System

Abstract

To provide a system.SOLUTION: A system includes: means for analyzing the content of a comment input by a user, and recognizing the content the user wants to communicate; means for combining expressions related to the content the user wants to communicate, which are suggested by generation AI that has learned the user's values, hobbies, and preferences, based on the user's selection, to generate text; and means for outputting the generated text.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 a description and related instruction sentence regarding 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] JP 2022-180282 A Summary of the Invention [Problem to be solved by the invention]

[0004] In UGC such as SNS, comments, and reviews, it is difficult to correctly recognize what users actually want to convey and generate content using expressions that are not defamatory. It is also difficult for AI to learn the writer's values, hobbies, and preferences and acquire the writer's unique expression style. Furthermore, with current technology, it is difficult for a writer to enter a comment, generate text by selecting and combining expressions suggested by AI, and create expressions that take the reader into consideration in the same way as in face-to-face communication. [Means for solving the problem]

[0005] The present invention receives comments entered by users, analyzes their contents, and recognizes what the user wants to convey. It then suggests non-abusive expressions and learns the user's values, hobbies, and preferences. Based on the results of this learning, the AI ​​masters the user's unique expression style. The user then selects and combines expressions suggested by the AI ​​to generate text. In this way, we provide a system that creates expressions that take the reader into consideration. [Brief description of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Diagram 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. FIG. [Diagram 3] FIG. 11 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Diagram 5] FIG. 13 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. 13 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 13 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a process flow of the data processing system in the first embodiment of the first form example. [Figure 12]FIG. 11 is a sequence diagram showing a process flow of the data processing system in the application example 1 of the first embodiment. [Figure 13] FIG. 11 is a sequence diagram showing a process flow of the data processing system in the second embodiment of the second embodiment. [Figure 14] FIG. 13 is a sequence diagram showing a process flow of the data processing system in application example 2 of embodiment 2. [Figure 15] FIG. 13 is a sequence diagram showing a process flow of a data processing system in a third embodiment of the present invention. [Figure 16] FIG. 13 is a sequence diagram showing a process flow of the data processing system in the third application example of the third embodiment. [Figure 17] FIG. 13 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 13 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when combined with an emotion engine. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

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

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

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

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

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

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

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

[0014] [First embodiment]

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

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

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

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

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

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

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54.

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

[0023] As shown in Fig. 2, 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. The specific process program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific process program 56 from the storage 32, and executes the read specific process 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 process program 56 executed on the RAM 30.

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

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

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

[0027] "Example 1"

[0028] In one embodiment of the present invention, when a user enters a comment on a social networking site, the system receives the comment and analyzes its contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The system then suggests expressions that are not defamatory. These suggestions are based on the user's values, hobbies, and preferences, learned from the user's past comments and actions. The AI ​​then learns the user's unique expression style based on the results of its learning.

[0029] "Example 2"

[0030] As a specific example, if a user enters the comment "This product is the worst," the system will analyze the comment and suggest non-defamatory expressions such as "This product did not meet my expectations." Also, if the user has often used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions. The user can select expressions suggested by the AI ​​and combine them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[0031] The process flow of each embodiment will be described below.

[0032] "Example 1"

[0033] Step 1: A user enters a comment on a social networking site.

[0034] Step 2: The system receives the comment and analyzes its content using natural language processing techniques to understand what the user is trying to convey.

[0035] Step 3: The system suggests non-abusive expressions based on the user's values, preferences, and tastes, learned from the user's past comments and behavior.

[0036] Step 4: Based on what it has learned, the AI ​​learns the user's unique expression style.

[0037] "Example 2"

[0038] Step 1: User types in a comment saying "This product is awful."

[0039] Step 2: The system analyzes the comment and suggests non-abusive wording, such as "This product did not meet my expectations."

[0040] Step 3: If the user has frequently used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions.

[0041] Step 4: The user selects expressions suggested by the AI ​​and combines them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[0042] Example 1

[0043] Next, a description will be given of Example 1 of the embodiment 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".

[0044] It has become a problem for users to post slanderous or negative comments on user-generated content such as social media and review sites. This can cause psychological damage to other users, the targeted individuals, or companies. There is also the issue that it is difficult for users to find appropriate expressions based on their own values ​​and hobbies. Furthermore, there is a demand for text to be generated with the same consideration as in face-to-face communication.

[0045] 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.

[0046] In this invention, the server includes a means for receiving comments entered by users, a means for analyzing the contents of the comments using natural language processing technology and recognizing the contents the users want to convey, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the users from their past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to post comments using appropriate expressions based on their own values ​​and hobbies while avoiding defamation.

[0047] "User" means any person or entity that utilizes the System to enter comments and review and select suggested expressions.

[0048] A "comment" is text information that a user enters into user-generated content such as on social media or review sites.

[0049] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[0050] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[0051] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[0052] "Values" refer to the beliefs and standards of judgment held by users.

[0053] "Hobbies and interests" refer to the activities and interests that a user prefers.

[0054] "Expression style" refers to the unique language and writing style used by a user when entering comments.

[0055] "Slander" refers to statements that speak ill of or harm others.

[0056] "User Generated Content" means content that is created and published on the Internet by users.

[0057] "Analysis" is the process of understanding the content of a comment and extracting meaning and sentiment.

[0058] A "suggestion" is an appropriate expression that the server provides to the user using a generative AI model.

[0059] "Text generation" is the process of combining user-selected expressions to create the final comment.

[0060] "Reader" means any other user or audience who reads User Generated Comments.

[0061] The present invention provides a system that suggests appropriate expressions while avoiding slander when a user enters a comment on user-generated content such as a social networking site or a review site. A specific embodiment of the system will be described below.

[0062] The server receives the comments entered by the user. The user uses his / her own device (smartphone or PC) to access the SNS application and enters a comment. For example, the user enters "This movie was boring."

[0063] The server receives comments sent from the social networking application via HTTP requests and stores them in a database.

[0064] The server then analyzes the received comments using natural language processing technology. Specifically, it analyzes the content of the comments using natural language processing libraries such as "spaCy" and "NLTK" through programs written in programming languages ​​such as Python and Java (registered trademark). During the analysis process, it tokenizes the comments, tags them by part of speech, and performs sentiment analysis. For example, it recognizes that the part "it was boring" indicates a negative sentiment.

[0065] Based on the analysis results, the server uses a generative AI model to suggest non-defamatory expressions. For example, "GPT-3 (registered trademark)" is used as the generative AI model. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, by inputting the prompt sentence "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression," the expression "This movie did not meet my expectations" is generated.

[0066] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device.

[0067] The user reviews the proposed expressions from the server, and if necessary, modifies the proposed expressions and posts them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations."

[0068] As a concrete example, consider the case where a user comments, "This restaurant is the worst." The server receives this comment and analyzes it using natural language processing technology. Based on the analysis results, the server inputs a prompt sentence to the generative AI model (GPT-3) saying, "The user commented, 'This restaurant is the worst.' Please suggest a non-defamatory expression." The server obtains the expression "This restaurant did not meet my expectations" from the generative AI model and returns it to the user. The user checks the proposed expression and posts the final comment, stating, "This restaurant did not meet my expectations."

[0069] In this way, the system helps users phrase their comments in a non-abusive manner.

[0070] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0071] Step 1:

[0072] A user enters a comment on a social networking site.

[0073] A user accesses an SNS application using his / her own device (smartphone or PC) and inputs a comment. For example, a user inputs "This movie was boring." The input comment is sent to the server via the SNS application.

[0074] Step 2:

[0075] The server receives the comment.

[0076] The server receives comments sent from the SNS application via HTTP requests. The received comments are stored in a database. The input is the user's comment and the output is the comment stored in the database.

[0077] Step 3:

[0078] The server analyzes the comments using natural language processing technology.

[0079] The server analyzes the received comments using a natural language processing library (e.g., spaCy or NLTK). Specifically, it performs tokenization of the comments, part-of-speech tagging, sentiment analysis, etc. For example, it recognizes that the part "it was boring" indicates a negative sentiment. The input is the comments stored in the database, and the output is the analysis result.

[0080] Step 4:

[0081] Based on the analysis results, the server uses a generative AI model to suggest expressions that are not defamatory.

[0082] The server uses a generative AI model (e.g., GPT-3) based on the analysis results to generate non-defamatory expressions. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, the prompt sentence is input as follows: "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression." The input is the analysis result and the prompt sentence, and the output is the generated expression.

[0083] Step 5:

[0084] The server returns suggestions to the user.

[0085] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device. The input is the generated representation and the output is the suggestions displayed on the user's device.

[0086] Step 6:

[0087] The user reviews the proposal and corrects it if necessary.

[0088] The user reviews the expressions suggested by the server. If necessary, he / she modifies the suggested expressions and submits them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations." The input is the server's suggestions and the output is the final comment.

[0089] (Application example 1)

[0090] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0091] It is becoming a problem that users unintentionally post slanderous comments on social media and messaging apps. This can make online communication unpleasant and lead to trouble and misunderstandings. There is also the issue that it is difficult for users to find appropriate expressions based on their own values, hobbies and preferences.

[0092] 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. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for functioning as an application installed on a smartphone, a means for analyzing and proposing comments in real time, a means for generating expressions that are not slanderous using a generative AI model, and a means for inputting a prompt sentence into the generative AI model and proposing appropriate expressions. This allows users to communicate online with peace of mind.

[0093] "User" means an individual or entity that uses the System to enter Comments.

[0094] A "comment" is a text message that a user enters on social media, messaging apps, etc.

[0095] "Analysis" is the process of using natural language processing technology to understand the content of the entered comment and recognize its meaning.

[0096] "Slander" is offensive language or language intended to injure others.

[0097] A "suggestion" is an act in which the system shows the user appropriate language that is not defamatory.

[0098] "Values" are the beliefs and ways of thinking held by users.

[0099] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[0100] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments to deepen its understanding.

[0101] "Expression style" refers to the unique phrasing and writing style used by a user when entering comments.

[0102] "AI" is a system that uses artificial intelligence technology to analyze data and make learning and suggestions.

[0103] "Text generation" is the process of combining user-selected expressions to create new comments.

[0104] "Reader" means an individual or entity that receives comments entered by a User.

[0105] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[0106] "Real-time" refers to the fact that analysis and suggestions are carried out the moment a user enters a comment.

[0107] A "generative AI model" is an algorithm or program that uses artificial intelligence technology to generate text.

[0108] A "prompt" is an instruction or question that is input into a generative AI model.

[0109] The system for carrying out the present invention functions as a smartphone application that analyzes comments entered by users in real time and suggests expressions that are not defamatory. A specific embodiment of this system will be described below.

[0110] System Configuration

[0111] The system consists of the following main components:

[0112] 1. User device (smartphone): The device where the user enters comments.

[0113] 2. Server: A central processing unit that analyzes comments and makes suggestions.

[0114] 3. Generative AI models: Algorithms that perform natural language processing and text generation.

[0115] Hardware and Software Used

[0116] Hardware: Smartphones, servers

[0117] Software: Python (registered trademark), OpenAI (registered trademark) API, spaCy (natural language processing library)

[0118] Data processing and calculation

[0119] 1. Receiving comments:

[0120] The user inputs comments through a smartphone application, and the input comments are sent to the server.

[0121] 2. Comment Analysis:

[0122] The server uses spaCy to analyze comments and understand their grammar and meaning, and the results are used to understand what the user is trying to say.

[0123] 3. Suggestions for non-abusive language:

[0124] The server uses OpenAI's generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[0125] Please rewrite your comment below so it is not abusive: You are such an idiot!

[0126] 4. Learning about user values, preferences, and tastes:

[0127] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. Based on this learning, the AI ​​will master the user's unique expression style.

[0128] 5. Text Generation:

[0129] Users can generate new comments by selecting and combining expressions suggested by the AI. The generated comments will be expressions that take the reader into consideration.

[0130] Examples

[0131] For example, if a user types a comment like, "You're such an idiot!", the system will analyze it and suggest non-abusive expressions such as:

[0132] I don't agree with that opinion, but I'd like you to think about it a bit more.

[0133] The user can select this suggestion and use it as the final comment.

[0134] In this way, users can communicate online with peace of mind.

[0135] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0136] Step 1:

[0137] A user inputs a comment through a smartphone application. The input comment is sent from the user terminal to a server. The input data is the text comment input by the user. The output data is the comment text sent to the server.

[0138] Step 2:

[0139] The server analyzes the received comments. Specifically, it uses spaCy to analyze the grammar and semantics of the comments. The input data is the comment text entered by the user. As data processing, spaCy tokenizes the comments and analyzes their grammatical structure. The output data is the grammatical structure and semantic information of the analyzed comments.

[0140] Step 3:

[0141] The server recognizes what the user wants to convey based on the analysis results. The input data is the grammatical structure and semantic information of the analyzed comment. The data calculation uses natural language processing technology to extract the meaning of the comment. The output data is the semantic information of the comment, including the user's intention.

[0142] Step 4:

[0143] The server uses a generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[0144] Please rewrite the following comment to be non-abusive: [User comment]

[0145] The input data is the semantic information of the comment, including the user's intention, and the prompt sentence. As a data calculation, the generative AI model analyzes the prompt sentence and generates an appropriate expression. The output data is an expression that is not defamatory.

[0146] Step 5:

[0147] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. The input data is the user's past comments and behavioral history. To process the data, machine learning algorithms are used to extract patterns and model the user's values, hobbies, and preferences. The output data is a model that reflects the user's values, hobbies, and preferences.

[0148] Step 6:

[0149] The server uses the AI ​​to learn the user's unique expression style based on the learning results. The input data is a model that reflects the user's values ​​and hobbies. As a data calculation, the AI ​​uses the model to learn the user's expression style. The output data is an AI model that reflects the user's expression style.

[0150] Step 7:

[0151] Users generate new comments by selecting and combining expressions suggested by the AI. The input data are non-abusive expressions suggested by the AI. In concrete terms, the user selects the suggested expressions on the smartphone application and generates the final comment. The output data is the non-abusive comment selected by the user.

[0152] Step 8:

[0153] The server saves the generated comments as expressions that take the reader into consideration, and posts them to social media or messaging apps as necessary. The input data are non-abusive comments selected by the user. Data processing involves converting the comments into an appropriate format and saving or posting them. The output data are the final comments that take the reader into consideration.

[0154] Example 2

[0155] Next, a second embodiment of the second embodiment will be described. 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."

[0156] It has become a problem that slander and offensive language is often seen in comments and reviews on the Internet. This can lead to a decline in the quality of communication and trouble between users. In addition, since it is difficult for users to use appropriate expressions based on their own values ​​and tastes, there is a demand for generating text that has the same consideration as face-to-face communication.

[0157] 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.

[0158] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for suggesting non-abusive expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions suggested by the AI, a means for creating expressions that take the reader into consideration, a means for sending a prompt sentence to the generative AI model, a means for the generative AI model to generate appropriate expressions, and a means for suggesting the generated expressions to the user. This enables the user to use non-abusive expressions and generate text with the same consideration as face-to-face communication.

[0159] "User" means any individual or entity that uses the System to enter comments and reviews.

[0160] A "comment" is an opinion or impression entered in text form by a user.

[0161] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[0162] "Analysis" is the process of understanding the content of comments entered by users and recognizing their emotions and intent.

[0163] "Slander" is offensive language intended to harm others.

[0164] "Expression" refers to words or phrases that adequately convey what the user wants to say.

[0165] "Values" are the beliefs and ways of thinking held by users.

[0166] "Hobbies and interests" are the activities and interests that a user prefers.

[0167] "Artificial intelligence" is a machine learning model that learns a user's values, hobbies, and preferences and suggests appropriate expressions.

[0168] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[0169] A "prompt sentence" is an instruction sentence that instructs a generative AI model to generate an appropriate expression.

[0170] "Text" refers to user-generated sentences and comments.

[0171] "User-Generated Content" is content, such as comments and reviews, that is created and published online by users.

[0172] A "social networking service" is an online platform that enables users to communicate with other users.

[0173] "Face-to-face communication" refers to communication that takes place by meeting face to face, and consideration for the other person is required.

[0174] This invention is a system that analyzes comments entered by users and suggests expressions that are not slanderous. The system uses a server, a terminal, and a generative AI model to receive and analyze comments entered by users and generate appropriate expressions.

[0175] The server receives the comments entered by the user. For example, if the user enters the comment "This product is the worst," the device sends this comment to the server. The server analyzes the received comments using natural language processing (NLP) techniques. For this analysis, it uses generative AI models such as Google's BERT and OpenAI's GPT-3.

[0176] Once the analysis is complete, the server generates non-abusive expressions based on the content of the comment. For example, it suggests expressions such as "This product did not meet my expectations." At this time, the server also refers to the user's past comment history and learns the user's frequently used expression style.

[0177] The user can select expressions suggested by the server and combine them to generate new text, which enables the user to create expressions that take the reader into consideration.

[0178] As a concrete example, if a user types a comment like "This product is the worst", the server will input the following prompt sentence into the generative AI model:

[0179] Example prompt:

[0180] User comment: "This product is awful."

[0181] Please revise this comment to something non-abusive.

[0182] Based on this prompt, the generative AI model generates the phrase, "This product did not meet my expectations." The user can accept this suggestion and use it as the new text.

[0183] The system helps users to use non-abusive language and generate text with the same level of consideration as in face-to-face communication, which is expected to improve the quality of communication on the Internet and reduce trouble between users.

[0184] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0185] Step 1:

[0186] The user enters a comment.

[0187] The user inputs a comment into an input field of the terminal, for example, "This product is the worst." The input comment is temporarily stored in the memory of the terminal.

[0188] Step 2:

[0189] The terminal transmits the comment to the server.

[0190] The terminal sends the comments entered by the user to the server. At this time, the terminal uses an HTTP request to send the comment data to the server. The input data is packaged in JSON format and sent to the server over the network. For example, the data {"comment": "This product is the worst"} is sent.

[0191] Step 3:

[0192] The server receives and parses the comment.

[0193] The server receives comments sent from the terminal. It analyzes the received comments using natural language processing (NLP) technology. Specifically, it performs sentiment analysis of the comments and extracts keywords. A Python NLP library (e.g., spaCy or NLTK) is used for the analysis. The input data is the comment text, and the output data is the analysis results (e.g., sentiment score and keyword list).

[0194] Step 4:

[0195] The server sends a prompt to the generative AI model.

[0196] The server sends a prompt to the generative AI model based on the analysis results. For example, it generates a prompt like, "User comment: 'This product is the worst.' Please convert this comment into a non-defamatory expression." The input data is the analysis results, and the output data is the prompt.

[0197] Step 5:

[0198] A generative AI model generates an appropriate representation.

[0199] A generative AI model (e.g., OpenAI's GPT-3) generates an appropriate expression based on a prompt sent from the server. For example, it generates an expression such as "This product did not meet my expectations." The input data is the prompt, and the output data is the generated expression.

[0200] Step 6:

[0201] The server proposes the generated representation to the user.

[0202] The server proposes the representations received from the generative AI model to the user and displays the proposed representations on the device screen. The input data is the generated representation, and the output data is the proposal displayed to the user.

[0203] Step 7:

[0204] The user selects a suggested expression to generate new text.

[0205] The user selects an expression suggested by the server, generates new text based on the selected expression, and posts it as a comment. The input data is the proposed expression, and the output data is the generated new text.

[0206] (Application example 2)

[0207] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0208] In conventional comment systems, the comments entered by users often contained slander or negative language, which caused discomfort to other users. It was also difficult for users to use language based on their own values ​​and preferences, making it difficult to generate text that was as considerate as face-to-face communication. Another problem was that it took time for users to find the appropriate language when posting a review.

[0209] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0210] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing non-libelous expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments using a generative AI model and generating non-libelous expressions, a means for proposing the generated expressions to the user and allowing the user to select them, and a means for posting the expressions selected by the user as a review. This allows the user to easily generate non-libelous expressions and post reviews that are considerate of other users.

[0211] "User" means any individual or entity that uses the System to enter comments and reviews.

[0212] A "comment" is an opinion or impression entered in text form by a user.

[0213] "Analysis" is the process of understanding the content of the entered comments and extracting meaning.

[0214] "Slander" is negative or offensive language that hurts others.

[0215] "Expression" is the way in which a user puts into words what they want to communicate.

[0216] A "suggestion" is an alternative representation or opinion that the system offers to the user.

[0217] "Values" are the beliefs and ways of thinking held by users.

[0218] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[0219] "Learning" is the process by which a system accumulates knowledge based on the user's past behavior and input.

[0220] A "generative AI model" is an algorithm that uses artificial intelligence to generate new text or expressions.

[0221] A "prompt sentence" is an instruction sentence input to a generative AI model.

[0222] A "review" is an evaluation or opinion a user gives about a product or service.

