System

Abstract

A system is provided.SOLUTION: A system comprising: means for automatically detecting an anti-comment; means for converting the detected anti-comment into a constructive advice form; means for replacing the original comment in a database with the converted comment; and means for displaying the replaced comment.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In recent years, aggressive anti-comments in online comment sections have increasingly caused significant psychological stress to users, and in some cases, serious psychological harm. If this problem is left unaddressed, it could lead to a serious situation in which the integrity of Internet communication is undermined and a safe and comfortable environment for users is lost. Conventional methods require users to manually hide or delete anti-comments, which does not provide a fundamental solution. Therefore, the objective of this invention is to provide an effective means for maintaining the integrity of comment sections while reducing users' psychological stress. [Means for solving the problem]

[0005] To solve the above problems, the present invention provides the following means. First, a means for automatically detecting anti-comments is provided. This means includes an AI model that determines whether a comment is an anti-comment. Next, a means for converting the detected anti-comments into constructive advice is provided. This conversion means is intended to improve offensive language and provide positive feedback to the user. Furthermore, a means for replacing the converted comment with the original comment in a database is provided, allowing the user to see the constructive comment. In addition, by providing a means for converting anti-comments into advice only if their confidence level exceeds a certain threshold and a means for flagging anti-comments that do not exceed a certain threshold for review, it is possible to achieve both detection accuracy and effective feedback. This realizes a system that improves the communication environment on the Internet and reduces the psychological burden on users.

[0006] "Anti-comments" are comments intended to hurt or attack others.

[0007] "Constructive advice" is a politely worded comment that is positive and includes specific suggestions for improvement.

[0008] An "AI model" is a collection of algorithms trained to perform a specific task using artificial intelligence techniques.

[0009] "Confidence" is the probability value that an AI model outputs as an analysis result for a certain input, and is a number that indicates the certainty of the result.

[0010] A "threshold" is a boundary value for determining a specific condition, and is a reference value at which processing branches depending on whether or not it is exceeded.

[0011] A "database" is a system for storing information in an organized manner and for efficiently searching and updating it.

[0012] A "flag" is a mark or symbol that is given to data that meets a specific condition and is used in subsequent processing.

[0013] A "flag for review" is a mark given to a comment that requires review, and is a sign that allows the comment to be managed separately from other comments.

[0014] "Loop processing" is a program control structure for repeatedly performing a specific process.

[0015] "UI" stands for user interface, and refers to the screen and input means that users use to interact with a system. [Brief explanation of the drawings]

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

[0017] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

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

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

[0020] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0021] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0022] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0023] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0024] [First embodiment]

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

[0026] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0027] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0028] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0029] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0030] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0031] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

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

[0033] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0034] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0035] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

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

[0037] System Overview

[0038] This invention is a system that automatically detects anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0039] Server processing

[0040] 1. Loading the AI ​​model

[0041] When the server initializes the system, it loads an AI model for detecting anti-comments, which is loaded from a specific location within the network.

[0042] 2. Comment Analysis

[0043] The server retrieves all comments from the database and inputs each comment into the AI ​​model for analysis. The analysis results determine whether the comment is an anti-comment. This determination result includes a probability value (confidence) that the comment is an anti-comment.

[0044] 3. Comment conversion

[0045] The server converts anti-comments into constructive advice if their confidence exceeds a certain threshold by replacing offensive language with polite language.

[0046] 4. Updating the database

[0047] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0048] Processing by the terminal

[0049] 1. Fetching comments

[0050] The device retrieves the new converted comments from the server and uses them to display in the user's browser or app UI.

[0051] 2. Viewing comments

[0052] The device UI will display the converted comments to the user, allowing them to receive more constructive feedback.

[0053] User Action

[0054] 1. Enter a comment

[0055] The user enters a comment, which is then saved in the database.

[0056] 2. Receiving Comments

[0057] Users can view comments from other users via their device, and the converted constructive comments are displayed, allowing users to receive more positive feedback.

[0058] Specific examples

[0059] Original comment:

[0060] "This video of yours is awful. It's uninspiring and not worth watching."

[0061] Post-processing comments:

[0062] "Your video could use some improvement. For example, if you improve the sound quality or picture quality, it might be easier for viewers to watch."

[0063] In this way, the present invention is a system that converts aggressive anti-comments into constructive advice, thereby reducing psychological stress for users and promoting healthy communication.

[0064] The processing flow will be explained below.

[0065] Step 1:

[0066] The server loads the AI ​​model

[0067] Operation: When the system starts up, the server loads the AI ​​model from the specified path and initializes the model, which prepares it for anti-comment detection.

[0068] Step 2:

[0069] The device connects to the comment database and fetches all comments.

[0070] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[0071] Step 3:

[0072] The server parses each comment

[0073] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model determines whether the comment is an anti-comment or not and returns a result including a confidence score (a value between 0 and 1).

[0074] Step 4:

[0075] The server checks the credibility of anti-comments

[0076] Operation: Determine if the confidence level of the analysis result is 75% or higher. If it is 75% or higher, proceed to the next step. If it is less than 75%, flag it for review.

[0077] Step 5:

[0078] The server converts hate comments into constructive advice

[0079] How it works: For anti-comments with a confidence level of 75% or higher, the server uses an AI model to convert the content of the comment into constructive language, specifically removing offensive language and including positive, specific suggestions for improvement.

[0080] Step 6:

[0081] The server updates the database with the converted comments

[0082] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[0083] Step 7:

[0084] The terminal fetches the converted comment

[0085] What it does: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[0086] Step 8:

[0087] The device will display the updated comment to the user.

[0088] What it does: Updates the device UI to show the user the converted constructive comment, ensuring they receive constructive feedback instead of abusive comments.

[0089] Example 1

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

[0091] In traditional online communication, offensive and anti-comments place a psychological burden on users. This often leads to a negative atmosphere in the community and hinders healthy communication. Furthermore, manually managing and censoring offensive comments requires a great deal of effort, making efficient management a key priority.

[0092] 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.

[0093] In this invention, the server includes means for loading an artificial intelligence model to automatically detect anti-comments, means for retrieving comments from a database and analyzing these comments, means for determining whether the analyzed comments are anti-comments, means for converting the anti-comments into constructive advice format based on the determination result, means for replacing the converted comments with the original comments in the database, and means for displaying the replaced comments on a user interface, thereby enabling the offensive comments to be automatically converted into constructive feedback.

[0094] "Anti-comments" are comments that contain negative, offensive, or insulting content and cause psychological stress to the target.

[0095] An "artificial intelligence model" is a set of algorithms and their parameters that can automatically process specific tasks using machine learning techniques.

[0096] A "database" is a system designed to store data in an organized manner and make it easy to access and manage.

[0097] "Analysis" is the process of examining the data obtained in detail and extracting useful information from it.

[0098] "Determining" refers to the act of determining whether or not a particular condition is met based on the analysis results.

[0099] "Constructive advice" is the conversion of negative feedback into positive suggestions for improvement that the recipient can take in a positive way.

[0100] "User interface" is a general term for the screens and operating methods that users use to interact with a system.

[0101] "Reliability" is a numerical index that indicates the accuracy and certainty of the judgment result.

[0102] A "threshold" is a value that sets a certain reference value and serves as a boundary for executing a specific action or process.

[0103] A "flag" is a mark or indicator that is set for identification or processing based on a specific condition.

[0104] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0105] Server processing

[0106] The server loads an artificial intelligence model for detecting anti-comments during system initialization. This model is loaded from a specific location within the network using a machine learning framework such as TensorFlow or PyTorch. The loaded artificial intelligence model is used to analyze and judge anti-comments.

[0107] The server then retrieves all comments from a database (e.g., MySQL or PostgreSQL) and analyzes this data. Each comment is fed into an artificial intelligence model, which determines whether it is an anti-comment or not. The result of this determination includes a probability value (confidence) that the comment is an anti-comment.

[0108] Based on the results, the server converts the anti-comment into a constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT) that replaces offensive language with polite language. This conversion process is performed using prompts such as:

[0109] Please convert the offensive comments below into constructive feedback:

[0110] "This video of yours is awful. It's uninspiring and not worth watching."

[0111] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is automatically flagged for review.

[0112] Processing by the terminal

[0113] The device fetches the new converted comments from the server and displays them in the user's browser or application's user interface, using front-end frameworks such as React or Vue.js to present the converted comments to the user in a visually understandable way.

[0114] User Action

[0115] Users can enter comments through a form on a website or application. These comments are immediately stored in a database and analyzed and converted by the server. After processing, users can receive comments from other users through their own devices and receive constructive feedback. This helps users reduce psychological stress and promote healthy communication.

[0116] The above is an embodiment of the present invention, and this system can automatically convert anti-comments into constructive feedback, thereby realizing positive communication between users.

[0117] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0118] Step 1: Loading the AI ​​model

[0119] When the system is initialized, the server loads an artificial intelligence model for detecting anti-comments. The input is the file path of the AI ​​model stored in a specific location within the network. Specifically, the model file is read into memory using TensorFlow or PyTorch. The output of this process is an AI model capable of detecting anti-comments loaded into memory.

[0120] Step 2: Obtaining and parsing comments

[0121] The server retrieves all comments from a database. The database uses MySQL or PostgreSQL, and executes SQL queries to extract comment data. The input is the access information to the database where the comments are stored. The retrieved comments are then input into an AI model to determine whether they are anti-comments. Specifically, each comment is input into the AI ​​model, and a confidence score is obtained as the output. The higher the confidence score, the more likely the comment is to be an anti-comment.

[0122] Step 3: Convert comments

[0123] The server converts anti-comments into constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT). The anti-comments and their confidence levels are used as input. Specifically, the server inputs a prompt sentence into the generative AI model to obtain constructive feedback. The following sentence is used as an example of this prompt sentence:

[0124] Please convert the offensive comments below into constructive feedback:

[0125] "This video of yours is awful. It's uninspiring and not worth watching."

[0126] The output is that offensive comments are transformed into constructive feedback.

[0127] Step 4: Update the database

[0128] The server replaces the original comment with the converted comment and stores it in the database. The inputs are the IDs of the converted comment and its original comment. Specifically, it executes an SQL query using an UPDATE statement to replace the original comment with the new comment. If the confidence does not exceed a threshold, the comment is flagged for review.

[0129] Step 5: Fetch comments

[0130] The device fetches the latest comments from the server. The server's API endpoint URL is used as input. Specifically, it sends an HTTP request to the server and receives data in JSON format from the server. The output of this process is the latest comment data retrieved by the device.

[0131] Step 6: View comments

[0132] The device's user interface displays the fetched comments. It uses the retrieved comment data as input, renders a component that visualizes the comment data using a front-end framework such as React or Vue.js, and displays the latest constructive feedback to the user as output.

[0133] Step 7: Enter a comment

[0134] Users enter comments through a form on a website or application. The actual comment text entered by the user is used as input. Specifically, the form data is sent to the server's API, which receives it and stores it in a database. The output of this process is that the new comment data is added to the database.

[0135] Step 8: Receiving comments

[0136] The user views the comments retrieved from the server via the device. The latest comment data retrieved by the device is used as input. The specific operation is to display the comments using a front-end framework. As output, the user can view constructive feedback.

[0137] (Application example 1)

[0138] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0139] On current content distribution platforms, user comments often contain offensive content, causing psychological stress for content providers and other users. Ignoring such anti-comments can hinder healthy communication and potentially lower the quality of the platform as a whole. Therefore, the present invention aims to provide a system that automatically detects offensive anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users and promoting healthy communication.

[0140] 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.

[0141] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice, means for determining whether a comment entered by a user is offensive, means for converting the offensive comment into constructive advice using a generative AI model, means for replacing the original comment in the database with the converted comment, and means for displaying the replaced comment. This allows negative comments from users to be instantly converted into positive feedback, enabling healthy communication on the content distribution platform.

[0142] "Anti-comments" are comments made by users that contain offensive, negative or harsh content about specific content or other users.

[0143] "Constructive advice" is a form of positive feedback that is beneficial to the recipient and includes specific improvements or suggestions.

[0144] A "generative AI model" is an artificial intelligence model that is trained on a large dataset and has the ability to analyze and convert input text data.

[0145] A "database" is an electronic data storage system that has a structure and allows efficient management, storage, searching, and retrieval of data.

[0146] An "offensive word list" is a collection of words that includes specific terms and phrases that, when used, may create a negative or offensive impression on others.

[0147] "Confidence" is a probability value that indicates whether a particular output (for example, a judgment of anti-comments) is correct based on the data input to the AI ​​model.

[0148] A "prompt" is the initial input data given to an AI model to perform a specific task.

[0149] A "content distribution service" is a service that provides digital content such as video, audio, and text to users via the Internet.

[0150] A "flag" is an identifier that is assigned to a data item or a processing step based on a specific condition.

[0151] The system for implementing the present invention is composed of a server, a terminal, and a user. The specific configuration and operation method of the system will be described below.

[0152] Server processing

[0153] Loading an AI model

[0154] When the system is initialized, the server loads a generative AI model for detecting anti-comments, which is loaded from a specific repository on the Internet.

[0155] Comment parsing and transformation

[0156] The server analyzes all comments retrieved from the database, checking them against a list of offensive words to detect anti-comments, which are then fed into a generative AI model and transformed into constructive advice using prompts.

[0157] Examples of prompts:

[0158] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[0159] Update the database with the conversion results

[0160] The server replaces the original comment with the converted comment and stores only those whose confidence exceeds a certain threshold in the database, or flags the comment for review if its confidence does not exceed the threshold.

[0161] Processing by the terminal

[0162] Fetching and displaying comments

[0163] The device retrieves the new converted comments from the server and displays them in the user's browser or app UI, allowing users to receive more constructive feedback.

[0164] User Action

[0165] Entering and receiving comments

[0166] Users can use the content distribution service's application to input and submit comments. The input comments are stored in a database, and users can also view constructive comments converted into feedback from other users.

[0167] Hardware and software used

[0168] Hardware: Smartphone (iOS, Android), Head-Mounted Display (HMD)

[0169] software:

[0170] Server-side AI model: OpenAI's generative AI model

[0171] Database: A relational database such as MySQL or PostgreSQL

[0172] Comment analysis and transformation: Scripting with Python

[0173] Explanation of specific steps in the process

[0174] Users can enter comments on videos.

[0175] The device will send the entered comment to the server.

[0176] The server receives the comments and checks them against a list of offensive words to detect them.

[0177] Using AI models, offensive comments are transformed using generative AI models.

[0178] The converted comment replaces the original comment in the database and is flagged if it has low confidence.

[0179] The terminal displays the converted comment to the user.

[0180] This invention instantly converts negative comments from users into positive feedback, promoting healthy communication on content distribution platforms.

[0181] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0182] Step 1:

[0183] The server loads the generative AI model during system initialization. Specifically, it downloads the model file from a specific repository on the internet and extracts it into memory. At this point, the generative AI model is ready for comment analysis and conversion.

[0184] Step 2:

[0185] A user inputs a comment for a video on a device. The input comment is temporarily stored in the device's memory and then sent to the server. The device then sends the comment data to the server as an HTTP request.

[0186] Step 3:

[0187] The server receives comments from clients and checks them against a list of offensive words to detect anti-comments. In this process, each word in the comment is checked sequentially to see if it is included in the offensive words list. If it is, the comment is flagged as an anti-comment.

[0188] Step 4:

[0189] The server inputs the detected anti-comments into the generative AI model, providing the model with a prompt like this:

[0190] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[0191] The generative AI model uses this prompt to convert offensive comments into constructive advice.

[0192] Step 5:

[0193] The server receives the converted comment and checks whether its confidence exceeds a certain threshold. If it does, it replaces the original comment with the converted comment and stores it in the database. If it does not, it flags the original comment for review. This process involves assigning a confidence score to the output of the generative AI model and determining whether it exceeds the threshold.

