system

Abstract

Provide a system. [Solution] A means for receiving comments sent from a user terminal; means for analyzing the received comments based on natural language processing techniques; A means for determining whether a comment is malicious based on the analysis result; A means for hiding comments determined to be malicious on a user terminal; means for transmitting a warning message to a user terminal when the message is determined to be malicious; a means for notifying an administrator terminal of comments determined to be malicious; A system including:

Description

[Technical Field]

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

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

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

[0004] In recent years, the comment sections of social media and video streaming platforms have become increasingly filled with violent and abusive language, placing an increasing mental burden on users. In particular, there have been reported cases where malicious comments have driven users to the brink of mental breakdown, leading to withdrawal from social activities and, in the worst cases, suicide. This situation not only undermines the culture of free communication on the Internet, but also poses a serious threat to human life. Therefore, there is a strong demand for a system that can automatically detect and appropriately address malicious comments. [Means for solving the problem]

[0005] The present invention provides a system that receives comments sent from user terminals and analyzes them based on natural language processing technology. Specifically, a server analyzes the received comments and determines whether the comments are malicious based on the results. Comments determined to be malicious are hidden on the user terminal and a warning message is sent to the user. Furthermore, if necessary, a notification is sent to an administrator terminal to prompt appropriate action. This system quickly detects malicious comments, enabling users and administrators to take appropriate action and maintaining the integrity of online communication.

[0006] A "user terminal" is a device that can access social media and video distribution platforms via the Internet and post and view comments.

[0007] A "server" is a computer system that receives comments from user terminals via a network, analyzes them using natural language processing technology, hides comments, sends warning messages, and notifies administrators.

[0008] A "comment" is a text message entered by a user on a social networking site or video distribution platform and posted in a format that can be viewed by other users.

[0009] "Natural language processing technology" is a technology that enables computers to understand, interpret, and manipulate human language, and is used to analyze the context and meaning of comments.

[0010] "Male comments" are comments that contain insulting, offensive, or harmful content to other users, and may have a negative psychological or social impact.

[0011] The "means for setting to not display" is a process for changing the settings so that the received comment is not displayed on the user terminal or to other users.

[0012] A "warning message" is a message that notifies and warns a user who posted a comment that has been determined to be malicious that the comment is inappropriate.

[0013] An "administrator device" is a device used by the operator of a social media or video distribution platform to receive notifications of malicious comments and take appropriate action. [Brief explanation of the drawings]

[0014] [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

[0015] 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.

[0016] First, the terms used in the following description will be explained.

[0017] 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).

[0018] 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.

[0019] 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.

[0020] 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.

[0021] 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."

[0022] [First embodiment]

[0023] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0024] 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.

[0025] 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).

[0026] 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.

[0027] 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.

[0028] 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.

[0029] 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.

[0030] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] 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.

[0032] 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.

[0033] 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.

[0034] 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."

[0035] The system based on this invention automatically analyzes and evaluates the content of comments and takes appropriate action against malicious comments in order to maintain the integrity of comment sections on social media and video distribution platforms. This system is composed of user terminals, a server, and an administrator terminal.

[0036] 1. Posting and Receiving Comments

[0037] When a user inputs a comment and presses the send button on the user terminal, the comment is sent to the server, and the server receives the comment.

[0038] 2. Comment Analysis

[0039] The server analyzes the content of the received comments using natural language processing technology. Specifically, it uses a natural language processing engine to analyze the context and meaning of the comments and determine whether the comments are malicious. This analysis takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[0040] 3. Identifying and Responding to Malicious Comments

[0041] The server determines whether a comment is malicious based on the analysis results obtained using natural language processing technology. For example, if a comment contains the expression "Go away," it will be judged as malicious if the expression is determined to be malicious based on the context.

[0042] 4. Hide comments

[0043] If the server determines that a comment is malicious, it will be instantly hidden from the user's device, preventing other users from viewing the malicious comment.

[0044] 5. Sending warning messages

[0045] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[0046] 6. Notification to Administrator

[0047] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[0048] example

[0049] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[0050] 1. The user terminal sends the comment to the server.

[0051] 2. The server receives the comments and analyzes them using natural language processing technology.

[0052] 3. The comment is determined to be malicious.

[0053] 4. The server hides the comment on the user's device and sends a warning message to User A.

[0054] 5. The server notifies the administrator terminal and reports details of the malicious comments.

[0055] In this way, the system based on the present invention can encourage users and administrators to take prompt and appropriate action, and provide a safe and sound online communication environment.

[0056] The processing flow will be explained below.

[0057] Step 1:

[0058] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[0059] Step 2:

[0060] Server: The server receives the comments sent from the user terminal.

[0061] Step 3:

[0062] Server: Passes received comments to a natural language processing engine, which analyzes the context and meaning of the comments.

[0063] Step 4:

[0064] Server: Based on the analysis results, the server determines whether the comment is malicious. For example, if the keyword "disappear" is included, the server evaluates whether the expression is malicious based on the context.

[0065] Step 5:

[0066] Server: If a comment is determined to be malicious, it will be hidden, preventing other users from viewing the malicious comment.

[0067] Step 6:

[0068] Server: Send a warning message to the user who posted the comment that was determined to be malicious. The message includes a detailed description of the inappropriateness of the comment.

[0069] Step 7:

[0070] Server: Details of malicious comments are sent to the administrator's terminal, allowing the administrator to understand the situation and take appropriate action.

[0071] Step 8:

[0072] Administrator device: Administrators will receive notifications and can delete comments or take additional action against users if necessary.

[0073] Example 1

[0074] 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."

[0075] Conventional comment management systems for social media and video streaming platforms manually determine whether a comment is malicious, making it difficult to respond quickly and ultimately making it difficult to maintain a healthy communication environment. Even automated systems sometimes fail to take appropriate action against comments judged to be malicious, potentially damaging the credibility of the entire platform. Furthermore, the process for notifying administrators of details of malicious comments is often insufficient, leaving administrators with insufficient information to take appropriate action.

[0076] 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.

[0077] In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments using natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, means for notifying an administrator terminal of comments determined to be malicious, and means for taking into account semantic analysis, keyword frequency, and context dependency in the natural language processing technology. This enables the content of comments to be quickly and accurately evaluated automatically and appropriate measures to be taken against malicious comments. Furthermore, detailed notifications to the administrator allow the administrator to properly understand the situation and take additional measures.

[0078] A "user device" is an electronic device used to access social media or video distribution platforms and input and send comments.

[0079] "Server" refers to a central system that receives comments sent by users, analyzes them, and takes appropriate action.

[0080] "Comments" are text messages that users enter and send on social media or video streaming platforms.

[0081] "Natural language processing technology" is technology used by computers to understand and analyze human language.

[0082] "Semantic analysis" is a method for understanding the meaning of text, and is a technology that analyzes the context of comments and the meaning of words.

[0083] "Keyword frequency" is a method of calculating how frequently a particular keyword appears in a piece of text.

[0084] "Context sensitivity" refers to the property of the meaning of a particular word or phrase changing depending on the context surrounding it.

[0085] A "malicious comment" is a comment entered by a user that contains offensive, discriminatory, or offensive content to other users.

[0086] A "warning message" is a message sent to a user who has posted a comment that is deemed malicious, informing them of a violation of the terms of service or inappropriate behavior.

[0087] An "administrator device" is an electronic device used by an administrator who manages and operates a social networking site or video distribution platform.

[0088] "Notification" is a means of informing interested parties in real time about a particular event or situation.

[0089] The present invention provides a system for maintaining the integrity of comment sections on social networking sites and video distribution platforms. The system includes a user terminal, a server, and an administrator terminal.

[0090] First, the user enters a comment on the user terminal and presses the send button to send the comment to the server. This transmission is done over the Internet. The server receives the comment sent from the user terminal.

[0091] The server then analyzes the received comments. This process uses natural language processing technology. For example, semantic analysis is performed to understand the context of the comment, taking into account the frequency of keyword occurrences and context-dependency. This is done using software such as Google® Cloud Natural Language API. Based on the results of this analysis, the server determines whether the comment is malicious. For example, a comment containing the phrase "Go away" is deemed malicious based on the context.

[0092] If the server determines that a comment is malicious based on the analysis results, it will hide the comment on the user's device. The server sends a command to hide the comment to the user's device, and the front-end application reflects this and hides the comment.

[0093] Furthermore, the server sends a warning message to users who post malicious comments. The warning message contains information about the inappropriateness of the comment and the penalties for repeated violations, thereby alerting the user.

[0094] The server also notifies the administrator terminal of details of malicious comments. This notification includes the content of the comment, information about the poster, and analysis results, allowing the administrator to take appropriate action based on this information. Notifications are sent in real time via email or a dedicated alert system.

[0095] Specific examples

[0096] For example, if user A posts a comment on a social networking site saying "I wish you'd just disappear," the following happens:

[0097] 1. The user terminal sends a comment to the server.

[0098] 2. The server receives the comments and parses them using the Google Cloud Natural Language API.

[0099] 3. The comment is deemed to contain the malicious phrase "Go away."

[0100] 4. The server instructs the user to hide the comment on the user's device and sends a warning message to User A.

[0101] 5. The server notifies the administrator terminal of details of the malicious comments.

[0102] By implementing this system, we can maintain a safe and healthy online communication environment.

[0103] Example prompts for generative AI models

[0104] "Please explain the program that analyzes comments posted on social media using natural language processing technology and automatically detects malicious comments. Please also include specific steps and the name of the technology used."

[0105] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0106] Step 1: Posting and receiving comments

[0107] When a user enters a comment and presses the send button on the user terminal, the comment data is sent to the server. The input is the text data entered by the user, and the output is a message to the server indicating that transmission is complete. The sent comment data is temporarily stored in a database on the server.

[0108] Step 2: Initial processing of comments

[0109] The server temporarily stores the received comment data in a database and then preprocesses it. The input is the comment data sent by the user, and the output is preprocessed text data. Specifically, it deletes unnecessary spaces and special characters and normalizes characters.

[0110] Step 3: Natural Language Processing Analysis of Comments

[0111] The server sends the preprocessed comment data to a natural language processing engine (e.g., Google Cloud Natural Language API) and obtains the analysis results. The input is the preprocessed text data, and the output is the analysis results (e.g., semantic analysis results and keyword frequency). Specifically, the comments are first tokenized, and then the meaning of each token is analyzed.

[0112] Step 4: Identifying malicious comments

[0113] The server determines whether a comment is malicious based on the analysis results from the natural language processing engine. The input is the analysis results obtained from the natural language processing engine, and the output is the maliciousness score of the comment. Specifically, it compares the score with a set reference value, and if the score exceeds the threshold, it is determined to be malicious.

[0114] Step 5: Hide comments

[0115] If the server judges a comment to be malicious, it issues an instruction to hide the comment on the user's device. The input is the comment data judged to be malicious and the judgment result, and the output is a command to hide the comment on the user's device. Specifically, the server sends a list of hidden comments to the front-end application.

[0116] Step 6: Sending a warning message

[0117] The server sends a warning message to the user who posted the malicious comment. The input is the data of the comment that was determined to be malicious and the user information, and the output is the warning message. Specifically, the server sends a data packet to the user's device to display the warning message as a pop-up.

[0118] Step 7: Notify administrators

[0119] The server notifies the administrator terminal of details of malicious comments. The input is the comment data determined to be malicious and the analysis results, and the output is a notification message to the administrator. Specifically, the server notifies the administrator in real time using email or a dedicated alert system. This allows the administrator to understand the situation and take appropriate action.

[0120] (Application example 1)

[0121] 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."

[0122] On modern social media and video streaming platforms, malicious comments can damage the health of the community. Especially when many comments are posted in real time, it is difficult to immediately detect and address malicious comments. Furthermore, if malicious comments are left unattended, they can cause discomfort and trouble for other users. Furthermore, manual monitoring and response requires too much time and effort, placing a significant burden on the system, so an efficient system is needed.

[0123] 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.

[0124] In this invention, the server includes: means for receiving comments sent from a user terminal; means for analyzing the received comments using natural language processing technology; means for determining whether the comments are malicious based on the analysis results; means for hiding comments determined to be malicious on the user terminal; means for sending a warning message to the user terminal when a comment is determined to be malicious; means for notifying an administrator terminal of details of the comments determined to be malicious; means for performing semantic analysis on comments acquired from the user terminal and hiding malicious comments in real time based on the results of the analysis; and means having a communication means for notifying the administrator terminal of details of the malicious comments. This automatically maintains the integrity of comments, allowing other users to enjoy safe and sound online communication.

[0125] A "user terminal" is a device used to input and send comments, and includes personal computers, smartphones, etc.

[0126] The "server" is a central computer system that receives comments sent from user terminals, analyzes them, and responds to them.

[0127] "Natural language processing technology" refers to a set of computational techniques and methodologies for analyzing human language and understanding its meaning.

[0128] "Semantic analysis" is a part of natural language processing techniques that allows for understanding the meaning and context of text.

[0129] "Real-time" refers to near-instant processing, meaning immediate responses to user-submitted comments.

[0130] "Hateful comments" are comments that contain offensive, inappropriate, or harmful content that offends other users.

[0131] A "warning message" is a notification sent to a user who posts a malicious comment, warning them that the content is inappropriate.

[0132] "Administrator terminal" refers to a device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[0133] A description will now be given of a form for realizing a system based on the present invention. This system includes a user terminal, a server, and an administrator terminal.

[0134] System Overview

[0135] The system of the present invention automatically analyzes and evaluates the content of comments to maintain the integrity of comment sections on social media and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of user terminals, a server, and an administrator terminal.

[0136] Hardware and software used

[0137] User terminal: A device used to input and send comments, such as a personal computer or smartphone.

[0138] Server: A central computer system that receives comments sent from user terminals, analyzes them using natural language processing technology, and responds accordingly.

[0139] Administrator device: A device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[0140] Software and Technology

[0141] Natural Language Processing (NLP): This technology uses the transformers library to detect malicious comments, specifically leveraging generative AI models to perform semantic analysis.

[0142] SNS API integration: Provides a means of communication using the requests library to retrieve, analyze, and hide comments from SNS APIs.

[0143] Sending warning messages: Using Apple Push Notification Service (APNs), we send warning messages to users who post inappropriate comments.

[0144] Admin Notification: Admin notifications are done with a simple POST request.

[0145] Specific examples of processing

[0146] For example, if a user comments "This app is the worst, get rid of it!", the system works like this:

[0147] 1. The user enters a comment on the terminal and sends it to the server.

[0148] 2. The server receives the comments and analyzes them using natural language processing technology.

[0149] 3. The server determines that the comment is malicious. Specifically, semantic analysis detects that the comment is inappropriate.

[0150] 4. The server sets the comment to be hidden on the user's device, generates a warning message based on the analysis results, and sends the warning message to the user.

[0151] 5. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes action as necessary.

[0152] An example of using a generative AI model to set prompts is:

[0153] Prompt: Please rate the following comment as malicious: "Go away."

[0154] In this way, the system of the present invention can automatically detect malicious comments in real time and take appropriate measures to maintain the health of the comment section, providing an effective solution for enhancing the safety and health of online communication.

[0155] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0156] Step 1:

[0157] The user terminal inputs and transmits a comment.

[0158] Input: Comment entered by the user (e.g. "I wish you'd just disappear")

[0159] Output: Request data with comments sent to the server

[0160] How it works: When a user enters a comment and presses the submit button, the comment data is sent to the server as a request.

[0161] Step 2:

[0162] The server receives the comment.

[0163] Input: Comment data sent from the user's device

[0164] Output: Received comment data

[0165] Operation: The server receives a request to send a comment from a user's device and stores the comment data in a buffer for analysis.

[0166] Step 3:

[0167] The server analyzes the comments using natural language processing technology.

[0168] Input: Received comment data

[0169] Output: Comment analysis results (e.g., maliciousness score, context information)

[0170] How it works: A natural language processing engine on the server analyzes the meaning of the comment and performs a semantic analysis of the comment. A generative AI model is used to calculate a severity score based on the prompt.

[0171] Step 4:

[0172] The server determines whether the comment is malicious.

[0173] Input: Comment analysis results

[0174] Output: Malicious or not (e.g., malicious)

[0175] How it works: The server determines a comment as malicious based on the generated maliciousness score if the score exceeds a certain threshold.

[0176] Step 5:

[0177] Comments that the server determines to be malicious are set to be hidden on the user's device.

[0178] Input: Comment data judged to be malicious, and judgment result

[0179] Output: Updated comment display state (not displayed)

[0180] How it works: The server immediately sends data to the user's device to hide any comments that are determined to be malicious, and the comments are hidden.

[0181] Step 6:

[0182] If the server is determined to be malicious, a warning message is sent to the user terminal.

[0183] Input: Comment data judged to be malicious, and judgment result

[0184] Output: A warning message displayed on the user's terminal.

[0185] How it works: The server generates a warning message and uses the Apple Push Notification Service to send it to the device of the user who posted the comment.

[0186] Step 7:

[0187] The server notifies the administrator terminal of details of the malicious comments.

