System

Abstract

A system is provided.SOLUTION: A system, comprising: means for obtaining advertising content; means for analyzing the advertising content with a generative model; means for filtering the advertising content based on the analysis; means for collecting feedback from users; and means for updating the generative model using the collected feedback.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In recent years, the increase in advertisements and videos that misuse generative models has become a serious problem. Younger and older generations are particularly susceptible to such fake information, which increases the risk of misinformation spreading easily. In emergencies such as disasters, there is also the risk that false information will cause confusion and prevent correct information from reaching the public. Therefore, there is an urgent need to provide a system that can detect and appropriately filter fake information that misuses generative models. [Means for solving the problem]

[0005] The present invention relates to a system that includes a means for acquiring advertising content, a means for analyzing the advertising content using a generative model, a means for filtering the advertising content based on the analysis results, a means for collecting user feedback, and a means for updating the generative model using the collected feedback. The system also includes a means for acquiring advertising content through API integration with other services, providing API integration, and a means for penalizing advertisers who place inappropriate advertisements. Furthermore, because the advertising content is text, image, audio, or video, and the generative model is trained based on official information, it is possible to detect and block fake information with high accuracy. This prevents the spread of inappropriate information, ensures the safety of end users, and improves the reliability of the service.

[0006] "Advertising content" means data, including text, images, audio, and video, of advertisements displayed on the Internet.

[0007] A "generative model" is an artificial intelligence model that learns from large amounts of data and generates new data or analyzes existing data using certain patterns and characteristics.

[0008] "API integration" is a technology that provides an interface for exchanging functions and data between different systems.

[0009] "Feedback" refers to the evaluations and opinions provided by system users regarding the results of their operations, and is used to improve the performance and accuracy of the system.

[0010] A "penalty" is a fine or sanction imposed for a specific act, and in the present invention, it is imposed on an advertiser who has placed an inappropriate advertisement.

[0011] "Official information" refers to data and information provided by reliable sources such as government agencies and authoritative organizations. [Brief explanation of the drawings]

[0012] [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

[0013] 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.

[0014] First, the terms used in the following description will be explained.

[0015] 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).

[0016] 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.

[0017] 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.

[0018] 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.

[0019] 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."

[0020] [First embodiment]

[0021] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0022] 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.

[0023] 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).

[0024] 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.

[0025] 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.

[0026] 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.

[0027] 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.

[0028] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0029] 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.

[0030] 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.

[0031] 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.

[0032] 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."

[0033] The present invention provides a system for acquiring advertising content, analyzing it using a generative model, filtering it, collecting feedback, and updating the model. Specific embodiments of each element are described below.

[0034] The server retrieves advertising content from websites or video streaming services. For example, it specifies a specific URL and downloads the advertising data from there. The retrieved data, which may be in the form of text, images, audio, or video, is then prepared for analysis by the generative model.

[0035] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The analysis results are used as an indicator to determine whether the advertisement is appropriate.

[0036] Based on the analysis of the ad content, the server decides whether to display it or not. If it is determined to contain inappropriate information, the ad is blocked and not shown to the user. On the other hand, if it is determined to be appropriate, the ad is displayed.

[0037] A dedicated UI is provided on the device to allow users to provide feedback on incorrectly displayed ads. Users fill out a feedback form, and this information is sent from the device to the server. This feedback data is used to update the generative model.

[0038] The server uses the collected feedback as part of model updates. By adding feedback data and retraining the generative model, it can detect fake information with high accuracy even on new data.

[0039] Furthermore, the server will provide other services with the ability to check advertising content through API integration. For example, it will provide ad checking functionality to platforms such as video streaming services and blogs, and analyze advertising data from outside. It will also improve the reliability of the entire system by imposing penalties on advertisers who deliver inappropriate ads.

[0040] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it will not be displayed; conversely, if it is determined to be appropriate, it will be displayed. Furthermore, appropriate feedback is provided to the advertiser to improve the model's performance.

[0041] As a result, the system of the present invention can effectively detect and block advertisements containing fake information, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0042] The processing flow will be explained below.

[0043] Step 1:

[0044] The server retrieves advertising content from websites and video streaming services. It sends an HTTP request from a specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[0045] Step 2:

[0046] The server sends the acquired advertising content to a generative model. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, which begins analysis.

[0047] Step 3:

[0048] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate, and uses this information to decide what to do next.

[0049] Step 4:

[0050] The server filters the ads based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the information is appropriate, the ad is displayed. The server records the filtering results in a log.

[0051] Step 5:

[0052] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button.

[0053] Step 6:

[0054] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[0055] Step 7:

[0056] The server receives the feedback and adds the feedback data to the dataset for model updates, which are then used as new training data to retrain the generative model.

[0057] Step 8:

[0058] The server retrains the generative model to detect fake information and inappropriate ads with higher accuracy. The retrained model is then used for the next ad analysis.

[0059] Step 9:

[0060] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[0061] Step 10:

[0062] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. In addition, advertisers who display appropriate ads are charged a service fee.

[0063] Through these steps, the system effectively analyzes and filters advertising content, providing users with a safe and reliable advertising experience.

[0064] Example 1

[0065] 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."

[0066] In today's digital advertising market, there are many advertisements containing fake information or inappropriate content, which not only impair the user experience but also reduce brand credibility. Conventional systems have had difficulty efficiently detecting and blocking these inappropriate advertisements. Furthermore, they lack mechanisms for effectively utilizing user feedback to improve ad quality. The present invention aims to solve these issues and provide a safe and reliable advertising experience.

[0067] 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.

[0068] In this invention, the server includes a device for acquiring advertising content, a device for analyzing the advertising content using a generative model, a device for filtering the advertising content based on the analysis results from the generative model, a device for collecting user feedback, and a device for updating the generative model using the collected feedback. This makes it possible to effectively detect and block advertisements containing fake information or inappropriate content. Furthermore, by reflecting user feedback, the accuracy of the generative model can be continuously improved, thereby providing a safer and more reliable advertising experience.

[0069] "Advertising content" refers to promotional information expressed in the form of text, images, audio, video, etc.

[0070] A "generative model" is an artificial intelligence model that uses an algorithm trained in advance based on official information to analyze and judge advertising content.

[0071] "Filtering" is the process of determining the appropriateness of advertising content and removing inappropriate advertisements based on the analysis results from the generative model.

[0072] "Feedback" refers to opinions and evaluation information provided by users regarding the display results of advertising content.

[0073] A "device" is a mechanism including hardware and software for performing each function of the system.

[0074] "API integration" is an interface that allows functions and data to be shared and linked between different software systems.

[0075] "Penalty" means a sanction or penalty imposed on an advertiser who places an inappropriate advertisement.

[0076] MODE FOR CARRYING OUT THE INVENTION

[0077] This system acquires advertising content, analyzes it using a generative model, filters it, collects feedback, and updates the model. This system consists of three main entities: a server, a terminal, and a user.