[0194] Step 6:

[0195] The device retrieves the new converted comments from the server. Specifically, the client application periodically sends a request to the server to retrieve the updated comment data. The retrieved data is stored in the device's memory.

[0196] Step 7:

[0197] The terminal displays the retrieved and converted comments to the user. The user interface reflects new comments in real time and allows the user to view them. During this process, the retrieved comment data is rendered into the UI component.

[0198] These steps will instantly transform negative comments from users into positive feedback, promoting healthy communication on the content distribution platform.

[0199] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0200] System Overview

[0201] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes user emotions. This system is mainly composed of a server, a terminal, and a user, each of which plays a specific role.

[0202] Server processing

[0203] 1. Loading the AI ​​model and emotion engine

[0204] When the server initializes the system, it loads the AI ​​model for detecting anti-comments and the emotion engine for recognizing user emotions, which prepares the system for anti-comment detection and emotion recognition.

[0205] 2. Comment Analysis

[0206] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[0207] 3. Emotional Recognition

[0208] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, which may include, for example, anger, sadness, and joy.

[0209] 4. Comment Conversion

[0210] If the reliability of the anti-comment exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted to encourage calmness.

[0211] 5. Updating the database

[0212] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0213] Processing by the terminal

[0214] 1. Getting comments

[0215] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[0216] 2. Viewing comments

[0217] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[0218] User Action

[0219] 1. Enter a comment

[0220] The user enters a comment, which is then saved in the database.

[0221] 2. Receiving Comments

[0222] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[0223] Specific examples

[0224] Original comment:

[0225] "This video of yours is awful. It's uninspiring and not worth watching."

[0226] User sentiment:

[0227] The AI ​​model detects this comment and recognizes that the user's emotion is "anger."

[0228] Post-processing comments:

[0229] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[0230] In this way, the present invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[0231] The processing flow will be explained below.

[0232] Step 1:

[0233] The server loads the AI ​​model and emotion engine.

[0234] Operation: When the server initializes the system, it loads the AI ​​model and emotion engine from the specified path and initializes each model. This prepares the system for anti-comment detection and emotion recognition.

[0235] Step 2:

[0236] The device connects to the comment database and fetches all comments.

[0237] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[0238] Step 3:

[0239] The server parses each comment

[0240] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model analyzes the comment, determines whether it is an anti-comment, and returns a result including a confidence score.

[0241] Step 4:

[0242] The server recognizes the user's emotions

[0243] How it works: The server uses the emotion engine to analyze the emotional state of the user who posted the comment, and obtains emotional data such as anger, sadness, and joy as the analysis result.

[0244] Step 5:

[0245] The server checks the credibility of anti-comments

[0246] How it works: The server uses the AI ​​model's analysis results to determine whether the confidence level exceeds a certain threshold (e.g., 75%). If it does, it proceeds to the next step; if it does not, it flags it for review.

[0247] Step 6:

[0248] The server converts hate comments into constructive advice

[0249] How it works: The server converts anti-comments with a confidence level above a threshold into constructive advice based on the analysis results of the emotion engine. For example, if anger is detected, the advice includes encouraging people to stay calm.

[0250] Step 7:

[0251] The server updates the database with the converted comments

[0252] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[0253] Step 8:

[0254] The terminal fetches the converted comment from the database.

[0255] Operation: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[0256] Step 9:

[0257] The device displays the updated comment to the user.

[0258] What it does: Updates the device UI to show the converted constructive comment to the user, ensuring they receive constructive and sensitive feedback instead of offensive comments.

[0259] Example 2

[0260] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0261] In today's Internet, users can freely post comments to each other, leading to the proliferation of anti-comments and making healthy communication difficult. Anti-comments, in particular, can increase the psychological burden on the recipient and potentially worsen the overall communication environment. While detecting anti-comments and converting them into constructive advice is important in this situation, there is also a need for more effective communication support that takes into account the user's emotions.

[0262] 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.

[0263] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice format, and means for taking user emotion data into consideration when converting. This makes it possible to convert anti-comments into constructive advice format while taking user emotion into consideration, thereby promoting a healthy communication environment.

[0264] "Anti-comments" are comments that contain negative content such as attacks, insults, or criticism of others.

[0265] "Constructive advice format" refers to comments that point out problems and include specific suggestions and advice for improving them.

[0266] "User emotion data" is data that indicates the user's emotional state at the time of posting a comment, and typically includes emotions such as anger, sadness, and joy.

[0267] "Confidence" is a score that indicates the degree of certainty with which the AI ​​model detects a comment as an anti-comment.

[0268] A "threshold" is a boundary value of the reliability score for performing a specific process or judgment.

[0269] A "database" is a system for systematically storing and managing information such as comments, their conversion results, and user emotional data.

[0270] A "flag" is a mark or indication that is set for a particular condition, in this case indicating that a review is required.

[0271] The "server" is a central computer system that performs comment analysis, emotion recognition, conversion processing, etc.

[0272] A "terminal" is a device that allows a user to input comments and view the displayed comments.

[0273] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0274] Server processing

[0275] The server is the center of the system and performs processing in the following procedure.

[0276] 1. Loading the AI ​​model and emotion engine

[0277] The server loads an AI model (e.g., generative AI model) for detecting anti-comments and an emotion engine (e.g., emotion analysis software) for recognizing user emotions during system initialization, which prepares the system for anti-comment detection and emotion recognition.

[0278] 2. Comment Analysis

[0279] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[0280] 3. Emotional Recognition

[0281] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, such as anger, sadness, and joy.

[0282] 4. Comment Conversion

[0283] If the confidence level exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine.

[0284] 5. Updating the database

[0285] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0286] Processing by the terminal

[0287] The terminal functions as an interface with the user and performs processing in the following procedure.

[0288] 1. Getting comments

[0289] The device retrieves the updated comments from the server. The latest comment information is retrieved from the server via an HTTP request.

[0290] 2. Viewing comments

[0291] The user interface of the terminal displays the updated comments to the user in a visually easy to understand format.

[0292] User Action

[0293] The user is the entity that directly interacts with the system and performs the following process.

[0294] 1. Enter a comment

[0295] Users enter and post comments. The posted comments are saved in a database and sent to the server.

[0296] 2. Receiving Comments

[0297] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[0298] Specific examples

[0299] For example, if the original comment was, "This video of yours is really terrible. It's unmotivating and not worth watching," the AI ​​model and emotion engine would analyze this comment and recognize that the user's emotion was "anger."

[0300] An example prompt that takes this into account is:

[0301] "Your video is terrible. It's uninspiring and not worth watching." The emotion in this comment is anger. Convert this comment into constructive advice that encourages calm.

[0302] The resulting comment after processing looks like this:

[0303] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[0304] In this way, the present invention not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[0305] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0306] Program processing flow

[0307] Step 1:

[0308] The server loads the AI ​​model and emotion engine during system initialization.

[0309] Specific behavior:

[0310] The server uses the APIs and authentication information of the generative AI model (e.g., GPT-4) and sentiment analysis engine (e.g., general sentiment analysis software) to launch these models. As input, it obtains the necessary information from the configuration file and environment variables set when the server is started, and as output, it obtains the state when the model has been loaded.

[0311] Step 2:

[0312] The server retrieves all comments from the database.

[0313] Specific behavior:

[0314] The server retrieves all comments from the comment database using an SQL query, etc. Database connection information and a query are required as input, and a text data list of comments is obtained as output.

[0315] Step 3:

[0316] Each comment received by the server is input into an AI model to determine whether it is an anti-comment.

[0317] Specific behavior:

[0318] The server inputs each comment as a prompt into the generative AI model and parses the model's response. The input is the comment text, and the output is whether the comment is an anti-comment or not, along with its confidence score.

[0319] Step 4:

[0320] The server uses an emotion engine to recognize the user's emotions when posting a comment.

[0321] Specific behavior:

[0322] The server sends each comment along with the user's emotional data to the emotion engine and receives the analysis results from the emotion engine. The input is the comment text and the emotion analysis request, and the output is the emotional data (e.g., anger, sadness, joy, etc.).

[0323] Step 5:

[0324] The server converts the anti-comments into constructive advice.

[0325] Specific behavior:

[0326] The server generates a prompt sentence for the AI ​​model and converts the anti-comment. At this time, the prompt also includes the user's emotional data. The input is the anti-comment and the prompt sentence including the emotional data, and the output is the comment converted into a constructive advice format.

[0327] Step 6:

[0328] The server replaces the converted comment with the original comment in the database.

[0329] Specific behavior:

[0330] The server replaces the original comment with the transformed comment using an SQL update query, taking as input the comment ID and the transformed comment text, and obtaining the updated database result as output.

[0331] Step 7:

[0332] Comments that do not exceed a confidence threshold are flagged for review.

[0333] Specific behavior:

[0334] If the reliability does not exceed the threshold, the server sets a review-needed flag for the comment in the database. The input is the comment ID and the review-needed flag, and the output is the updated database result.

[0335] Step 8:

[0336] The terminal obtains the updated comments from the server and reflects them in the user interface.

[0337] Specific behavior:

[0338] The terminal obtains the updated comments from the server via an HTTP request and displays them on the user interface. The input is the HTTP request and the comment data as a response, and the output is the comment display on the user interface.

[0339] Step 9:

[0340] The user enters a comment, which is then saved in the database.

[0341] Specific behavior:

[0342] The user enters a comment in the comment input form and submits it, and the server saves it in the database. The input is the comment text entered by the user and a submission request, and the output is the result of saving the new comment in the database.

[0343] Step 10:

[0344] A user views comments from other users via a terminal.

[0345] Specific behavior:

[0346] The user views the latest comment list displayed on the terminal screen and receives the converted constructive comments. The input is the comment data obtained from the server, and the output is the screen display that the user views.

[0347] (Application example 2)

[0348] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0349] In conventional systems, it is difficult for online content distribution services to respond appropriately to anti-comments among user comments, and the continued negative feedback can cause psychological stress. Furthermore, a simple negative comment detection system does not provide constructive feedback that takes into account the user's feelings, which means that healthy communication cannot be achieved.

[0350] 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.

[0351] In this invention, the server includes a means for automatically detecting anti-comments, a means for converting anti-comments into constructive advice using an AI model, and a means for recognizing the user's emotions and reflecting them in the converted comments. This allows the anti-comments to be converted into constructive feedback while taking the user's emotions into consideration, thereby promoting healthy communication while reducing psychological burden.

[0352] "Anti-comments" are comments that contain negative or offensive content about online content or individuals.

[0353] An "AI model" is a model for performing data analysis or predictive batch processing using artificial intelligence.

[0354] "Constructive advice" is feedback that contains negative content but is reframed as a positive suggestion for improvement.

[0355] "User emotions" is data obtained by analyzing the user's psychological state at the time of posting a comment using an emotion engine.

[0356] An "emotion engine" is an algorithm or software for identifying emotions from a user's text data.

[0357] "Confidence" is a score that indicates the degree of certainty of the analysis results that the AI ​​model assigns when detecting anti-comments.

[0358] The "specific threshold" is a reference value for determining whether or not a comment judgment is converted into constructive advice.

[0359] "Database" means an information management system for storing and managing comments and other data.

[0360] "Flag for review" is a marking of anti-comments that do not exceed a certain threshold, indicating that further review or human verification is required.

[0361] System Overview

[0362] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes users' emotions. The system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0363] Server processing

[0364] The server is the central element of the system and performs the following functions:

[0365] 1. Loading the AI ​​model and emotion engine:

[0366] When the server initializes the system, it loads an AI model for detecting anti-comments and an emotion engine for recognizing user emotions. This prepares the system for anti-comment detection and emotion recognition. The Python transformers library can be used as the AI ​​model.

[0367] 2. Comment Analysis:

[0368] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and the AI ​​model determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[0369] 3. Emotion Recognition:

[0370] The emotion engine recognizes the user's emotion when posting a comment and retrieves the emotion data. Emotion data can include, for example, anger, sadness, and joy. The vaderSentiment library can be used for this.

[0371] 4. Comment conversion:

[0372] If the confidence level exceeds a certain threshold, the anti-comment is converted into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted into a message encouraging them to stay calm.

[0373] 5. Database Update:

[0374] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0375] Processing by the terminal

[0376] The device acts as a link between the server and the user and performs the following tasks:

[0377] 1. Getting comments:

[0378] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[0379] 2. View comments:

[0380] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[0381] User Action

[0382] Users interact with the system in the following ways:

[0383] 1. Enter your comment:

[0384] The user enters a comment, which is then saved in the database.

[0385] 2. Receiving Comments:

[0386] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[0387] Specific examples

[0388] Original comment: "This video is so boring and a waste of time."

[0389] Recognized emotion: "Anger"

[0390] Converted comment: "There may be some room for improvement in this video. For example, if you put more thought into the content, it might be more interesting for viewers. I recommend you think about it calmly."

[0391] Prompt Sentence Examples

[0392] Type: "The sound in this video is so bad I can't hear you."

[0393] Prompt: "This comment is negative feedback. Please convert it into constructive feedback with suggestions for improvement."

[0394] In this way, this invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[0395] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0396] Step 1:

[0397] The server loads the AI ​​model and emotion engine when the system is initialized. Specifically, it loads the AI ​​model for detecting anti-comments using the Python transformers library and initializes the vaderSentiment library for emotion analysis. This enables the system to detect anti-comments and recognize user emotions.

[0398] Input: None

[0399] Data processing / calculation: Initialization of AI model and emotion engine

[0400] Output: Initialized AI model and emotion engine

[0401] Step 2:

[0402] The server retrieves all comments from the database. Each comment is then input into an AI model. The AI ​​model determines whether the comment is an anti-comment and returns a result including its confidence level. Specifically, it uses a transformers model to classify the text.

[0403] Input: Database comment

[0404] Data processing / calculation: Comment analysis using AI models

[0405] Output: Anti-comment judgment results and confidence score

[0406] Step 3:

[0407] The server inputs the comment text into the emotion engine to recognize the user's emotion at the same time as the analyzed comment. The emotion engine analyzes the comment's emotional state (anger, joy, sadness, etc.) and returns the results. Specifically, emotion analysis is performed using the vaderSentiment library.

[0408] Input: Comment text

[0409] Data processing / calculation: Emotion analysis using emotion engine

[0410] Output: User emotion data

[0411] Step 4:

[0412] Only if the reliability exceeds a certain threshold will the server convert the anti-comment into constructive advice. The conversion process takes into account the user's emotions as recognized by the emotion engine. Specifically, if the user's emotion is "anger," the server converts the message into a message encouraging them to stay calm, and if the user's emotion is "sadness," the server converts the message into a message containing encouragement.

[0413] Input: Anti-comment judgment results, user emotion data

[0414] Data processing / calculation: Comment conversion

[0415] Output: Comments converted into constructive advice format

[0416] Step 5:

[0417] The server replaces the original comment with the converted comment and stores it in the database. The server then deletes the pre-conversion comment data and inserts the new comment data. This updates the comment information in the database.

[0418] Input: Translated comment

[0419] Data processing / calculation: Database update

[0420] Output: Updated comment data

[0421] Step 6:

[0422] The device retrieves the latest comment data from the server and reflects it on the user's UI. The device's UI then displays the converted comments to the user, allowing the user to receive constructive and emotionally sensitive feedback.

[0423] Input: Updated comment data

[0424] Data processing / calculation: Acquisition and display of comment data

[0425] Output: Constructive comments displayed in the user UI

[0426] Step 7:

[0427] Users can view comments from other users via their devices, and can also enter their own comments, which are then saved in the database, after which the system returns to the loop of analyzing and converting new comments.