[0223] "Posting" refers to the act of publishing comments or reviews created by a user.

[0224] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and suggesting non-abusive expressions. Specific embodiments are described below.

[0225] Hardware and Software Used

[0226] Hardware: Smartphones, servers

[0227] Software: Python, OpenAI API

[0228] System Configuration

[0229] 1. User terminal: Provides an interface for users to input comments. Implemented as a smartphone application.

[0230] 2. Server: Receives the comments, analyzes them, generates non-abusive expressions, suggests them to the user, and posts the expression the user selects as a review.

[0231] Data processing and calculation

[0232] 1. Receiving comments: A user inputs a comment through a smartphone application. The input comment is sent to the server.

[0233] 2. Comment analysis: The server uses the OpenAI API to analyze the entered comments. Specifically, it uses a generative AI model (e.g. GPT-3) to understand the content of the comment and generate non-defamatory expressions.

[0234] 3. Representation Suggestion: The generated representations are suggested to the user, who selects the appropriate one from the suggested representations.

[0235] 4. Posting a Review: The expression selected by the user will be posted as the final review.

[0236] Examples

[0237] If a user enters "This product is the worst," the system will input the following prompt sentence into the generative AI model:

[0238] Example prompt:

[0239] User comment: This product is awful

[0240] Please revise this comment to something non-abusive.

[0241] Based on this prompt, the generative AI model generates a phrase such as, "This product did not meet my expectations." The generated phrase is suggested to the user, and if the user selects it, it is posted as a review.

[0242] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[0243] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0244] Step 1:

[0245] Users enter comments through a smartphone application.

[0246] Input: Comment entered by the user (e.g. "This product is awful")

[0247] Output: The entered comment is sent to the server.

[0248] Specific behavior: When a user enters a comment into the text box in the application and presses the "Send" button, the comment is sent to the server.

[0249] Step 2:

[0250] The server receives the comments and sends the prompt to the generative AI model for analysis.

[0251] Input: Comments submitted by users

[0252] Output: The prompt sent to the generative AI model

[0253] Specific behavior: The server receives the comment and generates the following prompt text:

[0254] User comment: This product is awful

[0255] Please revise this comment to something non-abusive.

[0256] The generated prompt sentence is sent to the generative AI model via the OpenAI API.

[0257] Step 3:

[0258] A generative AI model analyzes the prompt and generates non-abusive language.

[0259] Input: prompt statement

[0260] Output: Non-derogatory statements (e.g. "This product did not meet my expectations")

[0261] How it works: The generative AI model analyzes the prompt and generates an appropriate expression, which is then returned to the server via the OpenAI API.

[0262] Step 4:

[0263] The server proposes the generated representation to the user.

[0264] Input: Representations returned by a generative AI model

[0265] Output: Suggested representations for the user

[0266] Specific operation: The server receives the generated representation and proposes it to the user through the smartphone application. The user confirms the proposed representation.

[0267] Step 5:

[0268] The user selects a suggested expression and posts it as a review.

[0269] Input: Proposed expression

[0270] Output: The submitted reviews

[0271] Specific operation: The user selects the appropriate expression from the suggested expressions and presses the "Submit" button. The selected expression is sent to the server for final review and is published.

[0272] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[0273] Furthermore, an emotion engine that estimates the emotion of the user may be combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[0274] "Example 1"

[0275] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system receives comments entered by a user and analyzes their contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The emotion engine then analyzes the user's emotions and suggests expressions based on the results. For example, if a user enters a comment such as "This product is the worst," the emotion engine analyzes the user's anger or dissatisfaction from the comment. Then, taking the emotion into consideration, it suggests non-defamatory expressions such as "This product did not meet my expectations."

[0276] "Example 2"

[0277] The emotion engine also learns changes in the user's emotions and suggests expressions that adapt to the user's emotions based on the learning results. For example, if a user has often used the expression "did not meet my expectations" in the past, the emotion engine will learn that expression style. If the user has a similar emotion, it will suggest a similar expression. This allows the user to express their emotions appropriately and provides expressions that are considerate to the reader.

[0278] The process flow of each embodiment will be described below.

[0279] "Example 1"

[0280] Step 1: The user enters a comment into the system.

[0281] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[0282] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[0283] Step 4: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[0284] "Example 2"

[0285] Step 1: The user enters a comment into the system.

[0286] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[0287] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[0288] Step 4: The emotion engine learns from the user's past comments and emotional patterns, and based on that learning, suggests expressions that are adapted to the user's emotions.

[0289] Step 5: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[0290] Example 1

[0291] Next, a description will be given of Example 1 of the embodiment 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".

[0292] In modern social networking services (SNS) and user-generated content (UGC), it is a problem that users post slanderous or offensive comments. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. In addition, users may not be able to express their feelings appropriately, leading to misunderstandings. To solve these problems, a system is needed to help users use constructive and considerate expressions.

[0293] 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.

[0294] In this invention, the server includes a means for receiving comments entered by a user, a means for analyzing the content of the comment using natural language processing technology and recognizing the content the user wants to convey, a means for analyzing the emotions contained in the comment, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the user from past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to use constructive and considerate expressions on SNS and UGC.

[0295] "User" refers to an individual who posts comments or reviews using social networking services and user-generated content.

[0296] "Comment" refers to a text message entered by a user on a social networking service or user-generated content.

[0297] "Natural language processing technology" refers to technology that enables computers to understand, analyze, and generate human language.

[0298] "Sentiment analysis" refers to the technology of extracting user emotions (e.g., anger, dissatisfaction, joy, etc.) from text data.

[0299] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new text or expressions.

[0300] "Values" refer to the beliefs and standards of judgment held by users.

[0301] "Hobbies and interests" refers to a user's preferred activities and interests.

[0302] "Expression style" refers to the unique language and writing style used by a user when creating text.

[0303] "Slander" refers to offensive comments or statements made with the intent to harm others.

[0304] "User-Generated Content" refers to content that users create and share online (e.g., comments, reviews, blog posts, etc.).

[0305] "Reader" refers to any other individual who reads the comments and text created by a User.

[0306] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0307] The present invention is a system that supports users in using constructive and considerate expressions in SNS and user-generated content (UGC). A specific embodiment of the system is described below.

[0308] Hardware and Software Used

[0309] Hardware: Servers, terminals (PCs, smartphones)

[0310] Software: Natural language processing libraries (e.g., NLTK and spaCy), sentiment analysis engines (e.g., IBM Watson® Natural Language Understanding), generative AI models (e.g., GPT-3)

[0311] System Overview

[0312] When a user enters a comment on an SNS, the device sends the comment to a server. The server receives the comment and analyzes the content using natural language processing technology. Natural language processing libraries such as NLTK and spaCy are used for the analysis. The server then uses a sentiment analysis engine to analyze the sentiment contained in the comment. For example, IBM Watson Natural Language Understanding can be used to extract the user's sentiment (anger, dissatisfaction, etc.) from the comment.

[0313] The server then uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from past comments and behavior. The generated expressions are sent to the device and suggested to the user. The user can review the suggested expressions and correct them if necessary. Finally, the corrected comments are posted on the social networking site.

[0314] Examples

[0315] For example, if a user types a comment such as "This product is the worst," the device sends this comment to the server. The server receives the comment, analyzes it using a natural language processing library, and uses a sentiment analysis engine to detect "anger" or "dissatisfaction." It then uses a generative AI model to generate a non-defamatory expression such as "This product did not meet my expectations." The generated expression is sent to the device and suggested to the user. The user reviews the proposed expression, modifies it if necessary to "This product greatly disappointed my expectations," and finally posts it on social media.

[0316] Examples of prompt statements

[0317] "When a user enters a comment on a social networking site, the system must receive the comment, analyze the content using natural language processing technology, and suggest expressions that are not defamatory."

[0318] The system will help users use constructive and considerate language on social media and in UGC.

[0319] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0320] Program processing flow

[0321] Step 1:

[0322] The user enters a comment.

[0323] A user inputs a comment into the comment input field of a SNS application. For example, the user inputs "This product is the worst." The input comment is saved on the device.

[0324] Step 2:

[0325] The terminal transmits the comment to the server.

[0326] The terminal uses an HTTP POST request to send the entered comment data to the server. The sent data includes the comment text and the user ID.

[0327] Step 3:

[0328] The server receives the comment.

[0329] The server receives the HTTP request and stores the comment data in a database for further analysis.

[0330] Step 4:

[0331] The server analyzes the comments using natural language processing.

[0332] The server uses natural language processing libraries (e.g., NLTK or spaCy) to analyze the grammatical structure and semantics of the comments. Specifically, it performs tokenization, part-of-speech tagging, dependency analysis, etc. The input is the comment text, and the output is the analysis result (e.g., "This / DET product / NOUN is / AUX the worst / ADJ / AUX").

[0333] Step 5:

[0334] The server performs sentiment analysis.

[0335] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comments. The input is the parsed comment text, and the output is the sentiment data (e.g., "angry" or "unhappy").

[0336] Step 6:

[0337] The server generates non-abusive language.

[0338] The server uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from the user's past comments and behavior. The input is emotion data and the user's past comment data, and the output is the generated expression (e.g., "This product did not meet my expectations").

[0339] Step 7:

[0340] The server proposes the generated representation to the user.

[0341] The server sends the generated representation to the terminal and suggests it to the user. The input is the generated representation and the output is the suggested text that is displayed on the user's terminal.

[0342] Step 8:

[0343] The user reviews the proposal and corrects it if necessary.

[0344] The user checks the suggested expression and modifies it if necessary. For example, the user modifies it to "This product greatly exceeded my expectations." Finally, the modified comment is posted on the SNS. The input is the proposed text, and the output is the comment text that is finally posted.

[0345] In this way, users can be more constructive and considerate in their use of language on social media and in UGC.

[0346] (Application example 1)

[0347] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0348] In modern social networking sites and online forums, users posting abusive or offensive comments is a problem. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. Users themselves can also unintentionally use offensive language, which can result in trouble. To solve these problems, a system is needed that suggests expressions to avoid abusive comments in real time as the user types them.

[0349] 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.

[0350] In this invention, the server includes a means for receiving comments input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments input by a user in real time and detecting the possibility of slander, a means for analyzing emotions contained in the comment using an emotion analysis engine, and a means for notifying the user of the proposed expressions and providing options. This makes it possible for the user to receive suggestions of expressions that avoid slander in real time, improving the quality of communication.

[0351] "User" means an individual or entity that uses the System to enter Comments.

[0352] A "comment" is a text message entered by a user on a platform such as a social networking site or online forum.

[0353] "Analysis" is the process of understanding the content of the entered comments and recognizing the user's intent and emotions.

[0354] "Slander" refers to expressions or actions that attack or damage the reputation of others.

[0355] "Expression suggestions" refers to providing alternative expressions to comments entered by users to avoid defamation.

[0356] "Values" refer to the beliefs and standards of judgment held by users.

[0357] "Hobbies" refer to a user's preferred activities and areas of interest.

[0358] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments and uses them to inform future suggestions.

[0359] "Expression style" refers to the unique language and expression used by a user when entering comments.

[0360] "Real-time" refers to instant processing and suggestions occurring the moment a user enters a comment.

[0361] An "sentiment analysis engine" is software that analyzes the emotions contained in comments and suggests appropriate expressions based on the results.

[0362] "Notification" refers to the act of the system informing the user of a suggested expression or option.

[0363] "Choices" are multiple alternatives from which a user can choose among proposed representations.

[0364] The system for implementing the present invention proposes expressions to avoid slander in real time when a user enters a comment on a social networking site or online forum. A specific embodiment of the system will be described below.

[0365] Hardware and Software Used

[0366] Hardware: Smartphones, smart glasses

[0367] Software: Natural language processing libraries (e.g. spaCy, NLTK), sentiment analysis engines (e.g. IBM Watson Natural Language Understanding), machine learning models (e.g. TENSORFLOW, PyTorch)

[0368] Data processing and calculation

[0369] Comments received

[0370] When a user inputs a comment in the SNS application, the comment is transmitted to the server, which receives the comment in real time using a means for receiving the comment input by the user.

[0371] Natural Language Processing

[0372] The server uses natural language processing libraries to analyze the content of the received comments. Specifically, it uses libraries such as spaCy and NLTK to analyze the grammatical structure and meaning of the comments and recognize what the user is trying to convey.

[0373] Emotion analysis

[0374] The server then uses a sentiment analysis engine to analyze the emotions contained in the comments, for example using IBM Watson Natural Language Understanding to detect emotions such as "anger" or "dissatisfaction" from the comments.

[0375] Expression suggestion generation

[0376] The server uses machine learning models to generate non-abusive expressions based on the values ​​and preferences of users learned from their past comments and behavior. It uses frameworks such as TensorFlow and PyTorch to learn the unique expression style of users and suggest appropriate expressions.

[0377] User Notification

[0378] The proposed expressions are sent from the server to the user's device, and the user can select an option from the suggested expressions to correct the comment.

[0379] Examples

[0380] For example, if a user enters a comment such as "This product is the worst," the server processes it as follows:

[0381] 1. Comment received: "This product is awful."

[0382] 2. Natural Language Processing: Analyze comments to detect potential abusive content

[0383] 3. Sentiment analysis: Detect "anger" using a sentiment analysis engine

[0384] 4. Expression suggestion generation: Generate the expression "This product did not meet my expectations"

[0385] 5. User notification: Notify users of proposed wording

[0386] Examples of prompt statements

[0387] A user wrote in the comment, "This product is the worst." Please use a sentiment analysis engine to detect "anger" and suggest expressions that are not defamatory.

[0388] In this way, a system can be realized that suggests expressions that will help users avoid slander on social media.

[0389] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0390] Step 1:

[0391] A user inputs a comment into an SNS application. The input comment is sent from the user's device to the server. The input data is a comment in text format.

[0392] Step 2:

[0393] The server receives comments entered by users. The received comments are analyzed using a natural language processing library (e.g., spaCy, NLTK). As a result of the analysis, the grammatical structure and meaning of the comments are extracted. The input data is the comments in text format, and the output data is the analyzed grammatical structure and semantic information.

[0394] Step 3:

[0395] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comment. As a result of the analysis, the sentiment contained in the comment (e.g., "anger" or "dissatisfaction") is detected. The input data is the analyzed grammatical structure and semantic information, and the output data is the detected sentiment information.

[0396] Step 4:

[0397] The server uses machine learning models (e.g., TensorFlow, PyTorch) to generate non-abusive expressions based on the values ​​and tastes learned from the user's past comments and behavior. The generated expressions reflect the user's unique expression style. The input data are the detected emotional information and the user's past comments and behavior data, and the output data is the generated non-abusive expressions.

[0398] Step 5:

[0399] The server notifies the user of the generated expression to the user's terminal. The user can select an option from the notified expressions and modify the comment. The input data is the generated non-abusive expression, and the output data is the option notified to the user.

[0400] Step 6:

[0401] The user selects a suggested expression and confirms the revised comment. The confirmed comment is sent to the server again and posted as the final comment to the SNS application. The input data is the expression selected by the user, and the output data is the final comment posted.

[0402] Example 2

[0403] Next, a second embodiment of the second embodiment will be described. 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."

[0404] In online comments and reviews, it is difficult for users to express their feelings and opinions appropriately while avoiding slander. In addition, there is a lack of systems that suggest expressions that are adapted to the user's expression style and feelings, making it difficult to communicate with the reader in a considerate manner.

[0405] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's expression style, a means for proposing expressions that are adapted to the user's emotions based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables the user to provide comments that appropriately express their own emotions while also taking the reader into consideration.

[0406] "User" means any individual or entity that uses the System to enter comments and reviews.

[0407] A "comment" is text data including an opinion or impression entered by a user.

[0408] "Analysis" is the process of understanding the content of entered comments and identifying their meaning and sentiment.

[0409] "Slander" is offensive language or language intended to injure others.

[0410] "Expression style" refers to the patterns of words and phrases that a user prefers to use.

[0411] An "emotion engine" is a software component that analyzes a user's emotions and suggests expressions that are appropriate to those emotions.

[0412] A "generative AI model" is an artificial intelligence algorithm that generates appropriate representations based on user input.

[0413] A "suggestion" is an alternative wording or phrase that the system offers to the user.

[0414] "Selection" is the act of the user choosing the appropriate expression from those suggested by the system.

[0415] "Text generation" is the process of combining user-selected expressions to create new text.

[0416] "Reader" means any other individual or entity that reads user-generated comments or reviews.

[0417] "User-Generated Content" refers to content that users themselves create and post on the Internet, such as through social networking services, comments, and reviews.

[0418] This invention is a system that provides appropriate expressions while avoiding slander when users enter comments or reviews on the Internet. The system works in cooperation with three parties: a server, a terminal, and a user.

[0419] Hardware and Software Used

[0420] Hardware: Servers, terminals (user devices)

[0421] Software: Generative AI models (e.g., GPT-4®), sentiment engines, natural language processing (NLP) libraries (e.g., spaCy, NLTK)

[0422] Program Processing

[0423] 1. User comments

[0424] The user uses the terminal to type in a comment, for example, "This product is awful."

[0425] 2. Comment Analysis

[0426] The server uses generative AI models and natural language processing libraries to analyze user comments. Specifically, it uses NLP libraries to perform sentiment analysis on comments and determine whether they are potentially abusive.

[0427] 3. Suggested expressions

[0428] The server uses a generative AI model to generate non-abusive expressions, such as "This product did not meet my expectations."

[0429] 4. Learning the user's expression style

[0430] The emotion engine learns from the expressions the user has used in the past. If the user often uses the expression "did not meet my expectations," the emotion engine will learn that expression style.

[0431] 5. Viewing and selecting offers

[0432] The server generates suggestions and displays them on the user's device, where the user can select suggested expressions and combine them to generate new text.

[0433] 6. Suggesting expressions that adapt to emotions

[0434] The emotion engine learns the changes in the user's emotions and suggests expressions that are adapted to the user's emotions based on the learning results. For example, if the user has a similar emotion, it will suggest the expression "did not meet my expectations" that was used in the past.

[0435] Examples and prompts

[0436] Examples

[0437] User input: "This product is awful."

[0438] Server analysis results: Possible slander

[0439] Suggested phrase: "This product did not meet my expectations."

[0440] Examples of prompt statements

[0441] Prompt: "Please suggest appropriate wording to make the user's comment less abusive. For example, in response to the comment 'This product is terrible,' suggest 'This product did not meet my expectations.'"

[0442] This system enables users to provide comments that are considerate of the reader while also expressing their own feelings appropriately.

[0443] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0444] Step 1: User comments

[0445] Action: A user uses a terminal to enter a comment.

[0446] Input: User types "This product is awful."

[0447] Output: The entered comment data is sent to the server.

[0448] Step 2: Analyzing comments

[0449] How it works: The server analyzes user comments using generative AI models and natural language processing libraries.

[0450] Input: Comment data received by the server: "This product is awful."

[0451] Data processing: Perform sentiment analysis of comments using NLP libraries.

[0452] Output: Analysis indicates the comment is negative and potentially abusive.

[0453] Step 3: Propose a representation

[0454] How it works: The server uses a generative AI model to generate non-abusive remarks.

[0455] Input: Comment data that, based on the analysis results, is potentially abusive.

[0456] Data computation: Generative AI models generate appropriate representations.

[0457] Output: A suggested wording: "This product did not meet my expectations."

[0458] Step 4: Learning the user's expression style

[0459] How it works: The emotion engine learns from expressions the user has used in the past.

[0460] Input: Comment data used by the user in the past.

[0461] Data processing: The emotion engine analyzes and learns from the user's expression style.

[0462] Output: Learning results about the user's expression style.

[0463] Step 5: View and select suggestions

[0464] How it works: The server generates suggestions and displays them on the user's device.

[0465] Input: Suggested phrase "This product did not meet my expectations."

[0466] Output: The suggestion is displayed on the user's terminal. The user selects the suggestion.

[0467] Step 6: Suggest expressions that match emotions

[0468] How it works: The emotion engine learns the changes in the user's emotions and, based on that learning, suggests expressions that are adapted to the user's emotions.

[0469] Input: User emotion data and past expression style learning results.

[0470] Data computation: The emotion engine generates adapted expressions.

[0471] Output: Expressions adapted to the user's emotions are suggested.

[0472] (Application example 2)

[0473] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0474] When users post product reviews on online shopping sites, they often include slanderous or inappropriate language, which can cause discomfort to other users. There is also the issue that it is difficult for users to generate comments that express their feelings appropriately while also being considerate of the reader.

[0475] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0476] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, and a means for a review posting system on an online shopping site, when a user posts a product review, a means for the system to automatically analyze the comment and suggest expressions that are not slanderous, a means for the user to select and combine the proposed expressions to generate new text, a means for generating expressions that are not slanderous using a generation AI model, and a means for inputting prompt sentences into the generation AI model. This makes it possible for the user to generate comments that take the reader into consideration while appropriately expressing their own feelings.

