing text data that has been corrected by the user based on the suggestion and confirming its appropriateness.

[1685] The "transmission means" is a device or program having a function for transmitting the final confirmed text data to the recipient.

[1686] An "alert message" is a message that warns the user when the analysis means detects inappropriate content or harassment.

[1687] "Appropriate language" refers to language that is respectful to the recipient and does not contain inappropriate or harassing content.

[1688] The system for implementing this invention consists of a server connected to the Internet and a user terminal. The user inputs and sends text messages using a terminal such as a smartphone or PC. The terminal includes an interface for receiving the input text data and transmitting it to the server.

[1689] The server performs various processes to analyze the received text data. Specifically, the server preprocesses the received message and passes it to a generative model for contextual analysis. The generative model can be the Transformers library from Hugging Face. Specifically, a natural language processing model (e.g., BERT or GPT-3) is used.

[1690] The generative model analyzes text data to detect inappropriate content and harassment. At this stage, the analysis results are scored and the problem areas are identified based on the score. For example, a message such as "This is an emergency and please act immediately!" is deemed to be imperative and inappropriate.

[1691] Based on the analysis results, the server makes appropriate suggestions to the user. The suggestions are specific and are displayed in the form of, for example, "This message contains imperative language. We recommend that you correct it as follows: 'Please respond as soon as possible.'"

[1692] The user reviews the suggestions and modifies the message. The modified message is then sent back to the server for re-analysis. This re-analysis ultimately verifies the message is clean. After verification, the modified message is finally sent to the recipient.

[1693] As a concrete example, consider the case of an imperative message sent late at night. A user types the message "This is urgent, please respond immediately!" at 2 a.m. and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message and detects that it contains the imperative expression "This is urgent, please respond immediately!" and that it was sent late at night. The server generates an alert stating, "This message was sent late at night and contains an imperative expression. We recommend sending it the next morning and revising it to 'Please respond as soon as possible tomorrow morning.'" The device displays the alert to the user as a pop-up. The user edits the message to "Please respond as soon as possible tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message again to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The next morning, the server automatically sends the message without any problems to the recipient.

[1694] An example of a prompt is as follows:

[1695] "Analyze the following sentence and determine if it is an imperative: 'Please respond immediately!'"

[1696] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1697] Step 1:

[1698] The user enters a text message into the terminal and presses the send button. At this time, the terminal prepares the entered text data and meta information (user ID, sending time, etc.). The terminal sends the text data and meta information to the server. The input is the "text message" and "meta information," and the output is the "text data and meta information sent to the server."

[1699] Step 2:

[1700] The server preprocesses the text data it receives. Specifically, it tokenizes and filters the text data. The input is the received text data and meta information, and the output is the preprocessed text data.

[1701] Step 3:

[1702] The server passes the preprocessed text data to a generative model for contextual analysis. Specifically, it uses a natural language processing model based on the Hugging Face Transformers library (e.g., BERT or GPT-3) to analyze inappropriate language and harassment in the text. The input is the preprocessed text data, and the output is the analysis results (scores for inappropriate language and harassment).

[1703] Step 4:

[1704] The server detects inappropriate content or harassment based on the analysis results of the generative model and generates an alert message or correction suggestions. The input is the "analysis results (scores for inappropriate expressions and harassment)" and the output is the "alert message or correction suggestions." Specifically, the message generated is "This message contains imperative language. We recommend correcting it as follows: 'Please respond as soon as possible.'"

[1705] Step 5:

[1706] The server sends an alert message or correction suggestion to the terminal, which then displays it to the user in a popup format. The input is the "alert message or correction suggestion" and the output is the "displayed alert message or correction suggestion." The user then corrects the message based on the suggested suggestions.

[1707] Step 6:

[1708] The user resends the corrected text message, and the terminal sends it to the server. The input is the "corrected text message" and the output is the "corrected text data sent to the server."

[1709] Step 7:

[1710] The server re-analyzes the corrected text message and performs a final check. The input is the "corrected text data" and the output is the "final analysis result." It checks to see if there are any problems.

[1711] Step 8:

[1712] The server sends the verified modified text message to the final recipient. The input is the "verified modified text message" and the output is the "text message sent to the final recipient".