[0428] Input: User comment

[0429] Data processing / calculation: Entering and saving comments

[0430] Output: New comment saved in the database

[0431] In this way, the system converts anti-comments into constructive feedback, promoting healthy communication while reducing the psychological burden on users.

[0432] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0433] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0434] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0435] [Second embodiment]

[0436] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0437] 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.

[0438] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0439] 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.

[0440] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0441] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0442] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0443] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0444] 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.

[0445] 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.

[0446] 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.

[0447] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0448] System Overview

[0449] This invention is a system that automatically detects anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0450] Server processing

[0451] 1. Loading the AI ​​model

[0452] When the server initializes the system, it loads an AI model for detecting anti-comments, which is loaded from a specific location within the network.

[0453] 2. Comment Analysis

[0454] The server retrieves all comments from the database and inputs each comment into the AI ​​model for analysis. The analysis results determine whether the comment is an anti-comment. This determination result includes a probability value (confidence) that the comment is an anti-comment.

[0455] 3. Comment conversion

[0456] The server converts anti-comments into constructive advice if their confidence exceeds a certain threshold by replacing offensive language with polite language.

[0457] 4. Updating the database

[0458] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0459] Processing by the terminal

[0460] 1. Fetching comments

[0461] The device retrieves the new converted comments from the server and uses them to display in the user's browser or app UI.

[0462] 2. Viewing comments

[0463] The device UI will display the converted comments to the user, allowing them to receive more constructive feedback.

[0464] User Action

[0465] 1. Enter a comment

[0466] The user enters a comment, which is then saved in the database.

[0467] 2. Receiving Comments

[0468] Users can view comments from other users via their device, and the converted constructive comments are displayed, allowing users to receive more positive feedback.

[0469] Specific examples

[0470] Original comment:

[0471] "This video of yours is awful. It's uninspiring and not worth watching."

[0472] Post-processing comments:

[0473] "Your video could use some improvement. For example, if you improve the sound quality or picture quality, it might be easier for viewers to watch."

[0474] In this way, the present invention is a system that converts aggressive anti-comments into constructive advice, thereby reducing psychological stress for users and promoting healthy communication.

[0475] The processing flow will be explained below.

[0476] Step 1:

[0477] The server loads the AI ​​model

[0478] Operation: When the system starts up, the server loads the AI ​​model from the specified path and initializes the model, which prepares it for anti-comment detection.

[0479] Step 2:

[0480] The device connects to the comment database and fetches all comments.

[0481] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[0482] Step 3:

[0483] The server parses each comment

[0484] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model determines whether the comment is an anti-comment or not and returns a result including a confidence score (a value between 0 and 1).

[0485] Step 4:

[0486] The server checks the credibility of anti-comments

[0487] Operation: Determine if the confidence level of the analysis result is 75% or higher. If it is 75% or higher, proceed to the next step. If it is less than 75%, flag it for review.

[0488] Step 5:

[0489] The server converts hate comments into constructive advice

[0490] How it works: For anti-comments with a confidence level of 75% or higher, the server uses an AI model to convert the content of the comment into constructive language, specifically removing offensive language and including positive, specific suggestions for improvement.

[0491] Step 6:

[0492] The server updates the database with the converted comments

[0493] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[0494] Step 7:

[0495] The terminal fetches the converted comment

[0496] What it does: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[0497] Step 8:

[0498] The device will display the updated comment to the user.

[0499] What it does: Updates the device UI to show the user the converted constructive comment, ensuring they receive constructive feedback instead of abusive comments.

[0500] Example 1

[0501] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0502] In traditional online communication, offensive and anti-comments place a psychological burden on users. This often leads to a negative atmosphere in the community and hinders healthy communication. Furthermore, manually managing and censoring offensive comments requires a great deal of effort, making efficient management a key priority.

[0503] 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.

[0504] In this invention, the server includes means for loading an artificial intelligence model to automatically detect anti-comments, means for retrieving comments from a database and analyzing these comments, means for determining whether the analyzed comments are anti-comments, means for converting the anti-comments into constructive advice format based on the determination result, means for replacing the converted comments with the original comments in the database, and means for displaying the replaced comments on a user interface, thereby enabling the offensive comments to be automatically converted into constructive feedback.

[0505] "Anti-comments" are comments that contain negative, offensive, or insulting content and cause psychological stress to the target.

[0506] An "artificial intelligence model" is a set of algorithms and their parameters that can automatically process specific tasks using machine learning techniques.

[0507] A "database" is a system designed to store data in an organized manner and make it easy to access and manage.

[0508] "Analysis" is the process of examining the data obtained in detail and extracting useful information from it.

[0509] "Determining" refers to the act of determining whether or not a particular condition is met based on the analysis results.

[0510] "Constructive advice" is the conversion of negative feedback into positive suggestions for improvement that the recipient can take in a positive way.

[0511] "User interface" is a general term for the screens and operating methods that users use to interact with a system.

[0512] "Reliability" is a numerical index that indicates the accuracy and certainty of the judgment result.

[0513] A "threshold" is a value that sets a certain reference value and serves as a boundary for executing a specific action or process.

[0514] A "flag" is a mark or indicator that is set for identification or processing based on a specific condition.

[0515] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0516] Server processing

[0517] The server loads an artificial intelligence model for detecting anti-comments during system initialization. This model is loaded from a specific location within the network using a machine learning framework such as TensorFlow or PyTorch. The loaded artificial intelligence model is used to analyze and judge anti-comments.

[0518] The server then retrieves all comments from a database (e.g., MySQL or PostgreSQL) and analyzes this data. Each comment is fed into an artificial intelligence model, which determines whether it is an anti-comment or not. The result of this determination includes a probability value (confidence) that the comment is an anti-comment.

[0519] Based on the results, the server converts the anti-comment into a constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT) that replaces offensive language with polite language. This conversion process is performed using prompts such as:

[0520] Please convert the offensive comments below into constructive feedback:

[0521] "This video of yours is awful. It's uninspiring and not worth watching."

[0522] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is automatically flagged for review.

[0523] Processing by the terminal

[0524] The device fetches the new converted comments from the server and displays them in the user's browser or application's user interface, using front-end frameworks such as React or Vue.js to present the converted comments to the user in a visually understandable way.

[0525] User Action

[0526] Users can enter comments through a form on a website or application. These comments are immediately stored in a database and analyzed and converted by the server. After processing, users can receive comments from other users through their own devices and receive constructive feedback. This helps users reduce psychological stress and promote healthy communication.

[0527] The above is an embodiment of the present invention, and this system can automatically convert anti-comments into constructive feedback, thereby realizing positive communication between users.

[0528] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0529] Step 1: Loading the AI ​​model

[0530] When the system is initialized, the server loads an artificial intelligence model for detecting anti-comments. The input is the file path of the AI ​​model stored in a specific location within the network. Specifically, the model file is read into memory using TensorFlow or PyTorch. The output of this process is an AI model capable of detecting anti-comments loaded into memory.

[0531] Step 2: Obtaining and parsing comments

[0532] The server retrieves all comments from a database. The database uses MySQL or PostgreSQL, and executes SQL queries to extract comment data. The input is the access information to the database where the comments are stored. The retrieved comments are then input into an AI model to determine whether they are anti-comments. Specifically, each comment is input into the AI ​​model, and a confidence score is obtained as the output. The higher the confidence score, the more likely the comment is to be an anti-comment.

[0533] Step 3: Convert comments

[0534] The server converts anti-comments into constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT). The anti-comments and their confidence levels are used as input. Specifically, the server inputs a prompt sentence into the generative AI model to obtain constructive feedback. The following sentence is used as an example of this prompt sentence:

[0535] Please convert the offensive comments below into constructive feedback:

[0536] "This video of yours is awful. It's uninspiring and not worth watching."

[0537] The output is that offensive comments are transformed into constructive feedback.

[0538] Step 4: Update the database

[0539] The server replaces the original comment with the converted comment and stores it in the database. The inputs are the IDs of the converted comment and its original comment. Specifically, it executes an SQL query using an UPDATE statement to replace the original comment with the new comment. If the confidence does not exceed a threshold, the comment is flagged for review.

[0540] Step 5: Fetch comments

[0541] The device fetches the latest comments from the server. The server's API endpoint URL is used as input. Specifically, it sends an HTTP request to the server and receives data in JSON format from the server. The output of this process is the latest comment data retrieved by the device.

[0542] Step 6: View comments

[0543] The device's user interface displays the fetched comments. It uses the retrieved comment data as input, renders a component that visualizes the comment data using a front-end framework such as React or Vue.js, and displays the latest constructive feedback to the user as output.

[0544] Step 7: Enter a comment

[0545] Users enter comments through a form on a website or application. The actual comment text entered by the user is used as input. Specifically, the form data is sent to the server's API, which receives it and stores it in a database. The output of this process is that the new comment data is added to the database.

[0546] Step 8: Receiving comments

[0547] The user views the comments retrieved from the server via the device. The latest comment data retrieved by the device is used as input. The specific operation is to display the comments using a front-end framework. As output, the user can view constructive feedback.

[0548] (Application example 1)

[0549] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0550] On current content distribution platforms, user comments often contain offensive content, causing psychological stress for content providers and other users. Ignoring such anti-comments can hinder healthy communication and potentially lower the quality of the platform as a whole. Therefore, the present invention aims to provide a system that automatically detects offensive anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users and promoting healthy communication.

[0551] 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.

[0552] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice, means for determining whether a comment entered by a user is offensive, means for converting the offensive comment into constructive advice using a generative AI model, means for replacing the original comment in the database with the converted comment, and means for displaying the replaced comment. This allows negative comments from users to be instantly converted into positive feedback, enabling healthy communication on the content distribution platform.

[0553] "Anti-comments" are comments made by users that contain offensive, negative or harsh content about specific content or other users.

[0554] "Constructive advice" is a form of positive feedback that is beneficial to the recipient and includes specific improvements or suggestions.

[0555] A "generative AI model" is an artificial intelligence model that is trained on a large dataset and has the ability to analyze and convert input text data.

[0556] A "database" is an electronic data storage system that has a structure and allows efficient management, storage, searching, and retrieval of data.

[0557] An "offensive word list" is a collection of words that includes specific terms and phrases that, when used, may create a negative or offensive impression on others.

[0558] "Confidence" is a probability value that indicates whether a particular output (for example, a judgment of anti-comments) is correct based on the data input to the AI ​​model.

[0559] A "prompt" is the initial input data given to an AI model to perform a specific task.

[0560] A "content distribution service" is a service that provides digital content such as video, audio, and text to users via the Internet.

[0561] A "flag" is an identifier that is assigned to a data item or a processing step based on a specific condition.

[0562] The system for implementing the present invention is composed of a server, a terminal, and a user. The specific configuration and operation method of the system will be described below.

[0563] Server processing

[0564] Loading an AI model

[0565] When the system is initialized, the server loads a generative AI model for detecting anti-comments, which is loaded from a specific repository on the Internet.

[0566] Comment parsing and transformation

[0567] The server analyzes all comments retrieved from the database, checking them against a list of offensive words to detect anti-comments, which are then fed into a generative AI model and transformed into constructive advice using prompts.

[0568] Examples of prompts:

[0569] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[0570] Update the database with the conversion results

[0571] The server replaces the original comment with the converted comment and stores only those whose confidence exceeds a certain threshold in the database, or flags the comment for review if its confidence does not exceed the threshold.

[0572] Processing by the terminal

[0573] Fetching and displaying comments

[0574] The device retrieves the new converted comments from the server and displays them in the user's browser or app UI, allowing users to receive more constructive feedback.

[0575] User Action

[0576] Entering and receiving comments

[0577] Users can use the content distribution service's application to input and submit comments. The input comments are stored in a database, and users can also view constructive comments converted into feedback from other users.

[0578] Hardware and software used

[0579] Hardware: Smartphone (iOS, Android), Head-Mounted Display (HMD)

[0580] software:

[0581] Server-side AI model: OpenAI's generative AI model

[0582] Database: A relational database such as MySQL or PostgreSQL

[0583] Comment analysis and transformation: Scripting with Python

[0584] Explanation of specific steps in the process

[0585] Users can enter comments on videos.

[0586] The device will send the entered comment to the server.

[0587] The server receives the comments and checks them against a list of offensive words to detect them.

[0588] Using AI models, offensive comments are transformed using generative AI models.

[0589] The converted comment replaces the original comment in the database and is flagged if it has low confidence.

[0590] The terminal displays the converted comment to the user.

[0591] This invention instantly converts negative comments from users into positive feedback, promoting healthy communication on content distribution platforms.

[0592] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0593] Step 1:

[0594] The server loads the generative AI model during system initialization. Specifically, it downloads the model file from a specific repository on the internet and extracts it into memory. At this point, the generative AI model is ready for comment analysis and conversion.

[0595] Step 2:

[0596] A user inputs a comment for a video on a device. The input comment is temporarily stored in the device's memory and then sent to the server. The device then sends the comment data to the server as an HTTP request.

[0597] Step 3:

[0598] The server receives comments from clients and checks them against a list of offensive words to detect anti-comments. In this process, each word in the comment is checked sequentially to see if it is included in the offensive words list. If it is, the comment is flagged as an anti-comment.

[0599] Step 4:

[0600] The server inputs the detected anti-comments into the generative AI model, providing the model with a prompt like this:

[0601] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[0602] The generative AI model uses this prompt to convert offensive comments into constructive advice.

[0603] Step 5:

[0604] The server receives the converted comment and checks whether its confidence exceeds a certain threshold. If it does, it replaces the original comment with the converted comment and stores it in the database. If it does not, it flags the original comment for review. This process involves assigning a confidence score to the output of the generative AI model and determining whether it exceeds the threshold.

[0605] Step 6:

[0606] The device retrieves the new converted comments from the server. Specifically, the client application periodically sends a request to the server to retrieve the updated comment data. The retrieved data is stored in the device's memory.

[0607] Step 7:

[0608] The terminal displays the retrieved and converted comments to the user. The user interface reflects new comments in real time and allows the user to view them. During this process, the retrieved comment data is rendered into the UI component.

[0609] These steps will instantly transform negative comments from users into positive feedback, promoting healthy communication on the content distribution platform.

[0610] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0611] System Overview

[0612] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes user emotions. This system is mainly composed of a server, a terminal, and a user, each of which plays a specific role.

[0613] Server processing

[0614] 1. Loading the AI ​​model and emotion engine

[0615] When the server initializes the system, it loads the AI ​​model for detecting anti-comments and the emotion engine for recognizing user emotions, which prepares the system for anti-comment detection and emotion recognition.

[0616] 2. Comment Analysis

[0617] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[0618] 3. Emotional Recognition

[0619] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, which may include, for example, anger, sadness, and joy.

[0620] 4. Comment Conversion

[0621] If the reliability of the anti-comment exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted to encourage calmness.

[0622] 5. Updating the database

[0623] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0624] Processing by the terminal

[0625] 1. Getting comments

[0626] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[0627] 2. Viewing comments

[0628] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[0629] User Action

[0630] 1. Enter a comment

[0631] The user enters a comment, which is then saved in the database.

[0632] 2. Receiving Comments

[0633] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[0634] Specific examples

[0635] Original comment:

[0636] "This video of yours is awful. It's uninspiring and not worth watching."

[0637] User sentiment:

[0638] The AI ​​model detects this comment and recognizes that the user's emotion is "anger."

[0639] Post-processing comments:

[0640] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[0641] In this way, the present invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[0642] The processing flow will be explained below.

[0643] Step 1:

[0644] The server loads the AI ​​model and emotion engine.

[0645] Operation: When the server initializes the system, it loads the AI ​​model and emotion engine from the specified path and initializes each model. This prepares the system for anti-comment detection and emotion recognition.