[0078] Server Features

[0079] The server acts as a device for acquiring advertising content, downloading advertising data from websites and video streaming services. Specifically, it uses the Python requests library to send HTTP requests to acquire advertising content. For example, it downloads advertising data from the specified URL https: / / example.com / ad and saves it in the form of text, images, audio, and video.

[0080] The server then sends the acquired advertising content to a generative model for analysis. This generative model is a generative AI model such as OpenAI's GPT, which determines the ad appropriateness by inputting specific prompts. The model is highly reliable because it is trained based on official information.

[0081] Specifically, the following prompt sentences are input to the generative AI model:

[0082] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[0083] Based on the analysis results of the generative model, the system decides whether to display an ad. If the ad contains inappropriate information, it is blocked, and if it is deemed appropriate, it is displayed.

[0084] User operations

[0085] The device is provided with a dedicated UI for users to provide feedback on advertisements. Users enter their opinions and ratings in a feedback form, and this information is sent from the device to the server. For example, a user may send feedback such as "This advertisement is misleading."

[0086] Server model update function

[0087] The server retrains the generative model based on the feedback collected from users. By adding this feedback data, the generative model improves its detection accuracy for new data. Retraining is the process of adding new feedback information to an existing dataset to improve the performance of the generative model.

[0088] API integration function

[0089] Furthermore, the server provides an API integration function, allowing it to obtain advertising content in cooperation with other services. This function allows it to link with external systems such as video streaming services and blog platforms to optimize advertising. It can also impose penalties on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0090] Specific examples

[0091] For example, a video streaming platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If an ad is determined to be inappropriate, it will not be displayed; only if it is appropriate will the ad be displayed. Appropriate feedback is also provided to the advertiser, contributing to improving the model's performance.

[0092] As a result, the system of the present invention can effectively detect and block advertisements containing fake information or inappropriate content. By using this system, it is possible to provide users with a safe and reliable advertising experience and improve the overall quality of services.

[0093] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0094] Processing Steps

[0095] Step 1:

[0096] The server retrieves advertising content from websites and video streaming services. Specifically, it uses the Python requests library to send an HTTP request and download the advertising data.

[0097] Input: Specific URL https: / / example.com / ad

[0098] Data processing: Sending HTTP requests, downloading advertising data

[0099] Output: Advertising data stored in the form of text, images, audio, and video

[0100] Specific operation: Specify a URL, send a request, and save the response data in local storage.

[0101] Step 2:

[0102] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained on public information and includes algorithms for detecting inappropriate content.

[0103] Input: Ad data obtained and saved in Step 1

[0104] Data processing: Formatting advertising data, creating prompts, and inputting them into the generative model

[0105] Output: Analysis results of advertisement content by generative model

[0106] Specific operation: Convert the ad data into text format, generate a prompt sentence like the following, and input it into the generative model:

[0107] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[0108] Step 3:

[0109] The server determines whether to display advertising content based on the analysis results from the generative model. If the content is deemed inappropriate, the ad is blocked, and if it is deemed appropriate, the ad is displayed to the user.

[0110] Input: Analysis results of the generative model

[0111] Data processing: Analysis of analysis results, filtering of advertisements deemed inappropriate

[0112] Output: Filtered ad content (displayable or not)

[0113] Specific operations: Analyze the analysis results, add inappropriate ads to a block list, and display appropriate ads on the user's screen.

[0114] Step 4:

[0115] Users provide feedback on advertising content through a dedicated UI, which is then sent from the device to the server.

[0116] Input: User feedback (e.g., "This ad is misleading")

[0117] Data Processing: Feedback data collection and transmission

[0118] Output: Feedback data is saved on the server

[0119] Specific operation: The user fills in the feedback form and presses the submit button. The data is sent from the device to the server.

[0120] Step 5:

[0121] The server retrains the generative model based on the collected feedback.

[0122] Input: Collected feedback data

[0123] Data processing: Integrating feedback data and retraining the generative model

[0124] Output: Updated generative model

[0125] What it does: Add the feedback data to the existing training dataset and retrain the generative model.

[0126] Step 6:

[0127] The server provides API integration functionality, connects with other services to acquire and analyze advertising content, and penalizes advertisers who deliver inappropriate ads.

[0128] Input: API request from external service

[0129] Data processing: Acquire external advertising data and analyze it with a generative model

[0130] Output: Sending analysis results back to external services, notifying inappropriate advertisers of penalties

[0131] Specific operation: Receives API requests from video streaming services, analyzes advertising data, returns analysis results to external services, and notifies advertisers of penalties if necessary.

[0132] (Application example 1)

[0133] 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."

[0134] In today's digital advertising environment, displaying inappropriate ads to users is a frequent problem. Displaying ads containing inappropriate content or fake information can ruin the user experience and reduce the overall reliability of the service. Another challenge is the inefficient collection of feedback and subsequent model updates. Furthermore, ads are presented in a variety of formats (text, image, audio, video), requiring appropriate filtering for each.

[0135] 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.

[0136] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis result, means for collecting feedback from users, means for updating the generative model using the collected feedback, and means for executing an advertising filter application that uses the collected feedback to update the model, thereby making it possible to effectively filter inappropriate advertisements and update the generative model in real time based on user feedback.

[0137] "Advertising Content" means advertising information provided in the form of text, images, audio, or video.

[0138] A "generative model" is an algorithm trained based on official information, and is an artificial intelligence model used to detect inappropriate information.

[0139] "Feedback" refers to the opinions and ratings users provide about ads, which are used to update the model.

[0140] "Format" refers to the text, image, audio, or video form of the advertising content.

[0141] An "advertising filter application" is an application that has the functionality to capture, analyze, and filter advertising content and collect user feedback.

[0142] "Server" refers to a computer system that performs analysis and filtering of advertising content on a network and operates and updates generative models.

[0143] "API integration" means providing and using an application program interface to obtain and share advertising content with external services.

[0144] A "penalty" is a punishment imposed on advertisers who post inappropriate advertisements, and is a measure taken to maintain the reliability of the service.

[0145] "Smartphone application" means software that runs on a smartphone and acquires, analyzes, and filters advertising content.

[0146] The present invention is a system that acquires advertising content, analyzes it using a generative model, filters it, collects feedback from users, and updates the model. Specific embodiments of each element are described below.

[0147] System Configuration

[0148] server:

[0149] The server plays a central role in retrieving, analyzing, and filtering advertising content, as well as updating the generative model. The software uses Flask as the API server, and the Hugging Face transformers library is used to run the generative model. A computer system with sufficient computing power is recommended.

[0150] Device:

[0151] The user's smartphone runs an ad filter application, which sends API requests to the server and receives analysis results of ad content. It also collects user feedback via a feedback form and sends it to the server.

[0152] Acquiring and analyzing advertising content

[0153] 1. The server retrieves advertising content from the web page or app the user is viewing, which can be in the form of text, images, audio, or video.

[0154] 2. The acquired advertising content is sent to a generative model, which is trained in advance on official information and detects whether the content contains inappropriate content.