[0188] Input: Comment data judged to be malicious, and judgment result

[0189] Output: Notification data to the administrator's terminal

[0190] Operation: The server generates data to notify the administrator terminal of details of the comments that were determined to be malicious, and sends a POST request to the administrator's terminal.

[0191] In this way, the system of the present invention can maintain the integrity of the comment section, automatically detect malicious comments, and respond appropriately.

[0192] 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.

[0193] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[0194] 1. Posting and Receiving Comments

[0195] The user inputs a comment on the user terminal and presses the send button, which sends the comment to the server.

[0196] 2. Receipt and analysis of comments

[0197] The server receives comments sent from user devices. It then passes the received comments to a natural language processing engine, which analyzes the context and meaning of the words in the comments. Specifically, it takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[0198] 3. Determining whether a comment is malicious

[0199] The server uses natural language processing technology to determine whether a comment is malicious. For example, if a comment contains the phrase "disappear," it evaluates whether the phrase is malicious given the context.

[0200] 4. Emotion analysis using an emotion engine

[0201] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The emotion engine recognizes the user's emotions, and that emotional information is reflected in determining whether the comment is malicious. For example, comments that express positive emotions will have their score reduced, and comments that express negative emotions will have their score increased.

[0202] 5. Emotion-based correction

[0203] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. If a positive emotion is recognized, the malicious judgment score is reduced, and conversely, if a negative emotion is recognized, the score is increased. This correction makes the comment judgment more accurate.

[0204] 6. Hide comments

[0205] If the server determines that a comment is malicious, it will be hidden from the user's device, preventing other users from viewing the malicious comment.

[0206] 7. Sending warning messages

[0207] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[0208] 8. Notification to Administrator

[0209] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[0210] example

[0211] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[0212] 1. The user terminal sends the comment to the server.

[0213] 2. The server receives the comments and analyzes them using natural language processing technology.

[0214] 3. The comment is determined to be malicious.

[0215] 4. The server analyzes the user's emotional information along with the comments using an emotion engine.

[0216] 5. If the emotional information is negative, the malicious judgment score increases, and if it is positive, the score decreases.

[0217] 6. The server hides the comment on the user's device and sends a warning message to User A.

[0218] 7. The server notifies the administrator terminal and reports details of the malicious comments.

[0219] In this way, the system based on the present invention can recognize the user's emotions and reflect them in its malicious judgment, thereby encouraging more accurate and appropriate responses and providing a safe and healthy online communication environment.

[0220] The processing flow will be explained below.

[0221] Step 1:

[0222] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[0223] Step 2:

[0224] Server: The server receives the comments sent from the user terminal.

[0225] Step 3:

[0226] Server: Passes the received comment to a natural language processing engine, which analyzes the context and meaning of the comment. For example, it analyzes the sentence "Go away" and evaluates whether the expression contains malicious intent.

[0227] Step 4:

[0228] Server: Based on the analysis results, the server determines whether the comment is malicious or not, by checking whether it contains specific keywords or phrases.

[0229] Step 5:

[0230] Server: Analyzes the user's emotional information sent along with the comment using an emotion engine. The emotion engine analyzes data such as user input and voice, and classifies it as positive, negative, or neutral.

[0231] Step 6:

[0232] Server: Based on the emotional information, the judgement score of malicious comments is adjusted. For example, if the emotional sentiment is strong, the score is increased, and if the emotional sentiment is strong, the score is decreased.

[0233] Step 7:

[0234] Server: Comments that are judged to be malicious will be hidden, so other users will not be able to see them.

[0235] Step 8:

[0236] Server: Sends a warning message to the user who posted the comment that was determined to be malicious. The warning message states that the comment was inappropriate and explains why.

[0237] Step 9:

[0238] Server: Detailed information about malicious comments is sent to the administrator's terminal, allowing the administrator to understand the situation and take necessary measures.

[0239] Step 10:

[0240] Administrator terminal: Administrators will receive notifications and can take more severe measures in accordance with platform rules, such as completely deleting comments or freezing user accounts.

[0241] Example 2

[0242] 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."

[0243] Conventional social media and video streaming platforms lack sufficient detection and response to inappropriate comments in their comment sections, resulting in the rampant proliferation of malicious comments, which can disrupt user communication. In particular, systems that rely on simple keyword filtering or manual moderation without considering the context of the comment or the user's sentiment can have low accuracy in their judgments and result in incorrect responses. Furthermore, systems often fail to adequately warn users or provide administrators with information. This makes it difficult to maintain the integrity of online communication.

[0244] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for analyzing the user's emotional information along with the comments, means for correcting the maliciousness determination score based on the emotional information, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, and means for notifying an administrator terminal of a comment determined to be malicious. This enables highly accurate maliciousness determination based on context analysis of the comments and user's emotional information, and by sending appropriate warning messages and notifying the administrator, it becomes possible to effectively maintain the soundness of online communication.

[0245] A "user device" is a device used by a user to access social media or video distribution platforms and post comments.

[0246] "Server" refers to a central processing unit that receives and analyzes comments sent from user terminals and performs appropriate processing.

[0247] A "comment" is a text message that a user posts to other users on social media or video streaming platforms.

[0248] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[0249] "Emotion information" is data that indicates the emotional state of the user who posted the comment, and includes emotions such as positive and negative.

[0250] The "maliciousness score" is a quantitative index for evaluating whether a comment is malicious or not.

[0251] The "means of setting to hidden" is a method for making comments that have been determined to be malicious invisible to other users.

[0252] A "warning message" is a warning message sent to a user who posted a comment that was determined to be malicious.

[0253] "Administrator Terminal" means a device used by the Platform Administrator to monitor and manage User comments and system notifications.

[0254] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[0255] When a user enters a comment and presses the send button on the user terminal, the comment is sent to the server. Specifically, the user posts a comment using a device such as a smartphone or PC. This comment is sent to the server via the Internet.

[0256] The server receives comments sent from user devices and analyzes them based on natural language processing technology. The natural language processing engine used here is, for example, OpenAI's GPT-3 (registered trademark) or Google NLP. This analyzes the context of the comment, the frequency of keyword occurrences, and the degree of emotion. Furthermore, the server uses an emotion engine (e.g., IBM Watson (registered trademark) Tone Analyzer) to analyze the user's emotional information and evaluate whether abusive language or discriminatory remarks have been detected.

[0257] The server then determines whether the comment is malicious based on the analysis results. If the comment content is likely to be malicious, the server calculates a judgment score, and if the score exceeds a certain threshold, the comment is deemed malicious. The server then corrects the malicious judgment score based on the emotional information provided by the emotion engine. If the comment contains positive emotions, the score is reduced, and if it contains negative emotions, the score is increased.

[0258] Comments that are judged to be malicious are hidden from the user's terminal by the server, which prevents other users from viewing the comment. Furthermore, the server sends a warning message to the user who posted the malicious comment. This message includes information that the comment is inappropriate and warns the user to be careful.

[0259] The server also notifies the administrator terminal of details of malicious comments. This notification includes the comment content, poster information, and maliciousness score, allowing the administrator to understand the situation and take appropriate action. For example, the administrator can suspend or delete the user's account as necessary.

[0260] Specific examples

[0261] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following flow will be executed.

[0262] 1. The user enters a comment and presses the send button to send it to the server.

[0263] 2. The server receives the comments and analyzes them using natural language processing technology (e.g., GPT-3).

[0264] 3. The comment is flagged as malicious (e.g., because it contains the phrase "go away").

[0265] 4. The server analyzes the user's emotional information along with the comments using an emotion engine (e.g., IBM Watson Tone Analyzer). Negative emotional information is detected.

[0266] 5. The maliciousness score is adjusted (increased in maliciousness), and the final score exceeds the threshold, so the comment is judged to be malicious.

[0267] 6. The server hides the comment and sends a warning message to User A.

[0268] 7. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes appropriate measures.

[0269] This will enable the system to efficiently and effectively detect and address malicious comments to maintain the health of online communication.

[0270] Prompt Sentence Examples

[0271] "You should just disappear."

[0272] Sentiment score: Negative

[0273] Processing step: Based on the comment content and sentiment score, the comment is judged to be malicious and a warning message is sent if necessary.

[0274] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0275] Step 1:

[0276] The user enters a comment and presses the send button. The user enters text into the comment field on a social networking site or video streaming platform. As a result, the comment data is sent from the user's device to the server. Specifically, the comment entered by the user using a smartphone or computer is sent to the server as an HTTP POST request.

[0277] Step 2:

[0278] The server receives the comment. The input is the comment data sent from the user's device. The server receives this data and temporarily stores it in order to proceed to the next step. Specifically, the server's API endpoint receives the HTTP request and stores the comment content in the database.

[0279] Step 3:

[0280] The server analyzes the comments it receives using natural language processing technology. The input is the received comment data. This comment data is passed to a natural language processing engine, which analyzes the context of the text, the frequency of keyword occurrences, the degree of emotion, etc. The output is the analysis results, which include contextual information and an emotion score. Specifically, the server sends an API request to the natural language processing engine (e.g., SpaCy or Google NLP) and obtains the analysis results.

[0281] Step 4:

[0282] The server determines whether a comment is malicious based on the analysis results. The input is the analysis results of the natural language processing engine. The server calculates a judgment score based on this analysis result and evaluates whether it exceeds a threshold. The output is the result of the malicious judgment. Specifically, the server calculates the judgment score using a scoring algorithm that it has set up itself.

[0283] Step 5:

[0284] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the user's emotional information and the content of the comment. This is passed to the emotion engine, which analyzes the user's emotional state. The output is an analyzed emotion score. Specifically, the server sends an API request to an emotion engine such as IBM Watson Tone Analyzer and obtains the analysis results.

[0285] Step 6:

[0286] The server corrects the malicious judgment score based on the emotional information. The inputs are the malicious judgment score and the analysis results of the emotional engine. The server uses this emotional information to increase or decrease the malicious judgment score. The output is the corrected malicious judgment score. Specifically, it applies logic that decreases the score if the emotional score is positive, and increases the score if it is negative.

[0287] Step 7:

[0288] Comments that the server judges to be malicious are set to be hidden on the user's terminal. The input is the corrected maliciousness score. The server decides whether to hide the comment based on this score and sets a "hidden flag" for the comment in the database. The output is that the comment is hidden. Specifically, it updates a specific field in the database and enables the hidden setting.

[0289] Step 8:

[0290] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the result of the malicious judgment and user information. The server uses this information to generate and send a warning message to the user. The output is a warning message that is displayed on the user's device. Specifically, the warning message is generated using a notification service and sent according to the user's notification settings.

[0291] Step 9:

[0292] The server notifies the administrator device of comments that have been judged to be malicious. The input is detailed information about the malicious judgment (comment content, judgment score, user information, etc.). The server uses this information to generate and send a notification to the administrator device. The output is a detailed notification displayed on the administrator device. Specifically, it makes an API request to send a notification via email or to a dashboard dedicated to the administrator.

[0293] As described above, this system can efficiently detect and deal with malicious comments on social media and video distribution platforms through each processing step.

[0294] (Application example 2)

[0295] 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."

[0296] On online social media and video streaming platforms, inappropriate comments posted in the comment sections can cause unpleasant experiences for users. Furthermore, an increase in malicious comments can lead to a decline in the integrity of the platform as a whole, potentially resulting in a decline in the number of users. Furthermore, manual comment monitoring by administrators is labor-intensive and ineffective. To address these issues, a system is needed that can automatically analyze the content of comments and user sentiment, and quickly and accurately identify and hide malicious comments.

[0297] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for analyzing the user's emotional information sent together with the received comments using an emotion engine, means for correcting the judgment score of malicious comments based on the emotional information recognized by the emotion engine, means for determining whether the comments are malicious based on the analysis results, means for setting the user terminal to hide comments judged to be malicious, means for sending a warning message to the user terminal when a comment is judged to be malicious, and means for notifying an administrator terminal of a comment judged to be malicious. This makes it possible to appropriately judge and hide malicious comments by combining the comment content and user emotion analysis information.

[0298] A "user terminal" is an electronic device used to send comments, and includes smartphones, personal computers, tablets, etc.

[0299] A "server" is a computer system that receives, analyzes, and processes data sent from a user terminal, and cooperates with other terminals and systems as necessary.

[0300] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language, and includes semantic analysis of text and analysis of the frequency of keyword occurrences.

[0301] The "emotion engine" is a system that analyzes the emotional state of a user's comment from its context and determines whether the emotional information is positive or negative.

[0302] The "judgment score" is a numerical representation of the maliciousness and emotional state of a comment, and is used to evaluate the appropriateness of the comment.

[0303] "Bad comments" refer to comments that are deemed offensive, abusive, defamatory, or inappropriate, and that cause offense to other users.

[0304] The "hide setting" is a process of making a comment invisible to other users when the comment is judged to be malicious beyond a certain standard.

[0305] A "warning message" is a message that notifies a user who has posted an inappropriate comment that the comment is problematic and urges them to be careful.

[0306] An "administrator terminal" is a terminal used by a system administrator, which receives notifications about malicious comments and can take action as necessary.

[0307] "Comment context analysis" is a technology that understands the context in which a comment was posted and the text content before and after it, allowing for a more accurate understanding of the comment's intent and meaning.

[0308] To implement this invention, we will build a system for maintaining the integrity of comment sections on social media and video streaming platforms. The system mainly includes the following components:

[0309] 1. User terminal: An electronic device on which a user inputs comments and sends them to the server. This includes smartphones, PCs, tablets, etc.

[0310] 2. Server: A central computer system that receives comments, analyzes them using natural language processing technology and an emotion engine, and determines whether they are malicious or not. Specifically, the following software is used:

[0311] "Hugging Face Transformers" as a natural language processing engine

[0312] "Google Cloud Natural Language API" as a sentiment analysis engine

[0313] "MySQL (registered trademark)" for database management

[0314] "Firebase Cloud Messaging" for notification system

[0315] 3. Processing details

[0316] Receiving comments: The server receives the comments sent from the user terminal.

[0317] Natural language processing analysis: The received comments are passed to a natural language processing engine to analyze the content and context of the text.

[0318] Sentiment analysis: The emotional information of users sent along with their comments is analyzed using an emotion engine.

[0319] Correction of judgment score: The judgment score of malicious comments is corrected based on the emotional information recognized by the emotion engine.

[0320] Maliciousness determination: Based on the adjusted score, a final determination is made as to whether the comment is malicious.

[0321] Comment hiding setting: Comments that are judged to be malicious are set to be hidden on the user's device.

[0322] Sending warning messages: The server sends a warning message to users who post malicious comments.

[0323] Notify administrator: Details of malicious comments will be sent to the administrator's device.

[0324] For example, if a user posts a comment on a social media platform saying "Go away," the following steps will occur:

[0325] 1. The user terminal sends the comment to the server.

[0326] 2. The server receives the comment and analyzes its context and content using natural language processing techniques.

[0327] 3. The server analyzes the emotional information sent along with the comment using an emotion engine.

[0328] 4. If the emotional information is negative, the maliciousness score increases and the message is ultimately judged to be malicious.

[0329] 5. The server marks the comment as hidden and sends a warning message to the user.

[0330] 6. The server notifies the administrator terminal and reports details of the malicious comments.

[0331] Examples of prompts include:

[0332] "A user comments, 'This place is awful.' The system does sentiment analysis and determines that the comment is negative. What should we do next?"

[0333] "User commented 'Go away'. Hide and send a warning."

[0334] In this way, the system can maintain a high level of integrity in comments based on user sentiment analysis information.

[0335] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0336] Step 1:

[0337] The user terminal inputs and sends a comment. The input is performed by the user entering text in the input field and clicking the send button. The output is the user's comment data being sent to the server.

[0338] Step 2:

[0339] The server receives comments sent from user terminals. The input is the comment data sent by the user, and the output is the storage of the received comment data. Specifically, the comments are recorded in a database.

[0340] Step 3:

[0341] The server analyzes the received comments using natural language processing technology. The input is the received comment data, and the output is the analysis results (information about the meaning and context of the comment). The server analyzes the content and context of the text using a natural language processing engine (Hugging Face Transformers).

[0342] Step 4:

[0343] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the comment data and the user's emotional information, and the output is the emotional judgment result. Specifically, it uses the Google Cloud Natural Language API to determine whether the user's emotion is positive or negative.

[0344] Step 5:

[0345] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. The input is the analysis result of natural language processing and the emotion judgment result, and the output is the corrected malicious judgment score. The score decreases if the emotion is positive, and increases if the emotion is negative.

[0346] Step 6:

[0347] The server finally judges whether the comment is malicious or not based on the corrected score. The input is the corrected malicious judgment score, and the output is the judgment result of whether the comment is malicious or not. If it is judged to be malicious, proceed to the next step.

[0348] Step 7:

[0349] The server sets comments that are judged to be malicious to be hidden on the user's terminal. The input is the judgment result of malicious comments, and the output is that the comment will be set to hidden. Specifically, the "hidden" flag for the relevant comment is set in the database.

[0350] Step 8:

[0351] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the judgment result of the malicious comment, and the output is the warning message to be sent to the user. The warning notification is sent to the user's device using Firebase Cloud Messaging.

[0352] Step 9:

[0353] The server notifies the administrator terminal of comments that are judged to be malicious. The input is the judgment result of malicious comments, and the output is notification information sent to the administrator. Based on this information, the administrator can consider taking more severe measures.