[0646] Step 2:

[0647] The device connects to the comment database and fetches all comments.

[0648] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[0649] Step 3:

[0650] The server parses each comment

[0651] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model analyzes the comment, determines whether it is an anti-comment, and returns a result including a confidence score.

[0652] Step 4:

[0653] The server recognizes the user's emotions

[0654] How it works: The server uses the emotion engine to analyze the emotional state of the user who posted the comment, and obtains emotional data such as anger, sadness, and joy as the analysis result.

[0655] Step 5:

[0656] The server checks the credibility of anti-comments

[0657] How it works: The server uses the AI ​​model's analysis results to determine whether the confidence level exceeds a certain threshold (e.g., 75%). If it does, it proceeds to the next step; if it does not, it flags it for review.

[0658] Step 6:

[0659] The server converts hate comments into constructive advice

[0660] How it works: The server converts anti-comments with a confidence level above a threshold into constructive advice based on the analysis results of the emotion engine. For example, if anger is detected, the advice includes encouraging people to stay calm.

[0661] Step 7:

[0662] The server updates the database with the converted comments

[0663] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[0664] Step 8:

[0665] The terminal fetches the converted comment from the database.

[0666] Operation: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[0667] Step 9:

[0668] The device displays the updated comment to the user.

[0669] What it does: Updates the device UI to show the converted constructive comment to the user, ensuring they receive constructive and sensitive feedback instead of offensive comments.

[0670] Example 2

[0671] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0672] In today's Internet, users can freely post comments to each other, leading to the proliferation of anti-comments and making healthy communication difficult. Anti-comments, in particular, can increase the psychological burden on the recipient and potentially worsen the overall communication environment. While detecting anti-comments and converting them into constructive advice is important in this situation, there is also a need for more effective communication support that takes into account the user's emotions.

[0673] 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.

[0674] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice format, and means for taking user emotion data into consideration when converting. This makes it possible to convert anti-comments into constructive advice format while taking user emotion into consideration, thereby promoting a healthy communication environment.

[0675] "Anti-comments" are comments that contain negative content such as attacks, insults, or criticism of others.

[0676] "Constructive advice format" refers to comments that point out problems and include specific suggestions and advice for improving them.

[0677] "User emotion data" is data that indicates the user's emotional state at the time of posting a comment, and typically includes emotions such as anger, sadness, and joy.

[0678] "Confidence" is a score that indicates the degree of certainty with which the AI ​​model detects a comment as an anti-comment.

[0679] A "threshold" is a boundary value of the reliability score for performing a specific process or judgment.

[0680] A "database" is a system for systematically storing and managing information such as comments, their conversion results, and user emotional data.

[0681] A "flag" is a mark or indication that is set for a particular condition, in this case indicating that a review is required.

[0682] The "server" is a central computer system that performs comment analysis, emotion recognition, conversion processing, etc.

[0683] A "terminal" is a device that allows a user to input comments and view the displayed comments.

[0684] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0685] Server processing

[0686] The server is the center of the system and performs processing in the following procedure.

[0687] 1. Loading the AI ​​model and emotion engine

[0688] The server loads an AI model (e.g., generative AI model) for detecting anti-comments and an emotion engine (e.g., emotion analysis software) for recognizing user emotions during system initialization, which prepares the system for anti-comment detection and emotion recognition.

[0689] 2. Comment Analysis

[0690] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[0691] 3. Emotional Recognition

[0692] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, such as anger, sadness, and joy.

[0693] 4. Comment Conversion

[0694] If the confidence level exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine.

[0695] 5. Updating the database

[0696] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0697] Processing by the terminal

[0698] The terminal functions as an interface with the user and performs processing in the following procedure.

[0699] 1. Getting comments

[0700] The device retrieves the updated comments from the server. The latest comment information is retrieved from the server via an HTTP request.

[0701] 2. Viewing comments

[0702] The user interface of the terminal displays the updated comments to the user in a visually easy to understand format.

[0703] User Action

[0704] The user is the entity that directly interacts with the system and performs the following process.

[0705] 1. Enter a comment

[0706] Users enter and post comments. The posted comments are saved in a database and sent to the server.

[0707] 2. Receiving Comments

[0708] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[0709] Specific examples

[0710] For example, if the original comment was, "This video of yours is really terrible. It's unmotivating and not worth watching," the AI ​​model and emotion engine would analyze this comment and recognize that the user's emotion was "anger."

[0711] An example prompt that takes this into account is:

[0712] "Your video is terrible. It's uninspiring and not worth watching." The emotion in this comment is anger. Convert this comment into constructive advice that encourages calm.

[0713] The resulting comment after processing looks like this:

[0714] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[0715] In this way, the present invention not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[0716] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0717] Program processing flow

[0718] Step 1:

[0719] The server loads the AI ​​model and emotion engine during system initialization.

[0720] Specific behavior:

[0721] The server uses the APIs and authentication information of the generative AI model (e.g., GPT-4) and sentiment analysis engine (e.g., general sentiment analysis software) to launch these models. As input, it obtains the necessary information from the configuration file and environment variables set when the server is started, and as output, it obtains the state when the model has been loaded.

[0722] Step 2:

[0723] The server retrieves all comments from the database.

[0724] Specific behavior:

[0725] The server retrieves all comments from the comment database using an SQL query, etc. Database connection information and a query are required as input, and a text data list of comments is obtained as output.

[0726] Step 3:

[0727] Each comment received by the server is input into an AI model to determine whether it is an anti-comment.

[0728] Specific behavior:

[0729] The server inputs each comment as a prompt into the generative AI model and parses the model's response. The input is the comment text, and the output is whether the comment is an anti-comment or not, along with its confidence score.

[0730] Step 4:

[0731] The server uses an emotion engine to recognize the user's emotions when posting a comment.

[0732] Specific behavior:

[0733] The server sends each comment along with the user's emotional data to the emotion engine and receives the analysis results from the emotion engine. The input is the comment text and the emotion analysis request, and the output is the emotional data (e.g., anger, sadness, joy, etc.).

[0734] Step 5:

[0735] The server converts the anti-comments into constructive advice.

[0736] Specific behavior:

[0737] The server generates a prompt sentence for the AI ​​model and converts the anti-comment. At this time, the prompt also includes the user's emotional data. The input is the anti-comment and the prompt sentence including the emotional data, and the output is the comment converted into a constructive advice format.

[0738] Step 6:

[0739] The server replaces the converted comment with the original comment in the database.

[0740] Specific behavior:

[0741] The server replaces the original comment with the transformed comment using an SQL update query, taking as input the comment ID and the transformed comment text, and obtaining the updated database result as output.

[0742] Step 7:

[0743] Comments that do not exceed a confidence threshold are flagged for review.

[0744] Specific behavior:

[0745] If the reliability does not exceed the threshold, the server sets a review-needed flag for the comment in the database. The input is the comment ID and the review-needed flag, and the output is the updated database result.

[0746] Step 8:

[0747] The terminal obtains the updated comments from the server and reflects them in the user interface.

[0748] Specific behavior:

[0749] The terminal obtains the updated comments from the server via an HTTP request and displays them on the user interface. The input is the HTTP request and the comment data as a response, and the output is the comment display on the user interface.

[0750] Step 9:

[0751] The user enters a comment, which is then saved in the database.

[0752] Specific behavior:

[0753] The user enters a comment in the comment input form and submits it, and the server saves it in the database. The input is the comment text entered by the user and a submission request, and the output is the result of saving the new comment in the database.

[0754] Step 10:

[0755] A user views comments from other users via a terminal.

[0756] Specific behavior:

[0757] The user views the latest comment list displayed on the terminal screen and receives the converted constructive comments. The input is the comment data obtained from the server, and the output is the screen display that the user views.

[0758] (Application example 2)

[0759] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0760] In conventional systems, it is difficult for online content distribution services to respond appropriately to anti-comments among user comments, and the continued negative feedback can cause psychological stress. Furthermore, a simple negative comment detection system does not provide constructive feedback that takes into account the user's feelings, which means that healthy communication cannot be achieved.

[0761] 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.

[0762] In this invention, the server includes a means for automatically detecting anti-comments, a means for converting anti-comments into constructive advice using an AI model, and a means for recognizing the user's emotions and reflecting them in the converted comments. This allows the anti-comments to be converted into constructive feedback while taking the user's emotions into consideration, thereby promoting healthy communication while reducing psychological burden.

[0763] "Anti-comments" are comments that contain negative or offensive content about online content or individuals.

[0764] An "AI model" is a model for performing data analysis or predictive batch processing using artificial intelligence.

[0765] "Constructive advice" is feedback that contains negative content but is reframed as a positive suggestion for improvement.

[0766] "User emotions" is data obtained by analyzing the user's psychological state at the time of posting a comment using an emotion engine.

[0767] An "emotion engine" is an algorithm or software for identifying emotions from a user's text data.

[0768] "Confidence" is a score that indicates the degree of certainty of the analysis results that the AI ​​model assigns when detecting anti-comments.

[0769] The "specific threshold" is a reference value for determining whether or not a comment judgment is converted into constructive advice.

[0770] "Database" means an information management system for storing and managing comments and other data.

[0771] "Flag for review" is a marking of anti-comments that do not exceed a certain threshold, indicating that further review or human verification is required.

[0772] System Overview

[0773] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes users' emotions. The system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0774] Server processing

[0775] The server is the central element of the system and performs the following functions:

[0776] 1. Loading the AI ​​model and emotion engine:

[0777] When the server initializes the system, it loads an AI model for detecting anti-comments and an emotion engine for recognizing user emotions. This prepares the system for anti-comment detection and emotion recognition. The Python transformers library can be used as the AI ​​model.

[0778] 2. Comment Analysis:

[0779] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and the AI ​​model determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[0780] 3. Emotion Recognition:

[0781] The emotion engine recognizes the user's emotion when posting a comment and retrieves the emotion data. Emotion data can include, for example, anger, sadness, and joy. The vaderSentiment library can be used for this.

[0782] 4. Comment conversion:

[0783] If the confidence level exceeds a certain threshold, the anti-comment is converted into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted into a message encouraging them to stay calm.

[0784] 5. Database Update:

[0785] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0786] Processing by the terminal

[0787] The device acts as a link between the server and the user and performs the following tasks:

[0788] 1. Getting comments:

[0789] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[0790] 2. View comments:

[0791] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[0792] User Action

[0793] Users interact with the system in the following ways:

[0794] 1. Enter your comment:

[0795] The user enters a comment, which is then saved in the database.

[0796] 2. Receiving Comments:

[0797] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[0798] Specific examples

[0799] Original comment: "This video is so boring and a waste of time."

[0800] Recognized emotion: "Anger"

[0801] Converted comment: "There may be some room for improvement in this video. For example, if you put more thought into the content, it might be more interesting for viewers. I recommend you think about it calmly."

[0802] Prompt Sentence Examples

[0803] Type: "The sound in this video is so bad I can't hear you."

[0804] Prompt: "This comment is negative feedback. Please convert it into constructive feedback with suggestions for improvement."

[0805] In this way, this invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[0806] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0807] Step 1:

[0808] The server loads the AI ​​model and emotion engine when the system is initialized. Specifically, it loads the AI ​​model for detecting anti-comments using the Python transformers library and initializes the vaderSentiment library for emotion analysis. This enables the system to detect anti-comments and recognize user emotions.

[0809] Input: None

[0810] Data processing / calculation: Initialization of AI model and emotion engine

[0811] Output: Initialized AI model and emotion engine

[0812] Step 2:

[0813] The server retrieves all comments from the database. Each comment is then input into an AI model. The AI ​​model determines whether the comment is an anti-comment and returns a result including its confidence level. Specifically, it uses a transformers model to classify the text.

[0814] Input: Database comment

[0815] Data processing / calculation: Comment analysis using AI models

[0816] Output: Anti-comment judgment results and confidence score

[0817] Step 3:

[0818] The server inputs the comment text into the emotion engine to recognize the user's emotion at the same time as the analyzed comment. The emotion engine analyzes the comment's emotional state (anger, joy, sadness, etc.) and returns the results. Specifically, emotion analysis is performed using the vaderSentiment library.

[0819] Input: Comment text

[0820] Data processing / calculation: Emotion analysis using emotion engine

[0821] Output: User emotion data

[0822] Step 4:

[0823] Only if the reliability exceeds a certain threshold will the server convert the anti-comment into constructive advice. The conversion process takes into account the user's emotions as recognized by the emotion engine. Specifically, if the user's emotion is "anger," the server converts the message into a message encouraging them to stay calm, and if the user's emotion is "sadness," the server converts the message into a message containing encouragement.

[0824] Input: Anti-comment judgment results, user emotion data

[0825] Data processing / calculation: Comment conversion

[0826] Output: Comments converted into constructive advice format

[0827] Step 5:

[0828] The server replaces the original comment with the converted comment and stores it in the database. The server then deletes the pre-conversion comment data and inserts the new comment data. This updates the comment information in the database.

[0829] Input: Translated comment

[0830] Data processing / calculation: Database update

[0831] Output: Updated comment data

[0832] Step 6:

[0833] The device retrieves the latest comment data from the server and reflects it on the user's UI. The device's UI then displays the converted comments to the user, allowing the user to receive constructive and emotionally sensitive feedback.

[0834] Input: Updated comment data

[0835] Data processing / calculation: Acquisition and display of comment data

[0836] Output: Constructive comments displayed in the user UI

[0837] Step 7:

[0838] Users can view comments from other users via their devices, and can also enter their own comments, which are then saved in the database, after which the system returns to the loop of analyzing and converting new comments.

[0839] Input: User comment

[0840] Data processing / calculation: Entering and saving comments

[0841] Output: New comment saved in the database

[0842] In this way, the system converts anti-comments into constructive feedback, promoting healthy communication while reducing the psychological burden on users.

[0843] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0844] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0845] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0846] [Third embodiment]

[0847] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0848] 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.

[0849] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0850] 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.

[0851] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0852] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0853] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0854] 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.

[0855] 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.

[0856] 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.

[0857] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0858] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0859] System Overview

[0860] This invention is a system that automatically detects anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0861] Server processing

[0862] 1. Loading the AI ​​model

[0863] When the server initializes the system, it loads an AI model for detecting anti-comments, which is loaded from a specific location within the network.

[0864] 2. Comment Analysis

[0865] The server retrieves all comments from the database and inputs each comment into the AI ​​model for analysis. The analysis results determine whether the comment is an anti-comment. This determination result includes a probability value (confidence) that the comment is an anti-comment.

[0866] 3. Comment conversion

[0867] The server converts anti-comments into constructive advice if their confidence exceeds a certain threshold by replacing offensive language with polite language.

[0868] 4. Updating the database

[0869] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[0870] Processing by the terminal

[0871] 1. Fetching comments

[0872] The device retrieves the new converted comments from the server and uses them to display in the user's browser or app UI.

[0873] 2. Viewing comments

[0874] The device UI will display the converted comments to the user, allowing them to receive more constructive feedback.

[0875] User Action

[0876] 1. Enter a comment

[0877] The user enters a comment, which is then saved in the database.

[0878] 2. Receiving Comments

[0879] Users can view comments from other users via their device, and the converted constructive comments are displayed, allowing users to receive more positive feedback.

[0880] Specific examples

[0881] Original comment:

[0882] "This video of yours is awful. It's uninspiring and not worth watching."

[0883] Post-processing comments:

[0884] "Your video could use some improvement. For example, if you improve the sound quality or picture quality, it might be easier for viewers to watch."

[0885] In this way, the present invention is a system that converts aggressive anti-comments into constructive advice, thereby reducing psychological stress for users and promoting healthy communication.

[0886] The processing flow will be explained below.