[1713] 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.

[1714] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining this system with an emotion engine, it is possible to realize more accurate alerts and suggestions based on the user's emotions.

[1715] System Configuration

[1716] 1. Device:

[1717] A device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text.

[1718] 2. Server:

[1719] It receives text data sent from the device and analyzes it using a generative model and emotion engine.

[1720] 3. Generative Model:

[1721] This algorithm uses natural language processing technology to analyze text data, understand context and meaning, and detect inappropriate content and harassment.

[1722] 4. Emotion Engine:

[1723] This is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, and neutral as analysis results.

[1724] Program processing

[1725] Enter and send a message

[1726] User:

[1727] The user types the text message "You're really useless" into a chat app.

[1728] Device:

[1729] When the send button is pressed, the terminal sends the text data and related meta information (user ID, sending time, etc.) to the server.

[1730] Message Parsing

[1731] server:

[1732] Preprocessing the received text data before passing it to the generative model includes tokenizing the message and escaping any necessary characters.

[1733] Generative model:

[1734] The system analyzes tokenized text data to detect negative content such as "unusable," and then scores it to quantify its importance.

[1735] emotion recognition

[1736] Emotion Engine:

[1737] It analyzes the user's emotions from text data and detects negative emotions, such as "anger."

[1738] Alerts and Suggestions

[1739] server:

[1740] The inappropriate content detection results from the generative model and the analysis results from the emotion engine are integrated to generate alert messages and suggestions.

[1741] For example, it generates a suggestion such as, "This message contains negative emotions and contains inappropriate language. Please try revising it to 'It would be helpful if you could provide specific feedback.'"

[1742] Device:

[1743] Displays generated alert and suggestion messages to the user in a popup format, allowing the user to review the message before sending.

[1744] User Selection

[1745] User:

[1746] Review the alert or suggestion and choose to modify the text message to "Specific feedback would be helpful" or send it as is.

[1747] Resend and final confirmation of the revised message

[1748] User:

[1749] If you have corrected the message, click the resend button.

[1750] Device:

[1751] The corrected text data is sent to the server again.

[1752] server:

[1753] Re-analyze the corrected text data to verify that the problem has been resolved.

[1754] Sending a message

[1755] server:

[1756] After final confirmation, the text data is sent to the actual recipient without any problems.

[1757] Message storage

[1758] server:

[1759] The sent message and its analysis results are recorded in a log and saved for future reference.

[1760] Specific examples

[1761] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[1762] User:

[1763] "At 2 a.m., I type a message saying, 'Deal with this right now,' and press the send button."

[1764] Device:

[1765] Sends text and meta information to the server.

[1766] server:

[1767] Receives messages, preprocesses them, and passes them to the generative model.

[1768] Generative model:

[1769] The message is analyzed to detect whether it is an "imperative expression" and is scored.

[1770] Emotion Engine:

[1771] Detects the emotion "anger" and sets a negative emotion score.

[1772] server:

[1773] Generate an alert saying, "This message was sent late at night and contains commanding language and angry sentiment. We recommend sending it the next morning and amending it to 'Please respond by tomorrow morning.'"

[1774] Device:

[1775] The alert is displayed to the user in a popup.

[1776] User:

[1777] Change the message to "Please respond by tomorrow morning" and set it to be sent the next morning.

[1778] Device:

[1779] The corrected message is sent back to the server for final confirmation.

[1780] server:

[1781] Re-analyze and check that there are no problems.

[1782] server:

[1783] The next morning, a clean message is automatically sent to the recipient. This invention provides a technological means to prevent harassment and promote better communication.

[1784] The processing flow will be explained below.

[1785] Step 1:

[1786] User: The user types a text message to send via email, chat, or social media.

[1787] Step 2:

[1788] User: Checks the message and clicks the send button.

[1789] Step 3:

[1790] Device: Sends the text message and related meta information (user ID, time sent, etc.) to the server.

[1791] Step 4:

[1792] Server: Preprocesses the received text messages and prepares them for the generative model. Preprocessing includes tokenizing the messages and normalizing the input data.