[0155] Filtering and Feedback

[0156] 1. The server filters advertising content based on the analysis results of the generative model. Advertising content deemed inappropriate is blocked, and only advertising content deemed appropriate is displayed to the user.

[0157] 2. Users can provide feedback on the ads they see. They can easily enter their opinions and ratings using the feedback form provided in the smartphone application.

[0158] Model Update

[0159] 1. The server stores the feedback collected from users and uses it to update the generative model. The feedback data is used to retrain the model, enabling it to analyze new data with high accuracy.

[0160] Specific examples

[0161] For example, a user browsing a news app can submit their opinion about an ad they see using a feedback form asking, "Do you think this ad is inappropriate?" This information is sent to the server and used as data for retraining the generative model.

[0162] Prompt Sentence Examples

[0163] "Analyze the text of this ad to determine if it contains inappropriate information. Ad text: 'Download your favorite characters now! 100% authentic!'"

[0164] This allows users to view more appropriate advertisements than before, improving the reliability of the service as a whole and user satisfaction.

[0165] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0166] Step 1:

[0167] The server obtains advertising content from web pages and apps viewed by users. Specifically, it downloads advertising content from specific URLs and data sources, converts it into text, images, audio, and video formats, and saves it. The input is the URL and data source information, and the output is the raw advertising content data.

[0168] Step 2:

[0169] The server sends the acquired advertising content to the generative model for analysis. The generative model uses a specific algorithm to detect whether the advertising content is inappropriate. The input is the raw data of the advertising content, and the output is the analysis result (appropriate / inappropriate).

[0170] Step 3:

[0171] The server filters the ad content based on the analysis results of the generative model. If inappropriate content is detected, the ad content is removed from the display list. The input is the analysis results of the generative model and the ad content itself, and the output is the filtered list.

[0172] Step 4:

[0173] The server sends the filtered advertising content to the terminal, which displays it to the user and saves the display results. The input is the filtered advertising content list, and the output is the advertisements displayed to the user.

[0174] Step 5:

[0175] Users provide feedback on the displayed advertisements by entering their opinions and ratings using a feedback form on their device and sending the data to the server. The input is the feedback content, and the output is the feedback data stored on the server.

[0176] Step 6:

[0177] The server stores the collected feedback data and uses it to retrain the generative model. New feedback data is used to update the generative model, improving the accuracy of the next analysis. The input is the feedback data, and the output is the updated generative model.

[0178] Step 7:

[0179] The server also connects with external services via API to provide advertising content filtering functionality. It retrieves external advertising content in response to API requests and analyzes it using a generative model. The input is advertising content information from external services, and the output is the analysis results and filtered advertising content.

[0180] Step 8:

[0181] The server imposes penalties on advertisers who have posted inappropriate ads. Based on the analysis results, it notifies the advertiser of a warning or penalty. The input is a list of ad content that is judged to be inappropriate, and the output is a warning or penalty notice to the advertiser.

[0182] 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.

[0183] The present invention combines a system that acquires advertising content, analyzes it using a generative model, filters it, collects user feedback, and updates the model with an emotion engine that recognizes user emotions. Specific embodiments of each element are described below.

[0184] The server retrieves advertising content from websites and video streaming services. The server sends an HTTP request from the specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[0185] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, where the analysis begins.

[0186] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[0187] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The emotion engine analyzes data acquired from the camera and microphone to recognize emotions from the user's facial expressions and tone of voice. The recognized emotion data is used in the filtering process of advertising content.

[0188] If a user feels that an ad is inappropriate, they can provide feedback using the device's feedback form. The user enters their opinion or problem with the ad and presses the submit button. This feedback also includes the user's emotional data.

[0189] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[0190] The server receives the feedback and adds the feedback data to the dataset for model update. The feedback data and the user's sentiment data are used as new training data to retrain the generative model. The retrained generative model is used for the next advertising analysis.

[0191] Furthermore, the server will provide ad checking functionality through API integration with other services. For example, the ad checking API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. Furthermore, penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0192] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it is not displayed; if it is determined to be appropriate, it is displayed. In addition, the accuracy of ad filtering is improved based on user emotion data, providing a more appropriate advertising experience for users.

[0193] In this way, the system of the present invention can effectively analyze and filter advertising content, and improve accuracy by incorporating user sentiment feedback, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0194] The processing flow will be explained below.

[0195] Step 1:

[0196] The server retrieves advertising content from websites and video streaming services via URLs, sending HTTP requests and downloading text, images, audio, and video data.

[0197] Step 2:

[0198] The server sends the acquired advertising content to a generative model for analysis. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The advertising data is passed to the model's analysis engine, and analysis begins.

[0199] Step 3:

[0200] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate. The server uses this information to decide what to do next.

[0201] Step 4:

[0202] The server filters the ad content based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it is displayed. The filtering results are recorded in a log.

[0203] Step 5:

[0204] While the ad is being displayed, the emotion engine recognizes the user's emotions in real time by analyzing data acquired from the device's camera and microphone, and recognizing emotions from the user's facial expressions and tone of voice.

[0205] Step 6:

[0206] The emotion engine recognizes the user's emotion data and sends it to the server. The emotion data is used for feedback and to evaluate the effectiveness of advertising displays.

[0207] Step 7:

[0208] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button. The feedback also includes the user's emotional data.

[0209] Step 8:

[0210] The device sends the user's feedback to the server. The device compiles the feedback data into an appropriate format, creates an HTTP request to the server, and sends it.

[0211] Step 9:

[0212] The server receives the feedback and adds it to the dataset for updating the generative model. The feedback data and the user's emotion data are used as new training data to retrain the generative model.

[0213] Step 10:

[0214] The server retrains the generative model to detect fake information and inappropriate ads with high accuracy. The retrained model is then used for the next ad analysis.

[0215] Step 11:

[0216] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[0217] Step 12:

[0218] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. Service fees are collected from companies that display appropriate ads.

[0219] As described above, this system acquires, analyzes, and filters advertising content, and utilizes user sentiment data to provide more accurate feedback and update the model, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0220] Example 2

[0221] 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."

[0222] Conventional ad delivery systems lacked sufficient means to prevent the display of inappropriate ads, resulting in a poor quality ad experience for users. Furthermore, they lacked a mechanism for collecting feedback that took user sentiment into account and improving the accuracy of ad filtering based on that feedback. As a result, users were unable to enjoy a safe and reliable ad experience. This led to issues such as a decline in the reliability of the service and user satisfaction.

[0223] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0224] In this invention, the server includes a means for acquiring advertising content, a means for analyzing the advertising content using a generative model, and a means for filtering the advertising content based on the analysis result, thereby enabling analysis of advertising content and blocking of inappropriate advertisements.

[0225] The server further includes means for recognizing user emotions, means for collecting user feedback, and means for updating the generative model using the collected feedback and emotion data, thereby improving the accuracy of advertisement filtering based on the user emotion data and providing a safer and more appropriate advertising experience for the user.