[0354] 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.

[0355] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0356] 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.

[0357] [Second embodiment]

[0358] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0359] 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.

[0360] 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).

[0361] 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.

[0362] 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.

[0363] 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).

[0364] 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. 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.

[0365] 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.

[0366] 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.

[0367] 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.

[0368] 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.

[0369] 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."

[0370] The system based on this invention automatically analyzes and evaluates the content of comments and takes appropriate action against malicious comments in order to maintain the integrity of comment sections on social media and video distribution platforms. This system is composed of user terminals, a server, and an administrator terminal.

[0371] 1. Posting and Receiving Comments

[0372] When a user inputs a comment and presses the send button on the user terminal, the comment is sent to the server, and the server receives the comment.

[0373] 2. Comment Analysis

[0374] The server analyzes the content of the received comments using natural language processing technology. Specifically, it uses a natural language processing engine to analyze the context and meaning of the comments and determine whether the comments are malicious. This analysis takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[0375] 3. Identifying and Responding to Malicious Comments

[0376] The server determines whether a comment is malicious based on the analysis results obtained using natural language processing technology. For example, if a comment contains the expression "Go away," it will be judged as malicious if the expression is determined to be malicious based on the context.

[0377] 4. Hide comments

[0378] If the server determines that a comment is malicious, it will be instantly hidden from the user's device, preventing other users from viewing the malicious comment.

[0379] 5. Sending warning messages

[0380] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[0381] 6. Notification to Administrator

[0382] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[0383] example

[0384] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[0385] 1. The user terminal sends the comment to the server.

[0386] 2. The server receives the comments and analyzes them using natural language processing technology.

[0387] 3. The comment is determined to be malicious.

[0388] 4. The server hides the comment on the user's device and sends a warning message to User A.

[0389] 5. The server notifies the administrator terminal and reports details of the malicious comments.

[0390] In this way, the system based on the present invention can encourage users and administrators to take prompt and appropriate action, and provide a safe and sound online communication environment.

[0391] The processing flow will be explained below.

[0392] Step 1:

[0393] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[0394] Step 2:

[0395] Server: The server receives the comments sent from the user terminal.

[0396] Step 3:

[0397] Server: Passes received comments to a natural language processing engine, which analyzes the context and meaning of the comments.

[0398] Step 4:

[0399] Server: Based on the analysis results, the server determines whether the comment is malicious. For example, if the keyword "disappear" is included, the server evaluates whether the expression is malicious based on the context.

[0400] Step 5:

[0401] Server: If a comment is determined to be malicious, it will be hidden, preventing other users from viewing the malicious comment.

[0402] Step 6:

[0403] Server: Send a warning message to the user who posted the comment that was determined to be malicious. The message includes a detailed description of the inappropriateness of the comment.

[0404] Step 7:

[0405] Server: Details of malicious comments are sent to the administrator's terminal, allowing the administrator to understand the situation and take appropriate action.

[0406] Step 8:

[0407] Administrator device: Administrators will receive notifications and can delete comments or take additional action against users if necessary.

[0408] Example 1

[0409] 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."

[0410] Conventional comment management systems for social media and video streaming platforms manually determine whether a comment is malicious, making it difficult to respond quickly and ultimately making it difficult to maintain a healthy communication environment. Even automated systems sometimes fail to take appropriate action against comments judged to be malicious, potentially damaging the credibility of the entire platform. Furthermore, the process for notifying administrators of details of malicious comments is often insufficient, leaving administrators with insufficient information to take appropriate action.

[0411] 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.

[0412] In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments using natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, means for notifying an administrator terminal of comments determined to be malicious, and means for taking into account semantic analysis, keyword frequency, and context dependency in the natural language processing technology. This enables the content of comments to be quickly and accurately evaluated automatically and appropriate measures to be taken against malicious comments. Furthermore, detailed notifications to the administrator allow the administrator to properly understand the situation and take additional measures.

[0413] A "user device" is an electronic device used to access social media or video distribution platforms and input and send comments.

[0414] "Server" refers to a central system that receives comments sent by users, analyzes them, and takes appropriate action.

[0415] "Comments" are text messages that users enter and send on social media or video streaming platforms.

[0416] "Natural language processing technology" is technology used by computers to understand and analyze human language.

[0417] "Semantic analysis" is a method for understanding the meaning of text, and is a technology that analyzes the context of comments and the meaning of words.

[0418] "Keyword frequency" is a method of calculating how frequently a particular keyword appears in a piece of text.

[0419] "Context sensitivity" refers to the property of the meaning of a particular word or phrase changing depending on the context surrounding it.

[0420] A "malicious comment" is a comment entered by a user that contains offensive, discriminatory, or offensive content to other users.

[0421] A "warning message" is a message sent to a user who has posted a comment that is deemed malicious, informing them of a violation of the terms of service or inappropriate behavior.

[0422] An "administrator device" is an electronic device used by an administrator who manages and operates a social networking site or video distribution platform.

[0423] "Notification" is a means of informing interested parties in real time about a particular event or situation.

[0424] The present invention provides a system for maintaining the integrity of comment sections on social networking sites and video distribution platforms. The system includes a user terminal, a server, and an administrator terminal.

[0425] First, the user enters a comment on the user terminal and presses the send button to send the comment to the server. This transmission is done over the Internet. The server receives the comment sent from the user terminal.

[0426] The server then analyzes the received comments. This process uses natural language processing technology. For example, semantic analysis is performed to understand the context of the comment, taking into account the frequency of keyword occurrences and context-dependency. This is done using software such as Google Cloud Natural Language API. Based on the results of this analysis, the server determines whether the comment is malicious. For example, a comment containing the phrase "Go away" is deemed malicious based on the context.

[0427] If the server determines that a comment is malicious based on the analysis results, it will hide the comment on the user's device. The server sends a command to hide the comment to the user's device, and the front-end application reflects this and hides the comment.

[0428] Furthermore, the server sends a warning message to users who post malicious comments. The warning message contains information about the inappropriateness of the comment and the penalties for repeated violations, thereby alerting the user.

[0429] The server also notifies the administrator terminal of details of malicious comments. This notification includes the content of the comment, information about the poster, and analysis results, allowing the administrator to take appropriate action based on this information. Notifications are sent in real time via email or a dedicated alert system.

[0430] Specific examples

[0431] For example, if user A posts a comment on a social networking site saying "I wish you'd just disappear," the following happens:

[0432] 1. The user terminal sends a comment to the server.

[0433] 2. The server receives the comments and parses them using the Google Cloud Natural Language API.

[0434] 3. The comment is deemed to contain the malicious phrase "Go away."

[0435] 4. The server instructs the user to hide the comment on the user's device and sends a warning message to User A.

[0436] 5. The server notifies the administrator terminal of details of the malicious comments.

[0437] By implementing this system, we can maintain a safe and healthy online communication environment.

[0438] Example prompts for generative AI models

[0439] "Please explain the program that analyzes comments posted on social media using natural language processing technology and automatically detects malicious comments. Please also include specific steps and the name of the technology used."

[0440] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0441] Step 1: Posting and receiving comments

[0442] When a user enters a comment and presses the send button on the user terminal, the comment data is sent to the server. The input is the text data entered by the user, and the output is a message to the server indicating that transmission is complete. The sent comment data is temporarily stored in a database on the server.

[0443] Step 2: Initial processing of comments

[0444] The server temporarily stores the received comment data in a database and then preprocesses it. The input is the comment data sent by the user, and the output is preprocessed text data. Specifically, it deletes unnecessary spaces and special characters and normalizes characters.

[0445] Step 3: Natural Language Processing Analysis of Comments

[0446] The server sends the preprocessed comment data to a natural language processing engine (e.g., Google Cloud Natural Language API) and obtains the analysis results. The input is the preprocessed text data, and the output is the analysis results (e.g., semantic analysis results and keyword frequency). Specifically, the comments are first tokenized, and then the meaning of each token is analyzed.

[0447] Step 4: Identifying malicious comments

[0448] The server determines whether a comment is malicious based on the analysis results from the natural language processing engine. The input is the analysis results obtained from the natural language processing engine, and the output is the maliciousness score of the comment. Specifically, it compares the score with a set reference value, and if the score exceeds the threshold, it is determined to be malicious.

[0449] Step 5: Hide comments

[0450] If the server judges a comment to be malicious, it issues an instruction to hide the comment on the user's device. The input is the comment data judged to be malicious and the judgment result, and the output is a command to hide the comment on the user's device. Specifically, the server sends a list of hidden comments to the front-end application.

[0451] Step 6: Sending a warning message

[0452] The server sends a warning message to the user who posted the malicious comment. The input is the data of the comment that was determined to be malicious and the user information, and the output is the warning message. Specifically, the server sends a data packet to the user's device to display the warning message as a pop-up.

[0453] Step 7: Notify administrators

[0454] The server notifies the administrator terminal of details of malicious comments. The input is the comment data determined to be malicious and the analysis results, and the output is a notification message to the administrator. Specifically, the server notifies the administrator in real time using email or a dedicated alert system. This allows the administrator to understand the situation and take appropriate action.

[0455] (Application example 1)

[0456] 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."

[0457] On modern social media and video streaming platforms, malicious comments can damage the health of the community. Especially when many comments are posted in real time, it is difficult to immediately detect and address malicious comments. Furthermore, if malicious comments are left unattended, they can cause discomfort and trouble for other users. Furthermore, manual monitoring and response requires too much time and effort, placing a significant burden on the system, so an efficient system is needed.

[0458] 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.

[0459] In this invention, the server includes: means for receiving comments sent from a user terminal; means for analyzing the received comments using natural language processing technology; means for determining whether the comments are malicious based on the analysis results; means for hiding comments determined to be malicious on the user terminal; means for sending a warning message to the user terminal when a comment is determined to be malicious; means for notifying an administrator terminal of details of the comments determined to be malicious; means for performing semantic analysis on comments acquired from the user terminal and hiding malicious comments in real time based on the results of the analysis; and means having a communication means for notifying the administrator terminal of details of the malicious comments. This automatically maintains the integrity of comments, allowing other users to enjoy safe and sound online communication.

[0460] A "user terminal" is a device used to input and send comments, and includes personal computers, smartphones, etc.

[0461] The "server" is a central computer system that receives comments sent from user terminals, analyzes them, and responds to them.

[0462] "Natural language processing technology" refers to a set of computational techniques and methodologies for analyzing human language and understanding its meaning.

[0463] "Semantic analysis" is a part of natural language processing techniques that allows for understanding the meaning and context of text.

[0464] "Real-time" refers to near-instant processing, meaning immediate responses to user-submitted comments.

[0465] "Hateful comments" are comments that contain offensive, inappropriate, or harmful content that offends other users.

[0466] A "warning message" is a notification sent to a user who posts a malicious comment, warning them that the content is inappropriate.

[0467] "Administrator terminal" refers to a device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[0468] A description will now be given of a form for realizing a system based on the present invention. This system includes a user terminal, a server, and an administrator terminal.

[0469] System Overview

[0470] The system of the present invention automatically analyzes and evaluates the content of comments to maintain the integrity of comment sections on social media and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of user terminals, a server, and an administrator terminal.

[0471] Hardware and software used

[0472] User terminal: A device used to input and send comments, such as a personal computer or smartphone.

[0473] Server: A central computer system that receives comments sent from user terminals, analyzes them using natural language processing technology, and responds accordingly.

[0474] Administrator device: A device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[0475] Software and Technology

[0476] Natural Language Processing (NLP): This technology uses the transformers library to detect malicious comments, specifically leveraging generative AI models to perform semantic analysis.

[0477] SNS API integration: Provides a means of communication using the requests library to retrieve, analyze, and hide comments from SNS APIs.

[0478] Sending warning messages: Using Apple Push Notification Service (APNs), we send warning messages to users who post inappropriate comments.

[0479] Admin Notification: Admin notifications are done with a simple POST request.

[0480] Specific examples of processing

[0481] For example, if a user comments "This app is the worst, get rid of it!", the system works like this:

[0482] 1. The user enters a comment on the terminal and sends it to the server.

[0483] 2. The server receives the comments and analyzes them using natural language processing technology.

[0484] 3. The server determines that the comment is malicious. Specifically, semantic analysis detects that the comment is inappropriate.

[0485] 4. The server sets the comment to be hidden on the user's device, generates a warning message based on the analysis results, and sends the warning message to the user.

[0486] 5. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes action as necessary.

[0487] An example of using a generative AI model to set prompts is:

[0488] Prompt: Please rate the following comment as malicious: "Go away."

[0489] In this way, the system of the present invention can automatically detect malicious comments in real time and take appropriate measures to maintain the health of the comment section, providing an effective solution for enhancing the safety and health of online communication.

[0490] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0491] Step 1:

[0492] The user terminal inputs and transmits a comment.

[0493] Input: Comment entered by the user (e.g. "I wish you'd just disappear")

[0494] Output: Request data with comments sent to the server

[0495] How it works: When a user enters a comment and presses the submit button, the comment data is sent to the server as a request.

[0496] Step 2:

[0497] The server receives the comment.

[0498] Input: Comment data sent from the user's device

[0499] Output: Received comment data

[0500] Operation: The server receives a request to send a comment from a user's device and stores the comment data in a buffer for analysis.

[0501] Step 3:

[0502] The server analyzes the comments using natural language processing technology.

[0503] Input: Received comment data

[0504] Output: Comment analysis results (e.g., maliciousness score, context information)

[0505] How it works: A natural language processing engine on the server analyzes the meaning of the comment and performs a semantic analysis of the comment. A generative AI model is used to calculate a severity score based on the prompt.

[0506] Step 4:

[0507] The server determines whether the comment is malicious.

[0508] Input: Comment analysis results

[0509] Output: Malicious or not (e.g., malicious)

[0510] How it works: The server determines a comment as malicious based on the generated maliciousness score if the score exceeds a certain threshold.

[0511] Step 5:

[0512] Comments that the server determines to be malicious are set to be hidden on the user's device.

[0513] Input: Comment data judged to be malicious, and judgment result

[0514] Output: Updated comment display state (not displayed)

[0515] How it works: The server immediately sends data to the user's device to hide any comments that are determined to be malicious, and the comments are hidden.

[0516] Step 6:

[0517] If the server is determined to be malicious, a warning message is sent to the user terminal.

[0518] Input: Comment data judged to be malicious, and judgment result

[0519] Output: A warning message displayed on the user's terminal.

[0520] How it works: The server generates a warning message and uses the Apple Push Notification Service to send it to the device of the user who posted the comment.

[0521] Step 7:

[0522] The server notifies the administrator terminal of details of the malicious comments.

[0523] Input: Comment data judged to be malicious, and judgment result

[0524] Output: Notification data to the administrator's terminal

[0525] Operation: The server generates data to notify the administrator terminal of details of the comments that were determined to be malicious, and sends a POST request to the administrator's terminal.

[0526] In this way, the system of the present invention can maintain the integrity of the comment section, automatically detect malicious comments, and respond appropriately.

[0527] 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.

[0528] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[0529] 1. Posting and Receiving Comments

[0530] The user inputs a comment on the user terminal and presses the send button, which sends the comment to the server.

[0531] 2. Receipt and analysis of comments

[0532] The server receives comments sent from user devices. It then passes the received comments to a natural language processing engine, which analyzes the context and meaning of the words in the comments. Specifically, it takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[0533] 3. Determining whether a comment is malicious

[0534] The server uses natural language processing technology to determine whether a comment is malicious. For example, if a comment contains the phrase "disappear," it evaluates whether the phrase is malicious given the context.

[0535] 4. Emotion analysis using an emotion engine

[0536] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The emotion engine recognizes the user's emotions, and that emotional information is reflected in determining whether the comment is malicious. For example, comments that express positive emotions will have their score reduced, and comments that express negative emotions will have their score increased.

[0537] 5. Emotion-based correction

[0538] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. If a positive emotion is recognized, the malicious judgment score is reduced, and conversely, if a negative emotion is recognized, the score is increased. This correction makes the comment judgment more accurate.

[0539] 6. Hide comments

[0540] If the server determines that a comment is malicious, it will be hidden from the user's device, preventing other users from viewing the malicious comment.

[0541] 7. Sending warning messages

[0542] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[0543] 8. Notification to Administrator

[0544] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[0545] example

[0546] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[0547] 1. The user terminal sends the comment to the server.

[0548] 2. The server receives the comments and analyzes them using natural language processing technology.

[0549] 3. The comment is determined to be malicious.

[0550] 4. The server analyzes the user's emotional information along with the comments using an emotion engine.

[0551] 5. If the emotional information is negative, the malicious judgment score increases, and if it is positive, the score decreases.

[0552] 6. The server hides the comment on the user's device and sends a warning message to User A.

[0553] 7. The server notifies the administrator terminal and reports details of the malicious comments.

[0554] In this way, the system based on the present invention can recognize the user's emotions and reflect them in its malicious judgment, thereby encouraging more accurate and appropriate responses and providing a safe and healthy online communication environment.

[0555] The processing flow will be explained below.

[0556] Step 1:

[0557] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[0558] Step 2:

[0559] Server: The server receives the comments sent from the user terminal.