[1793] Step 5:

[1794] Generative Model: A generative model analyzes tokenized text data, understands the context, and detects inappropriate content and harassment. It then assigns a score to the inappropriate content.

[1795] Step 6:

[1796] Emotion Engine: The emotion engine analyzes the user's emotions from text data. For example, it detects emotional states such as "anger," "sadness," and "joy" and assigns a score to each.

[1797] Step 7:

[1798] Server: Integrates the analysis results of the generative model and the emotion engine to generate alert messages and suggestion messages. For example, if negative expressions and the emotion "anger" have high scores, the server will suggest corrections.

[1799] Step 8:

[1800] Server: Sends generated alert and suggestion messages to the device.

[1801] Step 9:

[1802] Terminal: Received alert messages and suggestion messages are displayed to the user in a pop-up format.

[1803] Step 10:

[1804] User: Review the alert or suggestion and choose to modify the text message or send it as is.

[1805] Step 11:

[1806] User: If you have corrected the message, click the resend button.

[1807] Step 12:

[1808] On your device: Send the modified text message back to the server.

[1809] Step 13:

[1810] Server: Re-parse the modified text message to verify that the issue has been resolved.

[1811] Step 14:

[1812] Server: After final verification, the text message is sent to the intended recipient.

[1813] Step 15:

[1814] Server: Logs the sent message and its analysis results and stores them for future reference.

[1815] Example 2

[1816] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1817] Inappropriate language and harassment are commonplace in modern communication tools, placing a heavy burden on users. Effective technology to prevent these issues has yet to be fully established. Furthermore, few systems can provide alerts or suggestions based on users' emotions, which can lead to unintended misunderstandings and conflicts. To address these issues, a system is needed that uses more advanced natural language processing technology to analyze users' message content and emotions and provide appropriate alerts and suggestions.

[1818] 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.

[1819] In this invention, the server includes means for receiving text data entered by a user, means for preprocessing the received text data, means for analyzing the preprocessed text data using a generative model to detect inappropriate content or content that constitutes harassment, means for analyzing the received text data using an emotion engine that recognizes the user's emotions, means for suggesting an alert message or appropriate expression to the user based on the analysis results of the generative model and the emotion engine, means for sending the text data with the problem resolved to the actual recipient, and means for recording the sent message and the analysis results in a log. This makes it possible to prevent harassment and promote smooth communication by detecting in advance whether a user's message is inappropriate and what emotion it contains, and correcting it to an appropriate expression.

[1820] A "User" is an individual or entity that utilizes the system to enter and send text messages.

[1821] "Text data" is digital data of character strings including sentences and words entered by the user.

[1822] "Terminal" refers to a device that allows a user to input and send text messages, including a PC, smartphone, tablet, etc.

[1823] "Server" refers to a central management device that receives, analyzes, and stores text data sent from devices and generates alerts and suggestions.

[1824] "Preprocessing" refers to a preparatory process such as tokenizing and escaping data that is performed before the server analyzes the text data.

[1825] A "generative model" is an algorithm that uses natural language processing technology to analyze text data and detect inappropriate content or harassment.

[1826] An "emotion engine" is an algorithm that recognizes a user's emotions from text data and detects emotions such as positive, negative, or neutral.

[1827] An "alert message" is a warning message that the server sends to the user based on the analysis results of the generative model and emotion engine to inform them of the possibility of inappropriate content or harassment.

[1828] A "suggestion" is a message that the server sends to the user, based on the analysis results of the generative model and emotion engine, instructing them on more appropriate expressions and actions.

[1829] A "log" is a collection of data that records sent messages and their analysis results and is saved for later reference.

[1830] This invention is a system that analyzes the content of text messages that users are about to send, detects inappropriate language and harassment in advance, and recognizes the user's emotions to provide more appropriate alerts and suggestions.By combining an emotion engine and a generative model, this system can realize highly accurate alerts and suggestions based on the user's emotions.

[1831] System Configuration

[1832] 1. Device:

[1833] A terminal is a device that allows a user to input and send text messages. The terminal includes an interface that displays alerts and suggestions from the server, as well as the user's ability to input and send text. Terminals include PCs, smartphones, and tablets.