[0887] Step 1:

[0888] The server loads the AI ​​model

[0889] Operation: When the system starts up, the server loads the AI ​​model from the specified path and initializes the model, which prepares it for anti-comment detection.

[0890] Step 2:

[0891] The device connects to the comment database and fetches all comments.

[0892] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[0893] Step 3:

[0894] The server parses each comment

[0895] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model determines whether the comment is an anti-comment or not and returns a result including a confidence score (a value between 0 and 1).

[0896] Step 4:

[0897] The server checks the credibility of anti-comments

[0898] Operation: Determine if the confidence level of the analysis result is 75% or higher. If it is 75% or higher, proceed to the next step. If it is less than 75%, flag it for review.

[0899] Step 5:

[0900] The server converts hate comments into constructive advice

[0901] How it works: For anti-comments with a confidence level of 75% or higher, the server uses an AI model to convert the content of the comment into constructive language, specifically removing offensive language and including positive, specific suggestions for improvement.

[0902] Step 6:

[0903] The server updates the database with the converted comments

[0904] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[0905] Step 7:

[0906] The terminal fetches the converted comment

[0907] What it does: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[0908] Step 8:

[0909] The device will display the updated comment to the user.

[0910] What it does: Updates the device UI to show the user the converted constructive comment, ensuring they receive constructive feedback instead of abusive comments.

[0911] Example 1

[0912] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0913] In traditional online communication, offensive and anti-comments place a psychological burden on users. This often leads to a negative atmosphere in the community and hinders healthy communication. Furthermore, manually managing and censoring offensive comments requires a great deal of effort, making efficient management a key priority.

[0914] 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.

[0915] In this invention, the server includes means for loading an artificial intelligence model to automatically detect anti-comments, means for retrieving comments from a database and analyzing these comments, means for determining whether the analyzed comments are anti-comments, means for converting the anti-comments into constructive advice format based on the determination result, means for replacing the converted comments with the original comments in the database, and means for displaying the replaced comments on a user interface, thereby enabling the offensive comments to be automatically converted into constructive feedback.

[0916] "Anti-comments" are comments that contain negative, offensive, or insulting content and cause psychological stress to the target.

[0917] An "artificial intelligence model" is a set of algorithms and their parameters that can automatically process specific tasks using machine learning techniques.

[0918] A "database" is a system designed to store data in an organized manner and make it easy to access and manage.

[0919] "Analysis" is the process of examining the data obtained in detail and extracting useful information from it.

[0920] "Determining" refers to the act of determining whether or not a particular condition is met based on the analysis results.

[0921] "Constructive advice" is the conversion of negative feedback into positive suggestions for improvement that the recipient can take in a positive way.

[0922] "User interface" is a general term for the screens and operating methods that users use to interact with a system.

[0923] "Reliability" is a numerical index that indicates the accuracy and certainty of the judgment result.

[0924] A "threshold" is a value that sets a certain reference value and serves as a boundary for executing a specific action or process.

[0925] A "flag" is a mark or indicator that is set for identification or processing based on a specific condition.

[0926] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[0927] Server processing

[0928] The server loads an artificial intelligence model for detecting anti-comments during system initialization. This model is loaded from a specific location within the network using a machine learning framework such as TensorFlow or PyTorch. The loaded artificial intelligence model is used to analyze and judge anti-comments.

[0929] The server then retrieves all comments from a database (e.g., MySQL or PostgreSQL) and analyzes this data. Each comment is fed into an artificial intelligence model, which determines whether it is an anti-comment or not. The result of this determination includes a probability value (confidence) that the comment is an anti-comment.

[0930] Based on the results, the server converts the anti-comment into a constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT) that replaces offensive language with polite language. This conversion process is performed using prompts such as:

[0931] Please convert the offensive comments below into constructive feedback:

[0932] "This video of yours is awful. It's uninspiring and not worth watching."

[0933] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is automatically flagged for review.

[0934] Processing by the terminal

[0935] The device fetches the new converted comments from the server and displays them in the user's browser or application's user interface, using front-end frameworks such as React or Vue.js to present the converted comments to the user in a visually understandable way.

[0936] User Action

[0937] Users can enter comments through a form on a website or application. These comments are immediately stored in a database and analyzed and converted by the server. After processing, users can receive comments from other users through their own devices and receive constructive feedback. This helps users reduce psychological stress and promote healthy communication.

[0938] The above is an embodiment of the present invention, and this system can automatically convert anti-comments into constructive feedback, thereby realizing positive communication between users.

[0939] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0940] Step 1: Loading the AI ​​model

[0941] When the system is initialized, the server loads an artificial intelligence model for detecting anti-comments. The input is the file path of the AI ​​model stored in a specific location within the network. Specifically, the model file is read into memory using TensorFlow or PyTorch. The output of this process is an AI model capable of detecting anti-comments loaded into memory.

[0942] Step 2: Obtaining and parsing comments

[0943] The server retrieves all comments from a database. The database uses MySQL or PostgreSQL, and executes SQL queries to extract comment data. The input is the access information to the database where the comments are stored. The retrieved comments are then input into an AI model to determine whether they are anti-comments. Specifically, each comment is input into the AI ​​model, and a confidence score is obtained as the output. The higher the confidence score, the more likely the comment is to be an anti-comment.

[0944] Step 3: Convert comments

[0945] The server converts anti-comments into constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT). The anti-comments and their confidence levels are used as input. Specifically, the server inputs a prompt sentence into the generative AI model to obtain constructive feedback. The following sentence is used as an example of this prompt sentence:

[0946] Please convert the offensive comments below into constructive feedback:

[0947] "This video of yours is awful. It's uninspiring and not worth watching."

[0948] The output is that offensive comments are transformed into constructive feedback.

[0949] Step 4: Update the database

[0950] The server replaces the original comment with the converted comment and stores it in the database. The inputs are the IDs of the converted comment and its original comment. Specifically, it executes an SQL query using an UPDATE statement to replace the original comment with the new comment. If the confidence does not exceed a threshold, the comment is flagged for review.

[0951] Step 5: Fetch comments

[0952] The device fetches the latest comments from the server. The server's API endpoint URL is used as input. Specifically, it sends an HTTP request to the server and receives data in JSON format from the server. The output of this process is the latest comment data retrieved by the device.

[0953] Step 6: View comments

[0954] The device's user interface displays the fetched comments. It uses the retrieved comment data as input, renders a component that visualizes the comment data using a front-end framework such as React or Vue.js, and displays the latest constructive feedback to the user as output.

[0955] Step 7: Enter a comment

[0956] Users enter comments through a form on a website or application. The actual comment text entered by the user is used as input. Specifically, the form data is sent to the server's API, which receives it and stores it in a database. The output of this process is that the new comment data is added to the database.

[0957] Step 8: Receiving comments

[0958] The user views the comments retrieved from the server via the device. The latest comment data retrieved by the device is used as input. The specific operation is to display the comments using a front-end framework. As output, the user can view constructive feedback.

[0959] (Application example 1)

[0960] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0961] On current content distribution platforms, user comments often contain offensive content, causing psychological stress for content providers and other users. Ignoring such anti-comments can hinder healthy communication and potentially lower the quality of the platform as a whole. Therefore, the present invention aims to provide a system that automatically detects offensive anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users and promoting healthy communication.

[0962] 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.

[0963] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice, means for determining whether a comment entered by a user is offensive, means for converting the offensive comment into constructive advice using a generative AI model, means for replacing the original comment in the database with the converted comment, and means for displaying the replaced comment. This allows negative comments from users to be instantly converted into positive feedback, enabling healthy communication on the content distribution platform.

[0964] "Anti-comments" are comments made by users that contain offensive, negative or harsh content about specific content or other users.

[0965] "Constructive advice" is a form of positive feedback that is beneficial to the recipient and includes specific improvements or suggestions.

[0966] A "generative AI model" is an artificial intelligence model that is trained on a large dataset and has the ability to analyze and convert input text data.

[0967] A "database" is an electronic data storage system that has a structure and allows efficient management, storage, searching, and retrieval of data.

[0968] An "offensive word list" is a collection of words that includes specific terms and phrases that, when used, may create a negative or offensive impression on others.

[0969] "Confidence" is a probability value that indicates whether a particular output (for example, a judgment of anti-comments) is correct based on the data input to the AI ​​model.

[0970] A "prompt" is the initial input data given to an AI model to perform a specific task.

[0971] A "content distribution service" is a service that provides digital content such as video, audio, and text to users via the Internet.

[0972] A "flag" is an identifier that is assigned to a data item or a processing step based on a specific condition.

[0973] The system for implementing the present invention is composed of a server, a terminal, and a user. The specific configuration and operation method of the system will be described below.

[0974] Server processing

[0975] Loading an AI model

[0976] When the system is initialized, the server loads a generative AI model for detecting anti-comments, which is loaded from a specific repository on the Internet.

[0977] Comment parsing and transformation

[0978] The server analyzes all comments retrieved from the database, checking them against a list of offensive words to detect anti-comments, which are then fed into a generative AI model and transformed into constructive advice using prompts.

[0979] Examples of prompts:

[0980] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[0981] Update the database with the conversion results

[0982] The server replaces the original comment with the converted comment and stores only those whose confidence exceeds a certain threshold in the database, or flags the comment for review if its confidence does not exceed the threshold.

[0983] Processing by the terminal

[0984] Fetching and displaying comments

[0985] The device retrieves the new converted comments from the server and displays them in the user's browser or app UI, allowing users to receive more constructive feedback.

[0986] User Action

[0987] Entering and receiving comments

[0988] Users can use the content distribution service's application to input and submit comments. The input comments are stored in a database, and users can also view constructive comments converted into feedback from other users.

[0989] Hardware and software used

[0990] Hardware: Smartphone (iOS, Android), Head-Mounted Display (HMD)

[0991] software:

[0992] Server-side AI model: OpenAI's generative AI model

[0993] Database: A relational database such as MySQL or PostgreSQL

[0994] Comment analysis and transformation: Scripting with Python

[0995] Explanation of specific steps in the process

[0996] Users can enter comments on videos.

[0997] The device will send the entered comment to the server.

[0998] The server receives the comments and checks them against a list of offensive words to detect them.

[0999] Using AI models, offensive comments are transformed using generative AI models.

[1000] The converted comment replaces the original comment in the database and is flagged if it has low confidence.

[1001] The terminal displays the converted comment to the user.

[1002] This invention instantly converts negative comments from users into positive feedback, promoting healthy communication on content distribution platforms.

[1003] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1004] Step 1:

[1005] The server loads the generative AI model during system initialization. Specifically, it downloads the model file from a specific repository on the internet and extracts it into memory. At this point, the generative AI model is ready for comment analysis and conversion.

[1006] Step 2:

[1007] A user inputs a comment for a video on a device. The input comment is temporarily stored in the device's memory and then sent to the server. The device then sends the comment data to the server as an HTTP request.

[1008] Step 3:

[1009] The server receives comments from clients and checks them against a list of offensive words to detect anti-comments. In this process, each word in the comment is checked sequentially to see if it is included in the offensive words list. If it is, the comment is flagged as an anti-comment.

[1010] Step 4:

[1011] The server inputs the detected anti-comments into the generative AI model, providing the model with a prompt like this:

[1012] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[1013] The generative AI model uses this prompt to convert offensive comments into constructive advice.

[1014] Step 5:

[1015] The server receives the converted comment and checks whether its confidence exceeds a certain threshold. If it does, it replaces the original comment with the converted comment and stores it in the database. If it does not, it flags the original comment for review. This process involves assigning a confidence score to the output of the generative AI model and determining whether it exceeds the threshold.

[1016] Step 6:

[1017] The device retrieves the new converted comments from the server. Specifically, the client application periodically sends a request to the server to retrieve the updated comment data. The retrieved data is stored in the device's memory.

[1018] Step 7:

[1019] The terminal displays the retrieved and converted comments to the user. The user interface reflects new comments in real time and allows the user to view them. During this process, the retrieved comment data is rendered into the UI component.

[1020] These steps will instantly transform negative comments from users into positive feedback, promoting healthy communication on the content distribution platform.

[1021] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1022] System Overview

[1023] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes user emotions. This system is mainly composed of a server, a terminal, and a user, each of which plays a specific role.

[1024] Server processing

[1025] 1. Loading the AI ​​model and emotion engine

[1026] When the server initializes the system, it loads the AI ​​model for detecting anti-comments and the emotion engine for recognizing user emotions, which prepares the system for anti-comment detection and emotion recognition.

[1027] 2. Comment Analysis

[1028] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[1029] 3. Emotional Recognition

[1030] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, which may include, for example, anger, sadness, and joy.

[1031] 4. Comment Conversion

[1032] If the reliability of the anti-comment exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted to encourage calmness.

[1033] 5. Updating the database

[1034] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1035] Processing by the terminal

[1036] 1. Getting comments

[1037] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[1038] 2. Viewing comments

[1039] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[1040] User Action

[1041] 1. Enter a comment

[1042] The user enters a comment, which is then saved in the database.

[1043] 2. Receiving Comments

[1044] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[1045] Specific examples

[1046] Original comment:

[1047] "This video of yours is awful. It's uninspiring and not worth watching."

[1048] User sentiment:

[1049] The AI ​​model detects this comment and recognizes that the user's emotion is "anger."

[1050] Post-processing comments:

[1051] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[1052] In this way, the present invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[1053] The processing flow will be explained below.

[1054] Step 1:

[1055] The server loads the AI ​​model and emotion engine.

[1056] Operation: When the server initializes the system, it loads the AI ​​model and emotion engine from the specified path and initializes each model. This prepares the system for anti-comment detection and emotion recognition.

[1057] Step 2:

[1058] The device connects to the comment database and fetches all comments.

[1059] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[1060] Step 3:

[1061] The server parses each comment

[1062] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model analyzes the comment, determines whether it is an anti-comment, and returns a result including a confidence score.

[1063] Step 4:

[1064] The server recognizes the user's emotions

[1065] How it works: The server uses the emotion engine to analyze the emotional state of the user who posted the comment, and obtains emotional data such as anger, sadness, and joy as the analysis result.

[1066] Step 5:

[1067] The server checks the credibility of anti-comments

[1068] How it works: The server uses the AI ​​model's analysis results to determine whether the confidence level exceeds a certain threshold (e.g., 75%). If it does, it proceeds to the next step; if it does not, it flags it for review.

[1069] Step 6:

[1070] The server converts hate comments into constructive advice

[1071] How it works: The server converts anti-comments with a confidence level above a threshold into constructive advice based on the analysis results of the emotion engine. For example, if anger is detected, the advice includes encouraging people to stay calm.

[1072] Step 7:

[1073] The server updates the database with the converted comments

[1074] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[1075] Step 8:

[1076] The terminal fetches the converted comment from the database.

[1077] Operation: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[1078] Step 9:

[1079] The device displays the updated comment to the user.

[1080] What it does: Updates the device UI to show the converted constructive comment to the user, ensuring they receive constructive and sensitive feedback instead of offensive comments.

[1081] Example 2

[1082] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1083] In today's Internet, users can freely post comments to each other, leading to the proliferation of anti-comments and making healthy communication difficult. Anti-comments, in particular, can increase the psychological burden on the recipient and potentially worsen the overall communication environment. While detecting anti-comments and converting them into constructive advice is important in this situation, there is also a need for more effective communication support that takes into account the user's emotions.

[1084] 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.

[1085] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice format, and means for taking user emotion data into consideration when converting. This makes it possible to convert anti-comments into constructive advice format while taking user emotion into consideration, thereby promoting a healthy communication environment.

[1086] "Anti-comments" are comments that contain negative content such as attacks, insults, or criticism of others.

[1087] "Constructive advice format" refers to comments that point out problems and include specific suggestions and advice for improving them.