[0560] Step 3:

[0561] Server: Passes the received comment to a natural language processing engine, which analyzes the context and meaning of the comment. For example, it analyzes the sentence "Go away" and evaluates whether the expression contains malicious intent.

[0562] Step 4:

[0563] Server: Based on the analysis results, the server determines whether the comment is malicious or not, by checking whether it contains specific keywords or phrases.

[0564] Step 5:

[0565] Server: Analyzes the user's emotional information sent along with the comment using an emotion engine. The emotion engine analyzes data such as user input and voice, and classifies it as positive, negative, or neutral.

[0566] Step 6:

[0567] Server: Based on the emotional information, the judgement score of malicious comments is adjusted. For example, if the emotional sentiment is strong, the score is increased, and if the emotional sentiment is strong, the score is decreased.

[0568] Step 7:

[0569] Server: Comments that are judged to be malicious will be hidden, so other users will not be able to see them.

[0570] Step 8:

[0571] Server: Sends a warning message to the user who posted the comment that was determined to be malicious. The warning message states that the comment was inappropriate and explains why.

[0572] Step 9:

[0573] Server: Detailed information about malicious comments is sent to the administrator's terminal, allowing the administrator to understand the situation and take necessary measures.

[0574] Step 10:

[0575] Administrator terminal: Administrators will receive notifications and can take more severe measures in accordance with platform rules, such as completely deleting comments or freezing user accounts.

[0576] Example 2

[0577] 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."

[0578] Conventional social media and video streaming platforms lack sufficient detection and response to inappropriate comments in their comment sections, resulting in the rampant proliferation of malicious comments, which can disrupt user communication. In particular, systems that rely on simple keyword filtering or manual moderation without considering the context of the comment or the user's sentiment can have low accuracy in their judgments and result in incorrect responses. Furthermore, systems often fail to adequately warn users or provide administrators with information. This makes it difficult to maintain the integrity of online communication.

[0579] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for analyzing the user's emotional information along with the comments, means for correcting the maliciousness determination score based on the emotional information, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, and means for notifying an administrator terminal of a comment determined to be malicious. This enables highly accurate maliciousness determination based on context analysis of the comments and user's emotional information, and by sending appropriate warning messages and notifying the administrator, it becomes possible to effectively maintain the soundness of online communication.

[0580] A "user device" is a device used by a user to access social media or video distribution platforms and post comments.

[0581] "Server" refers to a central processing unit that receives and analyzes comments sent from user terminals and performs appropriate processing.

[0582] A "comment" is a text message that a user posts to other users on social media or video streaming platforms.

[0583] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[0584] "Emotion information" is data that indicates the emotional state of the user who posted the comment, and includes emotions such as positive and negative.

[0585] The "maliciousness score" is a quantitative index for evaluating whether a comment is malicious or not.

[0586] The "means of setting to hidden" is a method for making comments that have been determined to be malicious invisible to other users.

[0587] A "warning message" is a warning message sent to a user who posted a comment that was determined to be malicious.

[0588] "Administrator Terminal" means a device used by the Platform Administrator to monitor and manage User comments and system notifications.

[0589] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[0590] When a user enters a comment and presses the send button on the user terminal, the comment is sent to the server. Specifically, the user posts a comment using a device such as a smartphone or PC. This comment is sent to the server via the Internet.

[0591] The server receives comments sent from user devices and analyzes them using natural language processing technology. The natural language processing engines used here are, for example, OpenAI's GPT-3 or Google NLP. This analyzes the context of the comment, the frequency of keyword occurrences, and the degree of emotion. Furthermore, the server uses an emotion engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional information and evaluate whether abusive language or discriminatory remarks have been detected.

[0592] The server then determines whether the comment is malicious based on the analysis results. If the comment content is likely to be malicious, the server calculates a judgment score, and if the score exceeds a certain threshold, the comment is deemed malicious. The server then corrects the malicious judgment score based on the emotional information provided by the emotion engine. If the comment contains positive emotions, the score is reduced, and if it contains negative emotions, the score is increased.

[0593] Comments that are judged to be malicious are hidden from the user's terminal by the server, which prevents other users from viewing the comment. Furthermore, the server sends a warning message to the user who posted the malicious comment. This message includes information that the comment is inappropriate and warns the user to be careful.

[0594] The server also notifies the administrator terminal of details of malicious comments. This notification includes the comment content, poster information, and maliciousness score, allowing the administrator to understand the situation and take appropriate action. For example, the administrator can suspend or delete the user's account as necessary.

[0595] Specific examples

[0596] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following flow will be executed.

[0597] 1. The user enters a comment and presses the send button to send it to the server.

[0598] 2. The server receives the comments and analyzes them using natural language processing technology (e.g., GPT-3).

[0599] 3. The comment is flagged as malicious (e.g., because it contains the phrase "go away").

[0600] 4. The server analyzes the user's emotional information along with the comments using an emotion engine (e.g., IBM Watson Tone Analyzer). Negative emotional information is detected.

[0601] 5. The maliciousness score is adjusted (increased in maliciousness), and the final score exceeds the threshold, so the comment is judged to be malicious.

[0602] 6. The server hides the comment and sends a warning message to User A.

[0603] 7. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes appropriate measures.

[0604] This will enable the system to efficiently and effectively detect and address malicious comments to maintain the health of online communication.

[0605] Prompt Sentence Examples

[0606] "You should just disappear."

[0607] Sentiment score: Negative

[0608] Processing step: Based on the comment content and sentiment score, the comment is judged to be malicious and a warning message is sent if necessary.

[0609] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0610] Step 1:

[0611] The user enters a comment and presses the send button. The user enters text into the comment field on a social networking site or video streaming platform. As a result, the comment data is sent from the user's device to the server. Specifically, the comment entered by the user using a smartphone or computer is sent to the server as an HTTP POST request.

[0612] Step 2:

[0613] The server receives the comment. The input is the comment data sent from the user's device. The server receives this data and temporarily stores it in order to proceed to the next step. Specifically, the server's API endpoint receives the HTTP request and stores the comment content in the database.

[0614] Step 3:

[0615] The server analyzes the comments it receives using natural language processing technology. The input is the received comment data. This comment data is passed to a natural language processing engine, which analyzes the context of the text, the frequency of keyword occurrences, the degree of emotion, etc. The output is the analysis results, which include contextual information and an emotion score. Specifically, the server sends an API request to the natural language processing engine (e.g., SpaCy or Google NLP) and obtains the analysis results.

[0616] Step 4:

[0617] The server determines whether a comment is malicious based on the analysis results. The input is the analysis results of the natural language processing engine. The server calculates a judgment score based on this analysis result and evaluates whether it exceeds a threshold. The output is the result of the malicious judgment. Specifically, the server calculates the judgment score using a scoring algorithm that it has set up itself.

[0618] Step 5:

[0619] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the user's emotional information and the content of the comment. This is passed to the emotion engine, which analyzes the user's emotional state. The output is an analyzed emotion score. Specifically, the server sends an API request to an emotion engine such as IBM Watson Tone Analyzer and obtains the analysis results.

[0620] Step 6:

[0621] The server corrects the malicious judgment score based on the emotional information. The inputs are the malicious judgment score and the analysis results of the emotional engine. The server uses this emotional information to increase or decrease the malicious judgment score. The output is the corrected malicious judgment score. Specifically, it applies logic that decreases the score if the emotional score is positive, and increases the score if it is negative.

[0622] Step 7:

[0623] Comments that the server judges to be malicious are set to be hidden on the user's terminal. The input is the corrected maliciousness score. The server decides whether to hide the comment based on this score and sets a "hidden flag" for the comment in the database. The output is that the comment is hidden. Specifically, it updates a specific field in the database and enables the hidden setting.

[0624] Step 8:

[0625] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the result of the malicious judgment and user information. The server uses this information to generate and send a warning message to the user. The output is a warning message that is displayed on the user's device. Specifically, the warning message is generated using a notification service and sent according to the user's notification settings.

[0626] Step 9:

[0627] The server notifies the administrator device of comments that have been judged to be malicious. The input is detailed information about the malicious judgment (comment content, judgment score, user information, etc.). The server uses this information to generate and send a notification to the administrator device. The output is a detailed notification displayed on the administrator device. Specifically, it makes an API request to send a notification via email or to a dashboard dedicated to the administrator.

[0628] As described above, this system can efficiently detect and deal with malicious comments on social media and video distribution platforms through each processing step.

[0629] (Application example 2)

[0630] 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."

[0631] On online social media and video streaming platforms, inappropriate comments posted in the comment sections can cause unpleasant experiences for users. Furthermore, an increase in malicious comments can lead to a decline in the integrity of the platform as a whole, potentially resulting in a decline in the number of users. Furthermore, manual comment monitoring by administrators is labor-intensive and ineffective. To address these issues, a system is needed that can automatically analyze the content of comments and user sentiment, and quickly and accurately identify and hide malicious comments.

[0632] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for analyzing the user's emotional information sent together with the received comments using an emotion engine, means for correcting the judgment score of malicious comments based on the emotional information recognized by the emotion engine, means for determining whether the comments are malicious based on the analysis results, means for setting the user terminal to hide comments judged to be malicious, means for sending a warning message to the user terminal when a comment is judged to be malicious, and means for notifying an administrator terminal of a comment judged to be malicious. This makes it possible to appropriately judge and hide malicious comments by combining the comment content and user emotion analysis information.

[0633] A "user terminal" is an electronic device used to send comments, and includes smartphones, personal computers, tablets, etc.

[0634] A "server" is a computer system that receives, analyzes, and processes data sent from a user terminal, and cooperates with other terminals and systems as necessary.

[0635] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language, and includes semantic analysis of text and analysis of the frequency of keyword occurrences.

[0636] The "emotion engine" is a system that analyzes the emotional state of a user's comment from its context and determines whether the emotional information is positive or negative.

[0637] The "judgment score" is a numerical representation of the maliciousness and emotional state of a comment, and is used to evaluate the appropriateness of the comment.

[0638] "Bad comments" refer to comments that are deemed offensive, abusive, defamatory, or inappropriate, and that cause offense to other users.

[0639] The "hide setting" is a process of making a comment invisible to other users when the comment is judged to be malicious beyond a certain standard.

[0640] A "warning message" is a message that notifies a user who has posted an inappropriate comment that the comment is problematic and urges them to be careful.

[0641] An "administrator terminal" is a terminal used by a system administrator, which receives notifications about malicious comments and can take action as necessary.

[0642] "Comment context analysis" is a technology that understands the context in which a comment was posted and the text content before and after it, allowing for a more accurate understanding of the comment's intent and meaning.

[0643] To implement this invention, we will build a system for maintaining the integrity of comment sections on social media and video streaming platforms. The system mainly includes the following components:

[0644] 1. User terminal: An electronic device on which a user inputs comments and sends them to the server. This includes smartphones, PCs, tablets, etc.

[0645] 2. Server: A central computer system that receives comments, analyzes them using natural language processing technology and an emotion engine, and determines whether they are malicious or not. Specifically, the following software is used:

[0646] "Hugging Face Transformers" as a natural language processing engine

[0647] "Google Cloud Natural Language API" as a sentiment analysis engine

[0648] MySQL for database management

[0649] "Firebase Cloud Messaging" for notification system

[0650] 3. Processing details

[0651] Receiving comments: The server receives the comments sent from the user terminal.

[0652] Natural language processing analysis: The received comments are passed to a natural language processing engine to analyze the content and context of the text.

[0653] Sentiment analysis: The emotional information of users sent along with their comments is analyzed using an emotion engine.

[0654] Correction of judgment score: The judgment score of malicious comments is corrected based on the emotional information recognized by the emotion engine.

[0655] Maliciousness determination: Based on the adjusted score, a final determination is made as to whether the comment is malicious.

[0656] Comment hiding setting: Comments that are judged to be malicious are set to be hidden on the user's device.

[0657] Sending warning messages: The server sends a warning message to users who post malicious comments.

[0658] Notify administrator: Details of malicious comments will be sent to the administrator's device.

[0659] For example, if a user posts a comment on a social media platform saying "Go away," the following steps will occur:

[0660] 1. The user terminal sends the comment to the server.

[0661] 2. The server receives the comment and analyzes its context and content using natural language processing techniques.

[0662] 3. The server analyzes the emotional information sent along with the comment using an emotion engine.

[0663] 4. If the emotional information is negative, the maliciousness score increases and the message is ultimately judged to be malicious.

[0664] 5. The server marks the comment as hidden and sends a warning message to the user.

[0665] 6. The server notifies the administrator terminal and reports details of the malicious comments.

[0666] Examples of prompts include:

[0667] "A user comments, 'This place is awful.' The system does sentiment analysis and determines that the comment is negative. What should we do next?"

[0668] "User commented 'Go away'. Hide and send a warning."

[0669] In this way, the system can maintain a high level of integrity in comments based on user sentiment analysis information.

[0670] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0671] Step 1:

[0672] The user terminal inputs and sends a comment. The input is performed by the user entering text in the input field and clicking the send button. The output is the user's comment data being sent to the server.

[0673] Step 2:

[0674] The server receives comments sent from user terminals. The input is the comment data sent by the user, and the output is the storage of the received comment data. Specifically, the comments are recorded in a database.

[0675] Step 3:

[0676] The server analyzes the received comments using natural language processing technology. The input is the received comment data, and the output is the analysis results (information about the meaning and context of the comment). The server analyzes the content and context of the text using a natural language processing engine (Hugging Face Transformers).

[0677] Step 4:

[0678] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the comment data and the user's emotional information, and the output is the emotional judgment result. Specifically, it uses the Google Cloud Natural Language API to determine whether the user's emotion is positive or negative.

[0679] Step 5:

[0680] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. The input is the analysis result of natural language processing and the emotion judgment result, and the output is the corrected malicious judgment score. The score decreases if the emotion is positive, and increases if the emotion is negative.

[0681] Step 6:

[0682] The server finally judges whether the comment is malicious or not based on the corrected score. The input is the corrected malicious judgment score, and the output is the judgment result of whether the comment is malicious or not. If it is judged to be malicious, proceed to the next step.

[0683] Step 7:

[0684] The server sets comments that are judged to be malicious to be hidden on the user's terminal. The input is the judgment result of malicious comments, and the output is that the comment will be set to hidden. Specifically, the "hidden" flag for the relevant comment is set in the database.

[0685] Step 8:

[0686] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the judgment result of the malicious comment, and the output is the warning message to be sent to the user. The warning notification is sent to the user's device using Firebase Cloud Messaging.

[0687] Step 9:

[0688] The server notifies the administrator terminal of comments that are judged to be malicious. The input is the judgment result of malicious comments, and the output is notification information sent to the administrator. Based on this information, the administrator can consider taking more severe measures.

[0689] 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.

[0690] 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.

[0691] 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.

[0692] [Third embodiment]

[0693] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0694] 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.

[0695] 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).

[0696] 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.

[0697] 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.

[0698] 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).

[0699] 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. 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.

[0700] 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.

[0701] 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.

[0702] 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.

[0703] 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.

[0704] 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."

[0705] The system based on this invention automatically analyzes and evaluates the content of comments and takes appropriate action against malicious comments in order to maintain the integrity of comment sections on social media and video distribution platforms. This system is composed of user terminals, a server, and an administrator terminal.

[0706] 1. Posting and Receiving Comments

[0707] When a user inputs a comment and presses the send button on the user terminal, the comment is sent to the server, and the server receives the comment.

[0708] 2. Comment Analysis

[0709] The server analyzes the content of the received comments using natural language processing technology. Specifically, it uses a natural language processing engine to analyze the context and meaning of the comments and determine whether the comments are malicious. This analysis takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[0710] 3. Identifying and Responding to Malicious Comments

[0711] The server determines whether a comment is malicious based on the analysis results obtained using natural language processing technology. For example, if a comment contains the expression "Go away," it will be judged as malicious if the expression is determined to be malicious based on the context.

[0712] 4. Hide comments

[0713] If the server determines that a comment is malicious, it will be instantly hidden from the user's device, preventing other users from viewing the malicious comment.

[0714] 5. Sending warning messages

[0715] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[0716] 6. Notification to Administrator

[0717] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[0718] example

[0719] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[0720] 1. The user terminal sends the comment to the server.

[0721] 2. The server receives the comments and analyzes them using natural language processing technology.

[0722] 3. The comment is determined to be malicious.

[0723] 4. The server hides the comment on the user's device and sends a warning message to User A.

[0724] 5. The server notifies the administrator terminal and reports details of the malicious comments.

[0725] In this way, the system based on the present invention can encourage users and administrators to take prompt and appropriate action, and provide a safe and sound online communication environment.

[0726] The processing flow will be explained below.

[0727] Step 1:

[0728] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[0729] Step 2:

[0730] Server: The server receives the comments sent from the user terminal.

[0731] Step 3:

[0732] Server: Passes received comments to a natural language processing engine, which analyzes the context and meaning of the comments.

[0733] Step 4:

[0734] Server: Based on the analysis results, the server determines whether the comment is malicious. For example, if the keyword "disappear" is included, the server evaluates whether the expression is malicious based on the context.