[1834] 2. Server:

[1835] The server receives text data sent from the device and analyzes it using a generative model and emotion engine. The server also generates alert messages and suggestions based on the analysis results and sends them to the device to notify the user.

[1836] 3. Generative Model:

[1837] A generative model is an algorithm that uses natural language processing techniques to analyze text data, understand context and meaning, and detect inappropriate content and harassment. Specific techniques used include BERT and GPT-3.

[1838] 4. Emotion Engine:

[1839] An emotion engine is an algorithm that recognizes user emotions from text data and detects emotional states such as positive, negative, neutral, etc. Sentiment Analysis API and Emotion Detection Model are used as emotion engines.

[1840] Program processing overview

[1841] User message entry and sending

[1842] The user types the text message "You're really useless" into the chat app, and once this action is completed, the device sends the message to the server.

[1843] Message analysis and emotion recognition

[1844] The server passes the received text data to the generative model for preprocessing, which includes tokenization and escaping. The generative model then analyzes the preprocessed data and detects the inappropriate expression "useless." At the same time, the emotion engine analyzes the user's emotions and detects the emotion "anger."

[1845] Alert generation and user notification

[1846] Based on the analysis results, the server generates an alert message and a suggestion, such as "This message contains negative sentiment and contains inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The device then notifies the user of this message.

[1847] Specific examples

[1848] As a concrete example, we will explain "modification of imperative messages and emotion recognition."

[1849] At 2 a.m., the user types a message saying, "Please take care of this now," and presses the send button. The device sends the text and meta information to the server. The server receives the message, performs preprocessing, and passes it to the generative model. The generative model analyzes the message, detects that it is an "imperative expression," and scores it. The emotion engine detects the emotion of "anger" and assigns a negative emotion score. The server generates an alert stating, "This message was sent late at night and contains imperative expressions and the emotion of anger. We recommend sending it the next morning and revising it to 'Please take care of this by tomorrow morning.'" The device displays the alert as a pop-up to the user. The user edits the message to "Please take care of this by tomorrow morning" and sets it to be sent the following morning. The device then sends the revised message back to the server for a final check. The server reanalyzes the message and confirms that there are no problems. The server automatically sends the clean message to the recipient the following morning.

[1850] This system will help prevent harassment and promote better communication.

[1851] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1852] Step 1:

[1853] User:

[1854] The user types the text message "You're really useless" into a chat app. The input is made to the device, and text data is generated.

[1855] Step 2:

[1856] Device:

[1857] When the user presses the send button, the text data and related metadata (user ID, time of sending, etc.) are sent to the server. The input here is the text data and metadata, and this is the output sent to the server.

[1858] Step 3:

[1859] server:

[1860] The server receives data sent from the terminal. It preprocesses the received data (input). Preprocessing includes tokenization (for example, breaking down "You really can't use it" into "o," "mae," "ha," "hontoku," and "tsukanenai") and character escaping (removing certain special characters). The preprocessed text data is the output.

[1861] Step 4:

[1862] server:

[1863] The preprocessed text data is passed as input to a generative model. The generative model then analyzes the text data to detect inappropriate content and harassment. The analysis involves understanding the context and extracting keywords. For example, negative expressions such as "useless" are detected. The analysis results output as inappropriate content detection results.

[1864] Step 5:

[1865] server:

[1866] Based on the analysis results from the generative model, the server passes the text data as input to the emotion engine. The emotion engine analyzes the user's emotion from the text data. For example, the emotion "anger" is detected. The emotion analysis results output the type of emotion and its intensity.

[1867] Step 6:

[1868] server:

[1869] The analysis results of the generative model and the emotion engine are integrated. Based on this integration result, the server generates an alert message and a suggestion message. For example, "This message contains negative emotions and appears to contain inappropriate language. Please try revising it to 'Specific feedback would be helpful.'" The generated alert message and suggestion message are output.

[1870] Step 7:

[1871] Device:

[1872] Receives alert messages and suggestions sent from the server and displays them to the user in a popup format. The displayed alert messages and suggestions become input information for the user to confirm. The displayed content becomes the output.

[1873] Step 8:

[1874] User:

[1875] The user reviews the alerts and suggestions and chooses whether to revise the message or send it as is. For example, they might revise it to "Specific feedback would be helpful." This choice is the input, and the revised message is the output.