[1088] "User emotion data" is data that indicates the user's emotional state at the time of posting a comment, and typically includes emotions such as anger, sadness, and joy.

[1089] "Confidence" is a score that indicates the degree of certainty with which the AI ​​model detects a comment as an anti-comment.

[1090] A "threshold" is a boundary value of the reliability score for performing a specific process or judgment.

[1091] A "database" is a system for systematically storing and managing information such as comments, their conversion results, and user emotional data.

[1092] A "flag" is a mark or indication that is set for a particular condition, in this case indicating that a review is required.

[1093] The "server" is a central computer system that performs comment analysis, emotion recognition, conversion processing, etc.

[1094] A "terminal" is a device that allows a user to input comments and view the displayed comments.

[1095] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[1096] Server processing

[1097] The server is the center of the system and performs processing in the following procedure.

[1098] 1. Loading the AI ​​model and emotion engine

[1099] The server loads an AI model (e.g., generative AI model) for detecting anti-comments and an emotion engine (e.g., emotion analysis software) for recognizing user emotions during system initialization, which prepares the system for anti-comment detection and emotion recognition.

[1100] 2. Comment Analysis

[1101] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[1102] 3. Emotional Recognition

[1103] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, such as anger, sadness, and joy.

[1104] 4. Comment Conversion

[1105] If the confidence level exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine.

[1106] 5. Updating the database

[1107] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1108] Processing by the terminal

[1109] The terminal functions as an interface with the user and performs processing in the following procedure.

[1110] 1. Getting comments

[1111] The device retrieves the updated comments from the server. The latest comment information is retrieved from the server via an HTTP request.

[1112] 2. Viewing comments

[1113] The user interface of the terminal displays the updated comments to the user in a visually easy to understand format.

[1114] User Action

[1115] The user is the entity that directly interacts with the system and performs the following process.

[1116] 1. Enter a comment

[1117] Users enter and post comments. The posted comments are saved in a database and sent to the server.

[1118] 2. Receiving Comments

[1119] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[1120] Specific examples

[1121] For example, if the original comment was, "This video of yours is really terrible. It's unmotivating and not worth watching," the AI ​​model and emotion engine would analyze this comment and recognize that the user's emotion was "anger."

[1122] An example prompt that takes this into account is:

[1123] "Your video is terrible. It's uninspiring and not worth watching." The emotion in this comment is anger. Convert this comment into constructive advice that encourages calm.

[1124] The resulting comment after processing looks like this:

[1125] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[1126] In this way, the present invention not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[1127] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1128] Program processing flow

[1129] Step 1:

[1130] The server loads the AI ​​model and emotion engine during system initialization.

[1131] Specific behavior:

[1132] The server uses the APIs and authentication information of the generative AI model (e.g., GPT-4) and sentiment analysis engine (e.g., general sentiment analysis software) to launch these models. As input, it obtains the necessary information from the configuration file and environment variables set when the server is started, and as output, it obtains the state when the model has been loaded.

[1133] Step 2:

[1134] The server retrieves all comments from the database.

[1135] Specific behavior:

[1136] The server retrieves all comments from the comment database using an SQL query, etc. Database connection information and a query are required as input, and a text data list of comments is obtained as output.

[1137] Step 3:

[1138] Each comment received by the server is input into an AI model to determine whether it is an anti-comment.

[1139] Specific behavior:

[1140] The server inputs each comment as a prompt into the generative AI model and parses the model's response. The input is the comment text, and the output is whether the comment is an anti-comment or not, along with its confidence score.

[1141] Step 4:

[1142] The server uses an emotion engine to recognize the user's emotions when posting a comment.

[1143] Specific behavior:

[1144] The server sends each comment along with the user's emotional data to the emotion engine and receives the analysis results from the emotion engine. The input is the comment text and the emotion analysis request, and the output is the emotional data (e.g., anger, sadness, joy, etc.).

[1145] Step 5:

[1146] The server converts the anti-comments into constructive advice.

[1147] Specific behavior:

[1148] The server generates a prompt sentence for the AI ​​model and converts the anti-comment. At this time, the prompt also includes the user's emotional data. The input is the anti-comment and the prompt sentence including the emotional data, and the output is the comment converted into a constructive advice format.

[1149] Step 6:

[1150] The server replaces the converted comment with the original comment in the database.

[1151] Specific behavior:

[1152] The server replaces the original comment with the transformed comment using an SQL update query, taking as input the comment ID and the transformed comment text, and obtaining the updated database result as output.

[1153] Step 7:

[1154] Comments that do not exceed a confidence threshold are flagged for review.

[1155] Specific behavior:

[1156] If the reliability does not exceed the threshold, the server sets a review-needed flag for the comment in the database. The input is the comment ID and the review-needed flag, and the output is the updated database result.

[1157] Step 8:

[1158] The terminal obtains the updated comments from the server and reflects them in the user interface.

[1159] Specific behavior:

[1160] The terminal obtains the updated comments from the server via an HTTP request and displays them on the user interface. The input is the HTTP request and the comment data as a response, and the output is the comment display on the user interface.

[1161] Step 9:

[1162] The user enters a comment, which is then saved in the database.

[1163] Specific behavior:

[1164] The user enters a comment in the comment input form and submits it, and the server saves it in the database. The input is the comment text entered by the user and a submission request, and the output is the result of saving the new comment in the database.

[1165] Step 10:

[1166] A user views comments from other users via a terminal.

[1167] Specific behavior:

[1168] The user views the latest comment list displayed on the terminal screen and receives the converted constructive comments. The input is the comment data obtained from the server, and the output is the screen display that the user views.

[1169] (Application example 2)

[1170] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1171] In conventional systems, it is difficult for online content distribution services to respond appropriately to anti-comments among user comments, and the continued negative feedback can cause psychological stress. Furthermore, a simple negative comment detection system does not provide constructive feedback that takes into account the user's feelings, which means that healthy communication cannot be achieved.

[1172] 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.

[1173] In this invention, the server includes a means for automatically detecting anti-comments, a means for converting anti-comments into constructive advice using an AI model, and a means for recognizing the user's emotions and reflecting them in the converted comments. This allows the anti-comments to be converted into constructive feedback while taking the user's emotions into consideration, thereby promoting healthy communication while reducing psychological burden.

[1174] "Anti-comments" are comments that contain negative or offensive content about online content or individuals.

[1175] An "AI model" is a model for performing data analysis or predictive batch processing using artificial intelligence.

[1176] "Constructive advice" is feedback that contains negative content but is reframed as a positive suggestion for improvement.

[1177] "User emotions" is data obtained by analyzing the user's psychological state at the time of posting a comment using an emotion engine.

[1178] An "emotion engine" is an algorithm or software for identifying emotions from a user's text data.

[1179] "Confidence" is a score that indicates the degree of certainty of the analysis results that the AI ​​model assigns when detecting anti-comments.

[1180] The "specific threshold" is a reference value for determining whether or not a comment judgment is converted into constructive advice.

[1181] "Database" means an information management system for storing and managing comments and other data.

[1182] "Flag for review" is a marking of anti-comments that do not exceed a certain threshold, indicating that further review or human verification is required.

[1183] System Overview

[1184] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes users' emotions. The system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[1185] Server processing

[1186] The server is the central element of the system and performs the following functions:

[1187] 1. Loading the AI ​​model and emotion engine:

[1188] When the server initializes the system, it loads an AI model for detecting anti-comments and an emotion engine for recognizing user emotions. This prepares the system for anti-comment detection and emotion recognition. The Python transformers library can be used as the AI ​​model.

[1189] 2. Comment Analysis:

[1190] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and the AI ​​model determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[1191] 3. Emotion Recognition:

[1192] The emotion engine recognizes the user's emotion when posting a comment and retrieves the emotion data. Emotion data can include, for example, anger, sadness, and joy. The vaderSentiment library can be used for this.

[1193] 4. Comment conversion:

[1194] If the confidence level exceeds a certain threshold, the anti-comment is converted into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted into a message encouraging them to stay calm.

[1195] 5. Database Update:

[1196] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1197] Processing by the terminal

[1198] The device acts as a link between the server and the user and performs the following tasks:

[1199] 1. Getting comments:

[1200] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[1201] 2. View comments:

[1202] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[1203] User Action

[1204] Users interact with the system in the following ways:

[1205] 1. Enter your comment:

[1206] The user enters a comment, which is then saved in the database.

[1207] 2. Receiving Comments:

[1208] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[1209] Specific examples

[1210] Original comment: "This video is so boring and a waste of time."

[1211] Recognized emotion: "Anger"

[1212] Converted comment: "There may be some room for improvement in this video. For example, if you put more thought into the content, it might be more interesting for viewers. I recommend you think about it calmly."

[1213] Prompt Sentence Examples

[1214] Type: "The sound in this video is so bad I can't hear you."

[1215] Prompt: "This comment is negative feedback. Please convert it into constructive feedback with suggestions for improvement."

[1216] In this way, this invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[1217] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1218] Step 1:

[1219] The server loads the AI ​​model and emotion engine when the system is initialized. Specifically, it loads the AI ​​model for detecting anti-comments using the Python transformers library and initializes the vaderSentiment library for emotion analysis. This enables the system to detect anti-comments and recognize user emotions.

[1220] Input: None

[1221] Data processing / calculation: Initialization of AI model and emotion engine

[1222] Output: Initialized AI model and emotion engine

[1223] Step 2:

[1224] The server retrieves all comments from the database. Each comment is then input into an AI model. The AI ​​model determines whether the comment is an anti-comment and returns a result including its confidence level. Specifically, it uses a transformers model to classify the text.

[1225] Input: Database comment

[1226] Data processing / calculation: Comment analysis using AI models

[1227] Output: Anti-comment judgment results and confidence score

[1228] Step 3:

[1229] The server inputs the comment text into the emotion engine to recognize the user's emotion at the same time as the analyzed comment. The emotion engine analyzes the comment's emotional state (anger, joy, sadness, etc.) and returns the results. Specifically, emotion analysis is performed using the vaderSentiment library.

[1230] Input: Comment text

[1231] Data processing / calculation: Emotion analysis using emotion engine

[1232] Output: User emotion data

[1233] Step 4:

[1234] Only if the reliability exceeds a certain threshold will the server convert the anti-comment into constructive advice. The conversion process takes into account the user's emotions as recognized by the emotion engine. Specifically, if the user's emotion is "anger," the server converts the message into a message encouraging them to stay calm, and if the user's emotion is "sadness," the server converts the message into a message containing encouragement.

[1235] Input: Anti-comment judgment results, user emotion data

[1236] Data processing / calculation: Comment conversion

[1237] Output: Comments converted into constructive advice format

[1238] Step 5:

[1239] The server replaces the original comment with the converted comment and stores it in the database. The server then deletes the pre-conversion comment data and inserts the new comment data. This updates the comment information in the database.

[1240] Input: Translated comment

[1241] Data processing / calculation: Database update

[1242] Output: Updated comment data

[1243] Step 6:

[1244] The device retrieves the latest comment data from the server and reflects it on the user's UI. The device's UI then displays the converted comments to the user, allowing the user to receive constructive and emotionally sensitive feedback.

[1245] Input: Updated comment data

[1246] Data processing / calculation: Acquisition and display of comment data

[1247] Output: Constructive comments displayed in the user UI

[1248] Step 7:

[1249] Users can view comments from other users via their devices, and can also enter their own comments, which are then saved in the database, after which the system returns to the loop of analyzing and converting new comments.

[1250] Input: User comment

[1251] Data processing / calculation: Entering and saving comments

[1252] Output: New comment saved in the database

[1253] In this way, the system converts anti-comments into constructive feedback, promoting healthy communication while reducing the psychological burden on users.

[1254] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1255] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1256] 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.

[1257] [Fourth embodiment]

[1258] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1259] 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.

[1260] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1261] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1262] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1263] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1264] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1265] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1266] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1267] 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.

[1268] 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.

[1269] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1270] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1271] System Overview

[1272] This invention is a system that automatically detects anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[1273] Server processing

[1274] 1. Loading the AI ​​model

[1275] When the server initializes the system, it loads an AI model for detecting anti-comments, which is loaded from a specific location within the network.

[1276] 2. Comment Analysis

[1277] The server retrieves all comments from the database and inputs each comment into the AI ​​model for analysis. The analysis results determine whether the comment is an anti-comment. This determination result includes a probability value (confidence) that the comment is an anti-comment.

[1278] 3. Comment conversion

[1279] The server converts anti-comments into constructive advice if their confidence exceeds a certain threshold by replacing offensive language with polite language.

[1280] 4. Updating the database

[1281] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1282] Processing by the terminal

[1283] 1. Fetching comments

[1284] The device retrieves the new converted comments from the server and uses them to display in the user's browser or app UI.

[1285] 2. Viewing comments

[1286] The device UI will display the converted comments to the user, allowing them to receive more constructive feedback.

[1287] User Action

[1288] 1. Enter a comment

[1289] The user enters a comment, which is then saved in the database.

[1290] 2. Receiving Comments

[1291] Users can view comments from other users via their device, and the converted constructive comments are displayed, allowing users to receive more positive feedback.

[1292] Specific examples

[1293] Original comment:

[1294] "This video of yours is awful. It's uninspiring and not worth watching."

[1295] Post-processing comments:

[1296] "Your video could use some improvement. For example, if you improve the sound quality or picture quality, it might be easier for viewers to watch."

[1297] In this way, the present invention is a system that converts aggressive anti-comments into constructive advice, thereby reducing psychological stress for users and promoting healthy communication.

[1298] The processing flow will be explained below.

[1299] Step 1:

[1300] The server loads the AI ​​model

[1301] Operation: When the system starts up, the server loads the AI ​​model from the specified path and initializes the model, which prepares it for anti-comment detection.

[1302] Step 2:

[1303] The device connects to the comment database and fetches all comments.

[1304] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[1305] Step 3:

[1306] The server parses each comment

[1307] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model determines whether the comment is an anti-comment or not and returns a result including a confidence score (a value between 0 and 1).

[1308] Step 4:

[1309] The server checks the credibility of anti-comments

[1310] Operation: Determine if the confidence level of the analysis result is 75% or higher. If it is 75% or higher, proceed to the next step. If it is less than 75%, flag it for review.

[1311] Step 5:

[1312] The server converts hate comments into constructive advice

[1313] How it works: For anti-comments with a confidence level of 75% or higher, the server uses an AI model to convert the content of the comment into constructive language, specifically removing offensive language and including positive, specific suggestions for improvement.

[1314] Step 6:

[1315] The server updates the database with the converted comments

[1316] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[1317] Step 7:

[1318] The terminal fetches the converted comment

[1319] What it does: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[1320] Step 8:

[1321] The device will display the updated comment to the user.

[1322] What it does: Updates the device UI to show the user the converted constructive comment, ensuring they receive constructive feedback instead of abusive comments.

[1323] Example 1

[1324] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1325] In traditional online communication, offensive and anti-comments place a psychological burden on users. This often leads to a negative atmosphere in the community and hinders healthy communication. Furthermore, manually managing and censoring offensive comments requires a great deal of effort, making efficient management a key priority.

[1326] 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.

[1327] In this invention, the server includes means for loading an artificial intelligence model to automatically detect anti-comments, means for retrieving comments from a database and analyzing these comments, means for determining whether the analyzed comments are anti-comments, means for converting the anti-comments into constructive advice format based on the determination result, means for replacing the converted comments with the original comments in the database, and means for displaying the replaced comments on a user interface, thereby enabling the offensive comments to be automatically converted into constructive feedback.

[1328] "Anti-comments" are comments that contain negative, offensive, or insulting content and cause psychological stress to the target.

[1329] An "artificial intelligence model" is a set of algorithms and their parameters that can automatically process specific tasks using machine learning techniques.