[0735] Step 5:

[0736] Server: If a comment is determined to be malicious, it will be hidden, preventing other users from viewing the malicious comment.

[0737] Step 6:

[0738] Server: Send a warning message to the user who posted the comment that was determined to be malicious. The message includes a detailed description of the inappropriateness of the comment.

[0739] Step 7:

[0740] Server: Details of malicious comments are sent to the administrator's terminal, allowing the administrator to understand the situation and take appropriate action.

[0741] Step 8:

[0742] Administrator device: Administrators will receive notifications and can delete comments or take additional action against users if necessary.

[0743] Example 1

[0744] 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."

[0745] Conventional comment management systems for social media and video streaming platforms manually determine whether a comment is malicious, making it difficult to respond quickly and ultimately making it difficult to maintain a healthy communication environment. Even automated systems sometimes fail to take appropriate action against comments judged to be malicious, potentially damaging the credibility of the entire platform. Furthermore, the process for notifying administrators of details of malicious comments is often insufficient, leaving administrators with insufficient information to take appropriate action.

[0746] 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.

[0747] In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments using natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, means for notifying an administrator terminal of comments determined to be malicious, and means for taking into account semantic analysis, keyword frequency, and context dependency in the natural language processing technology. This enables the content of comments to be quickly and accurately evaluated automatically and appropriate measures to be taken against malicious comments. Furthermore, detailed notifications to the administrator allow the administrator to properly understand the situation and take additional measures.

[0748] A "user device" is an electronic device used to access social media or video distribution platforms and input and send comments.

[0749] "Server" refers to a central system that receives comments sent by users, analyzes them, and takes appropriate action.

[0750] "Comments" are text messages that users enter and send on social media or video streaming platforms.

[0751] "Natural language processing technology" is technology used by computers to understand and analyze human language.

[0752] "Semantic analysis" is a method for understanding the meaning of text, and is a technology that analyzes the context of comments and the meaning of words.

[0753] "Keyword frequency" is a method of calculating how frequently a particular keyword appears in a piece of text.

[0754] "Context sensitivity" refers to the property of the meaning of a particular word or phrase changing depending on the context surrounding it.

[0755] A "malicious comment" is a comment entered by a user that contains offensive, discriminatory, or offensive content to other users.

[0756] A "warning message" is a message sent to a user who has posted a comment that is deemed malicious, informing them of a violation of the terms of service or inappropriate behavior.

[0757] An "administrator device" is an electronic device used by an administrator who manages and operates a social networking site or video distribution platform.

[0758] "Notification" is a means of informing interested parties in real time about a particular event or situation.

[0759] The present invention provides a system for maintaining the integrity of comment sections on social networking sites and video distribution platforms. The system includes a user terminal, a server, and an administrator terminal.

[0760] First, the user enters a comment on the user terminal and presses the send button to send the comment to the server. This transmission is done over the Internet. The server receives the comment sent from the user terminal.

[0761] The server then analyzes the received comments. This process uses natural language processing technology. For example, semantic analysis is performed to understand the context of the comment, taking into account the frequency of keyword occurrences and context-dependency. This is done using software such as Google Cloud Natural Language API. Based on the results of this analysis, the server determines whether the comment is malicious. For example, a comment containing the phrase "Go away" is deemed malicious based on the context.

[0762] If the server determines that a comment is malicious based on the analysis results, it will hide the comment on the user's device. The server sends a command to hide the comment to the user's device, and the front-end application reflects this and hides the comment.

[0763] Furthermore, the server sends a warning message to users who post malicious comments. The warning message contains information about the inappropriateness of the comment and the penalties for repeated violations, thereby alerting the user.

[0764] The server also notifies the administrator terminal of details of malicious comments. This notification includes the content of the comment, information about the poster, and analysis results, allowing the administrator to take appropriate action based on this information. Notifications are sent in real time via email or a dedicated alert system.

[0765] Specific examples

[0766] For example, if user A posts a comment on a social networking site saying "I wish you'd just disappear," the following happens:

[0767] 1. The user terminal sends a comment to the server.

[0768] 2. The server receives the comments and parses them using the Google Cloud Natural Language API.

[0769] 3. The comment is deemed to contain the malicious phrase "Go away."

[0770] 4. The server instructs the user to hide the comment on the user's device and sends a warning message to User A.

[0771] 5. The server notifies the administrator terminal of details of the malicious comments.

[0772] By implementing this system, we can maintain a safe and healthy online communication environment.

[0773] Example prompts for generative AI models

[0774] "Please explain the program that analyzes comments posted on social media using natural language processing technology and automatically detects malicious comments. Please also include specific steps and the name of the technology used."

[0775] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0776] Step 1: Posting and receiving comments

[0777] When a user enters a comment and presses the send button on the user terminal, the comment data is sent to the server. The input is the text data entered by the user, and the output is a message to the server indicating that transmission is complete. The sent comment data is temporarily stored in a database on the server.

[0778] Step 2: Initial processing of comments

[0779] The server temporarily stores the received comment data in a database and then preprocesses it. The input is the comment data sent by the user, and the output is preprocessed text data. Specifically, it deletes unnecessary spaces and special characters and normalizes characters.

[0780] Step 3: Natural Language Processing Analysis of Comments

[0781] The server sends the preprocessed comment data to a natural language processing engine (e.g., Google Cloud Natural Language API) and obtains the analysis results. The input is the preprocessed text data, and the output is the analysis results (e.g., semantic analysis results and keyword frequency). Specifically, the comments are first tokenized, and then the meaning of each token is analyzed.

[0782] Step 4: Identifying malicious comments

[0783] The server determines whether a comment is malicious based on the analysis results from the natural language processing engine. The input is the analysis results obtained from the natural language processing engine, and the output is the maliciousness score of the comment. Specifically, it compares the score with a set reference value, and if the score exceeds the threshold, it is determined to be malicious.

[0784] Step 5: Hide comments

[0785] If the server judges a comment to be malicious, it issues an instruction to hide the comment on the user's device. The input is the comment data judged to be malicious and the judgment result, and the output is a command to hide the comment on the user's device. Specifically, the server sends a list of hidden comments to the front-end application.

[0786] Step 6: Sending a warning message

[0787] The server sends a warning message to the user who posted the malicious comment. The input is the data of the comment that was determined to be malicious and the user information, and the output is the warning message. Specifically, the server sends a data packet to the user's device to display the warning message as a pop-up.

[0788] Step 7: Notify administrators

[0789] The server notifies the administrator terminal of details of malicious comments. The input is the comment data determined to be malicious and the analysis results, and the output is a notification message to the administrator. Specifically, the server notifies the administrator in real time using email or a dedicated alert system. This allows the administrator to understand the situation and take appropriate action.

[0790] (Application example 1)

[0791] 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."

[0792] On modern social media and video streaming platforms, malicious comments can damage the health of the community. Especially when many comments are posted in real time, it is difficult to immediately detect and address malicious comments. Furthermore, if malicious comments are left unattended, they can cause discomfort and trouble for other users. Furthermore, manual monitoring and response requires too much time and effort, placing a significant burden on the system, so an efficient system is needed.

[0793] 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.

[0794] In this invention, the server includes: means for receiving comments sent from a user terminal; means for analyzing the received comments using natural language processing technology; means for determining whether the comments are malicious based on the analysis results; means for hiding comments determined to be malicious on the user terminal; means for sending a warning message to the user terminal when a comment is determined to be malicious; means for notifying an administrator terminal of details of the comments determined to be malicious; means for performing semantic analysis on comments acquired from the user terminal and hiding malicious comments in real time based on the results of the analysis; and means having a communication means for notifying the administrator terminal of details of the malicious comments. This automatically maintains the integrity of comments, allowing other users to enjoy safe and sound online communication.

[0795] A "user terminal" is a device used to input and send comments, and includes personal computers, smartphones, etc.

[0796] The "server" is a central computer system that receives comments sent from user terminals, analyzes them, and responds to them.

[0797] "Natural language processing technology" refers to a set of computational techniques and methodologies for analyzing human language and understanding its meaning.

[0798] "Semantic analysis" is a part of natural language processing techniques that allows for understanding the meaning and context of text.

[0799] "Real-time" refers to near-instant processing, meaning immediate responses to user-submitted comments.

[0800] "Hateful comments" are comments that contain offensive, inappropriate, or harmful content that offends other users.

[0801] A "warning message" is a notification sent to a user who posts a malicious comment, warning them that the content is inappropriate.

[0802] "Administrator terminal" refers to a device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[0803] A description will now be given of a form for realizing a system based on the present invention. This system includes a user terminal, a server, and an administrator terminal.

[0804] System Overview

[0805] The system of the present invention automatically analyzes and evaluates the content of comments to maintain the integrity of comment sections on social media and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of user terminals, a server, and an administrator terminal.

[0806] Hardware and software used

[0807] User terminal: A device used to input and send comments, such as a personal computer or smartphone.

[0808] Server: A central computer system that receives comments sent from user terminals, analyzes them using natural language processing technology, and responds accordingly.

[0809] Administrator device: A device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[0810] Software and Technology

[0811] Natural Language Processing (NLP): This technology uses the transformers library to detect malicious comments, specifically leveraging generative AI models to perform semantic analysis.

[0812] SNS API integration: Provides a means of communication using the requests library to retrieve, analyze, and hide comments from SNS APIs.

[0813] Sending warning messages: Using Apple Push Notification Service (APNs), we send warning messages to users who post inappropriate comments.

[0814] Admin Notification: Admin notifications are done with a simple POST request.

[0815] Specific examples of processing

[0816] For example, if a user comments "This app is the worst, get rid of it!", the system works like this:

[0817] 1. The user enters a comment on the terminal and sends it to the server.

[0818] 2. The server receives the comments and analyzes them using natural language processing technology.

[0819] 3. The server determines that the comment is malicious. Specifically, semantic analysis detects that the comment is inappropriate.

[0820] 4. The server sets the comment to be hidden on the user's device, generates a warning message based on the analysis results, and sends the warning message to the user.

[0821] 5. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes action as necessary.

[0822] An example of using a generative AI model to set prompts is:

[0823] Prompt: Please rate the following comment as malicious: "Go away."

[0824] In this way, the system of the present invention can automatically detect malicious comments in real time and take appropriate measures to maintain the health of the comment section, providing an effective solution for enhancing the safety and health of online communication.

[0825] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0826] Step 1:

[0827] The user terminal inputs and transmits a comment.

[0828] Input: Comment entered by the user (e.g. "I wish you'd just disappear")

[0829] Output: Request data with comments sent to the server

[0830] How it works: When a user enters a comment and presses the submit button, the comment data is sent to the server as a request.

[0831] Step 2:

[0832] The server receives the comment.

[0833] Input: Comment data sent from the user's device

[0834] Output: Received comment data

[0835] Operation: The server receives a request to send a comment from a user's device and stores the comment data in a buffer for analysis.

[0836] Step 3:

[0837] The server analyzes the comments using natural language processing technology.

[0838] Input: Received comment data

[0839] Output: Comment analysis results (e.g., maliciousness score, context information)

[0840] How it works: A natural language processing engine on the server analyzes the meaning of the comment and performs a semantic analysis of the comment. A generative AI model is used to calculate a severity score based on the prompt.

[0841] Step 4:

[0842] The server determines whether the comment is malicious.

[0843] Input: Comment analysis results

[0844] Output: Malicious or not (e.g., malicious)

[0845] How it works: The server determines a comment as malicious based on the generated maliciousness score if the score exceeds a certain threshold.

[0846] Step 5:

[0847] Comments that the server determines to be malicious are set to be hidden on the user's device.

[0848] Input: Comment data judged to be malicious, and judgment result

[0849] Output: Updated comment display state (not displayed)

[0850] How it works: The server immediately sends data to the user's device to hide any comments that are determined to be malicious, and the comments are hidden.

[0851] Step 6:

[0852] If the server is determined to be malicious, a warning message is sent to the user terminal.

[0853] Input: Comment data judged to be malicious, and judgment result

[0854] Output: A warning message displayed on the user's terminal.

[0855] How it works: The server generates a warning message and uses the Apple Push Notification Service to send it to the device of the user who posted the comment.

[0856] Step 7:

[0857] The server notifies the administrator terminal of details of the malicious comments.

[0858] Input: Comment data judged to be malicious, and judgment result

[0859] Output: Notification data to the administrator's terminal

[0860] Operation: The server generates data to notify the administrator terminal of details of the comments that were determined to be malicious, and sends a POST request to the administrator's terminal.

[0861] In this way, the system of the present invention can maintain the integrity of the comment section, automatically detect malicious comments, and respond appropriately.

[0862] 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.

[0863] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[0864] 1. Posting and Receiving Comments

[0865] The user inputs a comment on the user terminal and presses the send button, which sends the comment to the server.

[0866] 2. Receipt and analysis of comments

[0867] The server receives comments sent from user devices. It then passes the received comments to a natural language processing engine, which analyzes the context and meaning of the words in the comments. Specifically, it takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[0868] 3. Determining whether a comment is malicious

[0869] The server uses natural language processing technology to determine whether a comment is malicious. For example, if a comment contains the phrase "disappear," it evaluates whether the phrase is malicious given the context.

[0870] 4. Emotion analysis using an emotion engine

[0871] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The emotion engine recognizes the user's emotions, and that emotional information is reflected in determining whether the comment is malicious. For example, comments that express positive emotions will have their score reduced, and comments that express negative emotions will have their score increased.

[0872] 5. Emotion-based correction

[0873] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. If a positive emotion is recognized, the malicious judgment score is reduced, and conversely, if a negative emotion is recognized, the score is increased. This correction makes the comment judgment more accurate.

[0874] 6. Hide comments

[0875] If the server determines that a comment is malicious, it will be hidden from the user's device, preventing other users from viewing the malicious comment.

[0876] 7. Sending warning messages

[0877] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[0878] 8. Notification to Administrator

[0879] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[0880] example

[0881] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[0882] 1. The user terminal sends the comment to the server.

[0883] 2. The server receives the comments and analyzes them using natural language processing technology.

[0884] 3. The comment is determined to be malicious.

[0885] 4. The server analyzes the user's emotional information along with the comments using an emotion engine.

[0886] 5. If the emotional information is negative, the malicious judgment score increases, and if it is positive, the score decreases.

[0887] 6. The server hides the comment on the user's device and sends a warning message to User A.

[0888] 7. The server notifies the administrator terminal and reports details of the malicious comments.

[0889] In this way, the system based on the present invention can recognize the user's emotions and reflect them in its malicious judgment, thereby encouraging more accurate and appropriate responses and providing a safe and healthy online communication environment.

[0890] The processing flow will be explained below.

[0891] Step 1:

[0892] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[0893] Step 2:

[0894] Server: The server receives the comments sent from the user terminal.

[0895] Step 3:

[0896] Server: Passes the received comment to a natural language processing engine, which analyzes the context and meaning of the comment. For example, it analyzes the sentence "Go away" and evaluates whether the expression contains malicious intent.

[0897] Step 4:

[0898] Server: Based on the analysis results, the server determines whether the comment is malicious or not, by checking whether it contains specific keywords or phrases.

[0899] Step 5:

[0900] Server: Analyzes the user's emotional information sent along with the comment using an emotion engine. The emotion engine analyzes data such as user input and voice, and classifies it as positive, negative, or neutral.

[0901] Step 6:

[0902] Server: Based on the emotional information, the judgement score of malicious comments is adjusted. For example, if the emotional sentiment is strong, the score is increased, and if the emotional sentiment is strong, the score is decreased.

[0903] Step 7:

[0904] Server: Comments that are judged to be malicious will be hidden, so other users will not be able to see them.

[0905] Step 8:

[0906] Server: Sends a warning message to the user who posted the comment that was determined to be malicious. The warning message states that the comment was inappropriate and explains why.

[0907] Step 9:

[0908] Server: Detailed information about malicious comments is sent to the administrator's terminal, allowing the administrator to understand the situation and take necessary measures.

[0909] Step 10:

[0910] Administrator terminal: Administrators will receive notifications and can take more severe measures in accordance with platform rules, such as completely deleting comments or freezing user accounts.

[0911] Example 2

[0912] 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."

[0913] Conventional social media and video streaming platforms lack sufficient detection and response to inappropriate comments in their comment sections, resulting in the rampant proliferation of malicious comments, which can disrupt user communication. In particular, systems that rely on simple keyword filtering or manual moderation without considering the context of the comment or the user's sentiment can have low accuracy in their judgments and result in incorrect responses. Furthermore, systems often fail to adequately warn users or provide administrators with information. This makes it difficult to maintain the integrity of online communication.

[0914] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for analyzing the user's emotional information along with the comments, means for correcting the maliciousness determination score based on the emotional information, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, and means for notifying an administrator terminal of a comment determined to be malicious. This enables highly accurate maliciousness determination based on context analysis of the comments and user's emotional information, and by sending appropriate warning messages and notifying the administrator, it becomes possible to effectively maintain the soundness of online communication.

[0915] A "user device" is a device used by a user to access social media or video distribution platforms and post comments.

[0916] "Server" refers to a central processing unit that receives and analyzes comments sent from user terminals and performs appropriate processing.

[0917] A "comment" is a text message that a user posts to other users on social media or video streaming platforms.