[0477] "User" means any individual or entity that uses the System to post comments and reviews.

[0478] A "comment" is an opinion or impression entered in text form by a user.

[0479] "Analysis" is the process of understanding the content of the entered comments and extracting their meaning and intent.

[0480] "Slander" is offensive language or language intended to injure others.

[0481] "Expression suggestions" means that the system suggests appropriate, non-defamatory expressions based on comments entered by the user.

[0482] "Values" are the beliefs and ways of thinking held by users.

[0483] "Hobbies" are the activities or things that a user likes.

[0484] "Learning" is the process by which a system understands patterns and trends based on a user's past actions and expressions.

[0485] "Style" refers to the characteristics of the words and phrases a user uses in a particular situation.

[0486] A "generative AI model" is an algorithm or system that uses artificial intelligence to generate new text or expressions.

[0487] A "prompt" is an instruction or question to be input into a generative AI model.

[0488] An "online shopping site" is a website that sells products over the Internet.

[0489] A "review posting system" is a system that has the function of allowing users to post their opinions and ratings on products and services.

[0490] "Text generation" is the process of combining user-selected expressions to create new sentences.

[0491] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and proposing expressions that are not slanderous. A specific embodiment of this system is described below.

[0492] System Configuration

[0493] The system consists of the following main components:

[0494] 1. User device: The device through which a user enters comments (e.g. smartphone, computer).

[0495] 2. Server: A central processing unit that performs comment analysis, expression suggestion, learning, and text generation.

[0496] 3. Generative AI models: Artificial intelligence models for generating non-abusive language (e.g., OpenAI’s GPT-3).

[0497] 4. Database: Storage for saving users' past comments and learning data.

[0498] Program Processing

[0499] 1. Receiving comments

[0500] Comments entered from the user's terminal are sent to the server, which receives the comments and prepares them for analysis.

[0501] 2. Comment Analysis

[0502] The server inputs the received comments into the generative AI model, which analyzes the content of the comments and recognizes what the user wants to convey.

[0503] 3. Expression suggestions

[0504] The generative AI model generates non-abusive expressions based on the analysis results, and the server sends these suggestions to the user's device.

[0505] 4. Learning

[0506] The server learns from the user's past comments and selected expressions, and acquires an expression style based on the user's values, hobbies, and preferences.

[0507] 5. Text Generation

[0508] The user selects and combines suggested expressions to generate new text, which the server stores and optionally displays to other users.

[0509] Hardware and software used

[0510] Hardware: General purpose PC or server

[0511] Software: Python, OpenAI API

[0512] Examples

[0513] If a user enters a comment such as "This product is terrible," the server sends this comment to the generative AI model, which then suggests a non-defamatory rendition of "This product did not meet my expectations." The user selects this suggestion and posts it as the final comment.

[0514] Examples of prompt statements

[0515] User comment: 'This product is awful'

[0516] Please suggest non-defamatory expressions.

[0517] In this way, users can generate comments that are considerate of the reader while still expressing their own feelings appropriately.

[0518] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0519] Step 1:

[0520] The user inputs a comment. The user terminal receives the comment input by the user and transmits it to the server. The input is the user's comment, and the output is the transmission of comment data to the server.

[0521] Step 2:

[0522] The server prepares the received comments for analysis. Specifically, it converts the comment data into a format suitable for the generative AI model. The input is the comment data sent by the user, and the output is a prompt sentence to be input to the generative AI model.

[0523] Step 3:

[0524] The server sends the prompt to the generative AI model. The generative AI model analyzes the content of the comment based on the prompt and recognizes what the user wants to convey. The input is the prompt, and the output is the analysis result.

[0525] Step 4:

[0526] The generative AI model generates non-abusive expressions based on the analysis results. The server receives the generated expressions and sends them to the user's device. The input is the analysis results, and the output is a suggestion for non-abusive expressions.

[0527] Step 5:

[0528] The user terminal displays the suggested expressions received from the server to the user. The user selects and optionally combines the suggested expressions. The input is the suggested expressions and the output is the user's selected expression.

[0529] Step 6:

[0530] The user sends the representation selected by the user to the server, which generates the final text based on the representation selected. The input is the user's representation and the output is the final text.

[0531] Step 7:

[0532] The server stores the generated final text in a database and displays it to other users if desired. The input is the final text and the output is storage in a database and display to other users.

[0533] In this way, a series of processes is achieved: the comments entered by the user are analyzed, non-abusive expressions are suggested, and the final text is generated.

[0534] 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 a voice indicating a user input for the result of the specific processing. The control unit 46A transmits the voice 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 voice data.

[0535] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to 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.

[0536] Another example of generative AI is Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0537] In the above embodiment, an example was given in which the specific process was performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0538] [Second embodiment]

[0539] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0540] 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.

[0541] 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 wide area network (WAN) and / or a local area network (LAN).

[0542] 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.

[0543] 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 the voice according to instructions from the processor 46.

[0544] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[0545] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[0546] Fig. 4 shows an example of 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.

[0547] 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.

[0548] 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.

[0549] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0550] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0551] "Example 1"

[0552] In one embodiment of the present invention, when a user enters a comment on a social networking site, the system receives the comment and analyzes its contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The system then suggests expressions that are not defamatory. These suggestions are based on the user's values, hobbies, and preferences, learned from the user's past comments and actions. The AI ​​then learns the user's unique expression style based on the results of its learning.

[0553] "Example 2"

[0554] As a specific example, if a user enters the comment "This product is the worst," the system will analyze the comment and suggest non-defamatory expressions such as "This product did not meet my expectations." Also, if the user has often used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions. The user can select expressions suggested by the AI ​​and combine them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[0555] The process flow of each embodiment will be described below.

[0556] "Example 1"

[0557] Step 1: A user enters a comment on a social networking site.

[0558] Step 2: The system receives the comment and analyzes its content using natural language processing techniques to understand what the user is trying to convey.

[0559] Step 3: The system suggests non-abusive expressions based on the user's values, preferences, and tastes, learned from the user's past comments and behavior.

[0560] Step 4: Based on what it has learned, the AI ​​learns the user's unique expression style.

[0561] "Example 2"

[0562] Step 1: User types in a comment saying "This product is awful."

[0563] Step 2: The system analyzes the comment and suggests non-abusive wording, such as "This product did not meet my expectations."

[0564] Step 3: If the user has frequently used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions.

[0565] Step 4: The user selects expressions suggested by the AI ​​and combines them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[0566] Example 1

[0567] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0568] It has become a problem for users to post slanderous or negative comments on user-generated content such as social media and review sites. This can cause psychological damage to other users, the targeted individuals, or companies. There is also the issue that it is difficult for users to find appropriate expressions based on their own values ​​and hobbies. Furthermore, there is a demand for text to be generated with the same consideration as in face-to-face communication.

[0569] 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.

[0570] In this invention, the server includes a means for receiving comments entered by users, a means for analyzing the contents of the comments using natural language processing technology and recognizing the contents the users want to convey, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the users from their past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to post comments using appropriate expressions based on their own values ​​and hobbies while avoiding defamation.

[0571] "User" means any person or entity that utilizes the System to enter comments and review and select suggested expressions.

[0572] A "comment" is text information that a user enters into user-generated content such as on social media or review sites.

[0573] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[0574] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[0575] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[0576] "Values" refer to the beliefs and standards of judgment held by users.

[0577] "Hobbies and interests" refer to the activities and interests that a user prefers.

[0578] "Expression style" refers to the unique language and writing style used by a user when entering comments.

[0579] "Slander" refers to statements that speak ill of or harm others.

[0580] "User Generated Content" means content that is created and published on the Internet by users.

[0581] "Analysis" is the process of understanding the content of a comment and extracting meaning and sentiment.

[0582] A "suggestion" is an appropriate expression that the server provides to the user using a generative AI model.

[0583] "Text generation" is the process of combining user-selected expressions to create the final comment.

[0584] "Reader" means any other user or audience who reads User Generated Comments.

[0585] The present invention provides a system that suggests appropriate expressions while avoiding slander when a user enters a comment on user-generated content such as a social networking site or a review site. A specific embodiment of the system will be described below.

[0586] The server receives the comments entered by the user. The user uses his / her own device (smartphone or PC) to access the SNS application and enters a comment. For example, the user enters "This movie was boring."

[0587] The server receives comments sent from the social networking application via HTTP requests and stores them in a database.

[0588] The server then analyzes the received comments using natural language processing technology. Specifically, it analyzes the content of the comments using natural language processing libraries such as "spaCy" and "NLTK" through programs written in programming languages ​​such as Python and Java. During the analysis process, it tokenizes the comments, tags them by part of speech, and performs sentiment analysis. For example, it recognizes that the part "it was boring" indicates a negative sentiment.

[0589] Based on the analysis results, the server uses a generative AI model to suggest non-defamatory expressions. For example, "GPT-3" is used as a generative AI model. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, by inputting the prompt sentence "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression," the expression "This movie did not meet my expectations" is generated.

[0590] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device.

[0591] The user reviews the proposed expressions from the server, and if necessary, modifies the proposed expressions and posts them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations."

[0592] As a concrete example, consider the case where a user comments, "This restaurant is the worst." The server receives this comment and analyzes it using natural language processing technology. Based on the analysis results, the server inputs a prompt sentence to the generative AI model (GPT-3) saying, "The user commented, 'This restaurant is the worst.' Please suggest a non-defamatory expression." The server obtains the expression "This restaurant did not meet my expectations" from the generative AI model and returns it to the user. The user checks the proposed expression and posts the final comment, stating, "This restaurant did not meet my expectations."

[0593] In this way, the system helps users phrase their comments in a non-abusive manner.

[0594] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0595] Step 1:

[0596] A user enters a comment on a social networking site.

[0597] A user accesses an SNS application using his / her own device (smartphone or PC) and inputs a comment. For example, a user inputs "This movie was boring." The input comment is sent to the server via the SNS application.

[0598] Step 2:

[0599] The server receives the comment.

[0600] The server receives comments sent from the SNS application via HTTP requests. The received comments are stored in a database. The input is the user's comment and the output is the comment stored in the database.

[0601] Step 3:

[0602] The server analyzes the comments using natural language processing technology.

[0603] The server analyzes the received comments using a natural language processing library (e.g., spaCy or NLTK). Specifically, it performs tokenization of the comments, part-of-speech tagging, sentiment analysis, etc. For example, it recognizes that the part "it was boring" indicates a negative sentiment. The input is the comments stored in the database, and the output is the analysis result.

[0604] Step 4:

[0605] Based on the analysis results, the server uses a generative AI model to suggest expressions that are not defamatory.

[0606] The server uses a generative AI model (e.g., GPT-3) based on the analysis results to generate non-defamatory expressions. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, the prompt sentence is input as follows: "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression." The input is the analysis result and the prompt sentence, and the output is the generated expression.

[0607] Step 5:

[0608] The server returns suggestions to the user.

[0609] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device. The input is the generated representation and the output is the suggestions displayed on the user's device.

[0610] Step 6:

[0611] The user reviews the proposal and corrects it if necessary.

[0612] The user reviews the expressions suggested by the server. If necessary, he / she modifies the suggested expressions and submits them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations." The input is the server's suggestions and the output is the final comment.

[0613] (Application example 1)

[0614] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0615] It is becoming a problem that users unintentionally post slanderous comments on social media and messaging apps. This can make online communication unpleasant and lead to trouble and misunderstandings. There is also the issue that it is difficult for users to find appropriate expressions based on their own values, hobbies and preferences.

[0616] 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. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for functioning as an application installed on a smartphone, a means for analyzing and proposing comments in real time, a means for generating expressions that are not slanderous using a generative AI model, and a means for inputting a prompt sentence into the generative AI model and proposing appropriate expressions. This allows users to communicate online with peace of mind.

[0617] "User" means an individual or entity that uses the System to enter Comments.

[0618] A "comment" is a text message that a user enters on social media, messaging apps, etc.

[0619] "Analysis" is the process of using natural language processing technology to understand the content of the entered comment and recognize its meaning.

[0620] "Slander" is offensive language or language intended to injure others.

[0621] A "suggestion" is an act in which the system shows the user appropriate language that is not defamatory.

[0622] "Values" are the beliefs and ways of thinking held by users.

[0623] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[0624] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments to deepen its understanding.

[0625] "Expression style" refers to the unique phrasing and writing style used by a user when entering comments.

[0626] "AI" is a system that uses artificial intelligence technology to analyze data and make learning and suggestions.

[0627] "Text generation" is the process of combining user-selected expressions to create new comments.

[0628] "Reader" means an individual or entity that receives comments entered by a User.

[0629] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[0630] "Real-time" refers to the fact that analysis and suggestions are carried out the moment a user enters a comment.

[0631] A "generative AI model" is an algorithm or program that uses artificial intelligence technology to generate text.

[0632] A "prompt" is an instruction or question that is input into a generative AI model.

[0633] The system for carrying out the present invention functions as a smartphone application that analyzes comments entered by users in real time and suggests expressions that are not defamatory. A specific embodiment of this system will be described below.

[0634] System Configuration

[0635] The system consists of the following main components:

[0636] 1. User device (smartphone): The device where the user enters comments.

[0637] 2. Server: A central processing unit that analyzes comments and makes suggestions.

[0638] 3. Generative AI models: Algorithms that perform natural language processing and text generation.

[0639] Hardware and Software Used

[0640] Hardware: Smartphones, servers

[0641] Software: Python, OpenAI API, spaCy (natural language processing library)

[0642] Data processing and calculation

[0643] 1. Receiving comments:

[0644] The user inputs comments through a smartphone application, and the input comments are sent to the server.

[0645] 2. Comment Analysis:

[0646] The server uses spaCy to analyze comments and understand their grammar and meaning, and the results are used to understand what the user is trying to say.

[0647] 3. Suggestions for non-abusive language:

[0648] The server uses OpenAI's generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[0649] Please rewrite your comment below so it is not abusive: You are such an idiot!

[0650] 4. Learning about user values, preferences, and tastes:

[0651] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. Based on this learning, the AI ​​will master the user's unique expression style.

[0652] 5. Text Generation:

[0653] Users can generate new comments by selecting and combining expressions suggested by the AI. The generated comments will be expressions that take the reader into consideration.

[0654] Examples

[0655] For example, if a user types a comment like, "You're such an idiot!", the system will analyze it and suggest non-abusive expressions such as:

[0656] I don't agree with that opinion, but I'd like you to think about it a bit more.

[0657] The user can select this suggestion and use it as the final comment.

[0658] In this way, users can communicate online with peace of mind.

[0659] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0660] Step 1:

[0661] A user inputs a comment through a smartphone application. The input comment is sent from the user terminal to a server. The input data is the text comment input by the user. The output data is the comment text sent to the server.

[0662] Step 2:

[0663] The server analyzes the received comments. Specifically, it uses spaCy to analyze the grammar and semantics of the comments. The input data is the comment text entered by the user. As data processing, spaCy tokenizes the comments and analyzes their grammatical structure. The output data is the grammatical structure and semantic information of the analyzed comments.

[0664] Step 3:

[0665] The server recognizes what the user wants to convey based on the analysis results. The input data is the grammatical structure and semantic information of the analyzed comment. The data calculation uses natural language processing technology to extract the meaning of the comment. The output data is the semantic information of the comment, including the user's intention.

[0666] Step 4:

[0667] The server uses a generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[0668] Please rewrite the following comment to be non-abusive: [User comment]

[0669] The input data is the semantic information of the comment, including the user's intention, and the prompt sentence. As a data calculation, the generative AI model analyzes the prompt sentence and generates an appropriate expression. The output data is an expression that is not defamatory.

[0670] Step 5:

[0671] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. The input data is the user's past comments and behavioral history. To process the data, machine learning algorithms are used to extract patterns and model the user's values, hobbies, and preferences. The output data is a model that reflects the user's values, hobbies, and preferences.

[0672] Step 6:

[0673] The server uses the AI ​​to learn the user's unique expression style based on the learning results. The input data is a model that reflects the user's values ​​and hobbies. As a data calculation, the AI ​​uses the model to learn the user's expression style. The output data is an AI model that reflects the user's expression style.

[0674] Step 7:

[0675] Users generate new comments by selecting and combining expressions suggested by the AI. The input data are non-abusive expressions suggested by the AI. In concrete terms, the user selects the suggested expressions on the smartphone application and generates the final comment. The output data is the non-abusive comment selected by the user.

[0676] Step 8:

[0677] The server saves the generated comments as expressions that take the reader into consideration, and posts them to social media or messaging apps as necessary. The input data are non-abusive comments selected by the user. Data processing involves converting the comments into an appropriate format and saving or posting them. The output data are the final comments that take the reader into consideration.

[0678] Example 2

[0679] Next, a description will be given of Example 2 of the embodiment 2. In the following description, the data processing device 12 is referred to as a "server" and the smart glasses 214 are referred to as a "terminal".

[0680] It has become a problem that slander and offensive language is often seen in comments and reviews on the Internet. This can lead to a decline in the quality of communication and trouble between users. In addition, since it is difficult for users to use appropriate expressions based on their own values ​​and tastes, there is a demand for generating text that has the same consideration as face-to-face communication.

[0681] 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.

[0682] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for suggesting non-abusive expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions suggested by the AI, a means for creating expressions that take the reader into consideration, a means for sending a prompt sentence to the generative AI model, a means for the generative AI model to generate appropriate expressions, and a means for suggesting the generated expressions to the user. This enables the user to use non-abusive expressions and generate text with the same consideration as face-to-face communication.

[0683] "User" means any individual or entity that uses the System to enter comments and reviews.

[0684] A "comment" is an opinion or impression entered in text form by a user.

[0685] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[0686] "Analysis" is the process of understanding the content of comments entered by users and recognizing their emotions and intent.

[0687] "Slander" is offensive language intended to harm others.

[0688] "Expression" refers to words or phrases that adequately convey what the user wants to say.

[0689] "Values" are the beliefs and ways of thinking held by users.

[0690] "Hobbies and interests" are the activities and interests that a user prefers.

[0691] "Artificial intelligence" is a machine learning model that learns a user's values, hobbies, and preferences and suggests appropriate expressions.

[0692] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[0693] A "prompt sentence" is an instruction sentence that instructs a generative AI model to generate an appropriate expression.

[0694] "Text" refers to user-generated sentences and comments.

[0695] "User-Generated Content" is content, such as comments and reviews, that is created and published online by users.

[0696] A "social networking service" is an online platform that enables users to communicate with other users.

[0697] "Face-to-face communication" refers to communication that takes place by meeting face to face, and consideration for the other person is required.

[0698] This invention is a system that analyzes comments entered by users and suggests expressions that are not slanderous. The system uses a server, a terminal, and a generative AI model to receive and analyze comments entered by users and generate appropriate expressions.

[0699] The server receives the comments entered by the user. For example, if the user enters the comment "This product is the worst," the device sends this comment to the server. The server analyzes the received comments using natural language processing (NLP) techniques. For this analysis, it uses generative AI models such as Google's BERT and OpenAI's GPT-3.

[0700] Once the analysis is complete, the server generates non-abusive expressions based on the content of the comment. For example, it suggests expressions such as "This product did not meet my expectations." At this time, the server also refers to the user's past comment history and learns the user's frequently used expression style.

[0701] The user can select expressions suggested by the server and combine them to generate new text, which enables the user to create expressions that take the reader into consideration.

[0702] As a concrete example, if a user types a comment like "This product is the worst", the server will input the following prompt sentence into the generative AI model:

[0703] Example prompt:

[0704] User comment: "This product is awful."

[0705] Please revise this comment to something non-abusive.

[0706] Based on this prompt, the generative AI model generates the phrase, "This product did not meet my expectations." The user can accept this suggestion and use it as the new text.

[0707] The system helps users to use non-abusive language and generate text with the same level of consideration as in face-to-face communication, which is expected to improve the quality of communication on the Internet and reduce trouble between users.

[0708] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0709] Step 1:

[0710] The user enters a comment.

[0711] The user inputs a comment into an input field of the terminal, for example, "This product is the worst." The input comment is temporarily stored in the memory of the terminal.

[0712] Step 2:

[0713] The terminal transmits the comment to the server.

[0714] The terminal sends the comments entered by the user to the server. At this time, the terminal uses an HTTP request to send the comment data to the server. The input data is packaged in JSON format and sent to the server over the network. For example, the data {"comment": "This product is the worst"} is sent.

[0715] Step 3:

[0716] The server receives and parses the comment.

[0717] The server receives comments sent from the terminal. It analyzes the received comments using natural language processing (NLP) technology. Specifically, it performs sentiment analysis of the comments and extracts keywords. A Python NLP library (e.g., spaCy or NLTK) is used for the analysis. The input data is the comment text, and the output data is the analysis results (e.g., sentiment score and keyword list).