[1876] Step 9:

[1877] User:

[1878] If the user modifies the message, he or she presses the resend button.

[1879] Step 10:

[1880] Device:

[1881] The corrected text data is sent to the server again. The sent corrected data becomes the input and the output sent to the server.

[1882] Step 11:

[1883] server:

[1884] The server receives the corrected text data again and re-analyzes it using the generative model. As before, it performs preprocessing, analysis, and emotion recognition, and then verifies that the problem has been solved. The verification result is output.

[1885] Step 12:

[1886] server:

[1887] The text data that has been confirmed to be problem-free is sent to the actual recipient. The sent text data becomes the output.

[1888] Step 13:

[1889] server:

[1890] The sent message and its analysis results are logged for future reference. This log is the output.

[1891] In this way, the system can provide appropriate analysis and suggestions for users' messages, helping to prevent harassment and promote smooth communication.

[1892] (Application example 2)

[1893] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1894] In recent years, with the development of communication tools, there has been an increase in inappropriate language and harassment via text messages. Such problems can cause psychological stress for the recipient and, in some cases, legal issues. Furthermore, senders themselves may send misleading messages in an emotionally charged state, which can lead to future problems. There is a need to prevent these from happening in advance and ensure healthy and safe communication.

[1895] The identification process by the identification 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 receiving text data entered by a user; analysis means using a generative model for analyzing the received text data to detect inappropriate content or content that constitutes harassment; means for suggesting alert messages or appropriate expressions to the user based on the analysis results; emotion recognition means using an emotion engine for detecting the user's emotional state based on the analysis results; and means for generating a specific correction suggestion message based on the inappropriate content, harassment, and emotional state. This allows users to prevent inappropriate messages from being sent in advance and maintain healthy and smooth communication.

[1896] "User" means any person or entity that uses the System to input and send text messages.

[1897] "Text data" refers to the text information entered by the user, and is text that is subject to detection of inappropriate expressions and harassment.

[1898] The "receiving means" refers to a function or device for transmitting text data entered by a user to a server and receiving the data.

[1899] "Analysis means" refers to functions or algorithms that analyze received text data using a generative model to detect inappropriate content or harassment.

[1900] A "generative model" is a model that uses natural language processing technology to analyze text data and understand context and meaning.

[1901] An "alert message" is a message that warns the user based on the analysis results.

[1902] "Means for suggesting" refers to functions and algorithms for suggesting alert messages and appropriate expressions to users.

[1903] "Emotion recognition means" refers to functions and algorithms for detecting the user's emotional state based on the analysis results and analyzing that state.

[1904] An "emotion engine" is a system that includes an algorithm that recognizes a user's emotions from text data and detects emotional states such as positive, negative, or neutral.

[1905] A "correction suggestion message" is a message that provides the user with specific suggestions for correction based on inappropriate content, harassment, and emotional state.

[1906] The present invention aims to enable users to communicate safely and appropriately by providing a system for analyzing the content of text messages and the emotions behind them in real time. Hereinafter, an embodiment of the present invention will be described in detail.

[1907] Program processing

[1908] This system consists of a terminal where users can input and send text messages, and a server that analyzes those messages. Specifically, the system uses the following hardware and software:

[1909] Hardware

[1910] Device: Any device that allows a user to type text messages, such as a smartphone, tablet, or PC.

[1911] Server: The server for receiving and analyzing messages is run on Apache or Nginx and uses a Python-based framework (Flask or Django).

[1912] software

[1913] Generative models: Generative AI models that use natural language processing techniques, such as GPT-3 and BERT, to analyze text data and detect inappropriate content and harassment.

[1914] Emotion engine: Detects the user's emotional state from text using technologies such as IBM Watson Natural Language Understanding (NLU).

[1915] Database: A database (e.g., MySQL, PostgreSQL) for storing analysis results and transmission history.

[1916] Data processing and calculation

[1917] Message input: The user types a text message into the device, for example, "Why are you always late!"

[1918] Data transmission: The entered text data and metadata (user ID, submission time, etc.) are sent to the server.