[0918] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[0919] "Emotion information" is data that indicates the emotional state of the user who posted the comment, and includes emotions such as positive and negative.

[0920] The "maliciousness score" is a quantitative index for evaluating whether a comment is malicious or not.

[0921] The "means of setting to hidden" is a method for making comments that have been determined to be malicious invisible to other users.

[0922] A "warning message" is a warning message sent to a user who posted a comment that was determined to be malicious.

[0923] "Administrator Terminal" means a device used by the Platform Administrator to monitor and manage User comments and system notifications.

[0924] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[0925] When a user enters a comment and presses the send button on the user terminal, the comment is sent to the server. Specifically, the user posts a comment using a device such as a smartphone or PC. This comment is sent to the server via the Internet.

[0926] The server receives comments sent from user devices and analyzes them using natural language processing technology. The natural language processing engines used here are, for example, OpenAI's GPT-3 or Google NLP. This analyzes the context of the comment, the frequency of keyword occurrences, and the degree of emotion. Furthermore, the server uses an emotion engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional information and evaluate whether abusive language or discriminatory remarks have been detected.

[0927] The server then determines whether the comment is malicious based on the analysis results. If the comment content is likely to be malicious, the server calculates a judgment score, and if the score exceeds a certain threshold, the comment is deemed malicious. The server then corrects the malicious judgment score based on the emotional information provided by the emotion engine. If the comment contains positive emotions, the score is reduced, and if it contains negative emotions, the score is increased.

[0928] Comments that are judged to be malicious are hidden from the user's terminal by the server, which prevents other users from viewing the comment. Furthermore, the server sends a warning message to the user who posted the malicious comment. This message includes information that the comment is inappropriate and warns the user to be careful.

[0929] The server also notifies the administrator terminal of details of malicious comments. This notification includes the comment content, poster information, and maliciousness score, allowing the administrator to understand the situation and take appropriate action. For example, the administrator can suspend or delete the user's account as necessary.

[0930] Specific examples

[0931] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following flow will be executed.

[0932] 1. The user enters a comment and presses the send button to send it to the server.

[0933] 2. The server receives the comments and analyzes them using natural language processing technology (e.g., GPT-3).

[0934] 3. The comment is flagged as malicious (e.g., because it contains the phrase "go away").

[0935] 4. The server analyzes the user's emotional information along with the comments using an emotion engine (e.g., IBM Watson Tone Analyzer). Negative emotional information is detected.

[0936] 5. The maliciousness score is adjusted (increased in maliciousness), and the final score exceeds the threshold, so the comment is judged to be malicious.

[0937] 6. The server hides the comment and sends a warning message to User A.

[0938] 7. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes appropriate measures.

[0939] This will enable the system to efficiently and effectively detect and address malicious comments to maintain the health of online communication.

[0940] Prompt Sentence Examples

[0941] "You should just disappear."

[0942] Sentiment score: Negative

[0943] Processing step: Based on the comment content and sentiment score, the comment is judged to be malicious and a warning message is sent if necessary.

[0944] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0945] Step 1:

[0946] The user enters a comment and presses the send button. The user enters text into the comment field on a social networking site or video streaming platform. As a result, the comment data is sent from the user's device to the server. Specifically, the comment entered by the user using a smartphone or computer is sent to the server as an HTTP POST request.

[0947] Step 2:

[0948] The server receives the comment. The input is the comment data sent from the user's device. The server receives this data and temporarily stores it in order to proceed to the next step. Specifically, the server's API endpoint receives the HTTP request and stores the comment content in the database.

[0949] Step 3:

[0950] The server analyzes the comments it receives using natural language processing technology. The input is the received comment data. This comment data is passed to a natural language processing engine, which analyzes the context of the text, the frequency of keyword occurrences, the degree of emotion, etc. The output is the analysis results, which include contextual information and an emotion score. Specifically, the server sends an API request to the natural language processing engine (e.g., SpaCy or Google NLP) and obtains the analysis results.

[0951] Step 4:

[0952] The server determines whether a comment is malicious based on the analysis results. The input is the analysis results of the natural language processing engine. The server calculates a judgment score based on this analysis result and evaluates whether it exceeds a threshold. The output is the result of the malicious judgment. Specifically, the server calculates the judgment score using a scoring algorithm that it has set up itself.

[0953] Step 5:

[0954] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the user's emotional information and the content of the comment. This is passed to the emotion engine, which analyzes the user's emotional state. The output is an analyzed emotion score. Specifically, the server sends an API request to an emotion engine such as IBM Watson Tone Analyzer and obtains the analysis results.

[0955] Step 6:

[0956] The server corrects the malicious judgment score based on the emotional information. The inputs are the malicious judgment score and the analysis results of the emotional engine. The server uses this emotional information to increase or decrease the malicious judgment score. The output is the corrected malicious judgment score. Specifically, it applies logic that decreases the score if the emotional score is positive, and increases the score if it is negative.

[0957] Step 7:

[0958] Comments that the server judges to be malicious are set to be hidden on the user's terminal. The input is the corrected maliciousness score. The server decides whether to hide the comment based on this score and sets a "hidden flag" for the comment in the database. The output is that the comment is hidden. Specifically, it updates a specific field in the database and enables the hidden setting.

[0959] Step 8:

[0960] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the result of the malicious judgment and user information. The server uses this information to generate and send a warning message to the user. The output is a warning message that is displayed on the user's device. Specifically, the warning message is generated using a notification service and sent according to the user's notification settings.

[0961] Step 9:

[0962] The server notifies the administrator device of comments that have been judged to be malicious. The input is detailed information about the malicious judgment (comment content, judgment score, user information, etc.). The server uses this information to generate and send a notification to the administrator device. The output is a detailed notification displayed on the administrator device. Specifically, it makes an API request to send a notification via email or to a dashboard dedicated to the administrator.

[0963] As described above, this system can efficiently detect and deal with malicious comments on social media and video distribution platforms through each processing step.

[0964] (Application example 2)

[0965] 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."

[0966] On online social media and video streaming platforms, inappropriate comments posted in the comment sections can cause unpleasant experiences for users. Furthermore, an increase in malicious comments can lead to a decline in the integrity of the platform as a whole, potentially resulting in a decline in the number of users. Furthermore, manual comment monitoring by administrators is labor-intensive and ineffective. To address these issues, a system is needed that can automatically analyze the content of comments and user sentiment, and quickly and accurately identify and hide malicious comments.

[0967] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for analyzing the user's emotional information sent together with the received comments using an emotion engine, means for correcting the judgment score of malicious comments based on the emotional information recognized by the emotion engine, means for determining whether the comments are malicious based on the analysis results, means for setting the user terminal to hide comments judged to be malicious, means for sending a warning message to the user terminal when a comment is judged to be malicious, and means for notifying an administrator terminal of a comment judged to be malicious. This makes it possible to appropriately judge and hide malicious comments by combining the comment content and user emotion analysis information.

[0968] A "user terminal" is an electronic device used to send comments, and includes smartphones, personal computers, tablets, etc.

[0969] A "server" is a computer system that receives, analyzes, and processes data sent from a user terminal, and cooperates with other terminals and systems as necessary.

[0970] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language, and includes semantic analysis of text and analysis of the frequency of keyword occurrences.

[0971] The "emotion engine" is a system that analyzes the emotional state of a user's comment from its context and determines whether the emotional information is positive or negative.

[0972] The "judgment score" is a numerical representation of the maliciousness and emotional state of a comment, and is used to evaluate the appropriateness of the comment.

[0973] "Bad comments" refer to comments that are deemed offensive, abusive, defamatory, or inappropriate, and that cause offense to other users.

[0974] The "hide setting" is a process of making a comment invisible to other users when the comment is judged to be malicious beyond a certain standard.

[0975] A "warning message" is a message that notifies a user who has posted an inappropriate comment that the comment is problematic and urges them to be careful.

[0976] An "administrator terminal" is a terminal used by a system administrator, which receives notifications about malicious comments and can take action as necessary.

[0977] "Comment context analysis" is a technology that understands the context in which a comment was posted and the text content before and after it, allowing for a more accurate understanding of the comment's intent and meaning.

[0978] To implement this invention, we will build a system for maintaining the integrity of comment sections on social media and video streaming platforms. The system mainly includes the following components:

[0979] 1. User terminal: An electronic device on which a user inputs comments and sends them to the server. This includes smartphones, PCs, tablets, etc.

[0980] 2. Server: A central computer system that receives comments, analyzes them using natural language processing technology and an emotion engine, and determines whether they are malicious or not. Specifically, the following software is used:

[0981] "Hugging Face Transformers" as a natural language processing engine

[0982] "Google Cloud Natural Language API" as a sentiment analysis engine

[0983] MySQL for database management

[0984] "Firebase Cloud Messaging" for notification system

[0985] 3. Processing details

[0986] Receiving comments: The server receives the comments sent from the user terminal.

[0987] Natural language processing analysis: The received comments are passed to a natural language processing engine to analyze the content and context of the text.

[0988] Sentiment analysis: The emotional information of users sent along with their comments is analyzed using an emotion engine.

[0989] Correction of judgment score: The judgment score of malicious comments is corrected based on the emotional information recognized by the emotion engine.

[0990] Maliciousness determination: Based on the adjusted score, a final determination is made as to whether the comment is malicious.

[0991] Comment hiding setting: Comments that are judged to be malicious are set to be hidden on the user's device.

[0992] Sending warning messages: The server sends a warning message to users who post malicious comments.

[0993] Notify administrator: Details of malicious comments will be sent to the administrator's device.

[0994] For example, if a user posts a comment on a social media platform saying "Go away," the following steps will occur:

[0995] 1. The user terminal sends the comment to the server.

[0996] 2. The server receives the comment and analyzes its context and content using natural language processing techniques.

[0997] 3. The server analyzes the emotional information sent along with the comment using an emotion engine.

[0998] 4. If the emotional information is negative, the maliciousness score increases and the message is ultimately judged to be malicious.

[0999] 5. The server marks the comment as hidden and sends a warning message to the user.

[1000] 6. The server notifies the administrator terminal and reports details of the malicious comments.

[1001] Examples of prompts include:

[1002] "A user comments, 'This place is awful.' The system does sentiment analysis and determines that the comment is negative. What should we do next?"

[1003] "User commented 'Go away'. Hide and send a warning."

[1004] In this way, the system can maintain a high level of integrity in comments based on user sentiment analysis information.

[1005] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1006] Step 1:

[1007] The user terminal inputs and sends a comment. The input is performed by the user entering text in the input field and clicking the send button. The output is the user's comment data being sent to the server.

[1008] Step 2:

[1009] The server receives comments sent from user terminals. The input is the comment data sent by the user, and the output is the storage of the received comment data. Specifically, the comments are recorded in a database.

[1010] Step 3:

[1011] The server analyzes the received comments using natural language processing technology. The input is the received comment data, and the output is the analysis results (information about the meaning and context of the comment). The server analyzes the content and context of the text using a natural language processing engine (Hugging Face Transformers).

[1012] Step 4:

[1013] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the comment data and the user's emotional information, and the output is the emotional judgment result. Specifically, it uses the Google Cloud Natural Language API to determine whether the user's emotion is positive or negative.

[1014] Step 5:

[1015] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. The input is the analysis result of natural language processing and the emotion judgment result, and the output is the corrected malicious judgment score. The score decreases if the emotion is positive, and increases if the emotion is negative.

[1016] Step 6:

[1017] The server finally judges whether the comment is malicious or not based on the corrected score. The input is the corrected malicious judgment score, and the output is the judgment result of whether the comment is malicious or not. If it is judged to be malicious, proceed to the next step.

[1018] Step 7:

[1019] The server sets comments that are judged to be malicious to be hidden on the user's terminal. The input is the judgment result of malicious comments, and the output is that the comment will be set to hidden. Specifically, the "hidden" flag for the relevant comment is set in the database.

[1020] Step 8:

[1021] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the judgment result of the malicious comment, and the output is the warning message to be sent to the user. The warning notification is sent to the user's device using Firebase Cloud Messaging.

[1022] Step 9:

[1023] The server notifies the administrator terminal of comments that are judged to be malicious. The input is the judgment result of malicious comments, and the output is notification information sent to the administrator. Based on this information, the administrator can consider taking more severe measures.

[1024] 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.

[1025] 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.

[1026] 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.

[1027] [Fourth embodiment]

[1028] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1029] 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.

[1030] 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).

[1031] 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.

[1032] 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.

[1033] 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).

[1034] 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. 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.

[1035] 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.

[1036] 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.

[1037] 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.

[1038] 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.

[1039] 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.

[1040] 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."

[1041] The system based on this invention automatically analyzes and evaluates the content of comments and takes appropriate action against malicious comments in order to maintain the integrity of comment sections on social media and video distribution platforms. This system is composed of user terminals, a server, and an administrator terminal.

[1042] 1. Posting and Receiving Comments

[1043] When a user inputs a comment and presses the send button on the user terminal, the comment is sent to the server, and the server receives the comment.

[1044] 2. Comment Analysis

[1045] The server analyzes the content of the received comments using natural language processing technology. Specifically, it uses a natural language processing engine to analyze the context and meaning of the comments and determine whether the comments are malicious. This analysis takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[1046] 3. Identifying and Responding to Malicious Comments

[1047] The server determines whether a comment is malicious based on the analysis results obtained using natural language processing technology. For example, if a comment contains the expression "Go away," it will be judged as malicious if the expression is determined to be malicious based on the context.

[1048] 4. Hide comments

[1049] If the server determines that a comment is malicious, it will be instantly hidden from the user's device, preventing other users from viewing the malicious comment.

[1050] 5. Sending warning messages

[1051] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[1052] 6. Notification to Administrator

[1053] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[1054] example

[1055] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[1056] 1. The user terminal sends the comment to the server.

[1057] 2. The server receives the comments and analyzes them using natural language processing technology.

[1058] 3. The comment is determined to be malicious.

[1059] 4. The server hides the comment on the user's device and sends a warning message to User A.

[1060] 5. The server notifies the administrator terminal and reports details of the malicious comments.

[1061] In this way, the system based on the present invention can encourage users and administrators to take prompt and appropriate action, and provide a safe and sound online communication environment.

[1062] The processing flow will be explained below.

[1063] Step 1:

[1064] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[1065] Step 2:

[1066] Server: The server receives the comments sent from the user terminal.

[1067] Step 3:

[1068] Server: Passes received comments to a natural language processing engine, which analyzes the context and meaning of the comments.

[1069] Step 4:

[1070] Server: Based on the analysis results, the server determines whether the comment is malicious. For example, if the keyword "disappear" is included, the server evaluates whether the expression is malicious based on the context.

[1071] Step 5:

[1072] Server: If a comment is determined to be malicious, it will be hidden, preventing other users from viewing the malicious comment.

[1073] Step 6:

[1074] Server: Send a warning message to the user who posted the comment that was determined to be malicious. The message includes a detailed description of the inappropriateness of the comment.

[1075] Step 7:

[1076] Server: Details of malicious comments are sent to the administrator's terminal, allowing the administrator to understand the situation and take appropriate action.

[1077] Step 8:

[1078] Administrator device: Administrators will receive notifications and can delete comments or take additional action against users if necessary.

[1079] Example 1

[1080] 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."

[1081] Conventional comment management systems for social media and video streaming platforms manually determine whether a comment is malicious, making it difficult to respond quickly and ultimately making it difficult to maintain a healthy communication environment. Even automated systems sometimes fail to take appropriate action against comments judged to be malicious, potentially damaging the credibility of the entire platform. Furthermore, the process for notifying administrators of details of malicious comments is often insufficient, leaving administrators with insufficient information to take appropriate action.

[1082] 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.

[1083] In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments using natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, means for notifying an administrator terminal of comments determined to be malicious, and means for taking into account semantic analysis, keyword frequency, and context dependency in the natural language processing technology. This enables the content of comments to be quickly and accurately evaluated automatically and appropriate measures to be taken against malicious comments. Furthermore, detailed notifications to the administrator allow the administrator to properly understand the situation and take additional measures.

[1084] A "user device" is an electronic device used to access social media or video distribution platforms and input and send comments.

[1085] "Server" refers to a central system that receives comments sent by users, analyzes them, and takes appropriate action.

[1086] "Comments" are text messages that users enter and send on social media or video streaming platforms.

[1087] "Natural language processing technology" is technology used by computers to understand and analyze human language.

[1088] "Semantic analysis" is a method for understanding the meaning of text, and is a technology that analyzes the context of comments and the meaning of words.

[1089] "Keyword frequency" is a method of calculating how frequently a particular keyword appears in a piece of text.

[1090] "Context sensitivity" refers to the property of the meaning of a particular word or phrase changing depending on the context surrounding it.

[1091] A "malicious comment" is a comment entered by a user that contains offensive, discriminatory, or offensive content to other users.

[1092] A "warning message" is a message sent to a user who has posted a comment that is deemed malicious, informing them of a violation of the terms of service or inappropriate behavior.

[1093] An "administrator device" is an electronic device used by an administrator who manages and operates a social networking site or video distribution platform.

[1094] "Notification" is a means of informing interested parties in real time about a particular event or situation.