[0718] Step 4:

[0719] The server sends a prompt to the generative AI model.

[0720] The server sends a prompt to the generative AI model based on the analysis results. For example, it generates a prompt like, "User comment: 'This product is the worst.' Please convert this comment into a non-defamatory expression." The input data is the analysis results, and the output data is the prompt.

[0721] Step 5:

[0722] A generative AI model generates an appropriate representation.

[0723] A generative AI model (e.g., OpenAI's GPT-3) generates an appropriate expression based on a prompt sent from the server. For example, it generates an expression such as "This product did not meet my expectations." The input data is the prompt, and the output data is the generated expression.

[0724] Step 6:

[0725] The server proposes the generated representation to the user.

[0726] The server proposes the representations received from the generative AI model to the user and displays the proposed representations on the device screen. The input data is the generated representation, and the output data is the proposal displayed to the user.

[0727] Step 7:

[0728] The user selects a suggested expression to generate new text.

[0729] The user selects an expression suggested by the server, generates new text based on the selected expression, and posts it as a comment. The input data is the proposed expression, and the output data is the generated new text.

[0730] (Application example 2)

[0731] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the smart glasses 214 will be referred to as a "terminal".

[0732] In conventional comment systems, the comments entered by users often contained slander or negative language, which caused discomfort to other users. It was also difficult for users to use language based on their own values ​​and preferences, making it difficult to generate text that was as considerate as face-to-face communication. Another problem was that it took time for users to find the appropriate language when posting a review.

[0733] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0734] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing non-libelous expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments using a generative AI model and generating non-libelous expressions, a means for proposing the generated expressions to the user and allowing the user to select them, and a means for posting the expressions selected by the user as a review. This allows the user to easily generate non-libelous expressions and post reviews that are considerate of other users.

[0735] "User" means any individual or entity that uses the System to enter comments and reviews.

[0736] A "comment" is an opinion or impression entered in text form by a user.

[0737] "Analysis" is the process of understanding the content of the entered comments and extracting meaning.

[0738] "Slander" is negative or offensive language that hurts others.

[0739] "Expression" is the way in which a user puts into words what they want to communicate.

[0740] A "suggestion" is an alternative representation or opinion that the system offers to the user.

[0741] "Values" are the beliefs and ways of thinking held by users.

[0742] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[0743] "Learning" is the process by which a system accumulates knowledge based on the user's past behavior and input.

[0744] A "generative AI model" is an algorithm that uses artificial intelligence to generate new text or expressions.

[0745] A "prompt sentence" is an instruction sentence input to a generative AI model.

[0746] A "review" is an evaluation or opinion a user gives about a product or service.

[0747] "Posting" refers to the act of publishing comments or reviews created by a user.

[0748] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and suggesting non-abusive expressions. Specific embodiments are described below.

[0749] Hardware and Software Used

[0750] Hardware: Smartphones, servers

[0751] Software: Python, OpenAI API

[0752] System Configuration

[0753] 1. User terminal: Provides an interface for users to input comments. Implemented as a smartphone application.

[0754] 2. Server: Receives the comments, analyzes them, generates non-abusive expressions, suggests them to the user, and posts the expression the user selects as a review.

[0755] Data processing and calculation

[0756] 1. Receiving comments: A user inputs a comment through a smartphone application. The input comment is sent to the server.

[0757] 2. Comment analysis: The server uses the OpenAI API to analyze the entered comments. Specifically, it uses a generative AI model (e.g. GPT-3) to understand the content of the comment and generate non-defamatory expressions.

[0758] 3. Representation Suggestion: The generated representations are suggested to the user, who selects the appropriate one from the suggested representations.

[0759] 4. Posting a Review: The expression selected by the user will be posted as the final review.

[0760] Examples

[0761] If a user enters "This product is the worst," the system will input the following prompt sentence into the generative AI model:

[0762] Example prompt:

[0763] User comment: This product is awful

[0764] Please revise this comment to something non-abusive.

[0765] Based on this prompt, the generative AI model generates a phrase such as, "This product did not meet my expectations." The generated phrase is suggested to the user, and if the user selects it, it is posted as a review.

[0766] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[0767] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0768] Step 1:

[0769] Users enter comments through a smartphone application.

[0770] Input: Comment entered by the user (e.g. "This product is awful")

[0771] Output: The entered comment is sent to the server.

[0772] Specific behavior: When a user enters a comment into the text box in the application and presses the "Send" button, the comment is sent to the server.

[0773] Step 2:

[0774] The server receives the comments and sends the prompt to the generative AI model for analysis.

[0775] Input: Comments submitted by users

[0776] Output: The prompt sent to the generative AI model

[0777] Specific behavior: The server receives the comment and generates the following prompt text:

[0778] User comment: This product is awful

[0779] Please revise this comment to something non-abusive.

[0780] The generated prompt sentence is sent to the generative AI model via the OpenAI API.

[0781] Step 3:

[0782] A generative AI model analyzes the prompt and generates non-abusive language.

[0783] Input: prompt statement

[0784] Output: Non-derogatory statements (e.g. "This product did not meet my expectations")

[0785] How it works: The generative AI model analyzes the prompt and generates an appropriate expression, which is then returned to the server via the OpenAI API.

[0786] Step 4:

[0787] The server proposes the generated representation to the user.

[0788] Input: Representations returned by a generative AI model

[0789] Output: Suggested representations for the user

[0790] Specific operation: The server receives the generated representation and proposes it to the user through the smartphone application. The user confirms the proposed representation.

[0791] Step 5:

[0792] The user selects a suggested expression and posts it as a review.

[0793] Input: Proposed expression

[0794] Output: The submitted reviews

[0795] Specific operation: The user selects the appropriate expression from the suggested expressions and presses the "Submit" button. The selected expression is sent to the server for final review and is published.

[0796] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[0797] In addition, an emotion engine that estimates the emotion of the user may be further combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[0798] "Example 1"

[0799] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system receives comments entered by a user and analyzes their contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The emotion engine then analyzes the user's emotions and suggests expressions based on the results. For example, if a user enters a comment such as "This product is the worst," the emotion engine analyzes the user's anger or dissatisfaction from the comment. Then, taking the emotion into consideration, it suggests non-defamatory expressions such as "This product did not meet my expectations."

[0800] "Example 2"

[0801] The emotion engine also learns changes in the user's emotions and suggests expressions that adapt to the user's emotions based on the learning results. For example, if a user has often used the expression "did not meet my expectations" in the past, the emotion engine will learn that expression style. If the user has a similar emotion, it will suggest a similar expression. This allows the user to express their emotions appropriately and provides expressions that are considerate to the reader.

[0802] The process flow of each embodiment will be described below.

[0803] "Example 1"

[0804] Step 1: The user enters a comment into the system.

[0805] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[0806] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[0807] Step 4: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[0808] "Example 2"

[0809] Step 1: The user enters a comment into the system.

[0810] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[0811] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[0812] Step 4: The emotion engine learns from the user's past comments and emotional patterns, and based on that learning, suggests expressions that are adapted to the user's emotions.

[0813] Step 5: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[0814] Example 1

[0815] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0816] In modern social networking services (SNS) and user-generated content (UGC), it is a problem that users post slanderous or offensive comments. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. In addition, users may not be able to express their feelings appropriately, leading to misunderstandings. To solve these problems, a system is needed to help users use constructive and considerate expressions.

[0817] 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.

[0818] In this invention, the server includes a means for receiving comments entered by a user, a means for analyzing the content of the comment using natural language processing technology and recognizing the content the user wants to convey, a means for analyzing the emotions contained in the comment, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the user from past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to use constructive and considerate expressions on SNS and UGC.

[0819] "User" refers to an individual who posts comments or reviews using social networking services and user-generated content.

[0820] "Comment" refers to a text message entered by a user on a social networking service or user-generated content.

[0821] "Natural language processing technology" refers to technology that enables computers to understand, analyze, and generate human language.

[0822] "Sentiment analysis" refers to the technology of extracting user emotions (e.g., anger, dissatisfaction, joy, etc.) from text data.

[0823] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new text or expressions.

[0824] "Values" refer to the beliefs and standards of judgment held by users.

[0825] "Hobbies and interests" refers to a user's preferred activities and interests.

[0826] "Expression style" refers to the unique language and writing style used by a user when creating text.

[0827] "Slander" refers to offensive comments or statements made with the intent to harm others.

[0828] "User-Generated Content" refers to content that users create and share online (e.g., comments, reviews, blog posts, etc.).

[0829] "Reader" refers to any other individual who reads the comments and text created by a User.

[0830] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0831] The present invention is a system that supports users in using constructive and considerate expressions in SNS and user-generated content (UGC). A specific embodiment of the system is described below.

[0832] Hardware and Software Used

[0833] Hardware: Servers, terminals (PCs, smartphones)

[0834] Software: Natural language processing libraries (e.g., NLTK and spaCy), sentiment analysis engines (e.g., IBM Watson Natural Language Understanding), generative AI models (e.g., GPT-3)

[0835] System Overview

[0836] When a user enters a comment on an SNS, the device sends the comment to a server. The server receives the comment and analyzes the content using natural language processing technology. Natural language processing libraries such as NLTK and spaCy are used for the analysis. The server then uses a sentiment analysis engine to analyze the sentiment contained in the comment. For example, IBM Watson Natural Language Understanding can be used to extract the user's sentiment (anger, dissatisfaction, etc.) from the comment.

[0837] The server then uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from past comments and behavior. The generated expressions are sent to the device and suggested to the user. The user can review the suggested expressions and correct them if necessary. Finally, the corrected comments are posted on the social networking site.

[0838] Examples

[0839] For example, if a user types a comment such as "This product is the worst," the device sends this comment to the server. The server receives the comment, analyzes it using a natural language processing library, and uses a sentiment analysis engine to detect "anger" or "dissatisfaction." It then uses a generative AI model to generate a non-defamatory expression such as "This product did not meet my expectations." The generated expression is sent to the device and suggested to the user. The user reviews the proposed expression, modifies it if necessary to "This product greatly disappointed my expectations," and finally posts it on social media.

[0840] Examples of prompt statements

[0841] "When a user enters a comment on a social networking site, the system must receive the comment, analyze the content using natural language processing technology, and suggest expressions that are not defamatory."

[0842] The system will help users use constructive and considerate language on social media and in UGC.

[0843] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0844] Program processing flow

[0845] Step 1:

[0846] The user enters a comment.

[0847] A user inputs a comment into the comment input field of a SNS application. For example, the user inputs "This product is the worst." The input comment is saved on the device.

[0848] Step 2:

[0849] The terminal transmits the comment to the server.

[0850] The terminal uses an HTTP POST request to send the entered comment data to the server. The sent data includes the comment text and the user ID.

[0851] Step 3:

[0852] The server receives the comment.

[0853] The server receives the HTTP request and stores the comment data in a database for further analysis.

[0854] Step 4:

[0855] The server analyzes the comments using natural language processing.

[0856] The server uses natural language processing libraries (e.g., NLTK or spaCy) to analyze the grammatical structure and semantics of the comments. Specifically, it performs tokenization, part-of-speech tagging, dependency analysis, etc. The input is the comment text, and the output is the analysis result (e.g., "This / DET product / NOUN is / AUX the worst / ADJ / AUX").

[0857] Step 5:

[0858] The server performs sentiment analysis.

[0859] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comments. The input is the parsed comment text, and the output is the sentiment data (e.g., "angry" or "unhappy").

[0860] Step 6:

[0861] The server generates non-abusive language.

[0862] The server uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from the user's past comments and behavior. The input is emotion data and the user's past comment data, and the output is the generated expression (e.g., "This product did not meet my expectations").

[0863] Step 7:

[0864] The server proposes the generated representation to the user.

[0865] The server sends the generated representation to the terminal and suggests it to the user. The input is the generated representation and the output is the suggested text that is displayed on the user's terminal.

[0866] Step 8:

[0867] The user reviews the proposal and corrects it if necessary.

[0868] The user checks the suggested expression and modifies it if necessary. For example, the user modifies it to "This product greatly exceeded my expectations." Finally, the modified comment is posted on the SNS. The input is the proposed text, and the output is the comment text that is finally posted.

[0869] In this way, users can be more constructive and considerate in their use of language on social media and in UGC.

[0870] (Application example 1)

[0871] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0872] In modern social networking sites and online forums, users posting abusive or offensive comments is a problem. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. Users themselves can also unintentionally use offensive language, which can result in trouble. To solve these problems, a system is needed that suggests expressions to avoid abusive comments in real time as the user types them.

[0873] 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.

[0874] In this invention, the server includes a means for receiving comments input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments input by a user in real time and detecting the possibility of slander, a means for analyzing emotions contained in the comment using an emotion analysis engine, and a means for notifying the user of the proposed expressions and providing options. This makes it possible for the user to receive suggestions of expressions that avoid slander in real time, improving the quality of communication.

[0875] "User" means an individual or entity that uses the System to enter Comments.

[0876] A "comment" is a text message entered by a user on a platform such as a social networking site or online forum.

[0877] "Analysis" is the process of understanding the content of the entered comments and recognizing the user's intent and emotions.

[0878] "Slander" refers to expressions or actions that attack or damage the reputation of others.

[0879] "Expression suggestions" refers to providing alternative expressions to comments entered by users to avoid defamation.

[0880] "Values" refer to the beliefs and standards of judgment held by users.

[0881] "Hobbies" refer to a user's preferred activities and areas of interest.

[0882] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments and uses them to inform future suggestions.

[0883] "Expression style" refers to the unique language and expression used by a user when entering comments.

[0884] "Real-time" refers to instant processing and suggestions occurring the moment a user enters a comment.

[0885] An "sentiment analysis engine" is software that analyzes the emotions contained in comments and suggests appropriate expressions based on the results.

[0886] "Notification" refers to the act of the system informing the user of a suggested expression or option.

[0887] "Choices" are multiple alternatives from which a user can choose among proposed representations.

[0888] The system for implementing the present invention proposes expressions to avoid slander in real time when a user enters a comment on a social networking site or online forum. A specific embodiment of the system will be described below.

[0889] Hardware and Software Used

[0890] Hardware: Smartphones, smart glasses

[0891] Software: Natural language processing libraries (e.g. spaCy, NLTK), sentiment analysis engines (e.g. IBM Watson Natural Language Understanding), machine learning models (e.g. TensorFlow, PyTorch)

[0892] Data processing and calculation

[0893] Comments received

[0894] When a user inputs a comment in the SNS application, the comment is transmitted to the server, which receives the comment in real time using a means for receiving the comment input by the user.

[0895] Natural Language Processing

[0896] The server uses natural language processing libraries to analyze the content of the received comments. Specifically, it uses libraries such as spaCy and NLTK to analyze the grammatical structure and meaning of the comments and recognize what the user is trying to convey.

[0897] Emotion analysis

[0898] The server then uses a sentiment analysis engine to analyze the emotions contained in the comments, for example using IBM Watson Natural Language Understanding to detect emotions such as "anger" or "dissatisfaction" from the comments.

[0899] Expression suggestion generation

[0900] The server uses machine learning models to generate non-abusive expressions based on the values ​​and preferences of users learned from their past comments and behavior. It uses frameworks such as TensorFlow and PyTorch to learn the unique expression style of users and suggest appropriate expressions.

[0901] User Notification

[0902] The proposed expressions are sent from the server to the user's device, and the user can select an option from the suggested expressions to correct the comment.

[0903] Examples

[0904] For example, if a user enters a comment such as "This product is the worst," the server processes it as follows:

[0905] 1. Comment received: "This product is awful."

[0906] 2. Natural Language Processing: Analyze comments to detect potential abusive content

[0907] 3. Sentiment analysis: Detect "anger" using a sentiment analysis engine

[0908] 4. Expression suggestion generation: Generate the expression "This product did not meet my expectations"

[0909] 5. User notification: Notify users of proposed wording

[0910] Examples of prompt statements

[0911] A user wrote in the comment, "This product is the worst." Please use a sentiment analysis engine to detect "anger" and suggest expressions that are not defamatory.

[0912] In this way, a system can be realized that suggests expressions that will help users avoid slander on social media.

[0913] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0914] Step 1:

[0915] A user inputs a comment into an SNS application. The input comment is sent from the user's device to the server. The input data is a comment in text format.

[0916] Step 2:

[0917] The server receives comments entered by users. The received comments are analyzed using a natural language processing library (e.g., spaCy, NLTK). As a result of the analysis, the grammatical structure and meaning of the comments are extracted. The input data is the comments in text format, and the output data is the analyzed grammatical structure and semantic information.

[0918] Step 3:

[0919] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comment. As a result of the analysis, the sentiment contained in the comment (e.g., "anger" or "dissatisfaction") is detected. The input data is the analyzed grammatical structure and semantic information, and the output data is the detected sentiment information.

[0920] Step 4:

[0921] The server uses machine learning models (e.g., TensorFlow, PyTorch) to generate non-abusive expressions based on the values ​​and tastes learned from the user's past comments and behavior. The generated expressions reflect the user's unique expression style. The input data are the detected emotional information and the user's past comments and behavior data, and the output data is the generated non-abusive expressions.

[0922] Step 5:

[0923] The server notifies the user of the generated expression to the user's terminal. The user can select an option from the notified expressions and modify the comment. The input data is the generated non-abusive expression, and the output data is the option notified to the user.

[0924] Step 6:

[0925] The user selects a suggested expression and confirms the revised comment. The confirmed comment is sent to the server again and posted as the final comment to the SNS application. The input data is the expression selected by the user, and the output data is the final comment posted.

[0926] Example 2

[0927] Next, a description will be given of Example 2 of the embodiment 2. In the following description, the data processing device 12 is referred to as a "server" and the smart glasses 214 are referred to as a "terminal".

[0928] In online comments and reviews, it is difficult for users to express their feelings and opinions appropriately while avoiding slander. In addition, there is a lack of systems that suggest expressions that are adapted to the user's expression style and feelings, making it difficult to communicate with the reader in a considerate manner.

[0929] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's expression style, a means for proposing expressions that are adapted to the user's emotions based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables the user to provide comments that appropriately express their own emotions while also taking the reader into consideration.

[0930] "User" means any individual or entity that uses the System to enter comments and reviews.

[0931] A "comment" is text data including an opinion or impression entered by a user.

[0932] "Analysis" is the process of understanding the content of entered comments and identifying their meaning and sentiment.

[0933] "Slander" is offensive language or language intended to injure others.

[0934] "Expression style" refers to the patterns of words and phrases that a user prefers to use.

[0935] An "emotion engine" is a software component that analyzes a user's emotions and suggests expressions that are appropriate to those emotions.

[0936] A "generative AI model" is an artificial intelligence algorithm that generates appropriate representations based on user input.

[0937] A "suggestion" is an alternative wording or phrase that the system offers to the user.

[0938] "Selection" is the act of the user choosing the appropriate expression from those suggested by the system.

[0939] "Text generation" is the process of combining user-selected expressions to create new text.

[0940] "Reader" means any other individual or entity that reads user-generated comments or reviews.

[0941] "User-Generated Content" refers to content that users themselves create and post on the Internet, such as through social networking services, comments, and reviews.

[0942] This invention is a system that provides appropriate expressions while avoiding slander when users enter comments or reviews on the Internet. The system works in cooperation with three parties: a server, a terminal, and a user.

[0943] Hardware and Software Used

[0944] Hardware: Servers, terminals (user devices)

[0945] Software: Generative AI models (e.g. GPT-4), sentiment engines, natural language processing (NLP) libraries (e.g. spaCy, NLTK)

[0946] Program Processing

[0947] 1. User comments

[0948] The user uses the terminal to type in a comment, for example, "This product is awful."

[0949] 2. Comment Analysis

[0950] The server uses generative AI models and natural language processing libraries to analyze user comments. Specifically, it uses NLP libraries to perform sentiment analysis on comments and determine whether they are potentially abusive.

[0951] 3. Suggested expressions

[0952] The server uses a generative AI model to generate non-abusive expressions, such as "This product did not meet my expectations."

[0953] 4. Learning the user's expression style

[0954] The emotion engine learns from the expressions the user has used in the past. If the user often uses the expression "did not meet my expectations," the emotion engine will learn that expression style.

[0955] 5. Viewing and selecting offers

[0956] The server generates suggestions and displays them on the user's device, where the user can select suggested expressions and combine them to generate new text.

[0957] 6. Suggesting expressions that adapt to emotions

[0958] The emotion engine learns the changes in the user's emotions and suggests expressions that are adapted to the user's emotions based on the learning results. For example, if the user has a similar emotion, it will suggest the expression "did not meet my expectations" that was used in the past.

[0959] Examples and prompts

[0960] Examples

[0961] User input: "This product is awful."