[1919] Preprocessing: The server tokenizes the received text data and performs any necessary character escaping.

[1920] Analysis: Preprocessed text data is fed into a generative model to analyze context and meaning, detecting inappropriate content and harassment.

[1921] Emotion recognition: The emotion engine detects the user's emotional state (e.g., "anger") from text data.

[1922] Alerts and Suggestions: Based on the analysis results, appropriate alert messages and suggestions for correction are generated and notified to the user.

[1923] Save: The sent message, its analysis results, and its revision history are saved in the database.

[1924] Specific examples

[1925] As a concrete example, consider the situation where a user tries to enter the following message:

[1926] "Why are you always late?"

[1927] When attempting to send this message, the generative AI model is prompted with the following:

[1928] Example prompt:

[1929] Message: "Why are you always late?"

[1930] Emotional state: "Anger"

[1931] Purpose: "Detect profanity and provide correction suggestions"

[1932] The generated alert and suggestion will notify the user, for example, "This message contains angry emotions and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late.'" By revising the message based on this suggestion, the user can achieve more appropriate communication.

[1933] The above has described in detail the embodiments of the present invention.

[1934] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1935] Step 1:

[1936] The user inputs a text message into the terminal, for example, "Why are you always late?" This input data is passed to subsequent processing.

[1937] Step 2:

[1938] The text message is sent from the device to the server, along with metadata such as the user ID and the time of sending, allowing the server to determine who is sending what message at what time.

[1939] Step 3:

[1940] The server preprocesses the text data it receives. Specifically, it tokenizes the text data and performs necessary character escaping. This preprocessing prepares the data in a format suitable for analysis. The input is text data and metadata, and the output is the preprocessed tokenized data.

[1941] Step 4:

[1942] The preprocessed tokenized text data is passed to a generative model. The server uses the generative model to analyze this data and detect inappropriate content and harassment while understanding the context and meaning. The input is the tokenized data, and the output is the detection results for inappropriate content and harassment.

[1943] Step 5:

[1944] The server uses an emotion engine to detect the user's emotional state from the tokenized text data. For example, emotional states such as "anger" or "joy" are analyzed. The input is the tokenized data, and the output is the detected emotional state.

[1945] Step 6:

[1946] The server combines the analysis results of the generative model with the results of the emotion engine, and generates a specific correction suggestion message based on the inappropriate content and emotional state. For example, if the analysis results determine that "anger is detected and contains inappropriate language," it generates a suggestion such as "This message contains anger and is inappropriate. Please revise it to say, 'It would be helpful if you could tell me why you will be late,' and send it." The input is the detection result and emotional state, and the output is a correction suggestion message.

[1947] Step 7:

[1948] The server sends a correction suggestion message to the terminal and notifies the user. The user checks the suggestion message and makes appropriate corrections. The input is the correction suggestion message, and the output is the text message corrected by the user.

[1949] Step 8:

[1950] If the user resubmits the corrected message, it is resent from the terminal to the server. The server reanalyzes it and verifies that there are no problems. The input is the corrected message, and the output is the analysis result.

[1951] Step 9:

[1952] The server finally sends the corrected text message to the recipient and saves the sending history in a database for future reference. The input is the corrected message and the analysis result, and the output is the sending history.

[1953] This series of processing steps allows users to prevent sending inappropriate messages in advance and maintain healthy and smooth communication.

[1954] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1955] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1956] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1957] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1958] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1959] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1960] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1961] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1962] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1963] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1964] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1965] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1966] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1967] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1968] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1969] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1970] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1971] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1972] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1973] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1974] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1975] The following is further disclosed regarding the above embodiment.

[1976] (Claim 1)

[1977] means for receiving text data entered by a user;

[1978] an analysis means using a generative model that analyzes received text data and detects inappropriate content or content that constitutes harassment;

[1979] A system including a suggestion means for suggesting alert messages and appropriate expressions to users based on the analysis results.

[1980] (Claim 2)

[1981] 10. The system according to claim 1, further comprising means for analyzing a transmission time of the received text data and proposing an appropriate transmission time.

[1982] (Claim 3)

[1983] The system according to claim 1, further comprising means for scoring the analysis results of the text data analyzed using the generative model and for issuing alerts or suggestions based on the scores.