[0226] "Advertising content" refers to advertising information provided in the form of text, images, audio, video, etc.

[0227] A "generative model" is an artificial intelligence model that is trained based on official information and has an algorithm for analyzing the content of advertising content.

[0228] "Filtering" is a process of classifying advertising content into appropriate and inappropriate content based on the analysis results, and preventing inappropriate content from being displayed to the user.

[0229] The "emotion engine" is a system that analyzes the user's facial expressions, tone of voice, etc. to recognize the user's emotions in real time.

[0230] "Feedback" refers to the opinions and impressions provided by users regarding advertising content, as well as the emotional data recognized at that time.

[0231] "Model updating" is the process of retraining a generative model using collected feedback and emotion data to improve the model's accuracy.

[0232] "API integration" is a method of exchanging data with other services through an application programming interface.

[0233] A "penalty" is a sanction imposed on advertisers who distribute inappropriate ads.

[0234] The present invention provides a system for effectively analyzing and filtering advertising content, and improving accuracy by reflecting feedback based on user sentiment. Specific embodiments of each element are described below.

[0235] Acquiring advertising content

[0236] The server collects advertising content from websites and video streaming services. To do this, the server sends an HTTP request to a specified URL and downloads advertising data such as text, images, audio, and video. This data is then stored in the server's database.

[0237] Ad content analysis

[0238] The server sends the acquired advertising content to a generative AI model for content analysis. This generative AI model is trained based on official information and has algorithms to detect inappropriate content and fake information. The server then sends the advertising data to the model's API endpoint and receives the analysis results.

[0239] Ad content filtering

[0240] Based on the analysis results, the server filters the ad content. Specifically, inappropriate ads are blocked and only appropriate ads are displayed. The filtering results are also recorded.

[0241] Emotion recognition by emotion engine

[0242] While a user is viewing an ad, the emotion engine recognizes the user's emotions in real time. The emotion engine obtains data from the camera and microphone and analyzes the user's facial expressions and tone of voice to recognize emotions. This recognition data is used in the ad filtering process.

[0243] User feedback

[0244] If a user feels that an advertisement is inappropriate after viewing it, they can provide their opinion using a feedback form. The feedback includes the user's emotional data and is sent from the device to the server.

[0245] Send feedback and update the model

[0246] The device sends user feedback to the server, which receives it and adds it to the retraining dataset for the generative model, which the server uses to retrain the model and improve its accuracy.

[0247] API integration with other services

[0248] Furthermore, the server will provide an ad checking function through API integration with other services. Specifically, the API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. The server will also improve reliability by imposing penalties on advertisers who display inappropriate ads.

[0249] Specific examples

[0250] As a concrete example, consider a video streaming platform that sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative AI model and returns the results to the platform. If the ad is deemed inappropriate, it will not be displayed; if it is deemed appropriate, it will be displayed. Furthermore, the accuracy of ad filtering can be improved based on user emotional data, providing a more appropriate advertising experience for users.

[0251] Prompt Sentence Examples

[0252] "Please use a generative model to analyze whether this ad contains inappropriate information and return the results."

[0253] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0254] Step 1:

[0255] The server collects advertising content from websites and video streaming services. As input, a specified URL is given, and the server sends an HTTP GET request to that URL. As output, advertising data such as text, images, audio, and video are stored in the server's database.

[0256] Specific behavior:

[0257] The server sends a GET request to "https: / / example.com / ad".

[0258] The server analyzes the JSON response and saves it as advertising data.

[0259] Step 2:

[0260] The server sends the acquired advertising content to the generative AI model for content analysis. As input, the advertising data is sent as a POST request to the API endpoint of the generative AI model. As output, the analysis results are returned and stored on the server.

[0261] Specific behavior:

[0262] The server sends the ad data as a POST request to "https: / / ai-model-service / api / analyze".

[0263] The server receives and stores the analysis results returned by the model.

[0264] Step 3:

[0265] The server filters the ad content based on the analysis results. The analysis results are the input, and the filtering blocks inappropriate ads. The output is that only appropriate ads are listed and recorded for display.

[0266] Specific behavior:

[0267] The server analyzes the analysis results it receives.

[0268] If it is determined to be fake news, the advertising data will be deleted.

[0269] Appropriate ad data is stored in the "approved_ads" table.

[0270] Step 4:

[0271] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The input is data acquired from the device's camera and microphone, and the output is analyzed emotion data sent to the server.

[0272] Specific behavior:

[0273] The emotion engine acquires and analyzes camera data in real time.

[0274] If the user smiles, it is recorded as "positive" emotion data.

[0275] The recognized emotion data is sent to the server.

[0276] Step 5:

[0277] After viewing an advertisement, if a user feels it is inappropriate, he or she can provide his or her opinion using a feedback form. As input, the user's opinion or thoughts are entered into the feedback form, including emotional data. As output, the feedback data is sent to the server.

[0278] Specific behavior:

[0279] A user types "This ad is violent" into a feedback form on their device.

[0280] Pressing the Send Feedback button will create and send a request that also includes emotion data.

[0281] Step 6:

[0282] The device sends user feedback to the server. The server receives the feedback data as input and adds it to the retraining dataset for the generative model. The output is a retrained generative model.

[0283] Specific behavior:

[0284] The device posts the feedback data to "https: / / server / api / feedback".

[0285] The server adds the feedback to the dataset and retrains the AI ​​model.

[0286] Step 7:

[0287] The server interacts with other services via API to provide external analysis of advertising data. As input, there is an API request from the other service, the server analyzes the advertising data, and provides the results. As output, the analysis results are returned to the other service.

[0288] Specific behavior:

[0289] The video streaming service sends an API request to the server.

[0290] The server analyzes the advertising data using a generative AI model and sends the results back to the video streaming service.

[0291] (Application example 2)

[0292] 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."

[0293] Conventional ad filtering systems simply analyze the content of text and images to filter out inappropriate ads, but do not fully utilize user sentiment and feedback, resulting in a suboptimal ad experience for users. In particular, users may be presented with ads that they find annoying, resulting in a poor ad experience. Furthermore, the lack of a mechanism for imposing appropriate penalties on advertisers who place inappropriate ads also poses a problem of reduced reliability for the system as a whole.

[0294] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0295] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis results, means for recognizing user emotions, means for determining whether to display an advertisement based on the filtering results and user emotion data, means for collecting user feedback, and means for updating the generative model using the collected feedback. This makes it possible to utilize user emotion data to improve the accuracy of advertising filtering and provide a more appropriate advertising experience for users. Furthermore, by collecting feedback data and updating the generative model, the system is always kept up to date, minimizing the display of inappropriate advertisements.

[0296] "Advertising Content" means promotional material displayed within a website or application, including in the form of text, images, audio, video, or other content.

[0297] A "generative model" is an algorithm that uses artificial intelligence techniques to learn and generate, particularly for use in analyzing and filtering advertising content.

[0298] "Analysis" is the process of evaluating advertising content to determine whether it contains inappropriate information.