[0962] Server analysis results: Possible slander

[0963] Suggested phrase: "This product did not meet my expectations."

[0964] Examples of prompt statements

[0965] Prompt: "Please suggest appropriate wording to make the user's comment less abusive. For example, in response to the comment 'This product is terrible,' suggest 'This product did not meet my expectations.'"

[0966] This system enables users to provide comments that are considerate of the reader while also expressing their own feelings appropriately.

[0967] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0968] Step 1: User comments

[0969] Action: A user uses a terminal to enter a comment.

[0970] Input: User types "This product is awful."

[0971] Output: The entered comment data is sent to the server.

[0972] Step 2: Analyzing comments

[0973] How it works: The server analyzes user comments using generative AI models and natural language processing libraries.

[0974] Input: Comment data received by the server: "This product is awful."

[0975] Data processing: Perform sentiment analysis of comments using NLP libraries.

[0976] Output: Analysis indicates the comment is negative and potentially abusive.

[0977] Step 3: Propose a representation

[0978] How it works: The server uses a generative AI model to generate non-abusive remarks.

[0979] Input: Comment data that, based on the analysis results, is potentially abusive.

[0980] Data computation: Generative AI models generate appropriate representations.

[0981] Output: A suggested wording: "This product did not meet my expectations."

[0982] Step 4: Learning the user's expression style

[0983] How it works: The emotion engine learns from expressions the user has used in the past.

[0984] Input: Comment data used by the user in the past.

[0985] Data processing: The emotion engine analyzes and learns from the user's expression style.

[0986] Output: Learning results about the user's expression style.

[0987] Step 5: View and select suggestions

[0988] How it works: The server generates suggestions and displays them on the user's device.

[0989] Input: Suggested phrase "This product did not meet my expectations."

[0990] Output: The suggestion is displayed on the user's terminal. The user selects the suggestion.

[0991] Step 6: Suggest expressions that match emotions

[0992] How it works: The emotion engine learns the changes in the user's emotions and, based on that learning, suggests expressions that are adapted to the user's emotions.

[0993] Input: User emotion data and past expression style learning results.

[0994] Data computation: The emotion engine generates adapted expressions.

[0995] Output: Expressions adapted to the user's emotions are suggested.

[0996] (Application example 2)

[0997] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the smart glasses 214 will be referred to as a "terminal".

[0998] When users post product reviews on online shopping sites, they often include slanderous or inappropriate language, which can cause discomfort to other users. There is also the issue that it is difficult for users to generate comments that express their feelings appropriately while also being considerate of the reader.

[0999] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1000] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, and a means for a review posting system on an online shopping site, when a user posts a product review, a means for the system to automatically analyze the comment and suggest expressions that are not slanderous, a means for the user to select and combine the proposed expressions to generate new text, a means for generating expressions that are not slanderous using a generation AI model, and a means for inputting prompt sentences into the generation AI model. This makes it possible for the user to generate comments that take the reader into consideration while appropriately expressing their own feelings.

[1001] "User" means any individual or entity that uses the System to post comments and reviews.

[1002] A "comment" is an opinion or impression entered in text form by a user.

[1003] "Analysis" is the process of understanding the content of the entered comments and extracting their meaning and intent.

[1004] "Slander" is offensive language or language intended to injure others.

[1005] "Expression suggestions" means that the system suggests appropriate, non-defamatory expressions based on comments entered by the user.

[1006] "Values" are the beliefs and ways of thinking held by users.

[1007] "Hobbies" are the activities or things that a user likes.

[1008] "Learning" is the process by which a system understands patterns and trends based on a user's past actions and expressions.

[1009] "Style" refers to the characteristics of the words and phrases a user uses in a particular situation.

[1010] A "generative AI model" is an algorithm or system that uses artificial intelligence to generate new text or expressions.

[1011] A "prompt" is an instruction or question to be input into a generative AI model.

[1012] An "online shopping site" is a website that sells products over the Internet.

[1013] A "review posting system" is a system that has the function of allowing users to post their opinions and ratings on products and services.

[1014] "Text generation" is the process of combining user-selected expressions to create new sentences.

[1015] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and proposing expressions that are not slanderous. A specific embodiment of this system is described below.

[1016] System Configuration

[1017] The system consists of the following main components:

[1018] 1. User device: The device through which a user enters comments (e.g. smartphone, computer).

[1019] 2. Server: A central processing unit that performs comment analysis, expression suggestion, learning, and text generation.

[1020] 3. Generative AI models: Artificial intelligence models for generating non-abusive language (e.g., OpenAI’s GPT-3).

[1021] 4. Database: Storage for saving users' past comments and learning data.

[1022] Program Processing

[1023] 1. Receiving comments

[1024] Comments entered from the user's terminal are sent to the server, which receives the comments and prepares them for analysis.

[1025] 2. Comment Analysis

[1026] The server inputs the received comments into the generative AI model, which analyzes the content of the comments and recognizes what the user wants to convey.

[1027] 3. Expression suggestions

[1028] The generative AI model generates non-abusive expressions based on the analysis results, and the server sends these suggestions to the user's device.

[1029] 4. Learning

[1030] The server learns from the user's past comments and selected expressions, and acquires an expression style based on the user's values, hobbies, and preferences.

[1031] 5. Text Generation

[1032] The user selects and combines suggested expressions to generate new text, which the server stores and optionally displays to other users.

[1033] Hardware and software used

[1034] Hardware: General purpose PC or server

[1035] Software: Python, OpenAI API

[1036] Examples

[1037] If a user enters a comment such as "This product is terrible," the server sends this comment to the generative AI model, which then suggests a non-defamatory rendition of "This product did not meet my expectations." The user selects this suggestion and posts it as the final comment.

[1038] Examples of prompt statements

[1039] User comment: 'This product is awful'

[1040] Please suggest non-defamatory expressions.

[1041] In this way, users can generate comments that are considerate of the reader while still expressing their own feelings appropriately.

[1042] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1043] Step 1:

[1044] The user inputs a comment. The user terminal receives the comment input by the user and transmits it to the server. The input is the user's comment, and the output is the transmission of comment data to the server.

[1045] Step 2:

[1046] The server prepares the received comments for analysis. Specifically, it converts the comment data into a format suitable for the generative AI model. The input is the comment data sent by the user, and the output is a prompt sentence to be input to the generative AI model.

[1047] Step 3:

[1048] The server sends the prompt to the generative AI model. The generative AI model analyzes the content of the comment based on the prompt and recognizes what the user wants to convey. The input is the prompt, and the output is the analysis result.

[1049] Step 4:

[1050] The generative AI model generates non-abusive expressions based on the analysis results. The server receives the generated expressions and sends them to the user's device. The input is the analysis results, and the output is a suggestion for non-abusive expressions.

[1051] Step 5:

[1052] The user terminal displays the suggested expressions received from the server to the user. The user selects and optionally combines the suggested expressions. The input is the suggested expressions and the output is the user's selected expression.

[1053] Step 6:

[1054] The user sends the representation selected by the user to the server, which generates the final text based on the representation selected. The input is the user's representation and the output is the final text.

[1055] Step 7:

[1056] The server stores the generated final text in a database and displays it to other users if desired. The input is the final text and the output is storage in a database and display to other users.

[1057] In this way, a series of processes is achieved: the comments entered by the user are analyzed, non-abusive expressions are suggested, and the final text is generated.

[1058] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires a voice indicating a user input for the result of the specific processing. The control unit 46A transmits the 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.

[1059] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to 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.

[1060] Another example of generative AI is Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1061] 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 smart glasses 214.

[1062] [Third embodiment]

[1063] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1064] 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.

[1065] 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 wide area network (WAN) and / or a local area network (LAN).

[1066] 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.

[1067] 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 the voice according to instructions from the processor 46.

[1068] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[1069] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[1070] 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.

[1071] 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.

[1072] 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.

[1073] In the headset type terminal 314, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1074] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1075] "Example 1"

[1076] In one embodiment of the present invention, when a user enters a comment on a social networking site, the system receives the comment and analyzes its contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The system then suggests expressions that are not defamatory. These suggestions are based on the user's values, hobbies, and preferences, learned from the user's past comments and actions. The AI ​​then learns the user's unique expression style based on the results of its learning.

[1077] "Example 2"

[1078] As a specific example, if a user enters the comment "This product is the worst," the system will analyze the comment and suggest non-defamatory expressions such as "This product did not meet my expectations." Also, if the user has often used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions. The user can select expressions suggested by the AI ​​and combine them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[1079] The process flow of each embodiment will be described below.

[1080] "Example 1"

[1081] Step 1: A user enters a comment on a social networking site.

[1082] Step 2: The system receives the comment and analyzes its content using natural language processing techniques to understand what the user is trying to convey.

[1083] Step 3: The system suggests non-abusive expressions based on the user's values, preferences, and tastes, learned from the user's past comments and behavior.

[1084] Step 4: Based on what it has learned, the AI ​​learns the user's unique expression style.

[1085] "Example 2"

[1086] Step 1: User types in a comment saying "This product is awful."

[1087] Step 2: The system analyzes the comment and suggests non-abusive wording, such as "This product did not meet my expectations."

[1088] Step 3: If the user has frequently used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions.

[1089] Step 4: The user selects expressions suggested by the AI ​​and combines them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[1090] Example 1

[1091] Next, a description will be given of Example 1 of the embodiment 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] It has become a problem for users to post slanderous or negative comments on user-generated content such as social media and review sites. This can cause psychological damage to other users, the targeted individuals, or companies. There is also the issue that it is difficult for users to find appropriate expressions based on their own values ​​and hobbies. Furthermore, there is a demand for text to be generated with the same consideration as in face-to-face communication.

[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 a means for receiving comments entered by users, a means for analyzing the contents of the comments using natural language processing technology and recognizing the contents the users want to convey, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the users from their past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to post comments using appropriate expressions based on their own values ​​and hobbies while avoiding defamation.

[1095] "User" means any person or entity that utilizes the System to enter comments and review and select suggested expressions.

[1096] A "comment" is text information that a user enters into user-generated content such as on social media or review sites.

[1097] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[1098] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[1099] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[1100] "Values" refer to the beliefs and standards of judgment held by users.

[1101] "Hobbies and interests" refer to the activities and interests that a user prefers.

[1102] "Expression style" refers to the unique language and writing style used by a user when entering comments.

[1103] "Slander" refers to statements that speak ill of or harm others.

[1104] "User Generated Content" means content that is created and published on the Internet by users.

[1105] "Analysis" is the process of understanding the content of a comment and extracting meaning and sentiment.

[1106] A "suggestion" is an appropriate expression that the server provides to the user using a generative AI model.

[1107] "Text generation" is the process of combining user-selected expressions to create the final comment.

[1108] "Reader" means any other user or audience who reads User Generated Comments.

[1109] The present invention provides a system that suggests appropriate expressions while avoiding slander when a user enters a comment on user-generated content such as a social networking site or a review site. A specific embodiment of the system will be described below.

[1110] The server receives the comments entered by the user. The user uses his / her own device (smartphone or PC) to access the SNS application and enters a comment. For example, the user enters "This movie was boring."

[1111] The server receives comments sent from the social networking application via HTTP requests and stores them in a database.

[1112] The server then analyzes the received comments using natural language processing technology. Specifically, it analyzes the content of the comments using natural language processing libraries such as "spaCy" and "NLTK" through programs written in programming languages ​​such as Python and Java. During the analysis process, it tokenizes the comments, tags them by part of speech, and performs sentiment analysis. For example, it recognizes that the part "it was boring" indicates a negative sentiment.

[1113] Based on the analysis results, the server uses a generative AI model to suggest non-defamatory expressions. For example, "GPT-3" is used as a generative AI model. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, by inputting the prompt sentence "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression," the expression "This movie did not meet my expectations" is generated.

[1114] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device.

[1115] The user reviews the proposed expressions from the server, and if necessary, modifies the proposed expressions and posts them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations."

[1116] As a concrete example, consider the case where a user comments, "This restaurant is the worst." The server receives this comment and analyzes it using natural language processing technology. Based on the analysis results, the server inputs a prompt sentence to the generative AI model (GPT-3) saying, "The user commented, 'This restaurant is the worst.' Please suggest a non-defamatory expression." The server obtains the expression "This restaurant did not meet my expectations" from the generative AI model and returns it to the user. The user checks the proposed expression and posts the final comment, stating, "This restaurant did not meet my expectations."

[1117] In this way, the system helps users phrase their comments in a non-abusive manner.

[1118] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1119] Step 1:

[1120] A user enters a comment on a social networking site.

[1121] A user accesses an SNS application using his / her own device (smartphone or PC) and inputs a comment. For example, a user inputs "This movie was boring." The input comment is sent to the server via the SNS application.

[1122] Step 2:

[1123] The server receives the comment.

[1124] The server receives comments sent from the SNS application via HTTP requests. The received comments are stored in a database. The input is the user's comment and the output is the comment stored in the database.

[1125] Step 3:

[1126] The server analyzes the comments using natural language processing technology.

[1127] The server analyzes the received comments using a natural language processing library (e.g., spaCy or NLTK). Specifically, it performs tokenization of the comments, part-of-speech tagging, sentiment analysis, etc. For example, it recognizes that the part "it was boring" indicates a negative sentiment. The input is the comments stored in the database, and the output is the analysis result.

[1128] Step 4:

[1129] Based on the analysis results, the server uses a generative AI model to suggest expressions that are not defamatory.

[1130] The server uses a generative AI model (e.g., GPT-3) based on the analysis results to generate non-defamatory expressions. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, the prompt sentence is input as follows: "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression." The input is the analysis result and the prompt sentence, and the output is the generated expression.

[1131] Step 5:

[1132] The server returns suggestions to the user.

[1133] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device. The input is the generated representation and the output is the suggestions displayed on the user's device.

[1134] Step 6:

[1135] The user reviews the proposal and corrects it if necessary.

[1136] The user reviews the expressions suggested by the server. If necessary, he / she modifies the suggested expressions and submits them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations." The input is the server's suggestions and the output is the final comment.

[1137] (Application example 1)

[1138] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1139] It is becoming a problem that users unintentionally post slanderous comments on social media and messaging apps. This can make online communication unpleasant and lead to trouble and misunderstandings. There is also the issue that it is difficult for users to find appropriate expressions based on their own values, hobbies and preferences.

[1140] 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. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for functioning as an application installed on a smartphone, a means for analyzing and proposing comments in real time, a means for generating expressions that are not slanderous using a generative AI model, and a means for inputting a prompt sentence into the generative AI model and proposing appropriate expressions. This allows users to communicate online with peace of mind.

[1141] "User" means an individual or entity that uses the System to enter Comments.

[1142] A "comment" is a text message that a user enters on social media, messaging apps, etc.

[1143] "Analysis" is the process of using natural language processing technology to understand the content of the entered comment and recognize its meaning.

[1144] "Slander" is offensive language or language intended to injure others.

[1145] A "suggestion" is an act in which the system shows the user appropriate language that is not defamatory.

[1146] "Values" are the beliefs and ways of thinking held by users.

[1147] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[1148] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments to deepen its understanding.

[1149] "Expression style" refers to the unique phrasing and writing style used by a user when entering comments.

[1150] "AI" is a system that uses artificial intelligence technology to analyze data and make learning and suggestions.

[1151] "Text generation" is the process of combining user-selected expressions to create new comments.

[1152] "Reader" means an individual or entity that receives comments entered by a User.

[1153] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[1154] "Real-time" refers to the fact that analysis and suggestions are carried out the moment a user enters a comment.

[1155] A "generative AI model" is an algorithm or program that uses artificial intelligence technology to generate text.

[1156] A "prompt" is an instruction or question that is input into a generative AI model.

[1157] The system for carrying out the present invention functions as a smartphone application that analyzes comments entered by users in real time and suggests expressions that are not defamatory. A specific embodiment of this system will be described below.

[1158] System Configuration

[1159] The system consists of the following main components:

[1160] 1. User device (smartphone): The device where the user enters comments.

[1161] 2. Server: A central processing unit that analyzes comments and makes suggestions.

[1162] 3. Generative AI models: Algorithms that perform natural language processing and text generation.

[1163] Hardware and Software Used

[1164] Hardware: Smartphones, servers

[1165] Software: Python, OpenAI API, spaCy (natural language processing library)

[1166] Data processing and calculation

[1167] 1. Receiving comments:

[1168] The user inputs comments through a smartphone application, and the input comments are sent to the server.

[1169] 2. Comment Analysis:

[1170] The server uses spaCy to analyze comments and understand their grammar and meaning, and the results are used to understand what the user is trying to say.

[1171] 3. Suggestions for non-abusive language:

[1172] The server uses OpenAI's generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[1173] Please rewrite your comment below so it is not abusive: You are such an idiot!

[1174] 4. Learning about user values, preferences, and tastes:

[1175] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. Based on this learning, the AI ​​will master the user's unique expression style.

[1176] 5. Text Generation:

[1177] Users can generate new comments by selecting and combining expressions suggested by the AI. The generated comments will be expressions that take the reader into consideration.

[1178] Examples

[1179] For example, if a user types a comment like, "You're such an idiot!", the system will analyze it and suggest non-abusive expressions such as:

[1180] I don't agree with that opinion, but I'd like you to think about it a bit more.

[1181] The user can select this suggestion and use it as the final comment.

[1182] In this way, users can communicate online with peace of mind.

[1183] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1184] Step 1:

[1185] A user inputs a comment through a smartphone application. The input comment is sent from the user terminal to a server. The input data is the text comment input by the user. The output data is the comment text sent to the server.

[1186] Step 2:

[1187] The server analyzes the received comments. Specifically, it uses spaCy to analyze the grammar and semantics of the comments. The input data is the comment text entered by the user. As data processing, spaCy tokenizes the comments and analyzes their grammatical structure. The output data is the grammatical structure and semantic information of the analyzed comments.

[1188] Step 3:

[1189] The server recognizes what the user wants to convey based on the analysis results. The input data is the grammatical structure and semantic information of the analyzed comment. The data calculation uses natural language processing technology to extract the meaning of the comment. The output data is the semantic information of the comment, including the user's intention.

[1190] Step 4:

[1191] The server uses a generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[1192] Please rewrite the following comment to be non-abusive: [User comment]

[1193] The input data is the semantic information of the comment, including the user's intention, and the prompt sentence. As a data calculation, the generative AI model analyzes the prompt sentence and generates an appropriate expression. The output data is an expression that is not defamatory.

[1194] Step 5:

[1195] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. The input data is the user's past comments and behavioral history. To process the data, machine learning algorithms are used to extract patterns and model the user's values, hobbies, and preferences. The output data is a model that reflects the user's values, hobbies, and preferences.

[1196] Step 6:

[1197] The server uses the AI ​​to learn the user's unique expression style based on the learning results. The input data is a model that reflects the user's values ​​and hobbies. As a data calculation, the AI ​​uses the model to learn the user's expression style. The output data is an AI model that reflects the user's expression style.

[1198] Step 7:

[1199] Users generate new comments by selecting and combining expressions suggested by the AI. The input data are non-abusive expressions suggested by the AI. In concrete terms, the user selects the suggested expressions on the smartphone application and generates the final comment. The output data is the non-abusive comment selected by the user.

[1200] Step 8:

[1201] The server saves the generated comments as expressions that take the reader into consideration, and posts them to social media or messaging apps as necessary. The input data are non-abusive comments selected by the user. Data processing involves converting the comments into an appropriate format and saving or posting them. The output data are the final comments that take the reader into consideration.

[1202] Example 2

[1203] Next, a second embodiment of the second form will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1204] It has become a problem that slander and offensive language is often seen in comments and reviews on the Internet. This can lead to a decline in the quality of communication and trouble between users. In addition, since it is difficult for users to use appropriate expressions based on their own values ​​and tastes, there is a demand for generating text that has the same consideration as face-to-face communication.

[1205] 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.

[1206] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for suggesting non-abusive expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions suggested by the AI, a means for creating expressions that take the reader into consideration, a means for sending a prompt sentence to the generative AI model, a means for the generative AI model to generate appropriate expressions, and a means for suggesting the generated expressions to the user. This enables the user to use non-abusive expressions and generate text with the same consideration as face-to-face communication.

[1207] "User" means any individual or entity that uses the System to enter comments and reviews.

[1208] A "comment" is an opinion or impression entered in text form by a user.

[1209] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[1210] "Analysis" is the process of understanding the content of comments entered by users and recognizing their emotions and intent.

[1211] "Slander" is offensive language intended to harm others.

[1212] "Expression" refers to words or phrases that adequately convey what the user wants to say.

[1213] "Values" are the beliefs and ways of thinking held by users.

[1214] "Hobbies and interests" are the activities and interests that a user prefers.

[1215] "Artificial intelligence" is a machine learning model that learns a user's values, hobbies, and preferences and suggests appropriate expressions.