[1984] "Example 1"

[1985] (Claim 1)

[1986] means for receiving text data entered by a user;

[1987] A means of preprocessing the received text data and passing it to a generative AI model;

[1988] A means for analyzing received text data using a generative AI model to detect inappropriate or harassing content; and

[1989] A means for suggesting an alert message or appropriate wording to a user based on the detected analysis results;

[1990] a means for displaying alert messages and suggestions to the user;

[1991] means for resubmitting and re-analyzing user-modified messages;

[1992] The system includes a means for transmitting the corrected message to the actual recipient after final confirmation.

[1993] (Claim 2)

[1994] 10. The system according to claim 1, further comprising means for analyzing a transmission time of the received text data and proposing an appropriate transmission time.

[1995] (Claim 3)

[1996] The system of claim 1, further comprising means for scoring the analysis results of the text data analyzed using the generative AI model and providing alerts or suggestions based on the scores.

[1997] "Application Example 1"

[1998] (Claim 1)

[1999] means for receiving text data entered by a user;

[2000] an analysis means using a generative model that analyzes received text data and detects inappropriate content or content that constitutes harassment;

[2001] A means for suggesting alert messages and appropriate expressions to users based on the analysis results;

[2002] A means for reanalyzing the text data corrected based on the proposed content and reconfirming its appropriateness;

[2003] The system includes means for transmitting the modified text data to the final recipient.

[2004] (Claim 2)

[2005] 10. The system according to claim 1, further comprising means for analyzing a transmission time of the received text data and proposing an appropriate transmission time.

[2006] (Claim 3)

[2007] The system according to claim 1, further comprising means for scoring the analysis results of the text data analyzed using the generative model and for issuing alerts or suggestions based on the scores.

[2008] "Example 2: Combining Emotion Engines"

[2009] (Claim 1)

[2010] means for receiving text data entered by a user;

[2011] means for preprocessing the received text data;

[2012] A means for analyzing the preprocessed text data using a generative model to detect inappropriate content or content that constitutes harassment;

[2013] analyzing the received text data using an emotion engine that recognizes the user's emotion;

[2014] A means for suggesting an alert message or an appropriate expression to a user based on the analysis results of the generative model and the emotion engine;

[2015] a means for transmitting the resolved text data to an actual recipient;

[2016] A system that includes a means for logging sent messages and their analysis results.

[2017] (Claim 2)

[2018] 10. The system according to claim 1, further comprising means for analyzing a transmission time of the received text data and proposing an appropriate transmission time.

[2019] (Claim 3)

[2020] The system according to claim 1, further comprising means for scoring the analysis results of the text data analyzed using the generative model and for issuing alerts or suggestions based on the scores.

[2021] "Application example 2 when combining emotion engines"

[2022] (Claim 1)

[2023] means for receiving text data entered by a user;

[2024] an analysis means using a generative model that analyzes received text data and detects inappropriate content or content that constitutes harassment;

[2025] A means for suggesting alert messages and appropriate expressions to users based on the analysis results;

[2026] an emotion recognition means using an emotion engine that detects the user's emotional state based on the analysis results;

[2027] The system includes a means for generating specific suggested correction messages based on inappropriate content, harassment, and emotional state.

[2028] (Claim 2)

[2029] 10. The system according to claim 1, further comprising means for analyzing a transmission time of the received text data and proposing an appropriate transmission time.

[2030] (Claim 3)

[2031] The system according to claim 1, further comprising means for scoring the analysis results of the text data analyzed using the generative model and for issuing alerts or suggestions based on the scores. [Explanation of symbols]

[2032] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for receiving text data entered by a user; an analysis means using a generative model that analyzes received text data and detects inappropriate content or content that constitutes harassment; A system including a suggestion means for suggesting alert messages and appropriate expressions to users based on the analysis results.

2. The system according to claim 1 , further comprising means for analyzing a transmission time of the received text data and proposing an appropriate transmission time.

3. The system according to claim 1 , further comprising means for scoring the analysis results of the text data analyzed using the generative model and for issuing alerts or suggestions based on the scores.

Images