[0299] "Filtering" is the process of filtering out or blocking inappropriate advertising content based on analysis results.

[0300] "User" means an individual or organization that uses the system to view advertising content.

[0301] "Emotion recognition" is a technology that uses a camera and microphone to analyze a user's facial expressions and tone of voice to identify their emotions.

[0302] "Feedback" refers to information such as opinions, impressions, and evaluations that users input regarding advertising content.

[0303] "Model updating" is the process of retraining a generative model with newly collected data to improve its accuracy and functionality.

[0304] "API integration" is a technology that provides an interface for exchanging data and functions between different software systems.

[0305] A "penalty" is a sanction imposed on advertisers who post inappropriate advertisements, and may include restrictions on advertisement distribution.

[0306] A system for implementing this invention retrieves advertising content, analyzes and filters it using a generative model, recognizes user emotions, collects feedback, and updates the generative model to provide a more effective advertising experience.

[0307] As a specific embodiment, consider the following system: First, a server obtains advertising content from a website or application. The advertising content is downloaded from a specified URL using an HTTP request, and may include formats such as text, images, audio, and video.

[0308] The server then sends the acquired advertising content to a generative model, which analyzes the advertising content. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the generative model's analysis engine and begins analysis.

[0309] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[0310] Furthermore, while users are viewing ads, the emotion engine recognizes their emotions in real time. The emotion engine analyzes data acquired from the smartphone's camera and microphone to recognize emotions from the user's facial expressions and tone of voice. This emotion data is used in the filtering process and for future ad analysis.

[0311] If a user feels that an ad is inappropriate after viewing it, they can provide feedback through a feedback form on their smartphone. Users enter their opinions or concerns about the ad and press the submit button. This feedback also includes the user's emotional data.

[0312] The device sends the user's feedback to the server. The device then formats the feedback data appropriately and creates and sends an HTTP request to the server. The server receives the feedback and adds it to a dataset for updating the model. This dataset is then used to retrain the generative model and improve its accuracy.

[0313] The server also provides ad checking functionality through API integration with other services. For example, it will expose the ad checking API to platforms such as video streaming services and blogs, allowing external analysis of ad data. Penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0314] As an example of this system, consider the following prompt:

[0315] Get Ad: "Get ad content from URL"

[0316] Content analysis: "Analyze using generative AI models to determine appropriateness"

[0317] Emotion Recognition: "Recognize user emotions in real time from camera footage"

[0318] Feedback collection: "User opinions and sentiment data are sent as feedback."

[0319] In practicing this invention, it is recommended to use hardware and software such as smartphone cameras, microphones, generative AI models, API interfaces, EmotionRecognizers, and AdAnalyzers, as appropriate, to provide users with a safe and reliable advertising experience.

[0320] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0321] Step 1:

[0322] The server retrieves advertising content from a website or application by using an HTTP request to download data from a specified URL. The input is a URL, and the output is the advertising data, which can be text, images, audio, or video.

[0323] Step 2:

[0324] The server sends the acquired advertising content to the generative model, which analyzes the advertising content. The input is the advertising data, and the generative model analyzes it to determine whether it contains inappropriate content. The output is a label indicating whether the advertisement is appropriate or inappropriate.

[0325] Step 3:

[0326] The server filters advertising content based on the analysis results of the generative model. If an ad is deemed inappropriate, it is blocked. If an ad is deemed appropriate, it is displayed. The input is the result of the judgment in step 2, and the output is an instruction to block or display the ad.

[0327] Step 4:

[0328] The device uses an emotion engine to recognize the user's emotions while they are viewing an advertisement. The emotion engine analyzes data acquired from the smartphone's camera and microphone. The input is camera video and audio data, and the output is the user's emotions (e.g., joy, anger, sadness, etc.).

[0329] Step 5:

[0330] After watching an advertisement, a user provides feedback using a feedback form on the device. The user enters their opinion or impression about the advertisement into the form and presses the submit button. The input is the user's text feedback and emotion data, and the output is the feedback data sent to the server.

[0331] Step 6:

[0332] The server receives the feedback data sent from the terminal and adds the collected data to the update dataset for the generative model. The input is the feedback data, and the output is the training dataset for the generative model.

[0333] Step 7:

[0334] The server uses the collected feedback data to retrain the generative model and improve its accuracy: the input is an updated training dataset, and the output is an updated generative model.

[0335] Step 8:

[0336] The server provides ad checking functionality through API integration with other services. The platform sends an API request, and the server performs external analysis of the ad data. The input is ad data from the external platform, and the output is the analysis results. If inappropriate ad delivery is confirmed, the server imposes penalties on the advertiser.

[0337] 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.

[0338] 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.

[0339] 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.

[0340] [Second embodiment]

[0341] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0342] 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.

[0343] 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).

[0344] 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.

[0345] 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.

[0346] 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).

[0347] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0348] 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.

[0349] 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.

[0350] 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.

[0351] 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.

[0352] 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."

[0353] The present invention provides a system for acquiring advertising content, analyzing it using a generative model, filtering it, collecting feedback, and updating the model. Specific embodiments of each element are described below.

[0354] The server retrieves advertising content from websites or video streaming services. For example, it specifies a specific URL and downloads the advertising data from there. The retrieved data, which may be in the form of text, images, audio, or video, is then prepared for analysis by the generative model.

[0355] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The analysis results are used as an indicator to determine whether the advertisement is appropriate.

[0356] Based on the analysis of the ad content, the server decides whether to display it or not. If it is determined to contain inappropriate information, the ad is blocked and not shown to the user. On the other hand, if it is determined to be appropriate, the ad is displayed.

[0357] A dedicated UI is provided on the device to allow users to provide feedback on incorrectly displayed ads. Users fill out a feedback form, and this information is sent from the device to the server. This feedback data is used to update the generative model.

[0358] The server uses the collected feedback as part of model updates. By adding feedback data and retraining the generative model, it can detect fake information with high accuracy even on new data.

[0359] Furthermore, the server will provide other services with the ability to check advertising content through API integration. For example, it will provide ad checking functionality to platforms such as video streaming services and blogs, and analyze advertising data from outside. It will also improve the reliability of the entire system by imposing penalties on advertisers who deliver inappropriate ads.

[0360] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it will not be displayed; conversely, if it is determined to be appropriate, it will be displayed. Furthermore, appropriate feedback is provided to the advertiser to improve the model's performance.

[0361] As a result, the system of the present invention can effectively detect and block advertisements containing fake information, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0362] The processing flow will be explained below.

[0363] Step 1:

[0364] The server retrieves advertising content from websites and video streaming services. It sends an HTTP request from a specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[0365] Step 2:

[0366] The server sends the acquired advertising content to a generative model. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, which begins analysis.

[0367] Step 3:

[0368] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate, and uses this information to decide what to do next.

[0369] Step 4:

[0370] The server filters the ads based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the information is appropriate, the ad is displayed. The server records the filtering results in a log.