[1216] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[1217] A "prompt sentence" is an instruction sentence that instructs a generative AI model to generate an appropriate expression.

[1218] "Text" refers to user-generated sentences and comments.

[1219] "User-Generated Content" is content, such as comments and reviews, that is created and published online by users.

[1220] A "social networking service" is an online platform that enables users to communicate with other users.

[1221] "Face-to-face communication" refers to communication that takes place by meeting face to face, and consideration for the other person is required.

[1222] This invention is a system that analyzes comments entered by users and suggests expressions that are not slanderous. The system uses a server, a terminal, and a generative AI model to receive and analyze comments entered by users and generate appropriate expressions.

[1223] The server receives the comments entered by the user. For example, if the user enters the comment "This product is the worst," the device sends this comment to the server. The server analyzes the received comments using natural language processing (NLP) techniques. For this analysis, it uses generative AI models such as Google's BERT and OpenAI's GPT-3.

[1224] Once the analysis is complete, the server generates non-abusive expressions based on the content of the comment. For example, it suggests expressions such as "This product did not meet my expectations." At this time, the server also refers to the user's past comment history and learns the user's frequently used expression style.

[1225] The user can select expressions suggested by the server and combine them to generate new text, which enables the user to create expressions that take the reader into consideration.

[1226] As a concrete example, if a user types a comment like "This product is the worst", the server will input the following prompt sentence into the generative AI model:

[1227] Example prompt:

[1228] User comment: "This product is awful."

[1229] Please revise this comment to something non-abusive.

[1230] Based on this prompt, the generative AI model generates the phrase, "This product did not meet my expectations." The user can accept this suggestion and use it as the new text.

[1231] The system helps users to use non-abusive language and generate text with the same level of consideration as in face-to-face communication, which is expected to improve the quality of communication on the Internet and reduce trouble between users.

[1232] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1233] Step 1:

[1234] The user enters a comment.

[1235] The user inputs a comment into an input field of the terminal, for example, "This product is the worst." The input comment is temporarily stored in the memory of the terminal.

[1236] Step 2:

[1237] The terminal transmits the comment to the server.

[1238] The terminal sends the comments entered by the user to the server. At this time, the terminal uses an HTTP request to send the comment data to the server. The input data is packaged in JSON format and sent to the server over the network. For example, the data {"comment": "This product is the worst"} is sent.

[1239] Step 3:

[1240] The server receives and parses the comment.

[1241] The server receives comments sent from the terminal. It analyzes the received comments using natural language processing (NLP) technology. Specifically, it performs sentiment analysis of the comments and extracts keywords. A Python NLP library (e.g., spaCy or NLTK) is used for the analysis. The input data is the comment text, and the output data is the analysis results (e.g., sentiment score and keyword list).

[1242] Step 4:

[1243] The server sends a prompt to the generative AI model.

[1244] The server sends a prompt to the generative AI model based on the analysis results. For example, it generates a prompt like, "User comment: 'This product is the worst.' Please convert this comment into a non-defamatory expression." The input data is the analysis results, and the output data is the prompt.

[1245] Step 5:

[1246] A generative AI model generates an appropriate representation.

[1247] A generative AI model (e.g., OpenAI's GPT-3) generates an appropriate expression based on a prompt sent from the server. For example, it generates an expression such as "This product did not meet my expectations." The input data is the prompt, and the output data is the generated expression.

[1248] Step 6:

[1249] The server proposes the generated representation to the user.

[1250] The server proposes the representations received from the generative AI model to the user and displays the proposed representations on the device screen. The input data is the generated representation, and the output data is the proposal displayed to the user.

[1251] Step 7:

[1252] The user selects a suggested expression to generate new text.

[1253] The user selects an expression suggested by the server, generates new text based on the selected expression, and posts it as a comment. The input data is the proposed expression, and the output data is the generated new text.

[1254] (Application example 2)

[1255] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1256] In conventional comment systems, the comments entered by users often contained slander or negative language, which caused discomfort to other users. It was also difficult for users to use language based on their own values ​​and preferences, making it difficult to generate text that was as considerate as face-to-face communication. Another problem was that it took time for users to find the appropriate language when posting a review.

[1257] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1258] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing non-libelous expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments using a generative AI model and generating non-libelous expressions, a means for proposing the generated expressions to the user and allowing the user to select them, and a means for posting the expressions selected by the user as a review. This allows the user to easily generate non-libelous expressions and post reviews that are considerate of other users.

[1259] "User" means any individual or entity that uses the System to enter comments and reviews.

[1260] A "comment" is an opinion or impression entered in text form by a user.

[1261] "Analysis" is the process of understanding the content of the entered comments and extracting meaning.

[1262] "Slander" is negative or offensive language that hurts others.

[1263] "Expression" is the way in which a user puts into words what they want to communicate.

[1264] A "suggestion" is an alternative representation or opinion that the system offers to the user.

[1265] "Values" are the beliefs and ways of thinking held by users.

[1266] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[1267] "Learning" is the process by which a system accumulates knowledge based on the user's past behavior and input.

[1268] A "generative AI model" is an algorithm that uses artificial intelligence to generate new text or expressions.

[1269] A "prompt sentence" is an instruction sentence input to a generative AI model.

[1270] A "review" is an evaluation or opinion a user gives about a product or service.

[1271] "Posting" refers to the act of publishing comments or reviews created by a user.

[1272] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and suggesting non-abusive expressions. Specific embodiments are described below.

[1273] Hardware and Software Used

[1274] Hardware: Smartphones, servers

[1275] Software: Python, OpenAI API

[1276] System Configuration

[1277] 1. User terminal: Provides an interface for users to input comments. Implemented as a smartphone application.

[1278] 2. Server: Receives the comments, analyzes them, generates non-abusive expressions, suggests them to the user, and posts the expression the user selects as a review.

[1279] Data processing and calculation

[1280] 1. Receiving comments: A user inputs a comment through a smartphone application. The input comment is sent to the server.

[1281] 2. Comment analysis: The server uses the OpenAI API to analyze the entered comments. Specifically, it uses a generative AI model (e.g. GPT-3) to understand the content of the comment and generate non-defamatory expressions.

[1282] 3. Representation Suggestion: The generated representations are suggested to the user, who selects the appropriate one from the suggested representations.

[1283] 4. Posting a Review: The expression selected by the user will be posted as the final review.

[1284] Examples

[1285] If a user enters "This product is the worst," the system will input the following prompt sentence into the generative AI model:

[1286] Example prompt:

[1287] User comment: This product is awful

[1288] Please revise this comment to something non-abusive.

[1289] Based on this prompt, the generative AI model generates a phrase such as, "This product did not meet my expectations." The generated phrase is suggested to the user, and if the user selects it, it is posted as a review.

[1290] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[1291] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1292] Step 1:

[1293] Users enter comments through a smartphone application.

[1294] Input: Comment entered by the user (e.g. "This product is awful")

[1295] Output: The entered comment is sent to the server.

[1296] Specific behavior: When a user enters a comment into the text box in the application and presses the "Send" button, the comment is sent to the server.

[1297] Step 2:

[1298] The server receives the comments and sends the prompt to the generative AI model for analysis.

[1299] Input: Comments submitted by users

[1300] Output: The prompt sent to the generative AI model

[1301] Specific behavior: The server receives the comment and generates the following prompt text:

[1302] User comment: This product is awful

[1303] Please revise this comment to something non-abusive.

[1304] The generated prompt sentence is sent to the generative AI model via the OpenAI API.

[1305] Step 3:

[1306] A generative AI model analyzes the prompt and generates non-abusive language.

[1307] Input: prompt statement

[1308] Output: Non-derogatory statements (e.g. "This product did not meet my expectations")

[1309] How it works: The generative AI model analyzes the prompt and generates an appropriate expression, which is then returned to the server via the OpenAI API.

[1310] Step 4:

[1311] The server proposes the generated representation to the user.

[1312] Input: Representations returned by a generative AI model

[1313] Output: Suggested representations for the user

[1314] Specific operation: The server receives the generated representation and proposes it to the user through the smartphone application. The user confirms the proposed representation.

[1315] Step 5:

[1316] The user selects a suggested expression and posts it as a review.

[1317] Input: Proposed expression

[1318] Output: The submitted reviews

[1319] Specific operation: The user selects the appropriate expression from the suggested expressions and presses the "Submit" button. The selected expression is sent to the server for final review and is published.

[1320] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[1321] In addition, an emotion engine that estimates the emotion of the user may be further combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[1322] "Example 1"

[1323] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system receives comments entered by a user and analyzes their contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The emotion engine then analyzes the user's emotions and suggests expressions based on the results. For example, if a user enters a comment such as "This product is the worst," the emotion engine analyzes the user's anger or dissatisfaction from the comment. Then, taking the emotion into consideration, it suggests non-defamatory expressions such as "This product did not meet my expectations."

[1324] "Example 2"

[1325] The emotion engine also learns changes in the user's emotions and suggests expressions that adapt to the user's emotions based on the learning results. For example, if a user has often used the expression "did not meet my expectations" in the past, the emotion engine will learn that expression style. If the user has a similar emotion, it will suggest a similar expression. This allows the user to express their emotions appropriately and provides expressions that are considerate to the reader.

[1326] The process flow of each embodiment will be described below.

[1327] "Example 1"

[1328] Step 1: The user enters a comment into the system.

[1329] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[1330] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[1331] Step 4: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[1332] "Example 2"

[1333] Step 1: The user enters a comment into the system.

[1334] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[1335] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[1336] Step 4: The emotion engine learns from the user's past comments and emotional patterns, and based on that learning, suggests expressions that are adapted to the user's emotions.

[1337] Step 5: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[1338] Example 1

[1339] Next, a description will be given of Example 1 of the embodiment 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".

[1340] In modern social networking services (SNS) and user-generated content (UGC), it is a problem that users post slanderous or offensive comments. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. In addition, users may not be able to express their feelings appropriately, leading to misunderstandings. To solve these problems, a system is needed to help users use constructive and considerate expressions.

[1341] 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.

[1342] In this invention, the server includes a means for receiving comments entered by a user, a means for analyzing the content of the comment using natural language processing technology and recognizing the content the user wants to convey, a means for analyzing the emotions contained in the comment, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the user from past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to use constructive and considerate expressions on SNS and UGC.

[1343] "User" refers to an individual who posts comments or reviews using social networking services and user-generated content.

[1344] "Comment" refers to a text message entered by a user on a social networking service or user-generated content.

[1345] "Natural language processing technology" refers to technology that enables computers to understand, analyze, and generate human language.

[1346] "Sentiment analysis" refers to the technology of extracting user emotions (e.g., anger, dissatisfaction, joy, etc.) from text data.

[1347] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new text or expressions.

[1348] "Values" refer to the beliefs and standards of judgment held by users.

[1349] "Hobbies and interests" refers to a user's preferred activities and interests.

[1350] "Expression style" refers to the unique language and writing style used by a user when creating text.

[1351] "Slander" refers to offensive comments or statements made with the intent to harm others.

[1352] "User-Generated Content" refers to content that users create and share online (e.g., comments, reviews, blog posts, etc.).

[1353] "Reader" refers to any other individual who reads the comments and text created by a User.

[1354] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[1355] The present invention is a system that supports users in using constructive and considerate expressions in SNS and user-generated content (UGC). A specific embodiment of the system is described below.

[1356] Hardware and Software Used

[1357] Hardware: Servers, terminals (PCs, smartphones)

[1358] Software: Natural language processing libraries (e.g., NLTK and spaCy), sentiment analysis engines (e.g., IBM Watson Natural Language Understanding), generative AI models (e.g., GPT-3)

[1359] System Overview

[1360] When a user enters a comment on an SNS, the device sends the comment to a server. The server receives the comment and analyzes the content using natural language processing technology. Natural language processing libraries such as NLTK and spaCy are used for the analysis. The server then uses a sentiment analysis engine to analyze the sentiment contained in the comment. For example, IBM Watson Natural Language Understanding can be used to extract the user's sentiment (anger, dissatisfaction, etc.) from the comment.

[1361] The server then uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from past comments and behavior. The generated expressions are sent to the device and suggested to the user. The user can review the suggested expressions and correct them if necessary. Finally, the corrected comments are posted on the social networking site.

[1362] Examples

[1363] For example, if a user types a comment such as "This product is the worst," the device sends this comment to the server. The server receives the comment, analyzes it using a natural language processing library, and uses a sentiment analysis engine to detect "anger" or "dissatisfaction." It then uses a generative AI model to generate a non-defamatory expression such as "This product did not meet my expectations." The generated expression is sent to the device and suggested to the user. The user reviews the proposed expression, modifies it if necessary to "This product greatly disappointed my expectations," and finally posts it on social media.

[1364] Examples of prompt statements

[1365] "When a user enters a comment on a social networking site, the system must receive the comment, analyze the content using natural language processing technology, and suggest expressions that are not defamatory."

[1366] The system will help users use constructive and considerate language on social media and in UGC.

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

[1368] Program processing flow

[1369] Step 1:

[1370] The user enters a comment.

[1371] A user inputs a comment into the comment input field of a SNS application. For example, the user inputs "This product is the worst." The input comment is saved on the device.

[1372] Step 2:

[1373] The terminal transmits the comment to the server.

[1374] The terminal uses an HTTP POST request to send the entered comment data to the server. The sent data includes the comment text and the user ID.

[1375] Step 3:

[1376] The server receives the comment.

[1377] The server receives the HTTP request and stores the comment data in a database for further analysis.

[1378] Step 4:

[1379] The server analyzes the comments using natural language processing.

[1380] The server uses natural language processing libraries (e.g., NLTK or spaCy) to analyze the grammatical structure and semantics of the comments. Specifically, it performs tokenization, part-of-speech tagging, dependency analysis, etc. The input is the comment text, and the output is the analysis result (e.g., "This / DET product / NOUN is / AUX the worst / ADJ / AUX").

[1381] Step 5:

[1382] The server performs sentiment analysis.

[1383] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comments. The input is the parsed comment text, and the output is the sentiment data (e.g., "angry" or "unhappy").

[1384] Step 6:

[1385] The server generates non-abusive language.

[1386] The server uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from the user's past comments and behavior. The input is emotion data and the user's past comment data, and the output is the generated expression (e.g., "This product did not meet my expectations").

[1387] Step 7:

[1388] The server proposes the generated representation to the user.

[1389] The server sends the generated representation to the terminal and suggests it to the user. The input is the generated representation and the output is the suggested text that is displayed on the user's terminal.

[1390] Step 8:

[1391] The user reviews the proposal and corrects it if necessary.

[1392] The user checks the suggested expression and modifies it if necessary. For example, the user modifies it to "This product greatly exceeded my expectations." Finally, the modified comment is posted on the SNS. The input is the proposed text, and the output is the comment text that is finally posted.

[1393] In this way, users can be more constructive and considerate in their use of language on social media and in UGC.

[1394] (Application example 1)

[1395] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1396] In modern social networking sites and online forums, users posting abusive or offensive comments is a problem. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. Users themselves can also unintentionally use offensive language, which can result in trouble. To solve these problems, a system is needed that suggests expressions to avoid abusive comments in real time as the user types them.

[1397] 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.

[1398] In this invention, the server includes a means for receiving comments input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments input by a user in real time and detecting the possibility of slander, a means for analyzing emotions contained in the comment using an emotion analysis engine, and a means for notifying the user of the proposed expressions and providing options. This makes it possible for the user to receive suggestions of expressions that avoid slander in real time, improving the quality of communication.

[1399] "User" means an individual or entity that uses the System to enter Comments.

[1400] A "comment" is a text message entered by a user on a platform such as a social networking site or online forum.

[1401] "Analysis" is the process of understanding the content of the entered comments and recognizing the user's intent and emotions.

[1402] "Slander" refers to expressions or actions that attack or damage the reputation of others.

[1403] "Expression suggestions" refers to providing alternative expressions to comments entered by users to avoid defamation.

[1404] "Values" refer to the beliefs and standards of judgment held by users.

[1405] "Hobbies" refer to a user's preferred activities and areas of interest.

[1406] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments and uses them to inform future suggestions.

[1407] "Expression style" refers to the unique language and expression used by a user when entering comments.

[1408] "Real-time" refers to instant processing and suggestions occurring the moment a user enters a comment.

[1409] An "sentiment analysis engine" is software that analyzes the emotions contained in comments and suggests appropriate expressions based on the results.

[1410] "Notification" refers to the act of the system informing the user of a suggested expression or option.

[1411] "Choices" are multiple alternatives from which a user can choose among proposed representations.

[1412] The system for implementing the present invention proposes expressions to avoid slander in real time when a user enters a comment on a social networking site or online forum. A specific embodiment of the system will be described below.

[1413] Hardware and Software Used

[1414] Hardware: Smartphones, smart glasses

[1415] Software: Natural language processing libraries (e.g. spaCy, NLTK), sentiment analysis engines (e.g. IBM Watson Natural Language Understanding), machine learning models (e.g. TensorFlow, PyTorch)

[1416] Data processing and calculation

[1417] Comments received

[1418] When a user inputs a comment in the SNS application, the comment is transmitted to the server, which receives the comment in real time using a means for receiving the comment input by the user.

[1419] Natural Language Processing

[1420] The server uses natural language processing libraries to analyze the content of the received comments. Specifically, it uses libraries such as spaCy and NLTK to analyze the grammatical structure and meaning of the comments and recognize what the user is trying to convey.

[1421] Emotion analysis

[1422] The server then uses a sentiment analysis engine to analyze the emotions contained in the comments, for example using IBM Watson Natural Language Understanding to detect emotions such as "anger" or "dissatisfaction" from the comments.

[1423] Expression suggestion generation

[1424] The server uses machine learning models to generate non-abusive expressions based on the values ​​and preferences of users learned from their past comments and behavior. It uses frameworks such as TensorFlow and PyTorch to learn the unique expression style of users and suggest appropriate expressions.

[1425] User Notification

[1426] The proposed expressions are sent from the server to the user's device, and the user can select an option from the suggested expressions to correct the comment.

[1427] Examples

[1428] For example, if a user enters a comment such as "This product is the worst," the server processes it as follows:

[1429] 1. Comment received: "This product is awful."

[1430] 2. Natural Language Processing: Analyze comments to detect potential abusive content

[1431] 3. Sentiment analysis: Detect "anger" using a sentiment analysis engine

[1432] 4. Expression suggestion generation: Generate the expression "This product did not meet my expectations"

[1433] 5. User notification: Notify users of proposed wording

[1434] Examples of prompt statements

[1435] A user wrote in the comment, "This product is the worst." Please use a sentiment analysis engine to detect "anger" and suggest expressions that are not defamatory.

[1436] In this way, a system can be realized that suggests expressions that will help users avoid slander on social media.

[1437] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1438] Step 1:

[1439] A user inputs a comment into an SNS application. The input comment is sent from the user's device to the server. The input data is a comment in text format.

[1440] Step 2:

[1441] The server receives comments entered by users. The received comments are analyzed using a natural language processing library (e.g., spaCy, NLTK). As a result of the analysis, the grammatical structure and meaning of the comments are extracted. The input data is the comments in text format, and the output data is the analyzed grammatical structure and semantic information.

[1442] Step 3:

[1443] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comment. As a result of the analysis, the sentiment contained in the comment (e.g., "anger" or "dissatisfaction") is detected. The input data is the analyzed grammatical structure and semantic information, and the output data is the detected sentiment information.

[1444] Step 4:

[1445] The server uses machine learning models (e.g., TensorFlow, PyTorch) to generate non-abusive expressions based on the values ​​and tastes learned from the user's past comments and behavior. The generated expressions reflect the user's unique expression style. The input data are the detected emotional information and the user's past comments and behavior data, and the output data is the generated non-abusive expressions.

[1446] Step 5:

[1447] The server notifies the user of the generated expression to the user's terminal. The user can select an option from the notified expressions and modify the comment. The input data is the generated non-abusive expression, and the output data is the option notified to the user.

[1448] Step 6:

[1449] The user selects a suggested expression and confirms the revised comment. The confirmed comment is sent to the server again and posted as the final comment to the SNS application. The input data is the expression selected by the user, and the output data is the final comment posted.

[1450] Example 2

[1451] Next, a second embodiment of the second form will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1452] In online comments and reviews, it is difficult for users to express their feelings and opinions appropriately while avoiding slander. In addition, there is a lack of systems that suggest expressions that are adapted to the user's expression style and feelings, making it difficult to communicate with the reader in a considerate manner.

[1453] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's expression style, a means for proposing expressions that are adapted to the user's emotions based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables the user to provide comments that appropriately express their own emotions while also taking the reader into consideration.

[1454] "User" means any individual or entity that uses the System to enter comments and reviews.

[1455] A "comment" is text data including an opinion or impression entered by a user.