[1095] The present invention provides a system for maintaining the integrity of comment sections on social networking sites and video distribution platforms. The system includes a user terminal, a server, and an administrator terminal.

[1096] First, the user enters a comment on the user terminal and presses the send button to send the comment to the server. This transmission is done over the Internet. The server receives the comment sent from the user terminal.

[1097] The server then analyzes the received comments. This process uses natural language processing technology. For example, semantic analysis is performed to understand the context of the comment, taking into account the frequency of keyword occurrences and context-dependency. This is done using software such as Google Cloud Natural Language API. Based on the results of this analysis, the server determines whether the comment is malicious. For example, a comment containing the phrase "Go away" is deemed malicious based on the context.

[1098] If the server determines that a comment is malicious based on the analysis results, it will hide the comment on the user's device. The server sends a command to hide the comment to the user's device, and the front-end application reflects this and hides the comment.

[1099] Furthermore, the server sends a warning message to users who post malicious comments. The warning message contains information about the inappropriateness of the comment and the penalties for repeated violations, thereby alerting the user.

[1100] The server also notifies the administrator terminal of details of malicious comments. This notification includes the content of the comment, information about the poster, and analysis results, allowing the administrator to take appropriate action based on this information. Notifications are sent in real time via email or a dedicated alert system.

[1101] Specific examples

[1102] For example, if user A posts a comment on a social networking site saying "I wish you'd just disappear," the following happens:

[1103] 1. The user terminal sends a comment to the server.

[1104] 2. The server receives the comments and parses them using the Google Cloud Natural Language API.

[1105] 3. The comment is deemed to contain the malicious phrase "Go away."

[1106] 4. The server instructs the user to hide the comment on the user's device and sends a warning message to User A.

[1107] 5. The server notifies the administrator terminal of details of the malicious comments.

[1108] By implementing this system, we can maintain a safe and healthy online communication environment.

[1109] Example prompts for generative AI models

[1110] "Please explain the program that analyzes comments posted on social media using natural language processing technology and automatically detects malicious comments. Please also include specific steps and the name of the technology used."

[1111] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1112] Step 1: Posting and receiving comments

[1113] When a user enters a comment and presses the send button on the user terminal, the comment data is sent to the server. The input is the text data entered by the user, and the output is a message to the server indicating that transmission is complete. The sent comment data is temporarily stored in a database on the server.

[1114] Step 2: Initial processing of comments

[1115] The server temporarily stores the received comment data in a database and then preprocesses it. The input is the comment data sent by the user, and the output is preprocessed text data. Specifically, it deletes unnecessary spaces and special characters and normalizes characters.

[1116] Step 3: Natural Language Processing Analysis of Comments

[1117] The server sends the preprocessed comment data to a natural language processing engine (e.g., Google Cloud Natural Language API) and obtains the analysis results. The input is the preprocessed text data, and the output is the analysis results (e.g., semantic analysis results and keyword frequency). Specifically, the comments are first tokenized, and then the meaning of each token is analyzed.

[1118] Step 4: Identifying malicious comments

[1119] The server determines whether a comment is malicious based on the analysis results from the natural language processing engine. The input is the analysis results obtained from the natural language processing engine, and the output is the maliciousness score of the comment. Specifically, it compares the score with a set reference value, and if the score exceeds the threshold, it is determined to be malicious.

[1120] Step 5: Hide comments

[1121] If the server judges a comment to be malicious, it issues an instruction to hide the comment on the user's device. The input is the comment data judged to be malicious and the judgment result, and the output is a command to hide the comment on the user's device. Specifically, the server sends a list of hidden comments to the front-end application.

[1122] Step 6: Sending a warning message

[1123] The server sends a warning message to the user who posted the malicious comment. The input is the data of the comment that was determined to be malicious and the user information, and the output is the warning message. Specifically, the server sends a data packet to the user's device to display the warning message as a pop-up.

[1124] Step 7: Notify administrators

[1125] The server notifies the administrator terminal of details of malicious comments. The input is the comment data determined to be malicious and the analysis results, and the output is a notification message to the administrator. Specifically, the server notifies the administrator in real time using email or a dedicated alert system. This allows the administrator to understand the situation and take appropriate action.

[1126] (Application example 1)

[1127] 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."

[1128] On modern social media and video streaming platforms, malicious comments can damage the health of the community. Especially when many comments are posted in real time, it is difficult to immediately detect and address malicious comments. Furthermore, if malicious comments are left unattended, they can cause discomfort and trouble for other users. Furthermore, manual monitoring and response requires too much time and effort, placing a significant burden on the system, so an efficient system is needed.

[1129] 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.

[1130] In this invention, the server includes: means for receiving comments sent from a user terminal; means for analyzing the received comments using natural language processing technology; means for determining whether the comments are malicious based on the analysis results; means for hiding comments determined to be malicious on the user terminal; means for sending a warning message to the user terminal when a comment is determined to be malicious; means for notifying an administrator terminal of details of the comments determined to be malicious; means for performing semantic analysis on comments acquired from the user terminal and hiding malicious comments in real time based on the results of the analysis; and means having a communication means for notifying the administrator terminal of details of the malicious comments. This automatically maintains the integrity of comments, allowing other users to enjoy safe and sound online communication.

[1131] A "user terminal" is a device used to input and send comments, and includes personal computers, smartphones, etc.

[1132] The "server" is a central computer system that receives comments sent from user terminals, analyzes them, and responds to them.

[1133] "Natural language processing technology" refers to a set of computational techniques and methodologies for analyzing human language and understanding its meaning.

[1134] "Semantic analysis" is a part of natural language processing techniques that allows for understanding the meaning and context of text.

[1135] "Real-time" refers to near-instant processing, meaning immediate responses to user-submitted comments.

[1136] "Hateful comments" are comments that contain offensive, inappropriate, or harmful content that offends other users.

[1137] A "warning message" is a notification sent to a user who posts a malicious comment, warning them that the content is inappropriate.

[1138] "Administrator terminal" refers to a device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[1139] A description will now be given of a form for realizing a system based on the present invention. This system includes a user terminal, a server, and an administrator terminal.

[1140] System Overview

[1141] The system of the present invention automatically analyzes and evaluates the content of comments to maintain the integrity of comment sections on social media and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of user terminals, a server, and an administrator terminal.

[1142] Hardware and software used

[1143] User terminal: A device used to input and send comments, such as a personal computer or smartphone.

[1144] Server: A central computer system that receives comments sent from user terminals, analyzes them using natural language processing technology, and responds accordingly.

[1145] Administrator device: A device used by the platform administrator to receive detailed notifications about and respond to malicious comments.

[1146] Software and Technology

[1147] Natural Language Processing (NLP): This technology uses the transformers library to detect malicious comments, specifically leveraging generative AI models to perform semantic analysis.

[1148] SNS API integration: Provides a means of communication using the requests library to retrieve, analyze, and hide comments from SNS APIs.

[1149] Sending warning messages: Using Apple Push Notification Service (APNs), we send warning messages to users who post inappropriate comments.

[1150] Admin Notification: Admin notifications are done with a simple POST request.

[1151] Specific examples of processing

[1152] For example, if a user comments "This app is the worst, get rid of it!", the system works like this:

[1153] 1. The user enters a comment on the terminal and sends it to the server.

[1154] 2. The server receives the comments and analyzes them using natural language processing technology.

[1155] 3. The server determines that the comment is malicious. Specifically, semantic analysis detects that the comment is inappropriate.

[1156] 4. The server sets the comment to be hidden on the user's device, generates a warning message based on the analysis results, and sends the warning message to the user.

[1157] 5. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes action as necessary.

[1158] An example of using a generative AI model to set prompts is:

[1159] Prompt: Please rate the following comment as malicious: "Go away."

[1160] In this way, the system of the present invention can automatically detect malicious comments in real time and take appropriate measures to maintain the health of the comment section, providing an effective solution for enhancing the safety and health of online communication.

[1161] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1162] Step 1:

[1163] The user terminal inputs and transmits a comment.

[1164] Input: Comment entered by the user (e.g. "I wish you'd just disappear")

[1165] Output: Request data with comments sent to the server

[1166] How it works: When a user enters a comment and presses the submit button, the comment data is sent to the server as a request.

[1167] Step 2:

[1168] The server receives the comment.

[1169] Input: Comment data sent from the user's device

[1170] Output: Received comment data

[1171] Operation: The server receives a request to send a comment from a user's device and stores the comment data in a buffer for analysis.

[1172] Step 3:

[1173] The server analyzes the comments using natural language processing technology.

[1174] Input: Received comment data

[1175] Output: Comment analysis results (e.g., maliciousness score, context information)

[1176] How it works: A natural language processing engine on the server analyzes the meaning of the comment and performs a semantic analysis of the comment. A generative AI model is used to calculate a severity score based on the prompt.

[1177] Step 4:

[1178] The server determines whether the comment is malicious.

[1179] Input: Comment analysis results

[1180] Output: Malicious or not (e.g., malicious)

[1181] How it works: The server determines a comment as malicious based on the generated maliciousness score if the score exceeds a certain threshold.

[1182] Step 5:

[1183] Comments that the server determines to be malicious are set to be hidden on the user's device.

[1184] Input: Comment data judged to be malicious, and judgment result

[1185] Output: Updated comment display state (not displayed)

[1186] How it works: The server immediately sends data to the user's device to hide any comments that are determined to be malicious, and the comments are hidden.

[1187] Step 6:

[1188] If the server is determined to be malicious, a warning message is sent to the user terminal.

[1189] Input: Comment data judged to be malicious, and judgment result

[1190] Output: A warning message displayed on the user's terminal.

[1191] How it works: The server generates a warning message and uses the Apple Push Notification Service to send it to the device of the user who posted the comment.

[1192] Step 7:

[1193] The server notifies the administrator terminal of details of the malicious comments.

[1194] Input: Comment data judged to be malicious, and judgment result

[1195] Output: Notification data to the administrator's terminal

[1196] Operation: The server generates data to notify the administrator terminal of details of the comments that were determined to be malicious, and sends a POST request to the administrator's terminal.

[1197] In this way, the system of the present invention can maintain the integrity of the comment section, automatically detect malicious comments, and respond appropriately.

[1198] 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.

[1199] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[1200] 1. Posting and Receiving Comments

[1201] The user inputs a comment on the user terminal and presses the send button, which sends the comment to the server.

[1202] 2. Receipt and analysis of comments

[1203] The server receives comments sent from user devices. It then passes the received comments to a natural language processing engine, which analyzes the context and meaning of the words in the comments. Specifically, it takes into account semantic analysis of the text, the frequency of keyword occurrences, and the context-dependency of the comments.

[1204] 3. Determining whether a comment is malicious

[1205] The server uses natural language processing technology to determine whether a comment is malicious. For example, if a comment contains the phrase "disappear," it evaluates whether the phrase is malicious given the context.

[1206] 4. Emotion analysis using an emotion engine

[1207] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The emotion engine recognizes the user's emotions, and that emotional information is reflected in determining whether the comment is malicious. For example, comments that express positive emotions will have their score reduced, and comments that express negative emotions will have their score increased.

[1208] 5. Emotion-based correction

[1209] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. If a positive emotion is recognized, the malicious judgment score is reduced, and conversely, if a negative emotion is recognized, the score is increased. This correction makes the comment judgment more accurate.

[1210] 6. Hide comments

[1211] If the server determines that a comment is malicious, it will be hidden from the user's device, preventing other users from viewing the malicious comment.

[1212] 7. Sending warning messages

[1213] Users who post comments that are judged to be malicious will receive a warning message from the server. The warning message will include information about the inappropriateness of the comment and will urge the user to be careful.

[1214] 8. Notification to Administrator

[1215] In addition, the server will notify the administrator terminal of the details of the malicious comments. This notification is important for the platform administrator to understand the situation and take appropriate action. After receiving the notification, the administrator can take more severe measures if necessary.

[1216] example

[1217] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following steps will be taken:

[1218] 1. The user terminal sends the comment to the server.

[1219] 2. The server receives the comments and analyzes them using natural language processing technology.

[1220] 3. The comment is determined to be malicious.

[1221] 4. The server analyzes the user's emotional information along with the comments using an emotion engine.

[1222] 5. If the emotional information is negative, the malicious judgment score increases, and if it is positive, the score decreases.

[1223] 6. The server hides the comment on the user's device and sends a warning message to User A.

[1224] 7. The server notifies the administrator terminal and reports details of the malicious comments.

[1225] In this way, the system based on the present invention can recognize the user's emotions and reflect them in its malicious judgment, thereby encouraging more accurate and appropriate responses and providing a safe and healthy online communication environment.

[1226] The processing flow will be explained below.

[1227] Step 1:

[1228] User terminal: The user enters a comment and presses the send button, which sends the comment to the server.

[1229] Step 2:

[1230] Server: The server receives the comments sent from the user terminal.

[1231] Step 3:

[1232] Server: Passes the received comment to a natural language processing engine, which analyzes the context and meaning of the comment. For example, it analyzes the sentence "Go away" and evaluates whether the expression contains malicious intent.

[1233] Step 4:

[1234] Server: Based on the analysis results, the server determines whether the comment is malicious or not, by checking whether it contains specific keywords or phrases.

[1235] Step 5:

[1236] Server: Analyzes the user's emotional information sent along with the comment using an emotion engine. The emotion engine analyzes data such as user input and voice, and classifies it as positive, negative, or neutral.

[1237] Step 6:

[1238] Server: Based on the emotional information, the judgement score of malicious comments is adjusted. For example, if the emotional sentiment is strong, the score is increased, and if the emotional sentiment is strong, the score is decreased.

[1239] Step 7:

[1240] Server: Comments that are judged to be malicious will be hidden, so other users will not be able to see them.

[1241] Step 8:

[1242] Server: Sends a warning message to the user who posted the comment that was determined to be malicious. The warning message states that the comment was inappropriate and explains why.

[1243] Step 9:

[1244] Server: Detailed information about malicious comments is sent to the administrator's terminal, allowing the administrator to understand the situation and take necessary measures.

[1245] Step 10:

[1246] Administrator terminal: Administrators will receive notifications and can take more severe measures in accordance with platform rules, such as completely deleting comments or freezing user accounts.

[1247] Example 2

[1248] 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."

[1249] Conventional social media and video streaming platforms lack sufficient detection and response to inappropriate comments in their comment sections, resulting in the rampant proliferation of malicious comments, which can disrupt user communication. In particular, systems that rely on simple keyword filtering or manual moderation without considering the context of the comment or the user's sentiment can have low accuracy in their judgments and result in incorrect responses. Furthermore, systems often fail to adequately warn users or provide administrators with information. This makes it difficult to maintain the integrity of online communication.

[1250] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for determining whether the comments are malicious based on the analysis results, means for analyzing the user's emotional information along with the comments, means for correcting the maliciousness determination score based on the emotional information, means for hiding comments determined to be malicious on the user terminal, means for sending a warning message to the user terminal when a comment is determined to be malicious, and means for notifying an administrator terminal of a comment determined to be malicious. This enables highly accurate maliciousness determination based on context analysis of the comments and user's emotional information, and by sending appropriate warning messages and notifying the administrator, it becomes possible to effectively maintain the soundness of online communication.

[1251] A "user device" is a device used by a user to access social media or video distribution platforms and post comments.

[1252] "Server" refers to a central processing unit that receives and analyzes comments sent from user terminals and performs appropriate processing.

[1253] A "comment" is a text message that a user posts to other users on social media or video streaming platforms.

[1254] "Natural language processing technology" is a technology that enables computers to understand and analyze human language.

[1255] "Emotion information" is data that indicates the emotional state of the user who posted the comment, and includes emotions such as positive and negative.

[1256] The "maliciousness score" is a quantitative index for evaluating whether a comment is malicious or not.

[1257] The "means of setting to hidden" is a method for making comments that have been determined to be malicious invisible to other users.

[1258] A "warning message" is a warning message sent to a user who posted a comment that was determined to be malicious.

[1259] "Administrator Terminal" means a device used by the Platform Administrator to monitor and manage User comments and system notifications.

[1260] The system based on this invention automatically analyzes and evaluates the content of comments and user sentiment to maintain the integrity of comment sections on social networking sites and video distribution platforms, and takes appropriate action against malicious comments. This system is composed of a user terminal, a server, an emotion engine, and an administrator terminal.

[1261] When a user enters a comment and presses the send button on the user terminal, the comment is sent to the server. Specifically, the user posts a comment using a device such as a smartphone or PC. This comment is sent to the server via the Internet.

[1262] The server receives comments sent from user devices and analyzes them using natural language processing technology. The natural language processing engines used here are, for example, OpenAI's GPT-3 or Google NLP. This analyzes the context of the comment, the frequency of keyword occurrences, and the degree of emotion. Furthermore, the server uses an emotion engine (e.g., IBM Watson Tone Analyzer) to analyze the user's emotional information and evaluate whether abusive language or discriminatory remarks have been detected.

[1263] The server then determines whether the comment is malicious based on the analysis results. If the comment content is likely to be malicious, the server calculates a judgment score, and if the score exceeds a certain threshold, the comment is deemed malicious. The server then corrects the malicious judgment score based on the emotional information provided by the emotion engine. If the comment contains positive emotions, the score is reduced, and if it contains negative emotions, the score is increased.