[0371] Step 5:

[0372] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button.

[0373] Step 6:

[0374] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[0375] Step 7:

[0376] The server receives the feedback and adds the feedback data to the dataset for model updates, which are then used as new training data to retrain the generative model.

[0377] Step 8:

[0378] The server retrains the generative model to detect fake information and inappropriate ads with higher accuracy. The retrained model is then used for the next ad analysis.

[0379] Step 9:

[0380] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[0381] Step 10:

[0382] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. In addition, advertisers who display appropriate ads are charged a service fee.

[0383] Through these steps, the system effectively analyzes and filters advertising content, providing users with a safe and reliable advertising experience.

[0384] Example 1

[0385] 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."

[0386] In today's digital advertising market, there are many advertisements containing fake information or inappropriate content, which not only impair the user experience but also reduce brand credibility. Conventional systems have had difficulty efficiently detecting and blocking these inappropriate advertisements. Furthermore, they lack mechanisms for effectively utilizing user feedback to improve ad quality. The present invention aims to solve these issues and provide a safe and reliable advertising experience.

[0387] 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.

[0388] In this invention, the server includes a device for acquiring advertising content, a device for analyzing the advertising content using a generative model, a device for filtering the advertising content based on the analysis results from the generative model, a device for collecting user feedback, and a device for updating the generative model using the collected feedback. This makes it possible to effectively detect and block advertisements containing fake information or inappropriate content. Furthermore, by reflecting user feedback, the accuracy of the generative model can be continuously improved, thereby providing a safer and more reliable advertising experience.

[0389] "Advertising content" refers to promotional information expressed in the form of text, images, audio, video, etc.

[0390] A "generative model" is an artificial intelligence model that uses an algorithm trained in advance based on official information to analyze and judge advertising content.

[0391] "Filtering" is the process of determining the appropriateness of advertising content and removing inappropriate advertisements based on the analysis results from the generative model.

[0392] "Feedback" refers to opinions and evaluation information provided by users regarding the display results of advertising content.

[0393] A "device" is a mechanism including hardware and software for performing each function of the system.

[0394] "API integration" is an interface that allows functions and data to be shared and linked between different software systems.

[0395] "Penalty" means a sanction or penalty imposed on an advertiser who places an inappropriate advertisement.

[0396] MODE FOR CARRYING OUT THE INVENTION

[0397] This system acquires advertising content, analyzes it using a generative model, filters it, collects feedback, and updates the model. This system consists of three main entities: a server, a terminal, and a user.

[0398] Server Features

[0399] The server acts as a device for acquiring advertising content, downloading advertising data from websites and video streaming services. Specifically, it uses the Python requests library to send HTTP requests to acquire advertising content. For example, it downloads advertising data from the specified URL https: / / example.com / ad and saves it in the form of text, images, audio, and video.

[0400] The server then sends the acquired advertising content to a generative model for analysis. This generative model is a generative AI model such as OpenAI's GPT, which determines the ad appropriateness by inputting specific prompts. The model is highly reliable because it is trained based on official information.

[0401] Specifically, the following prompt sentences are input to the generative AI model:

[0402] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[0403] Based on the analysis results of the generative model, the system decides whether to display an ad. If the ad contains inappropriate information, it is blocked, and if it is deemed appropriate, it is displayed.

[0404] User operations

[0405] The device is provided with a dedicated UI for users to provide feedback on advertisements. Users enter their opinions and ratings in a feedback form, and this information is sent from the device to the server. For example, a user may send feedback such as "This advertisement is misleading."

[0406] Server model update function

[0407] The server retrains the generative model based on the feedback collected from users. By adding this feedback data, the generative model improves its detection accuracy for new data. Retraining is the process of adding new feedback information to an existing dataset to improve the performance of the generative model.

[0408] API integration function

[0409] Furthermore, the server provides an API integration function, allowing it to obtain advertising content in cooperation with other services. This function allows it to link with external systems such as video streaming services and blog platforms to optimize advertising. It can also impose penalties on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0410] Specific examples

[0411] For example, a video streaming platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If an ad is determined to be inappropriate, it will not be displayed; only if it is appropriate will the ad be displayed. Appropriate feedback is also provided to the advertiser, contributing to improving the model's performance.

[0412] As a result, the system of the present invention can effectively detect and block advertisements containing fake information or inappropriate content. By using this system, it is possible to provide users with a safe and reliable advertising experience and improve the overall quality of services.

[0413] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0414] Processing Steps

[0415] Step 1:

[0416] The server retrieves advertising content from websites and video streaming services. Specifically, it uses the Python requests library to send an HTTP request and download the advertising data.

[0417] Input: Specific URL https: / / example.com / ad

[0418] Data processing: Sending HTTP requests, downloading advertising data

[0419] Output: Advertising data stored in the form of text, images, audio, and video

[0420] Specific operation: Specify a URL, send a request, and save the response data in local storage.

[0421] Step 2:

[0422] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained on public information and includes algorithms for detecting inappropriate content.

[0423] Input: Ad data obtained and saved in Step 1

[0424] Data processing: Formatting advertising data, creating prompts, and inputting them into the generative model

[0425] Output: Analysis results of advertisement content by generative model

[0426] Specific operation: Convert the ad data into text format, generate a prompt sentence like the following, and input it into the generative model:

[0427] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[0428] Step 3:

[0429] The server determines whether to display advertising content based on the analysis results from the generative model. If the content is deemed inappropriate, the ad is blocked, and if it is deemed appropriate, the ad is displayed to the user.

[0430] Input: Analysis results of the generative model

[0431] Data processing: Analysis of analysis results, filtering of advertisements deemed inappropriate

[0432] Output: Filtered ad content (displayable or not)

[0433] Specific operations: Analyze the analysis results, add inappropriate ads to a block list, and display appropriate ads on the user's screen.

[0434] Step 4:

[0435] Users provide feedback on advertising content through a dedicated UI, which is then sent from the device to the server.

[0436] Input: User feedback (e.g., "This ad is misleading")

[0437] Data Processing: Feedback data collection and transmission

[0438] Output: Feedback data is saved on the server

[0439] Specific operation: The user fills in the feedback form and presses the submit button. The data is sent from the device to the server.

[0440] Step 5:

[0441] The server retrains the generative model based on the collected feedback.

[0442] Input: Collected feedback data

[0443] Data processing: Integrating feedback data and retraining the generative model

[0444] Output: Updated generative model

[0445] What it does: Add the feedback data to the existing training dataset and retrain the generative model.

[0446] Step 6:

[0447] The server provides API integration functionality, connects with other services to acquire and analyze advertising content, and penalizes advertisers who deliver inappropriate ads.

[0448] Input: API request from external service

[0449] Data processing: Acquire external advertising data and analyze it with a generative model

[0450] Output: Sending analysis results back to external services, notifying inappropriate advertisers of penalties

[0451] Specific operation: Receives API requests from video streaming services, analyzes advertising data, returns analysis results to external services, and notifies advertisers of penalties if necessary.