[1456] "Analysis" is the process of understanding the content of entered comments and identifying their meaning and sentiment.

[1457] "Slander" is offensive language or language intended to injure others.

[1458] "Expression style" refers to the patterns of words and phrases that a user prefers to use.

[1459] An "emotion engine" is a software component that analyzes a user's emotions and suggests expressions that are appropriate to those emotions.

[1460] A "generative AI model" is an artificial intelligence algorithm that generates appropriate representations based on user input.

[1461] A "suggestion" is an alternative wording or phrase that the system offers to the user.

[1462] "Selection" is the act of the user choosing the appropriate expression from those suggested by the system.

[1463] "Text generation" is the process of combining user-selected expressions to create new text.

[1464] "Reader" means any other individual or entity that reads user-generated comments or reviews.

[1465] "User-Generated Content" refers to content that users themselves create and post on the Internet, such as through social networking services, comments, and reviews.

[1466] This invention is a system that provides appropriate expressions while avoiding slander when users enter comments or reviews on the Internet. The system works in cooperation with three parties: a server, a terminal, and a user.

[1467] Hardware and Software Used

[1468] Hardware: Servers, terminals (user devices)

[1469] Software: Generative AI models (e.g. GPT-4), sentiment engines, natural language processing (NLP) libraries (e.g. spaCy, NLTK)

[1470] Program Processing

[1471] 1. User comments

[1472] The user uses the terminal to type in a comment, for example, "This product is awful."

[1473] 2. Comment Analysis

[1474] The server uses generative AI models and natural language processing libraries to analyze user comments. Specifically, it uses NLP libraries to perform sentiment analysis on comments and determine whether they are potentially abusive.

[1475] 3. Suggested expressions

[1476] The server uses a generative AI model to generate non-abusive expressions, such as "This product did not meet my expectations."

[1477] 4. Learning the user's expression style

[1478] The emotion engine learns from the expressions the user has used in the past. If the user often uses the expression "did not meet my expectations," the emotion engine will learn that expression style.

[1479] 5. Viewing and selecting offers

[1480] The server generates suggestions and displays them on the user's device, where the user can select suggested expressions and combine them to generate new text.

[1481] 6. Suggesting expressions that adapt to emotions

[1482] The emotion engine learns the changes in the user's emotions and suggests expressions that are adapted to the user's emotions based on the learning results. For example, if the user has a similar emotion, it will suggest the expression "did not meet my expectations" that was used in the past.

[1483] Examples and prompts

[1484] Examples

[1485] User input: "This product is awful."

[1486] Server analysis results: Possible slander

[1487] Suggested phrase: "This product did not meet my expectations."

[1488] Examples of prompt statements

[1489] Prompt: "Please suggest appropriate wording to make the user's comment less abusive. For example, in response to the comment 'This product is terrible,' suggest 'This product did not meet my expectations.'"

[1490] This system enables users to provide comments that are considerate of the reader while also expressing their own feelings appropriately.

[1491] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1492] Step 1: User comments

[1493] Action: A user uses a terminal to enter a comment.

[1494] Input: User types "This product is awful."

[1495] Output: The entered comment data is sent to the server.

[1496] Step 2: Analyzing comments

[1497] How it works: The server analyzes user comments using generative AI models and natural language processing libraries.

[1498] Input: Comment data received by the server: "This product is awful."

[1499] Data processing: Perform sentiment analysis of comments using NLP libraries.

[1500] Output: Analysis indicates the comment is negative and potentially abusive.

[1501] Step 3: Propose a representation

[1502] How it works: The server uses a generative AI model to generate non-abusive remarks.

[1503] Input: Comment data that, based on the analysis results, is potentially abusive.

[1504] Data computation: Generative AI models generate appropriate representations.

[1505] Output: A suggested wording: "This product did not meet my expectations."

[1506] Step 4: Learning the user's expression style

[1507] How it works: The emotion engine learns from expressions the user has used in the past.

[1508] Input: Comment data used by the user in the past.

[1509] Data processing: The emotion engine analyzes and learns from the user's expression style.

[1510] Output: Learning results about the user's expression style.

[1511] Step 5: View and select suggestions

[1512] How it works: The server generates suggestions and displays them on the user's device.

[1513] Input: Suggested phrase "This product did not meet my expectations."

[1514] Output: The suggestion is displayed on the user's terminal. The user selects the suggestion.

[1515] Step 6: Suggest expressions that match emotions

[1516] How it works: The emotion engine learns the changes in the user's emotions and, based on that learning, suggests expressions that are adapted to the user's emotions.

[1517] Input: User emotion data and past expression style learning results.

[1518] Data computation: The emotion engine generates adapted expressions.

[1519] Output: Expressions adapted to the user's emotions are suggested.

[1520] (Application example 2)

[1521] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1522] When users post product reviews on online shopping sites, they often include slanderous or inappropriate language, which can cause discomfort to other users. There is also the issue that it is difficult for users to generate comments that express their feelings appropriately while also being considerate of the reader.

[1523] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1524] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, and a means for a review posting system on an online shopping site, when a user posts a product review, a means for the system to automatically analyze the comment and suggest expressions that are not slanderous, a means for the user to select and combine the proposed expressions to generate new text, a means for generating expressions that are not slanderous using a generation AI model, and a means for inputting prompt sentences into the generation AI model. This makes it possible for the user to generate comments that take the reader into consideration while appropriately expressing their own feelings.

[1525] "User" means any individual or entity that uses the System to post comments and reviews.

[1526] A "comment" is an opinion or impression entered in text form by a user.

[1527] "Analysis" is the process of understanding the content of the entered comments and extracting their meaning and intent.

[1528] "Slander" is offensive language or language intended to injure others.

[1529] "Expression suggestions" means that the system suggests appropriate, non-defamatory expressions based on comments entered by the user.

[1530] "Values" are the beliefs and ways of thinking held by users.

[1531] "Hobbies" are the activities or things that a user likes.

[1532] "Learning" is the process by which a system understands patterns and trends based on a user's past actions and expressions.

[1533] "Style" refers to the characteristics of the words and phrases a user uses in a particular situation.

[1534] A "generative AI model" is an algorithm or system that uses artificial intelligence to generate new text or expressions.

[1535] A "prompt" is an instruction or question to be input into a generative AI model.

[1536] An "online shopping site" is a website that sells products over the Internet.

[1537] A "review posting system" is a system that has the function of allowing users to post their opinions and ratings on products and services.

[1538] "Text generation" is the process of combining user-selected expressions to create new sentences.

[1539] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and proposing expressions that are not slanderous. A specific embodiment of this system is described below.

[1540] System Configuration

[1541] The system consists of the following main components:

[1542] 1. User device: The device through which a user enters comments (e.g. smartphone, computer).

[1543] 2. Server: A central processing unit that performs comment analysis, expression suggestion, learning, and text generation.

[1544] 3. Generative AI models: Artificial intelligence models for generating non-abusive language (e.g., OpenAI’s GPT-3).

[1545] 4. Database: Storage for saving users' past comments and learning data.

[1546] Program Processing

[1547] 1. Receiving comments

[1548] Comments entered from the user's terminal are sent to the server, which receives the comments and prepares them for analysis.

[1549] 2. Comment Analysis

[1550] The server inputs the received comments into the generative AI model, which analyzes the content of the comments and recognizes what the user wants to convey.

[1551] 3. Expression suggestions

[1552] The generative AI model generates non-abusive expressions based on the analysis results, and the server sends these suggestions to the user's device.

[1553] 4. Learning

[1554] The server learns from the user's past comments and selected expressions, and acquires an expression style based on the user's values, hobbies, and preferences.

[1555] 5. Text Generation

[1556] The user selects and combines suggested expressions to generate new text, which the server stores and optionally displays to other users.

[1557] Hardware and software used

[1558] Hardware: General purpose PC or server

[1559] Software: Python, OpenAI API

[1560] Examples

[1561] If a user enters a comment such as "This product is terrible," the server sends this comment to the generative AI model, which then suggests a non-defamatory rendition of "This product did not meet my expectations." The user selects this suggestion and posts it as the final comment.

[1562] Examples of prompt statements

[1563] User comment: 'This product is awful'

[1564] Please suggest non-defamatory expressions.

[1565] In this way, users can generate comments that are considerate of the reader while still expressing their own feelings appropriately.

[1566] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1567] Step 1:

[1568] The user inputs a comment. The user terminal receives the comment input by the user and transmits it to the server. The input is the user's comment, and the output is the transmission of comment data to the server.

[1569] Step 2:

[1570] The server prepares the received comments for analysis. Specifically, it converts the comment data into a format suitable for the generative AI model. The input is the comment data sent by the user, and the output is a prompt sentence to be input to the generative AI model.

[1571] Step 3:

[1572] The server sends the prompt to the generative AI model. The generative AI model analyzes the content of the comment based on the prompt and recognizes what the user wants to convey. The input is the prompt, and the output is the analysis result.

[1573] Step 4:

[1574] The generative AI model generates non-abusive expressions based on the analysis results. The server receives the generated expressions and sends them to the user's device. The input is the analysis results, and the output is a suggestion for non-abusive expressions.

[1575] Step 5:

[1576] The user terminal displays the suggested expressions received from the server to the user. The user selects and optionally combines the suggested expressions. The input is the suggested expressions and the output is the user's selected expression.

[1577] Step 6:

[1578] The user sends the representation selected by the user to the server, which generates the final text based on the representation selected. The input is the user's representation and the output is the final text.

[1579] Step 7:

[1580] The server stores the generated final text in a database and displays it to other users if desired. The input is the final text and the output is storage in a database and display to other users.

[1581] In this way, a series of processes is achieved: the comments entered by the user are analyzed, non-abusive expressions are suggested, and the final text is generated.

[1582] 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 voice indicating a user input for 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.

[1583] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to 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.

[1584] Another example of generative AI is Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1585] 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.

[1586] [Fourth embodiment]

[1587] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1588] 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.

[1589] 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 wide area network (WAN) and / or a local area network (LAN).

[1590] 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. In addition, the microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1591] 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 the voice according to instructions from the processor 46.

[1592] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[1593] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[1594] The control target 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, legs, etc. The posture and behavior of the robot 414 are controlled by controlling the motors of the arms, hands, legs, etc. 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.

[1595] Fig. 8 shows an example of 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.

[1596] 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.

[1597] 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.

[1598] In the robot 414, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1599] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1600] "Example 1"

[1601] In one embodiment of the present invention, when a user enters a comment on a social networking site, the system receives the comment and analyzes its contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The system then suggests expressions that are not defamatory. These suggestions are based on the user's values, hobbies, and preferences, learned from the user's past comments and actions. The AI ​​then learns the user's unique expression style based on the results of its learning.

[1602] "Example 2"

[1603] As a specific example, if a user enters the comment "This product is the worst," the system will analyze the comment and suggest non-defamatory expressions such as "This product did not meet my expectations." Also, if the user has often used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions. The user can select expressions suggested by the AI ​​and combine them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[1604] The process flow of each embodiment will be described below.

[1605] "Example 1"

[1606] Step 1: A user enters a comment on a social networking site.

[1607] Step 2: The system receives the comment and analyzes its content using natural language processing techniques to understand what the user is trying to convey.

[1608] Step 3: The system suggests non-abusive expressions based on the user's values, preferences, and tastes, learned from the user's past comments and behavior.

[1609] Step 4: Based on what it has learned, the AI ​​learns the user's unique expression style.

[1610] "Example 2"

[1611] Step 1: User types in a comment saying "This product is awful."

[1612] Step 2: The system analyzes the comment and suggests non-abusive wording, such as "This product did not meet my expectations."

[1613] Step 3: If the user has frequently used the expression "did not meet expectations" in the past, the AI ​​will learn that expression style and suggest similar expressions.

[1614] Step 4: The user selects expressions suggested by the AI ​​and combines them to generate new text. In this way, it is possible to create expressions that take the reader into consideration.

[1615] Example 1

[1616] Next, a description will be given of Example 1 of the embodiment 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".

[1617] It has become a problem for users to post slanderous or negative comments on user-generated content such as social media and review sites. This can cause psychological damage to other users, the targeted individuals, or companies. There is also the issue that it is difficult for users to find appropriate expressions based on their own values ​​and hobbies. Furthermore, there is a demand for text to be generated with the same consideration as in face-to-face communication.

[1618] 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.

[1619] In this invention, the server includes a means for receiving comments entered by users, a means for analyzing the contents of the comments using natural language processing technology and recognizing the contents the users want to convey, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the users from their past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to post comments using appropriate expressions based on their own values ​​and hobbies while avoiding defamation.

[1620] "User" means any person or entity that utilizes the System to enter comments and review and select suggested expressions.

[1621] A "comment" is text information that a user enters into user-generated content such as on social media or review sites.

[1622] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[1623] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language.

[1624] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[1625] "Values" refer to the beliefs and standards of judgment held by users.

[1626] "Hobbies and interests" refer to the activities and interests that a user prefers.

[1627] "Expression style" refers to the unique language and writing style used by a user when entering comments.

[1628] "Slander" refers to statements that speak ill of or harm others.

[1629] "User Generated Content" means content that is created and published on the Internet by users.

[1630] "Analysis" is the process of understanding the content of a comment and extracting meaning and sentiment.

[1631] A "suggestion" is an appropriate expression that the server provides to the user using a generative AI model.

[1632] "Text generation" is the process of combining user-selected expressions to create the final comment.

[1633] "Reader" means any other user or audience who reads User Generated Comments.

[1634] The present invention provides a system that suggests appropriate expressions while avoiding slander when a user enters a comment on user-generated content such as a social networking site or a review site. A specific embodiment of the system will be described below.

[1635] The server receives the comments entered by the user. The user uses his / her own device (smartphone or PC) to access the SNS application and enters a comment. For example, the user enters "This movie was boring."

[1636] The server receives comments sent from the social networking application via HTTP requests and stores them in a database.

[1637] The server then analyzes the received comments using natural language processing technology. Specifically, it analyzes the content of the comments using natural language processing libraries such as "spaCy" and "NLTK" through programs written in programming languages ​​such as Python and Java. During the analysis process, it tokenizes the comments, tags them by part of speech, and performs sentiment analysis. For example, it recognizes that the part "it was boring" indicates a negative sentiment.

[1638] Based on the analysis results, the server uses a generative AI model to suggest non-defamatory expressions. For example, "GPT-3" is used as a generative AI model. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, by inputting the prompt sentence "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression," the expression "This movie did not meet my expectations" is generated.

[1639] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device.

[1640] The user reviews the proposed expressions from the server, and if necessary, modifies the proposed expressions and posts them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations."

[1641] As a concrete example, consider the case where a user comments, "This restaurant is the worst." The server receives this comment and analyzes it using natural language processing technology. Based on the analysis results, the server inputs a prompt sentence to the generative AI model (GPT-3) saying, "The user commented, 'This restaurant is the worst.' Please suggest a non-defamatory expression." The server obtains the expression "This restaurant did not meet my expectations" from the generative AI model and returns it to the user. The user checks the proposed expression and posts the final comment, stating, "This restaurant did not meet my expectations."

[1642] In this way, the system helps users phrase their comments in a non-abusive manner.

[1643] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1644] Step 1:

[1645] A user enters a comment on a social networking site.

[1646] A user accesses an SNS application using his / her own device (smartphone or PC) and inputs a comment. For example, a user inputs "This movie was boring." The input comment is sent to the server via the SNS application.

[1647] Step 2:

[1648] The server receives the comment.

[1649] The server receives comments sent from the SNS application via HTTP requests. The received comments are stored in a database. The input is the user's comment and the output is the comment stored in the database.

[1650] Step 3:

[1651] The server analyzes the comments using natural language processing technology.

[1652] The server analyzes the received comments using a natural language processing library (e.g., spaCy or NLTK). Specifically, it performs tokenization of the comments, part-of-speech tagging, sentiment analysis, etc. For example, it recognizes that the part "it was boring" indicates a negative sentiment. The input is the comments stored in the database, and the output is the analysis result.

[1653] Step 4:

[1654] Based on the analysis results, the server uses a generative AI model to suggest expressions that are not defamatory.

[1655] The server uses a generative AI model (e.g., GPT-3) based on the analysis results to generate non-defamatory expressions. A prompt sentence is input to the generative AI model to obtain an appropriate expression. For example, the prompt sentence is input as follows: "The user commented that 'this movie was boring.' Please suggest a non-defamatory expression." The input is the analysis result and the prompt sentence, and the output is the generated expression.

[1656] Step 5:

[1657] The server returns suggestions to the user.

[1658] The server returns the generated representation to the user through the SNS application, which displays the suggested representations on the user's device. The input is the generated representation and the output is the suggestions displayed on the user's device.

[1659] Step 6:

[1660] The user reviews the proposal and corrects it if necessary.

[1661] The user reviews the expressions suggested by the server. If necessary, he / she modifies the suggested expressions and submits them as final comments. For example, the user may adopt an expression such as "This movie did not meet my expectations." The input is the server's suggestions and the output is the final comment.

[1662] (Application example 1)

[1663] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1664] It is becoming a problem that users unintentionally post slanderous comments on social media and messaging apps. This can make online communication unpleasant and lead to trouble and misunderstandings. There is also the issue that it is difficult for users to find appropriate expressions based on their own values, hobbies and preferences.

[1665] 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. In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content that the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for functioning as an application installed on a smartphone, a means for analyzing and proposing comments in real time, a means for generating expressions that are not slanderous using a generative AI model, and a means for inputting a prompt sentence into the generative AI model and proposing appropriate expressions. This allows users to communicate online with peace of mind.

[1666] "User" means an individual or entity that uses the System to enter Comments.

[1667] A "comment" is a text message that a user enters on social media, messaging apps, etc.

[1668] "Analysis" is the process of using natural language processing technology to understand the content of the entered comment and recognize its meaning.

[1669] "Slander" is offensive language or language intended to injure others.

[1670] A "suggestion" is an act in which the system shows the user appropriate language that is not defamatory.

[1671] "Values" are the beliefs and ways of thinking held by users.

[1672] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[1673] "Learning" is the process by which the system extracts patterns from a user's past behavior and comments to deepen its understanding.

[1674] "Expression style" refers to the unique phrasing and writing style used by a user when entering comments.

[1675] "AI" is a system that uses artificial intelligence technology to analyze data and make learning and suggestions.

[1676] "Text generation" is the process of combining user-selected expressions to create new comments.

[1677] "Reader" means an individual or entity that receives comments entered by a User.

[1678] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[1679] "Real-time" refers to the fact that analysis and suggestions are carried out the moment a user enters a comment.

[1680] A "generative AI model" is an algorithm or program that uses artificial intelligence technology to generate text.

[1681] A "prompt" is an instruction or question that is input into a generative AI model.

[1682] The system for carrying out the present invention functions as a smartphone application that analyzes comments entered by users in real time and suggests expressions that are not defamatory. A specific embodiment of this system will be described below.

[1683] System Configuration

[1684] The system consists of the following main components:

[1685] 1. User device (smartphone): The device where the user enters comments.

[1686] 2. Server: A central processing unit that analyzes comments and makes suggestions.

[1687] 3. Generative AI models: Algorithms that perform natural language processing and text generation.

[1688] Hardware and Software Used

[1689] Hardware: Smartphones, servers

[1690] Software: Python, OpenAI API, spaCy (natural language processing library)

[1691] Data processing and calculation

[1692] 1. Receiving comments:

[1693] The user inputs comments through a smartphone application, and the input comments are sent to the server.

[1694] 2. Comment Analysis:

[1695] The server uses spaCy to analyze comments and understand their grammar and meaning, and the results are used to understand what the user is trying to say.

[1696] 3. Suggestions for non-abusive language:

[1697] The server uses OpenAI's generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[1698] Please rewrite your comment below so it is not abusive: You are such an idiot!

[1699] 4. Learning about user values, preferences, and tastes:

[1700] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. Based on this learning, the AI ​​will master the user's unique expression style.

[1701] 5. Text Generation:

[1702] Users can generate new comments by selecting and combining expressions suggested by the AI. The generated comments will be expressions that take the reader into consideration.

[1703] Examples

[1704] For example, if a user types a comment like, "You're such an idiot!", the system will analyze it and suggest non-abusive expressions such as:

[1705] I don't agree with that opinion, but I'd like you to think about it a bit more.

[1706] The user can select this suggestion and use it as the final comment.

[1707] In this way, users can communicate online with peace of mind.

[1708] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1709] Step 1:

[1710] A user inputs a comment through a smartphone application. The input comment is sent from the user terminal to a server. The input data is the text comment input by the user. The output data is the comment text sent to the server.

[1711] Step 2:

[1712] The server analyzes the received comments. Specifically, it uses spaCy to analyze the grammar and semantics of the comments. The input data is the comment text entered by the user. As data processing, spaCy tokenizes the comments and analyzes their grammatical structure. The output data is the grammatical structure and semantic information of the analyzed comments.