[1330] A "database" is a system designed to store data in an organized manner and make it easy to access and manage.

[1331] "Analysis" is the process of examining the data obtained in detail and extracting useful information from it.

[1332] "Determining" refers to the act of determining whether or not a particular condition is met based on the analysis results.

[1333] "Constructive advice" is the conversion of negative feedback into positive suggestions for improvement that the recipient can take in a positive way.

[1334] "User interface" is a general term for the screens and operating methods that users use to interact with a system.

[1335] "Reliability" is a numerical index that indicates the accuracy and certainty of the judgment result.

[1336] A "threshold" is a value that sets a certain reference value and serves as a boundary for executing a specific action or process.

[1337] A "flag" is a mark or indicator that is set for identification or processing based on a specific condition.

[1338] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[1339] Server processing

[1340] The server loads an artificial intelligence model for detecting anti-comments during system initialization. This model is loaded from a specific location within the network using a machine learning framework such as TensorFlow or PyTorch. The loaded artificial intelligence model is used to analyze and judge anti-comments.

[1341] The server then retrieves all comments from a database (e.g., MySQL or PostgreSQL) and analyzes this data. Each comment is fed into an artificial intelligence model, which determines whether it is an anti-comment or not. The result of this determination includes a probability value (confidence) that the comment is an anti-comment.

[1342] Based on the results, the server converts the anti-comment into a constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT) that replaces offensive language with polite language. This conversion process is performed using prompts such as:

[1343] Please convert the offensive comments below into constructive feedback:

[1344] "This video of yours is awful. It's uninspiring and not worth watching."

[1345] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is automatically flagged for review.

[1346] Processing by the terminal

[1347] The device fetches the new converted comments from the server and displays them in the user's browser or application's user interface, using front-end frameworks such as React or Vue.js to present the converted comments to the user in a visually understandable way.

[1348] User Action

[1349] Users can enter comments through a form on a website or application. These comments are immediately stored in a database and analyzed and converted by the server. After processing, users can receive comments from other users through their own devices and receive constructive feedback. This helps users reduce psychological stress and promote healthy communication.

[1350] The above is an embodiment of the present invention, and this system can automatically convert anti-comments into constructive feedback, thereby realizing positive communication between users.

[1351] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1352] Step 1: Loading the AI ​​model

[1353] When the system is initialized, the server loads an artificial intelligence model for detecting anti-comments. The input is the file path of the AI ​​model stored in a specific location within the network. Specifically, the model file is read into memory using TensorFlow or PyTorch. The output of this process is an AI model capable of detecting anti-comments loaded into memory.

[1354] Step 2: Obtaining and parsing comments

[1355] The server retrieves all comments from a database. The database uses MySQL or PostgreSQL, and executes SQL queries to extract comment data. The input is the access information to the database where the comments are stored. The retrieved comments are then input into an AI model to determine whether they are anti-comments. Specifically, each comment is input into the AI ​​model, and a confidence score is obtained as the output. The higher the confidence score, the more likely the comment is to be an anti-comment.

[1356] Step 3: Convert comments

[1357] The server converts anti-comments into constructive advice format if the confidence level exceeds a certain threshold (e.g., 0.7). This conversion is performed using a generative AI model (e.g., GPT-3 or ChatGPT). The anti-comments and their confidence levels are used as input. Specifically, the server inputs a prompt sentence into the generative AI model to obtain constructive feedback. The following sentence is used as an example of this prompt sentence:

[1358] Please convert the offensive comments below into constructive feedback:

[1359] "This video of yours is awful. It's uninspiring and not worth watching."

[1360] The output is that offensive comments are transformed into constructive feedback.

[1361] Step 4: Update the database

[1362] The server replaces the original comment with the converted comment and stores it in the database. The inputs are the IDs of the converted comment and its original comment. Specifically, it executes an SQL query using an UPDATE statement to replace the original comment with the new comment. If the confidence does not exceed a threshold, the comment is flagged for review.

[1363] Step 5: Fetch comments

[1364] The device fetches the latest comments from the server. The server's API endpoint URL is used as input. Specifically, it sends an HTTP request to the server and receives data in JSON format from the server. The output of this process is the latest comment data retrieved by the device.

[1365] Step 6: View comments

[1366] The device's user interface displays the fetched comments. It uses the retrieved comment data as input, renders a component that visualizes the comment data using a front-end framework such as React or Vue.js, and displays the latest constructive feedback to the user as output.

[1367] Step 7: Enter a comment

[1368] Users enter comments through a form on a website or application. The actual comment text entered by the user is used as input. Specifically, the form data is sent to the server's API, which receives it and stores it in a database. The output of this process is that the new comment data is added to the database.

[1369] Step 8: Receiving comments

[1370] The user views the comments retrieved from the server via the device. The latest comment data retrieved by the device is used as input. The specific operation is to display the comments using a front-end framework. As output, the user can view constructive feedback.

[1371] (Application example 1)

[1372] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1373] On current content distribution platforms, user comments often contain offensive content, causing psychological stress for content providers and other users. Ignoring such anti-comments can hinder healthy communication and potentially lower the quality of the platform as a whole. Therefore, the present invention aims to provide a system that automatically detects offensive anti-comments and converts them into constructive advice, thereby reducing the psychological burden on users and promoting healthy communication.

[1374] 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.

[1375] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice, means for determining whether a comment entered by a user is offensive, means for converting the offensive comment into constructive advice using a generative AI model, means for replacing the original comment in the database with the converted comment, and means for displaying the replaced comment. This allows negative comments from users to be instantly converted into positive feedback, enabling healthy communication on the content distribution platform.

[1376] "Anti-comments" are comments made by users that contain offensive, negative or harsh content about specific content or other users.

[1377] "Constructive advice" is a form of positive feedback that is beneficial to the recipient and includes specific improvements or suggestions.

[1378] A "generative AI model" is an artificial intelligence model that is trained on a large dataset and has the ability to analyze and convert input text data.

[1379] A "database" is an electronic data storage system that has a structure and allows efficient management, storage, searching, and retrieval of data.

[1380] An "offensive word list" is a collection of words that includes specific terms and phrases that, when used, may create a negative or offensive impression on others.

[1381] "Confidence" is a probability value that indicates whether a particular output (for example, a judgment of anti-comments) is correct based on the data input to the AI ​​model.

[1382] A "prompt" is the initial input data given to an AI model to perform a specific task.

[1383] A "content distribution service" is a service that provides digital content such as video, audio, and text to users via the Internet.

[1384] A "flag" is an identifier that is assigned to a data item or a processing step based on a specific condition.

[1385] The system for implementing the present invention is composed of a server, a terminal, and a user. The specific configuration and operation method of the system will be described below.

[1386] Server processing

[1387] Loading an AI model

[1388] When the system is initialized, the server loads a generative AI model for detecting anti-comments, which is loaded from a specific repository on the Internet.

[1389] Comment parsing and transformation

[1390] The server analyzes all comments retrieved from the database, checking them against a list of offensive words to detect anti-comments, which are then fed into a generative AI model and transformed into constructive advice using prompts.

[1391] Examples of prompts:

[1392] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[1393] Update the database with the conversion results

[1394] The server replaces the original comment with the converted comment and stores only those whose confidence exceeds a certain threshold in the database, or flags the comment for review if its confidence does not exceed the threshold.

[1395] Processing by the terminal

[1396] Fetching and displaying comments

[1397] The device retrieves the new converted comments from the server and displays them in the user's browser or app UI, allowing users to receive more constructive feedback.

[1398] User Action

[1399] Entering and receiving comments

[1400] Users can use the content distribution service's application to input and submit comments. The input comments are stored in a database, and users can also view constructive comments converted into feedback from other users.

[1401] Hardware and software used

[1402] Hardware: Smartphone (iOS, Android), Head-Mounted Display (HMD)

[1403] software:

[1404] Server-side AI model: OpenAI's generative AI model

[1405] Database: A relational database such as MySQL or PostgreSQL

[1406] Comment analysis and transformation: Scripting with Python

[1407] Explanation of specific steps in the process

[1408] Users can enter comments on videos.

[1409] The device will send the entered comment to the server.

[1410] The server receives the comments and checks them against a list of offensive words to detect them.

[1411] Using AI models, offensive comments are transformed using generative AI models.

[1412] The converted comment replaces the original comment in the database and is flagged if it has low confidence.

[1413] The terminal displays the converted comment to the user.

[1414] This invention instantly converts negative comments from users into positive feedback, promoting healthy communication on content distribution platforms.

[1415] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1416] Step 1:

[1417] The server loads the generative AI model during system initialization. Specifically, it downloads the model file from a specific repository on the internet and extracts it into memory. At this point, the generative AI model is ready for comment analysis and conversion.

[1418] Step 2:

[1419] A user inputs a comment for a video on a device. The input comment is temporarily stored in the device's memory and then sent to the server. The device then sends the comment data to the server as an HTTP request.

[1420] Step 3:

[1421] The server receives comments from clients and checks them against a list of offensive words to detect anti-comments. In this process, each word in the comment is checked sequentially to see if it is included in the offensive words list. If it is, the comment is flagged as an anti-comment.

[1422] Step 4:

[1423] The server inputs the detected anti-comments into the generative AI model, providing the model with a prompt like this:

[1424] Convert this offensive comment into constructive advice: "This video of yours is absolutely awful. It's uninspiring and not worth watching."

[1425] The generative AI model uses this prompt to convert offensive comments into constructive advice.

[1426] Step 5:

[1427] The server receives the converted comment and checks whether its confidence exceeds a certain threshold. If it does, it replaces the original comment with the converted comment and stores it in the database. If it does not, it flags the original comment for review. This process involves assigning a confidence score to the output of the generative AI model and determining whether it exceeds the threshold.

[1428] Step 6:

[1429] The device retrieves the new converted comments from the server. Specifically, the client application periodically sends a request to the server to retrieve the updated comment data. The retrieved data is stored in the device's memory.

[1430] Step 7:

[1431] The terminal displays the retrieved and converted comments to the user. The user interface reflects new comments in real time and allows the user to view them. During this process, the retrieved comment data is rendered into the UI component.

[1432] These steps will instantly transform negative comments from users into positive feedback, promoting healthy communication on the content distribution platform.

[1433] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1434] System Overview

[1435] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes user emotions. This system is mainly composed of a server, a terminal, and a user, each of which plays a specific role.

[1436] Server processing

[1437] 1. Loading the AI ​​model and emotion engine

[1438] When the server initializes the system, it loads the AI ​​model for detecting anti-comments and the emotion engine for recognizing user emotions, which prepares the system for anti-comment detection and emotion recognition.

[1439] 2. Comment Analysis

[1440] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[1441] 3. Emotional Recognition

[1442] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, which may include, for example, anger, sadness, and joy.

[1443] 4. Comment Conversion

[1444] If the reliability of the anti-comment exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted to encourage calmness.

[1445] 5. Updating the database

[1446] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1447] Processing by the terminal

[1448] 1. Getting comments

[1449] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[1450] 2. Viewing comments

[1451] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[1452] User Action

[1453] 1. Enter a comment

[1454] The user enters a comment, which is then saved in the database.

[1455] 2. Receiving Comments

[1456] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[1457] Specific examples

[1458] Original comment:

[1459] "This video of yours is awful. It's uninspiring and not worth watching."

[1460] User sentiment:

[1461] The AI ​​model detects this comment and recognizes that the user's emotion is "anger."

[1462] Post-processing comments:

[1463] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[1464] In this way, the present invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[1465] The processing flow will be explained below.

[1466] Step 1:

[1467] The server loads the AI ​​model and emotion engine.

[1468] Operation: When the server initializes the system, it loads the AI ​​model and emotion engine from the specified path and initializes each model. This prepares the system for anti-comment detection and emotion recognition.

[1469] Step 2:

[1470] The device connects to the comment database and fetches all comments.

[1471] How it works: The device queries the comment database, retrieves all comments in list form, and passes them to the server.

[1472] Step 3:

[1473] The server parses each comment

[1474] How it works: The server reads the obtained comment list one by one and inputs it into the AI ​​model for analysis. The AI ​​model analyzes the comment, determines whether it is an anti-comment, and returns a result including a confidence score.

[1475] Step 4:

[1476] The server recognizes the user's emotions

[1477] How it works: The server uses the emotion engine to analyze the emotional state of the user who posted the comment, and obtains emotional data such as anger, sadness, and joy as the analysis result.

[1478] Step 5:

[1479] The server checks the credibility of anti-comments

[1480] How it works: The server uses the AI ​​model's analysis results to determine whether the confidence level exceeds a certain threshold (e.g., 75%). If it does, it proceeds to the next step; if it does not, it flags it for review.

[1481] Step 6:

[1482] The server converts hate comments into constructive advice

[1483] How it works: The server converts anti-comments with a confidence level above a threshold into constructive advice based on the analysis results of the emotion engine. For example, if anger is detected, the advice includes encouraging people to stay calm.

[1484] Step 7:

[1485] The server updates the database with the converted comments

[1486] What it does: The server uses the original comment ID to replace the comment in the database with the new constructive comment, which updates the comment displayed in the user interface.

[1487] Step 8:

[1488] The terminal fetches the converted comment from the database.

[1489] Operation: The device retrieves updated comments from the database and reflects the latest comment information in the user's UI.

[1490] Step 9:

[1491] The device displays the updated comment to the user.

[1492] What it does: Updates the device UI to show the converted constructive comment to the user, ensuring they receive constructive and sensitive feedback instead of offensive comments.

[1493] Example 2

[1494] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1495] In today's Internet, users can freely post comments to each other, leading to the proliferation of anti-comments and making healthy communication difficult. Anti-comments, in particular, can increase the psychological burden on the recipient and potentially worsen the overall communication environment. While detecting anti-comments and converting them into constructive advice is important in this situation, there is also a need for more effective communication support that takes into account the user's emotions.

[1496] 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.

[1497] In this invention, the server includes means for automatically detecting anti-comments, means for converting the detected anti-comments into constructive advice format, and means for taking user emotion data into consideration when converting. This makes it possible to convert anti-comments into constructive advice format while taking user emotion into consideration, thereby promoting a healthy communication environment.

[1498] "Anti-comments" are comments that contain negative content such as attacks, insults, or criticism of others.

[1499] "Constructive advice format" refers to comments that point out problems and include specific suggestions and advice for improving them.

[1500] "User emotion data" is data that indicates the user's emotional state at the time of posting a comment, and typically includes emotions such as anger, sadness, and joy.

[1501] "Confidence" is a score that indicates the degree of certainty with which the AI ​​model detects a comment as an anti-comment.

[1502] A "threshold" is a boundary value of the reliability score for performing a specific process or judgment.

[1503] A "database" is a system for systematically storing and managing information such as comments, their conversion results, and user emotional data.

[1504] A "flag" is a mark or indication that is set for a particular condition, in this case indicating that a review is required.

[1505] The "server" is a central computer system that performs comment analysis, emotion recognition, conversion processing, etc.

[1506] A "terminal" is a device that allows a user to input comments and view the displayed comments.

[1507] The present invention is a system that automatically detects anti-comments and converts them into constructive advice. This system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[1508] Server processing

[1509] The server is the center of the system and performs processing in the following procedure.

[1510] 1. Loading the AI ​​model and emotion engine

[1511] The server loads an AI model (e.g., generative AI model) for detecting anti-comments and an emotion engine (e.g., emotion analysis software) for recognizing user emotions during system initialization, which prepares the system for anti-comment detection and emotion recognition.

[1512] 2. Comment Analysis

[1513] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and then determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[1514] 3. Emotional Recognition

[1515] The emotion engine recognizes the user's emotion when posting a comment and acquires emotion data, such as anger, sadness, and joy.