[0452] (Application example 1)

[0453] 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."

[0454] In today's digital advertising environment, displaying inappropriate ads to users is a frequent problem. Displaying ads containing inappropriate content or fake information can ruin the user experience and reduce the overall reliability of the service. Another challenge is the inefficient collection of feedback and subsequent model updates. Furthermore, ads are presented in a variety of formats (text, image, audio, video), requiring appropriate filtering for each.

[0455] 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.

[0456] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis result, means for collecting feedback from users, means for updating the generative model using the collected feedback, and means for executing an advertising filter application that uses the collected feedback to update the model, thereby making it possible to effectively filter inappropriate advertisements and update the generative model in real time based on user feedback.

[0457] "Advertising Content" means advertising information provided in the form of text, images, audio, or video.

[0458] A "generative model" is an algorithm trained based on official information, and is an artificial intelligence model used to detect inappropriate information.

[0459] "Feedback" refers to the opinions and ratings users provide about ads, which are used to update the model.

[0460] "Format" refers to the text, image, audio, or video form of the advertising content.

[0461] An "advertising filter application" is an application that has the functionality to capture, analyze, and filter advertising content and collect user feedback.

[0462] "Server" refers to a computer system that performs analysis and filtering of advertising content on a network and operates and updates generative models.

[0463] "API integration" means providing and using an application program interface to obtain and share advertising content with external services.

[0464] A "penalty" is a punishment imposed on advertisers who post inappropriate advertisements, and is a measure taken to maintain the reliability of the service.

[0465] "Smartphone application" means software that runs on a smartphone and acquires, analyzes, and filters advertising content.

[0466] The present invention is a system that acquires advertising content, analyzes it using a generative model, filters it, collects feedback from users, and updates the model. Specific embodiments of each element are described below.

[0467] System Configuration

[0468] server:

[0469] The server plays a central role in retrieving, analyzing, and filtering advertising content, as well as updating the generative model. The software uses Flask as the API server, and the Hugging Face transformers library is used to run the generative model. A computer system with sufficient computing power is recommended.

[0470] Device:

[0471] The user's smartphone runs an ad filter application, which sends API requests to the server and receives analysis results of ad content. It also collects user feedback via a feedback form and sends it to the server.

[0472] Acquiring and analyzing advertising content

[0473] 1. The server retrieves advertising content from the web page or app the user is viewing, which can be in the form of text, images, audio, or video.

[0474] 2. The acquired advertising content is sent to a generative model, which is trained in advance on official information and detects whether the content contains inappropriate content.

[0475] Filtering and Feedback

[0476] 1. The server filters advertising content based on the analysis results of the generative model. Advertising content deemed inappropriate is blocked, and only advertising content deemed appropriate is displayed to the user.

[0477] 2. Users can provide feedback on the ads they see. They can easily enter their opinions and ratings using the feedback form provided in the smartphone application.

[0478] Model Update

[0479] 1. The server stores the feedback collected from users and uses it to update the generative model. The feedback data is used to retrain the model, enabling it to analyze new data with high accuracy.

[0480] Specific examples

[0481] For example, a user browsing a news app can submit their opinion about an ad they see using a feedback form asking, "Do you think this ad is inappropriate?" This information is sent to the server and used as data for retraining the generative model.

[0482] Prompt Sentence Examples

[0483] "Analyze the text of this ad to determine if it contains inappropriate information. Ad text: 'Download your favorite characters now! 100% authentic!'"

[0484] This allows users to view more appropriate advertisements than before, improving the reliability of the service as a whole and user satisfaction.

[0485] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0486] Step 1:

[0487] The server obtains advertising content from web pages and apps viewed by users. Specifically, it downloads advertising content from specific URLs and data sources, converts it into text, images, audio, and video formats, and saves it. The input is the URL and data source information, and the output is the raw advertising content data.

[0488] Step 2:

[0489] The server sends the acquired advertising content to the generative model for analysis. The generative model uses a specific algorithm to detect whether the advertising content is inappropriate. The input is the raw data of the advertising content, and the output is the analysis result (appropriate / inappropriate).

[0490] Step 3:

[0491] The server filters the ad content based on the analysis results of the generative model. If inappropriate content is detected, the ad content is removed from the display list. The input is the analysis results of the generative model and the ad content itself, and the output is the filtered list.

[0492] Step 4:

[0493] The server sends the filtered advertising content to the terminal, which displays it to the user and saves the display results. The input is the filtered advertising content list, and the output is the advertisements displayed to the user.

[0494] Step 5:

[0495] Users provide feedback on the displayed advertisements by entering their opinions and ratings using a feedback form on their device and sending the data to the server. The input is the feedback content, and the output is the feedback data stored on the server.

[0496] Step 6:

[0497] The server stores the collected feedback data and uses it to retrain the generative model. New feedback data is used to update the generative model, improving the accuracy of the next analysis. The input is the feedback data, and the output is the updated generative model.

[0498] Step 7:

[0499] The server also connects with external services via API to provide advertising content filtering functionality. It retrieves external advertising content in response to API requests and analyzes it using a generative model. The input is advertising content information from external services, and the output is the analysis results and filtered advertising content.

[0500] Step 8:

[0501] The server imposes penalties on advertisers who have posted inappropriate ads. Based on the analysis results, it notifies the advertiser of a warning or penalty. The input is a list of ad content that is judged to be inappropriate, and the output is a warning or penalty notice to the advertiser.

[0502] 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.

[0503] The present invention combines a system that acquires advertising content, analyzes it using a generative model, filters it, collects user feedback, and updates the model with an emotion engine that recognizes user emotions. Specific embodiments of each element are described below.

[0504] The server retrieves advertising content from websites and video streaming services. The server sends an HTTP request from the specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[0505] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, where the analysis begins.

[0506] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[0507] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The emotion engine analyzes data acquired from the camera and microphone to recognize emotions from the user's facial expressions and tone of voice. The recognized emotion data is used in the filtering process of advertising content.

[0508] If a user feels that an ad is inappropriate, they can provide feedback using the device's feedback form. The user enters their opinion or problem with the ad and presses the submit button. This feedback also includes the user's emotional data.

[0509] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[0510] The server receives the feedback and adds the feedback data to the dataset for model update. The feedback data and the user's sentiment data are used as new training data to retrain the generative model. The retrained generative model is used for the next advertising analysis.

[0511] Furthermore, the server will provide ad checking functionality through API integration with other services. For example, the ad checking API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. Furthermore, penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0512] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it is not displayed; if it is determined to be appropriate, it is displayed. In addition, the accuracy of ad filtering is improved based on user emotion data, providing a more appropriate advertising experience for users.

[0513] In this way, the system of the present invention can effectively analyze and filter advertising content, and improve accuracy by incorporating user sentiment feedback, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0514] The processing flow will be explained below.

[0515] Step 1:

[0516] The server retrieves advertising content from websites and video streaming services via URLs, sending HTTP requests and downloading text, images, audio, and video data.