[1264] Comments that are judged to be malicious are hidden from the user's terminal by the server, which prevents other users from viewing the comment. Furthermore, the server sends a warning message to the user who posted the malicious comment. This message includes information that the comment is inappropriate and warns the user to be careful.

[1265] The server also notifies the administrator terminal of details of malicious comments. This notification includes the comment content, poster information, and maliciousness score, allowing the administrator to understand the situation and take appropriate action. For example, the administrator can suspend or delete the user's account as necessary.

[1266] Specific examples

[1267] For example, if user A posts a comment on a social networking site saying, "I wish you'd just disappear," the following flow will be executed.

[1268] 1. The user enters a comment and presses the send button to send it to the server.

[1269] 2. The server receives the comments and analyzes them using natural language processing technology (e.g., GPT-3).

[1270] 3. The comment is flagged as malicious (e.g., because it contains the phrase "go away").

[1271] 4. The server analyzes the user's emotional information along with the comments using an emotion engine (e.g., IBM Watson Tone Analyzer). Negative emotional information is detected.

[1272] 5. The maliciousness score is adjusted (increased in maliciousness), and the final score exceeds the threshold, so the comment is judged to be malicious.

[1273] 6. The server hides the comment and sends a warning message to User A.

[1274] 7. The server notifies the administrator terminal of details of the malicious comments, and the administrator takes appropriate measures.

[1275] This will enable the system to efficiently and effectively detect and address malicious comments to maintain the health of online communication.

[1276] Prompt Sentence Examples

[1277] "You should just disappear."

[1278] Sentiment score: Negative

[1279] Processing step: Based on the comment content and sentiment score, the comment is judged to be malicious and a warning message is sent if necessary.

[1280] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1281] Step 1:

[1282] The user enters a comment and presses the send button. The user enters text into the comment field on a social networking site or video streaming platform. As a result, the comment data is sent from the user's device to the server. Specifically, the comment entered by the user using a smartphone or computer is sent to the server as an HTTP POST request.

[1283] Step 2:

[1284] The server receives the comment. The input is the comment data sent from the user's device. The server receives this data and temporarily stores it in order to proceed to the next step. Specifically, the server's API endpoint receives the HTTP request and stores the comment content in the database.

[1285] Step 3:

[1286] The server analyzes the comments it receives using natural language processing technology. The input is the received comment data. This comment data is passed to a natural language processing engine, which analyzes the context of the text, the frequency of keyword occurrences, the degree of emotion, etc. The output is the analysis results, which include contextual information and an emotion score. Specifically, the server sends an API request to the natural language processing engine (e.g., SpaCy or Google NLP) and obtains the analysis results.

[1287] Step 4:

[1288] The server determines whether a comment is malicious based on the analysis results. The input is the analysis results of the natural language processing engine. The server calculates a judgment score based on this analysis result and evaluates whether it exceeds a threshold. The output is the result of the malicious judgment. Specifically, the server calculates the judgment score using a scoring algorithm that it has set up itself.

[1289] Step 5:

[1290] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the user's emotional information and the content of the comment. This is passed to the emotion engine, which analyzes the user's emotional state. The output is an analyzed emotion score. Specifically, the server sends an API request to an emotion engine such as IBM Watson Tone Analyzer and obtains the analysis results.

[1291] Step 6:

[1292] The server corrects the malicious judgment score based on the emotional information. The inputs are the malicious judgment score and the analysis results of the emotional engine. The server uses this emotional information to increase or decrease the malicious judgment score. The output is the corrected malicious judgment score. Specifically, it applies logic that decreases the score if the emotional score is positive, and increases the score if it is negative.

[1293] Step 7:

[1294] Comments that the server judges to be malicious are set to be hidden on the user's terminal. The input is the corrected maliciousness score. The server decides whether to hide the comment based on this score and sets a "hidden flag" for the comment in the database. The output is that the comment is hidden. Specifically, it updates a specific field in the database and enables the hidden setting.

[1295] Step 8:

[1296] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the result of the malicious judgment and user information. The server uses this information to generate and send a warning message to the user. The output is a warning message that is displayed on the user's device. Specifically, the warning message is generated using a notification service and sent according to the user's notification settings.

[1297] Step 9:

[1298] The server notifies the administrator device of comments that have been judged to be malicious. The input is detailed information about the malicious judgment (comment content, judgment score, user information, etc.). The server uses this information to generate and send a notification to the administrator device. The output is a detailed notification displayed on the administrator device. Specifically, it makes an API request to send a notification via email or to a dashboard dedicated to the administrator.

[1299] As described above, this system can efficiently detect and deal with malicious comments on social media and video distribution platforms through each processing step.

[1300] (Application example 2)

[1301] 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."

[1302] On online social media and video streaming platforms, inappropriate comments posted in the comment sections can cause unpleasant experiences for users. Furthermore, an increase in malicious comments can lead to a decline in the integrity of the platform as a whole, potentially resulting in a decline in the number of users. Furthermore, manual comment monitoring by administrators is labor-intensive and ineffective. To address these issues, a system is needed that can automatically analyze the content of comments and user sentiment, and quickly and accurately identify and hide malicious comments.

[1303] The identification process by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving comments sent from a user terminal, means for analyzing the received comments based on natural language processing technology, means for analyzing the user's emotional information sent together with the received comments using an emotion engine, means for correcting the judgment score of malicious comments based on the emotional information recognized by the emotion engine, means for determining whether the comments are malicious based on the analysis results, means for setting the user terminal to hide comments judged to be malicious, means for sending a warning message to the user terminal when a comment is judged to be malicious, and means for notifying an administrator terminal of a comment judged to be malicious. This makes it possible to appropriately judge and hide malicious comments by combining the comment content and user emotion analysis information.

[1304] A "user terminal" is an electronic device used to send comments, and includes smartphones, personal computers, tablets, etc.

[1305] A "server" is a computer system that receives, analyzes, and processes data sent from a user terminal, and cooperates with other terminals and systems as necessary.

[1306] "Natural language processing technology" is a technology that allows computers to analyze, understand, and generate human language, and includes semantic analysis of text and analysis of the frequency of keyword occurrences.

[1307] The "emotion engine" is a system that analyzes the emotional state of a user's comment from its context and determines whether the emotional information is positive or negative.

[1308] The "judgment score" is a numerical representation of the maliciousness and emotional state of a comment, and is used to evaluate the appropriateness of the comment.

[1309] "Bad comments" refer to comments that are deemed offensive, abusive, defamatory, or inappropriate, and that cause offense to other users.

[1310] The "hide setting" is a process of making a comment invisible to other users when the comment is judged to be malicious beyond a certain standard.

[1311] A "warning message" is a message that notifies a user who has posted an inappropriate comment that the comment is problematic and urges them to be careful.

[1312] An "administrator terminal" is a terminal used by a system administrator, which receives notifications about malicious comments and can take action as necessary.

[1313] "Comment context analysis" is a technology that understands the context in which a comment was posted and the text content before and after it, allowing for a more accurate understanding of the comment's intent and meaning.

[1314] To implement this invention, we will build a system for maintaining the integrity of comment sections on social media and video streaming platforms. The system mainly includes the following components:

[1315] 1. User terminal: An electronic device on which a user inputs comments and sends them to the server. This includes smartphones, PCs, tablets, etc.

[1316] 2. Server: A central computer system that receives comments, analyzes them using natural language processing technology and an emotion engine, and determines whether they are malicious or not. Specifically, the following software is used:

[1317] "Hugging Face Transformers" as a natural language processing engine

[1318] "Google Cloud Natural Language API" as a sentiment analysis engine

[1319] MySQL for database management

[1320] "Firebase Cloud Messaging" for notification system

[1321] 3. Processing details

[1322] Receiving comments: The server receives the comments sent from the user terminal.

[1323] Natural language processing analysis: The received comments are passed to a natural language processing engine to analyze the content and context of the text.

[1324] Sentiment analysis: The emotional information of users sent along with their comments is analyzed using an emotion engine.

[1325] Correction of judgment score: The judgment score of malicious comments is corrected based on the emotional information recognized by the emotion engine.

[1326] Maliciousness determination: Based on the adjusted score, a final determination is made as to whether the comment is malicious.

[1327] Comment hiding setting: Comments that are judged to be malicious are set to be hidden on the user's device.

[1328] Sending warning messages: The server sends a warning message to users who post malicious comments.

[1329] Notify administrator: Details of malicious comments will be sent to the administrator's device.

[1330] For example, if a user posts a comment on a social media platform saying "Go away," the following steps will occur:

[1331] 1. The user terminal sends the comment to the server.

[1332] 2. The server receives the comment and analyzes its context and content using natural language processing techniques.

[1333] 3. The server analyzes the emotional information sent along with the comment using an emotion engine.

[1334] 4. If the emotional information is negative, the maliciousness score increases and the message is ultimately judged to be malicious.

[1335] 5. The server marks the comment as hidden and sends a warning message to the user.

[1336] 6. The server notifies the administrator terminal and reports details of the malicious comments.

[1337] Examples of prompts include:

[1338] "A user comments, 'This place is awful.' The system does sentiment analysis and determines that the comment is negative. What should we do next?"

[1339] "User commented 'Go away'. Hide and send a warning."

[1340] In this way, the system can maintain a high level of integrity in comments based on user sentiment analysis information.

[1341] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1342] Step 1:

[1343] The user terminal inputs and sends a comment. The input is performed by the user entering text in the input field and clicking the send button. The output is the user's comment data being sent to the server.

[1344] Step 2:

[1345] The server receives comments sent from user terminals. The input is the comment data sent by the user, and the output is the storage of the received comment data. Specifically, the comments are recorded in a database.

[1346] Step 3:

[1347] The server analyzes the received comments using natural language processing technology. The input is the received comment data, and the output is the analysis results (information about the meaning and context of the comment). The server analyzes the content and context of the text using a natural language processing engine (Hugging Face Transformers).

[1348] Step 4:

[1349] The server uses an emotion engine to analyze the user's emotional information sent along with the comment. The input is the comment data and the user's emotional information, and the output is the emotional judgment result. Specifically, it uses the Google Cloud Natural Language API to determine whether the user's emotion is positive or negative.

[1350] Step 5:

[1351] The server corrects the malicious comment judgment score based on the emotional information recognized by the emotion engine. The input is the analysis result of natural language processing and the emotion judgment result, and the output is the corrected malicious judgment score. The score decreases if the emotion is positive, and increases if the emotion is negative.

[1352] Step 6:

[1353] The server finally judges whether the comment is malicious or not based on the corrected score. The input is the corrected malicious judgment score, and the output is the judgment result of whether the comment is malicious or not. If it is judged to be malicious, proceed to the next step.

[1354] Step 7:

[1355] The server sets comments that are judged to be malicious to be hidden on the user's terminal. The input is the judgment result of malicious comments, and the output is that the comment will be set to hidden. Specifically, the "hidden" flag for the relevant comment is set in the database.

[1356] Step 8:

[1357] When a server is judged to be malicious, it sends a warning message to the user's device. The input is the judgment result of the malicious comment, and the output is the warning message to be sent to the user. The warning notification is sent to the user's device using Firebase Cloud Messaging.

[1358] Step 9:

[1359] The server notifies the administrator terminal of comments that are judged to be malicious. The input is the judgment result of malicious comments, and the output is notification information sent to the administrator. Based on this information, the administrator can consider taking more severe measures.

[1360] 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.

[1361] 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.

[1362] 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.

[1363] 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.

[1364] 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.

[1365] 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.

[1366] 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).

[1367] 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.

[1368] 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."

[1369] 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.

[1370] 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).

[1371] 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.

[1372] 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.

[1373] 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.

[1374] 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.

[1375] 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.

[1376] 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.

[1377] 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.

[1378] 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.

[1379] 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.

[1380] 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.

[1381] The following is further disclosed regarding the above embodiment.

[1382] (Claim 1)

[1383] A means for receiving comments sent from a user terminal;

[1384] means for analyzing the received comments based on natural language processing techniques;

[1385] A means for determining whether a comment is malicious based on the analysis result;

[1386] A means for hiding comments determined to be malicious on a user terminal;

[1387] means for transmitting a warning message to a user terminal when the message is determined to be malicious;

[1388] a means for notifying an administrator terminal of comments determined to be malicious;

[1389] A system including:

[1390] (Claim 2)

[1391] 2. The system according to claim 1, wherein the analysis means performs a context analysis of the comments.

[1392] (Claim 3)

[1393] 2. The system according to claim 1, further comprising means for suggesting to the user terminal to revise the comment when the comment determined to be malicious exceeds a specific score standard.

[1394] "Example 1"

[1395] (Claim 1)

[1396] A means for receiving comments sent from a user terminal;

[1397] means for analyzing the received comments based on natural language processing techniques;

[1398] A means for determining whether a comment is malicious based on the analysis result;

[1399] A means for hiding comments determined to be malicious on a user terminal;

[1400] means for transmitting a warning message to a user terminal when the message is determined to be malicious;

[1401] a means for notifying an administrator terminal of comments determined to be malicious;

[1402] In natural language processing technology, a means for taking into account semantic analysis, keyword frequency and context dependency;

[1403] A system including:

[1404] (Claim 2)

[1405] The system of claim 1, characterized in that the natural language processing technology for analyzing comments utilizes an artificial intelligence model.

[1406] (Claim 3)

[1407] 2. The system according to claim 1, further comprising means for issuing a warning message or suggesting corrections to the comment poster when a comment determined to be malicious exceeds a specific score standard.

[1408] "Application Example 1"

[1409] (Claim 1)

[1410] A means for receiving comments sent from a user terminal;

[1411] means for analyzing the received comments based on natural language processing techniques;

[1412] A means for determining whether a comment is malicious based on the analysis result;

[1413] A means for hiding comments determined to be malicious on a user terminal;

[1414] means for transmitting a warning message to a user terminal when the message is determined to be malicious;

[1415] a means for notifying an administrator terminal of comments determined to be malicious;

[1416] A means for performing semantic analysis on comments acquired from user devices and hiding malicious comments in real time based on the results of the analysis;

[1417] A means for providing a communication means for notifying an administrator terminal of details of malicious comments;

[1418] A system including:

[1419] (Claim 2)

[1420] The system according to claim 1, characterized in that it performs a contextual analysis of the comments and generates a prompt sentence to hide malicious comments based on the analysis results.

[1421] (Claim 3)

[1422] The system described in claim 1 further comprises a means for suggesting to the user terminal that the comment be revised when the comment determined to be malicious exceeds a specific score standard, and a means for generating a warning message based on a prompt text notified to the user terminal.

[1423] "Example 2: Combining Emotion Engines"

[1424] (Claim 1)

[1425] A means for receiving comments sent from a user terminal;

[1426] means for analyzing the received comments based on natural language processing techniques;

[1427] A means for determining whether a comment is malicious based on the analysis result;

[1428] A means for analyzing user emotional information together with the comments;

[1429] A means for correcting the maliciousness determination score based on emotion information;

[1430] A means for hiding comments determined to be malicious on a user terminal;

[1431] means for transmitting a warning message to a user terminal when the message is determined to be malicious;

[1432] a means for notifying an administrator terminal of comments determined to be malicious;

[1433] A system including:

[1434] (Claim 2)

[1435] The system of claim 1, further comprising: a contextual analysis of received comments.

[1436] (Claim 3)

[1437] 2. The system according to claim 1, further comprising means for suggesting to the user terminal to revise the comment when the comment determined to be malicious exceeds a specific score standard.

[1438] "Application example 2 when combining emotion engines"

[1439] (Claim 1)

[1440] A means for receiving comments sent from a user terminal;

[1441] means for analyzing the received comments based on natural language processing techniques;

[1442] A means for determining whether a comment is malicious based on the analysis result;

[1443] A means for hiding comments determined to be malicious on a user terminal;

[1444] means for transmitting a warning message to a user terminal when the message is determined to be malicious;

[1445] a means for notifying an administrator terminal of comments determined to be malicious;

[1446] means for analyzing, using an emotion engine, the user's emotion information sent together with the received comment;

[1447] A means for correcting a judgment score of malicious comments based on emotional information recognized by the emotion engine;

[1448] A system including:

[1449] (Claim 2)

[1450] The system according to claim 1, characterized in that it performs a contextual analysis of the comments.

[1451] (Claim 3)

[1452] 2. The system according to claim 1, further comprising means for suggesting to the user terminal to revise the comment when the comment determined to be malicious exceeds a specific score standard. [Explanation of symbols]

[1453] 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 for receiving comments sent from a user terminal; means for analyzing the received comments based on natural language processing techniques; A means for determining whether a comment is malicious based on the analysis result; A means for hiding comments determined to be malicious on a user terminal; means for transmitting a warning message to a user terminal when the message is determined to be malicious; a means for notifying an administrator terminal of comments determined to be malicious; A system including:

2. 2. The system according to claim 1, wherein the analyzing means performs a context analysis of the comments.

3. 2. The system according to claim 1, further comprising means for suggesting to the user terminal to correct the comment when the comment determined to be malicious exceeds a specific score standard.

Images