[1516] 4. Comment Conversion

[1517] If the confidence level exceeds a certain threshold, the server converts the anti-comment into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine.

[1518] 5. Updating the database

[1519] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1520] Processing by the terminal

[1521] The terminal functions as an interface with the user and performs processing in the following procedure.

[1522] 1. Getting comments

[1523] The device retrieves the updated comments from the server. The latest comment information is retrieved from the server via an HTTP request.

[1524] 2. Viewing comments

[1525] The user interface of the terminal displays the updated comments to the user in a visually easy to understand format.

[1526] User Action

[1527] The user is the entity that directly interacts with the system and performs the following process.

[1528] 1. Enter a comment

[1529] Users enter and post comments. The posted comments are saved in a database and sent to the server.

[1530] 2. Receiving Comments

[1531] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[1532] Specific examples

[1533] For example, if the original comment was, "This video of yours is really terrible. It's unmotivating and not worth watching," the AI ​​model and emotion engine would analyze this comment and recognize that the user's emotion was "anger."

[1534] An example prompt that takes this into account is:

[1535] "Your video is terrible. It's uninspiring and not worth watching." The emotion in this comment is anger. Convert this comment into constructive advice that encourages calm.

[1536] The resulting comment after processing looks like this:

[1537] "There may be some room for improvement in your video. For example, if you improve the sound quality or image quality a little more, it may be easier for viewers to watch. I recommend you think about it calmly."

[1538] In this way, the present invention not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[1539] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1540] Program processing flow

[1541] Step 1:

[1542] The server loads the AI ​​model and emotion engine during system initialization.

[1543] Specific behavior:

[1544] The server uses the APIs and authentication information of the generative AI model (e.g., GPT-4) and sentiment analysis engine (e.g., general sentiment analysis software) to launch these models. As input, it obtains the necessary information from the configuration file and environment variables set when the server is started, and as output, it obtains the state when the model has been loaded.

[1545] Step 2:

[1546] The server retrieves all comments from the database.

[1547] Specific behavior:

[1548] The server retrieves all comments from the comment database using an SQL query, etc. Database connection information and a query are required as input, and a text data list of comments is obtained as output.

[1549] Step 3:

[1550] Each comment received by the server is input into an AI model to determine whether it is an anti-comment.

[1551] Specific behavior:

[1552] The server inputs each comment as a prompt into the generative AI model and parses the model's response. The input is the comment text, and the output is whether the comment is an anti-comment or not, along with its confidence score.

[1553] Step 4:

[1554] The server uses an emotion engine to recognize the user's emotions when posting a comment.

[1555] Specific behavior:

[1556] The server sends each comment along with the user's emotional data to the emotion engine and receives the analysis results from the emotion engine. The input is the comment text and the emotion analysis request, and the output is the emotional data (e.g., anger, sadness, joy, etc.).

[1557] Step 5:

[1558] The server converts the anti-comments into constructive advice.

[1559] Specific behavior:

[1560] The server generates a prompt sentence for the AI ​​model and converts the anti-comment. At this time, the prompt also includes the user's emotional data. The input is the anti-comment and the prompt sentence including the emotional data, and the output is the comment converted into a constructive advice format.

[1561] Step 6:

[1562] The server replaces the converted comment with the original comment in the database.

[1563] Specific behavior:

[1564] The server replaces the original comment with the transformed comment using an SQL update query, taking as input the comment ID and the transformed comment text, and obtaining the updated database result as output.

[1565] Step 7:

[1566] Comments that do not exceed a confidence threshold are flagged for review.

[1567] Specific behavior:

[1568] If the reliability does not exceed the threshold, the server sets a review-needed flag for the comment in the database. The input is the comment ID and the review-needed flag, and the output is the updated database result.

[1569] Step 8:

[1570] The terminal obtains the updated comments from the server and reflects them in the user interface.

[1571] Specific behavior:

[1572] The terminal obtains the updated comments from the server via an HTTP request and displays them on the user interface. The input is the HTTP request and the comment data as a response, and the output is the comment display on the user interface.

[1573] Step 9:

[1574] The user enters a comment, which is then saved in the database.

[1575] Specific behavior:

[1576] The user enters a comment in the comment input form and submits it, and the server saves it in the database. The input is the comment text entered by the user and a submission request, and the output is the result of saving the new comment in the database.

[1577] Step 10:

[1578] A user views comments from other users via a terminal.

[1579] Specific behavior:

[1580] The user views the latest comment list displayed on the terminal screen and receives the converted constructive comments. The input is the comment data obtained from the server, and the output is the screen display that the user views.

[1581] (Application example 2)

[1582] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1583] In conventional systems, it is difficult for online content distribution services to respond appropriately to anti-comments among user comments, and the continued negative feedback can cause psychological stress. Furthermore, a simple negative comment detection system does not provide constructive feedback that takes into account the user's feelings, which means that healthy communication cannot be achieved.

[1584] 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.

[1585] In this invention, the server includes a means for automatically detecting anti-comments, a means for converting anti-comments into constructive advice using an AI model, and a means for recognizing the user's emotions and reflecting them in the converted comments. This allows the anti-comments to be converted into constructive feedback while taking the user's emotions into consideration, thereby promoting healthy communication while reducing psychological burden.

[1586] "Anti-comments" are comments that contain negative or offensive content about online content or individuals.

[1587] An "AI model" is a model for performing data analysis or predictive batch processing using artificial intelligence.

[1588] "Constructive advice" is feedback that contains negative content but is reframed as a positive suggestion for improvement.

[1589] "User emotions" is data obtained by analyzing the user's psychological state at the time of posting a comment using an emotion engine.

[1590] An "emotion engine" is an algorithm or software for identifying emotions from a user's text data.

[1591] "Confidence" is a score that indicates the degree of certainty of the analysis results that the AI ​​model assigns when detecting anti-comments.

[1592] The "specific threshold" is a reference value for determining whether or not a comment judgment is converted into constructive advice.

[1593] "Database" means an information management system for storing and managing comments and other data.

[1594] "Flag for review" is a marking of anti-comments that do not exceed a certain threshold, indicating that further review or human verification is required.

[1595] System Overview

[1596] This invention combines a system that automatically detects anti-comments and converts them into constructive advice with an emotion engine that recognizes users' emotions. The system is mainly composed of a server, terminals, and users, each of which plays a specific role.

[1597] Server processing

[1598] The server is the central element of the system and performs the following functions:

[1599] 1. Loading the AI ​​model and emotion engine:

[1600] When the server initializes the system, it loads an AI model for detecting anti-comments and an emotion engine for recognizing user emotions. This prepares the system for anti-comment detection and emotion recognition. The Python transformers library can be used as the AI ​​model.

[1601] 2. Comment Analysis:

[1602] The server retrieves all comments from the database, inputs each comment into the AI ​​model for analysis, and the AI ​​model determines whether the comment is an anti-comment or not, returning a result including a confidence score.

[1603] 3. Emotion Recognition:

[1604] The emotion engine recognizes the user's emotion when posting a comment and retrieves the emotion data. Emotion data can include, for example, anger, sadness, and joy. The vaderSentiment library can be used for this.

[1605] 4. Comment conversion:

[1606] If the confidence level exceeds a certain threshold, the anti-comment is converted into a constructive advice format, taking into account the user's emotional data recognized by the emotion engine. For example, if the poster is angry, the advice is converted into a message encouraging them to stay calm.

[1607] 5. Database Update:

[1608] The converted comment replaces the original comment and is stored in the database. If the confidence level does not exceed a threshold, the comment is flagged for review.

[1609] Processing by the terminal

[1610] The device acts as a link between the server and the user and performs the following tasks:

[1611] 1. Getting comments:

[1612] The device retrieves the updated comments from the server and reflects the latest comment information on the user's UI.

[1613] 2. View comments:

[1614] The device UI will display the updated comments to the user, allowing them to receive constructive and sensitive feedback.

[1615] User Action

[1616] Users interact with the system in the following ways:

[1617] 1. Enter your comment:

[1618] The user enters a comment, which is then saved in the database.

[1619] 2. Receiving Comments:

[1620] Users can view comments from other users via their devices, and the converted comments are displayed in a constructive and emotionally sensitive manner, allowing users to receive positive feedback.

[1621] Specific examples

[1622] Original comment: "This video is so boring and a waste of time."

[1623] Recognized emotion: "Anger"

[1624] Converted comment: "There may be some room for improvement in this video. For example, if you put more thought into the content, it might be more interesting for viewers. I recommend you think about it calmly."

[1625] Prompt Sentence Examples

[1626] Type: "The sound in this video is so bad I can't hear you."

[1627] Prompt: "This comment is negative feedback. Please convert it into constructive feedback with suggestions for improvement."

[1628] In this way, this invention is a system that not only converts anti-comments into constructive advice, but also takes the user's feelings into consideration, thereby further reducing the user's psychological burden and promoting healthy communication.

[1629] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1630] Step 1:

[1631] The server loads the AI ​​model and emotion engine when the system is initialized. Specifically, it loads the AI ​​model for detecting anti-comments using the Python transformers library and initializes the vaderSentiment library for emotion analysis. This enables the system to detect anti-comments and recognize user emotions.

[1632] Input: None

[1633] Data processing / calculation: Initialization of AI model and emotion engine

[1634] Output: Initialized AI model and emotion engine

[1635] Step 2:

[1636] The server retrieves all comments from the database. Each comment is then input into an AI model. The AI ​​model determines whether the comment is an anti-comment and returns a result including its confidence level. Specifically, it uses a transformers model to classify the text.

[1637] Input: Database comment

[1638] Data processing / calculation: Comment analysis using AI models

[1639] Output: Anti-comment judgment results and confidence score

[1640] Step 3:

[1641] The server inputs the comment text into the emotion engine to recognize the user's emotion at the same time as the analyzed comment. The emotion engine analyzes the comment's emotional state (anger, joy, sadness, etc.) and returns the results. Specifically, emotion analysis is performed using the vaderSentiment library.

[1642] Input: Comment text

[1643] Data processing / calculation: Emotion analysis using emotion engine

[1644] Output: User emotion data

[1645] Step 4:

[1646] Only if the reliability exceeds a certain threshold will the server convert the anti-comment into constructive advice. The conversion process takes into account the user's emotions as recognized by the emotion engine. Specifically, if the user's emotion is "anger," the server converts the message into a message encouraging them to stay calm, and if the user's emotion is "sadness," the server converts the message into a message containing encouragement.

[1647] Input: Anti-comment judgment results, user emotion data

[1648] Data processing / calculation: Comment conversion

[1649] Output: Comments converted into constructive advice format

[1650] Step 5:

[1651] The server replaces the original comment with the converted comment and stores it in the database. The server then deletes the pre-conversion comment data and inserts the new comment data. This updates the comment information in the database.

[1652] Input: Translated comment

[1653] Data processing / calculation: Database update

[1654] Output: Updated comment data

[1655] Step 6:

[1656] The device retrieves the latest comment data from the server and reflects it on the user's UI. The device's UI then displays the converted comments to the user, allowing the user to receive constructive and emotionally sensitive feedback.

[1657] Input: Updated comment data

[1658] Data processing / calculation: Acquisition and display of comment data

[1659] Output: Constructive comments displayed in the user UI

[1660] Step 7:

[1661] Users can view comments from other users via their devices, and can also enter their own comments, which are then saved in the database, after which the system returns to the loop of analyzing and converting new comments.

[1662] Input: User comment

[1663] Data processing / calculation: Entering and saving comments

[1664] Output: New comment saved in the database

[1665] In this way, the system converts anti-comments into constructive feedback, promoting healthy communication while reducing the psychological burden on users.

[1666] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1667] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1668] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1669] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1670] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1671] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1672] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1673] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1674] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1675] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1676] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1677] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1678] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1679] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1680] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1681] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1682] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1683] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1684] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1685] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1686] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1687] The following is further disclosed regarding the above embodiment.

[1688] (Claim 1)

[1689] A means of automatically detecting anti-comments;

[1690] A means of converting detected anti-comments into constructive advice; and

[1691] a means for replacing the converted comment with the original comment in the database;

[1692] a means for displaying the replaced comment;

[1693] A system including:

[1694] (Claim 2)

[1695] Having a means to convert anti-comments into constructive advice only if their credibility exceeds a certain threshold;

[1696] 10. The system of claim 1.

[1697] (Claim 3)

[1698] A means to flag anti-comments for review that do not exceed a certain threshold;

[1699] 10. The system of claim 1.

[1700] "Example 1"

[1701] New Claims

[1702] (Claim 1)

[1703] means for loading an artificial intelligence model to automatically detect anti-comments;

[1704] a means for retrieving comments from the database and analyzing those comments;

[1705] A means for determining whether the analyzed comment is an anti-comment;

[1706] A means for converting anti-comments into constructive advice based on the judgment results;

[1707] a means for replacing the converted comment with the original comment in the database;

[1708] means for displaying the replaced comment in a user interface;

[1709] A system including:

[1710] (Claim 2)

[1711] Having a means to convert anti-comments into constructive advice only if their credibility exceeds a certain threshold;

[1712] 10. The system of claim 1.

[1713] (Claim 3)

[1714] A means to flag anti-comments for review that do not exceed a certain threshold;

[1715] 10. The system of claim 1.

[1716] "Application Example 1"

[1717] (Claim 1)

[1718] A means of automatically detecting anti-comments;

[1719] A means of converting detected anti-comments into constructive advice; and

[1720] means for determining whether a comment entered by a user is offensive;

[1721] A means to transform offensive comments into constructive advice using generative AI models; and

[1722] a means for replacing the converted comment with the original comment in the database;

[1723] a means for displaying the replaced comment;

[1724] A system including:

[1725] (Claim 2)

[1726] Having a means to convert anti-comments into constructive advice only if their credibility exceeds a certain threshold;

[1727] 10. The system of claim 1.

[1728] (Claim 3)

[1729] A means to flag anti-comments for review that do not exceed a certain threshold;

[1730] 10. The system of claim 1.

[1731] "Example 2: Combining Emotion Engines"

[1732] (Claim 1)

[1733] A means of automatically detecting anti-comments;

[1734] A means of converting detected anti-comments into constructive advice; and

[1735] means for taking into account user emotion data when converting;

[1736] a means for replacing the converted comment with the original comment in the database;

[1737] a means for displaying the replaced comment;

[1738] A system including:

[1739] (Claim 2)

[1740] Having a means to convert anti-comments into constructive advice only if their credibility exceeds a certain threshold;

[1741] 10. The system of claim 1.

[1742] (Claim 3)

[1743] A means to flag anti-comments for review that do not exceed a certain threshold;

[1744] 10. The system of claim 1.

[1745] "Application example 2 when combining emotion engines"

[1746] (Claim 1)

[1747] A means of automatically detecting anti-comments;

[1748] A method to convert anti-comments into constructive advice using an AI model;

[1749] A means for recognizing a user's emotions and reflecting them in the converted comments;

[1750] a means for replacing the converted comment with the original comment in the database;

[1751] a means for displaying the replaced comment;

[1752] A system including:

[1753] (Claim 2)

[1754] Having a means to convert anti-comments into constructive advice only if their credibility exceeds a certain threshold;

[1755] 10. The system of claim 1.

[1756] (Claim 3)

[1757] A means to flag anti-comments for review that do not exceed a certain threshold;

[1758] 10. The system of claim 1. [Explanation of symbols]

[1759] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of automatically detecting anti-comments; A means of converting detected anti-comments into constructive advice; and a means for replacing the converted comment with the original comment in the database; a means for displaying the replaced comment; A system including:

2. Having a means to convert anti-comments into constructive advice only if their credibility exceeds a certain threshold; The system of claim 1 .

3. A means to flag anti-comments for review that do not exceed a certain threshold; The system of claim 1 .

Images