[1713] Step 3:

[1714] The server recognizes what the user wants to convey based on the analysis results. The input data is the grammatical structure and semantic information of the analyzed comment. The data calculation uses natural language processing technology to extract the meaning of the comment. The output data is the semantic information of the comment, including the user's intention.

[1715] Step 4:

[1716] The server uses a generative AI model to generate non-abusive expressions. Specifically, the server inputs the following prompt sentence into the generative AI model:

[1717] Please rewrite the following comment to be non-abusive: [User comment]

[1718] The input data is the semantic information of the comment, including the user's intention, and the prompt sentence. As a data calculation, the generative AI model analyzes the prompt sentence and generates an appropriate expression. The output data is an expression that is not defamatory.

[1719] Step 5:

[1720] The server analyzes the user's past comments and behavior to learn the user's values, hobbies, and preferences. The input data is the user's past comments and behavioral history. To process the data, machine learning algorithms are used to extract patterns and model the user's values, hobbies, and preferences. The output data is a model that reflects the user's values, hobbies, and preferences.

[1721] Step 6:

[1722] The server uses the AI ​​to learn the user's unique expression style based on the learning results. The input data is a model that reflects the user's values ​​and hobbies. As a data calculation, the AI ​​uses the model to learn the user's expression style. The output data is an AI model that reflects the user's expression style.

[1723] Step 7:

[1724] Users generate new comments by selecting and combining expressions suggested by the AI. The input data are non-abusive expressions suggested by the AI. In concrete terms, the user selects the suggested expressions on the smartphone application and generates the final comment. The output data is the non-abusive comment selected by the user.

[1725] Step 8:

[1726] The server saves the generated comments as expressions that take the reader into consideration, and posts them to social media or messaging apps as necessary. The input data are non-abusive comments selected by the user. Data processing involves converting the comments into an appropriate format and saving or posting them. The output data are the final comments that take the reader into consideration.

[1727] Example 2

[1728] Next, a second embodiment of the second embodiment will be described. 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".

[1729] It has become a problem that slander and offensive language is often seen in comments and reviews on the Internet. This can lead to a decline in the quality of communication and trouble between users. In addition, since it is difficult for users to use appropriate expressions based on their own values ​​and tastes, there is a demand for generating text that has the same consideration as face-to-face communication.

[1730] 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.

[1731] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for suggesting non-abusive expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions suggested by the AI, a means for creating expressions that take the reader into consideration, a means for sending a prompt sentence to the generative AI model, a means for the generative AI model to generate appropriate expressions, and a means for suggesting the generated expressions to the user. This enables the user to use non-abusive expressions and generate text with the same consideration as face-to-face communication.

[1732] "User" means any individual or entity that uses the System to enter comments and reviews.

[1733] A "comment" is an opinion or impression entered in text form by a user.

[1734] A "server" is a computer system that receives comments from users, analyzes them, and suggests appropriate expressions.

[1735] "Analysis" is the process of understanding the content of comments entered by users and recognizing their emotions and intent.

[1736] "Slander" is offensive language intended to harm others.

[1737] "Expression" refers to words or phrases that adequately convey what the user wants to say.

[1738] "Values" are the beliefs and ways of thinking held by users.

[1739] "Hobbies and interests" are the activities and interests that a user prefers.

[1740] "Artificial intelligence" is a machine learning model that learns a user's values, hobbies, and preferences and suggests appropriate expressions.

[1741] A "generative AI model" is an artificial intelligence model that generates appropriate expressions based on user comments.

[1742] A "prompt sentence" is an instruction sentence that instructs a generative AI model to generate an appropriate expression.

[1743] "Text" refers to user-generated sentences and comments.

[1744] "User-Generated Content" is content, such as comments and reviews, that is created and published online by users.

[1745] A "social networking service" is an online platform that enables users to communicate with other users.

[1746] "Face-to-face communication" refers to communication that takes place by meeting face to face, and consideration for the other person is required.

[1747] This invention is a system that analyzes comments entered by users and suggests expressions that are not slanderous. The system uses a server, a terminal, and a generative AI model to receive and analyze comments entered by users and generate appropriate expressions.

[1748] The server receives the comments entered by the user. For example, if the user enters the comment "This product is the worst," the device sends this comment to the server. The server analyzes the received comments using natural language processing (NLP) techniques. For this analysis, it uses generative AI models such as Google's BERT and OpenAI's GPT-3.

[1749] Once the analysis is complete, the server generates non-abusive expressions based on the content of the comment. For example, it suggests expressions such as "This product did not meet my expectations." At this time, the server also refers to the user's past comment history and learns the user's frequently used expression style.

[1750] The user can select expressions suggested by the server and combine them to generate new text, which enables the user to create expressions that take the reader into consideration.

[1751] As a concrete example, if a user types a comment like "This product is the worst", the server will input the following prompt sentence into the generative AI model:

[1752] Example prompt:

[1753] User comment: "This product is awful."

[1754] Please revise this comment to something non-abusive.

[1755] Based on this prompt, the generative AI model generates the phrase, "This product did not meet my expectations." The user can accept this suggestion and use it as the new text.

[1756] The system helps users to use non-abusive language and generate text with the same level of consideration as in face-to-face communication, which is expected to improve the quality of communication on the Internet and reduce trouble between users.

[1757] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1758] Step 1:

[1759] The user enters a comment.

[1760] The user inputs a comment into an input field of the terminal, for example, "This product is the worst." The input comment is temporarily stored in the memory of the terminal.

[1761] Step 2:

[1762] The terminal transmits the comment to the server.

[1763] The terminal sends the comments entered by the user to the server. At this time, the terminal uses an HTTP request to send the comment data to the server. The input data is packaged in JSON format and sent to the server over the network. For example, the data {"comment": "This product is the worst"} is sent.

[1764] Step 3:

[1765] The server receives and parses the comment.

[1766] The server receives comments sent from the terminal. It analyzes the received comments using natural language processing (NLP) technology. Specifically, it performs sentiment analysis of the comments and extracts keywords. A Python NLP library (e.g., spaCy or NLTK) is used for the analysis. The input data is the comment text, and the output data is the analysis results (e.g., sentiment score and keyword list).

[1767] Step 4:

[1768] The server sends a prompt to the generative AI model.

[1769] The server sends a prompt to the generative AI model based on the analysis results. For example, it generates a prompt like, "User comment: 'This product is the worst.' Please convert this comment into a non-defamatory expression." The input data is the analysis results, and the output data is the prompt.

[1770] Step 5:

[1771] A generative AI model generates an appropriate representation.

[1772] A generative AI model (e.g., OpenAI's GPT-3) generates an appropriate expression based on a prompt sent from the server. For example, it generates an expression such as "This product did not meet my expectations." The input data is the prompt, and the output data is the generated expression.

[1773] Step 6:

[1774] The server proposes the generated representation to the user.

[1775] The server proposes the representations received from the generative AI model to the user and displays the proposed representations on the device screen. The input data is the generated representation, and the output data is the proposal displayed to the user.

[1776] Step 7:

[1777] The user selects a suggested expression to generate new text.

[1778] The user selects an expression suggested by the server, generates new text based on the selected expression, and posts it as a comment. The input data is the proposed expression, and the output data is the generated new text.

[1779] (Application example 2)

[1780] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1781] In conventional comment systems, the comments entered by users often contained slander or negative language, which caused discomfort to other users. It was also difficult for users to use language based on their own values ​​and preferences, making it difficult to generate text that was as considerate as face-to-face communication. Another problem was that it took time for users to find the appropriate language when posting a review.

[1782] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1783] In this invention, the server includes a means for receiving a comment input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing non-libelous expressions, a means for learning the user's values ​​and hobbies, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, a means for creating expressions that take the reader into consideration, a means for analyzing comments using a generative AI model and generating non-libelous expressions, a means for proposing the generated expressions to the user and allowing the user to select them, and a means for posting the expressions selected by the user as a review. This allows the user to easily generate non-libelous expressions and post reviews that are considerate of other users.

[1784] "User" means any individual or entity that uses the System to enter comments and reviews.

[1785] A "comment" is an opinion or impression entered in text form by a user.

[1786] "Analysis" is the process of understanding the content of the entered comments and extracting meaning.

[1787] "Slander" is negative or offensive language that hurts others.

[1788] "Expression" is the way in which a user puts into words what they want to communicate.

[1789] A "suggestion" is an alternative representation or opinion that the system offers to the user.

[1790] "Values" are the beliefs and ways of thinking held by users.

[1791] "Hobbies and Interests" refers to a user's preferred activities and areas of interest.

[1792] "Learning" is the process by which a system accumulates knowledge based on the user's past behavior and input.

[1793] A "generative AI model" is an algorithm that uses artificial intelligence to generate new text or expressions.

[1794] A "prompt sentence" is an instruction sentence input to a generative AI model.

[1795] A "review" is an evaluation or opinion a user gives about a product or service.

[1796] "Posting" refers to the act of publishing comments or reviews created by a user.

[1797] The system for implementing the present invention includes a series of processes for receiving and analyzing comments entered by users and suggesting non-abusive expressions. Specific embodiments are described below.

[1798] Hardware and Software Used

[1799] Hardware: Smartphones, servers

[1800] Software: Python, OpenAI API

[1801] System Configuration

[1802] 1. User terminal: Provides an interface for users to input comments. Implemented as a smartphone application.

[1803] 2. Server: Receives the comments, analyzes them, generates non-abusive expressions, suggests them to the user, and posts the expression the user selects as a review.

[1804] Data processing and calculation

[1805] 1. Receiving comments: A user inputs a comment through a smartphone application. The input comment is sent to the server.

[1806] 2. Comment analysis: The server uses the OpenAI API to analyze the entered comments. Specifically, it uses a generative AI model (e.g. GPT-3) to understand the content of the comment and generate non-defamatory expressions.

[1807] 3. Representation Suggestion: The generated representations are suggested to the user, who selects the appropriate one from the suggested representations.

[1808] 4. Posting a Review: The expression selected by the user will be posted as the final review.

[1809] Examples

[1810] If a user enters "This product is the worst," the system will input the following prompt sentence into the generative AI model:

[1811] Example prompt:

[1812] User comment: This product is awful

[1813] Please revise this comment to something non-abusive.

[1814] Based on this prompt, the generative AI model generates a phrase such as, "This product did not meet my expectations." The generated phrase is suggested to the user, and if the user selects it, it is posted as a review.

[1815] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[1816] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1817] Step 1:

[1818] Users enter comments through a smartphone application.

[1819] Input: Comment entered by the user (e.g. "This product is awful")

[1820] Output: The entered comment is sent to the server.

[1821] Specific behavior: When a user enters a comment into the text box in the application and presses the "Send" button, the comment is sent to the server.

[1822] Step 2:

[1823] The server receives the comments and sends the prompt to the generative AI model for analysis.

[1824] Input: Comments submitted by users

[1825] Output: The prompt sent to the generative AI model

[1826] Specific behavior: The server receives the comment and generates the following prompt text:

[1827] User comment: This product is awful

[1828] Please revise this comment to something non-abusive.

[1829] The generated prompt sentence is sent to the generative AI model via the OpenAI API.

[1830] Step 3:

[1831] A generative AI model analyzes the prompt and generates non-abusive language.

[1832] Input: prompt statement

[1833] Output: Non-derogatory statements (e.g. "This product did not meet my expectations")

[1834] How it works: The generative AI model analyzes the prompt and generates an appropriate expression, which is then returned to the server via the OpenAI API.

[1835] Step 4:

[1836] The server proposes the generated representation to the user.

[1837] Input: Representations returned by a generative AI model

[1838] Output: Suggested representations for the user

[1839] Specific operation: The server receives the generated representation and proposes it to the user through the smartphone application. The user confirms the proposed representation.

[1840] Step 5:

[1841] The user selects a suggested expression and posts it as a review.

[1842] Input: Proposed expression

[1843] Output: The submitted reviews

[1844] Specific operation: The user selects the appropriate expression from the suggested expressions and presses the "Submit" button. The selected expression is sent to the server for final review and is published.

[1845] In this way, users can easily generate non-abusive language and post reviews that are considerate of other users.

[1846] In addition, an emotion engine that estimates the emotion of the user may be further combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[1847] "Example 1"

[1848] As one embodiment of the present invention, a system incorporating an emotion engine is provided. This system receives comments entered by a user and analyzes their contents. The analysis is performed using natural language processing technology to recognize the content that the user wants to convey. The emotion engine then analyzes the user's emotions and suggests expressions based on the results. For example, if a user enters a comment such as "This product is the worst," the emotion engine analyzes the user's anger or dissatisfaction from the comment. Then, taking the emotion into consideration, it suggests non-defamatory expressions such as "This product did not meet my expectations."

[1849] "Example 2"

[1850] The emotion engine also learns changes in the user's emotions and suggests expressions that adapt to the user's emotions based on the learning results. For example, if a user has often used the expression "did not meet my expectations" in the past, the emotion engine will learn that expression style. If the user has a similar emotion, it will suggest a similar expression. This allows the user to express their emotions appropriately and provides expressions that are considerate to the reader.

[1851] The process flow of each embodiment will be described below.

[1852] "Example 1"

[1853] Step 1: The user enters a comment into the system.

[1854] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[1855] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[1856] Step 4: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[1857] "Example 2"

[1858] Step 1: The user enters a comment into the system.

[1859] Step 2: The system receives the entered comment and analyzes its content using natural language processing techniques.

[1860] Step 3: The emotion engine extracts the user's emotion from the analysis results and suggests expressions based on that emotion.

[1861] Step 4: The emotion engine learns from the user's past comments and emotional patterns, and based on that learning, suggests expressions that are adapted to the user's emotions.

[1862] Step 5: The proposed expressions are presented to the user, who selects the appropriate expression and generates the text.

[1863] Example 1

[1864] Next, a description will be given of Example 1 of the embodiment 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".

[1865] In modern social networking services (SNS) and user-generated content (UGC), it is a problem that users post slanderous or offensive comments. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. In addition, users may not be able to express their feelings appropriately, leading to misunderstandings. To solve these problems, a system is needed to help users use constructive and considerate expressions.

[1866] 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.

[1867] In this invention, the server includes a means for receiving comments entered by a user, a means for analyzing the content of the comment using natural language processing technology and recognizing the content the user wants to convey, a means for analyzing the emotions contained in the comment, a means for proposing non-defamatory expressions using a generative AI model, a means for learning the values ​​and hobbies of the user from past comments and actions, a means for the AI ​​to learn the user's unique expression style based on the learning results, a means for the user to generate text by selecting and combining expressions proposed by the AI, and a means for creating expressions that take the reader into consideration. This enables users to use constructive and considerate expressions on SNS and UGC.

[1868] "User" refers to an individual who posts comments or reviews using social networking services and user-generated content.

[1869] "Comment" refers to a text message entered by a user on a social networking service or user-generated content.

[1870] "Natural language processing technology" refers to technology that enables computers to understand, analyze, and generate human language.

[1871] "Sentiment analysis" refers to the technology of extracting user emotions (e.g., anger, dissatisfaction, joy, etc.) from text data.

[1872] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new text or expressions.

[1873] "Values" refer to the beliefs and standards of judgment held by users.

[1874] "Hobbies and interests" refers to a user's preferred activities and interests.

[1875] "Expression style" refers to the unique language and writing style used by a user when creating text.

[1876] "Slander" refers to offensive comments or statements made with the intent to harm others.

[1877] "User-Generated Content" refers to content that users create and share online (e.g., comments, reviews, blog posts, etc.).

[1878] "Reader" refers to any other individual who reads the comments and text created by a User.

[1879] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[1880] The present invention is a system that supports users in using constructive and considerate expressions in SNS and user-generated content (UGC). A specific embodiment of the system is described below.

[1881] Hardware and Software Used

[1882] Hardware: Servers, terminals (PCs, smartphones)

[1883] Software: Natural language processing libraries (e.g., NLTK and spaCy), sentiment analysis engines (e.g., IBM Watson Natural Language Understanding), generative AI models (e.g., GPT-3)

[1884] System Overview

[1885] When a user enters a comment on an SNS, the device sends the comment to a server. The server receives the comment and analyzes the content using natural language processing technology. Natural language processing libraries such as NLTK and spaCy are used for the analysis. The server then uses a sentiment analysis engine to analyze the sentiment contained in the comment. For example, IBM Watson Natural Language Understanding can be used to extract the user's sentiment (anger, dissatisfaction, etc.) from the comment.

[1886] The server then uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from past comments and behavior. The generated expressions are sent to the device and suggested to the user. The user can review the suggested expressions and correct them if necessary. Finally, the corrected comments are posted on the social networking site.

[1887] Examples

[1888] For example, if a user types a comment such as "This product is the worst," the device sends this comment to the server. The server receives the comment, analyzes it using a natural language processing library, and uses a sentiment analysis engine to detect "anger" or "dissatisfaction." It then uses a generative AI model to generate a non-defamatory expression such as "This product did not meet my expectations." The generated expression is sent to the device and suggested to the user. The user reviews the proposed expression, modifies it if necessary to "This product greatly disappointed my expectations," and finally posts it on social media.

[1889] Examples of prompt statements

[1890] "When a user enters a comment on a social networking site, the system must receive the comment, analyze the content using natural language processing technology, and suggest expressions that are not defamatory."

[1891] The system will help users use constructive and considerate language on social media and in UGC.

[1892] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1893] Program processing flow

[1894] Step 1:

[1895] The user enters a comment.

[1896] A user inputs a comment into the comment input field of a SNS application. For example, the user inputs "This product is the worst." The input comment is saved on the device.

[1897] Step 2:

[1898] The terminal transmits the comment to the server.

[1899] The terminal uses an HTTP POST request to send the entered comment data to the server. The sent data includes the comment text and the user ID.

[1900] Step 3:

[1901] The server receives the comment.

[1902] The server receives the HTTP request and stores the comment data in a database for further analysis.

[1903] Step 4:

[1904] The server analyzes the comments using natural language processing.

[1905] The server uses natural language processing libraries (e.g., NLTK or spaCy) to analyze the grammatical structure and semantics of the comments. Specifically, it performs tokenization, part-of-speech tagging, dependency analysis, etc. The input is the comment text, and the output is the analysis result (e.g., "This / DET product / NOUN is / AUX the worst / ADJ / AUX").

[1906] Step 5:

[1907] The server performs sentiment analysis.

[1908] The server uses a sentiment analysis engine (e.g., IBM Watson Natural Language Understanding) to analyze the sentiment contained in the comments. The input is the parsed comment text, and the output is the sentiment data (e.g., "angry" or "unhappy").

[1909] Step 6:

[1910] The server generates non-abusive language.

[1911] The server uses a generative AI model (e.g., GPT-3) to generate non-abusive expressions, taking into account the user's values ​​and preferences learned from the user's past comments and behavior. The input is emotion data and the user's past comment data, and the output is the generated expression (e.g., "This product did not meet my expectations").

[1912] Step 7:

[1913] The server proposes the generated representation to the user.

[1914] The server sends the generated representation to the terminal and suggests it to the user. The input is the generated representation and the output is the suggested text that is displayed on the user's terminal.

[1915] Step 8:

[1916] The user reviews the proposal and corrects it if necessary.

[1917] The user checks the suggested expression and modifies it if necessary. For example, the user modifies it to "This product greatly exceeded my expectations." Finally, the modified comment is posted on the SNS. The input is the proposed text, and the output is the comment text that is finally posted.

[1918] In this way, users can be more constructive and considerate in their use of language on social media and in UGC.

[1919] (Application example 1)

[1920] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1921] In modern social networking sites and online forums, users posting abusive or offensive comments is a problem. This can make other users feel uncomfortable and can worsen the atmosphere of the entire community. Users themselves can also unintentionally use offensive language, which can result in trouble. To solve these problems, a system is needed that suggests expressions to avoid abusive comments in real time as the user types them.

[1922] 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.

[1923] In this invention, the server includes a means for receiving comments input by a user, a means for analyzing the content of the comment and recognizing the content the user wants to convey, a means for proposing expressions that are not slanderous, a means for learning the user's values ​​and hobbies and preferences, a means for the AI ​​to lear...

Claims

1. A means for analyzing the content of a comment input by a user and recognizing the content that the user wishes to convey; A means for generating text by combining expressions related to the content that the user wants to communicate, which are proposed by a generation AI that has learned the user's values ​​and hobbies and preferences, based on the user's selection; means for outputting the generated text; A system including:

2. The generating AI suggests expressions that avoid expressions that may be slanderous. The system of claim 1 .

3. means for analyzing emotions contained in the comments using a sentiment analysis engine; The generating means generates the text using the generation AI based on a prompt sentence corresponding to the analyzed emotion. The system of claim 1 .

Images