[0517] Step 2:

[0518] The server sends the acquired advertising content to a generative model for analysis. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The advertising data is passed to the model's analysis engine, and analysis begins.

[0519] Step 3:

[0520] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate. The server uses this information to decide what to do next.

[0521] Step 4:

[0522] The server filters the ad content based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it is displayed. The filtering results are recorded in a log.

[0523] Step 5:

[0524] While the ad is being displayed, the emotion engine recognizes the user's emotions in real time by analyzing data acquired from the device's camera and microphone, and recognizing emotions from the user's facial expressions and tone of voice.

[0525] Step 6:

[0526] The emotion engine recognizes the user's emotion data and sends it to the server. The emotion data is used for feedback and to evaluate the effectiveness of advertising displays.

[0527] Step 7:

[0528] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button. The feedback also includes the user's emotional data.

[0529] Step 8:

[0530] The device sends the user's feedback to the server. The device compiles the feedback data into an appropriate format, creates an HTTP request to the server, and sends it.

[0531] Step 9:

[0532] The server receives the feedback and adds it to the dataset for updating the generative model. The feedback data and the user's emotion data are used as new training data to retrain the generative model.

[0533] Step 10:

[0534] The server retrains the generative model to detect fake information and inappropriate ads with high accuracy. The retrained model is then used for the next ad analysis.

[0535] Step 11:

[0536] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[0537] Step 12:

[0538] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. Service fees are collected from companies that display appropriate ads.

[0539] As described above, this system acquires, analyzes, and filters advertising content, and utilizes user sentiment data to provide more accurate feedback and update the model, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0540] Example 2

[0541] 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."

[0542] Conventional ad delivery systems lacked sufficient means to prevent the display of inappropriate ads, resulting in a poor quality ad experience for users. Furthermore, they lacked a mechanism for collecting feedback that took user sentiment into account and improving the accuracy of ad filtering based on that feedback. As a result, users were unable to enjoy a safe and reliable ad experience. This led to issues such as a decline in the reliability of the service and user satisfaction.

[0543] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0544] In this invention, the server includes a means for acquiring advertising content, a means for analyzing the advertising content using a generative model, and a means for filtering the advertising content based on the analysis result, thereby enabling analysis of advertising content and blocking of inappropriate advertisements.

[0545] The server further includes means for recognizing user emotions, means for collecting user feedback, and means for updating the generative model using the collected feedback and emotion data, thereby improving the accuracy of advertisement filtering based on the user emotion data and providing a safer and more appropriate advertising experience for the user.

[0546] "Advertising content" refers to advertising information provided in the form of text, images, audio, video, etc.

[0547] A "generative model" is an artificial intelligence model that is trained based on official information and has an algorithm for analyzing the content of advertising content.

[0548] "Filtering" is a process of classifying advertising content into appropriate and inappropriate content based on the analysis results, and preventing inappropriate content from being displayed to the user.

[0549] The "emotion engine" is a system that analyzes the user's facial expressions, tone of voice, etc. to recognize the user's emotions in real time.

[0550] "Feedback" refers to the opinions and impressions provided by users regarding advertising content, as well as the emotional data recognized at that time.

[0551] "Model updating" is the process of retraining a generative model using collected feedback and emotion data to improve the model's accuracy.

[0552] "API integration" is a method of exchanging data with other services through an application programming interface.

[0553] A "penalty" is a sanction imposed on advertisers who distribute inappropriate ads.

[0554] The present invention provides a system for effectively analyzing and filtering advertising content, and improving accuracy by reflecting feedback based on user sentiment. Specific embodiments of each element are described below.

[0555] Acquiring advertising content

[0556] The server collects advertising content from websites and video streaming services. To do this, the server sends an HTTP request to a specified URL and downloads advertising data such as text, images, audio, and video. This data is then stored in the server's database.

[0557] Ad content analysis

[0558] The server sends the acquired advertising content to a generative AI model for content analysis. This generative AI model is trained based on official information and has algorithms to detect inappropriate content and fake information. The server then sends the advertising data to the model's API endpoint and receives the analysis results.

[0559] Ad content filtering

[0560] Based on the analysis results, the server filters the ad content. Specifically, inappropriate ads are blocked and only appropriate ads are displayed. The filtering results are also recorded.

[0561] Emotion recognition by emotion engine

[0562] While a user is viewing an ad, the emotion engine recognizes the user's emotions in real time. The emotion engine obtains data from the camera and microphone and analyzes the user's facial expressions and tone of voice to recognize emotions. This recognition data is used in the ad filtering process.

[0563] User feedback

[0564] If a user feels that an advertisement is inappropriate after viewing it, they can provide their opinion using a feedback form. The feedback includes the user's emotional data and is sent from the device to the server.

[0565] Send feedback and update the model

[0566] The device sends user feedback to the server, which receives it and adds it to the retraining dataset for the generative model, which the server uses to retrain the model and improve its accuracy.

[0567] API integration with other services

[0568] Furthermore, the server will provide an ad checking function through API integration with other services. Specifically, the API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. The server will also improve reliability by imposing penalties on advertisers who display inappropriate ads.

[0569] Specific examples

[0570] As a concrete example, consider a video streaming platform that sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative AI model and returns the results to the platform. If the ad is deemed inappropriate, it will not be displayed; if it is deemed appropriate, it will be displayed. Furthermore, the accuracy of ad filtering can be improved based on user emotional data, providing a more appropriate advertising experience for users.

[0571] Prompt Sentence Examples

[0572] "Please use a generative model to analyze whether this ad contains inappropriate information and return the results."

[0573] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0574] Step 1:

[0575] The server collects advertising content from websites and video streaming services. As input, a specified URL is given, and the server sends an HTTP GET request to that URL. As output, advertising data such as text, images, audio, and video are stored in the server's database.

[0576] Specific behavior:

[0577] The server sends a GET request to "https: / / example.com / ad".

[0578] The server analyzes the JSON response and saves it as advertising data.

[0579] Step 2:

[0580] The server sends the acquired advertising content to the generative AI model for content analysis. As input, the advertising data is sent as a POST request to the API endpoint of the generative AI model. As output, the analysis results are returned and stored on the server.

[0581] Specific behavior:

[0582] The server sends the ad data as a POST request to "https: / / ai-model-service / api / analyze".

[0583] The server receives and stores the analysis results returned by the model.

[0584] Step 3:

[0585] The server filters the ad content based on the analysis results. The analysis results are the input, and the filtering blocks inappropriate ads. The output is that only appropriate ads are listed and recorded for display.

[0586] Specific behavior:

[0587] The server analyzes the analysis results it receives.

[0588] If it is determined to be fake news, the advertising data will be deleted.

[0589] Appropriate ad data is stored in the "approved_ads" table.

[0590] Step 4:

[0591] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The input is data acquired from the device's camera and microphone, and the output is analyzed emotion data sent to the server.

[0592] Specific behavior:

[0593] The emotion engine acquires and analyzes camera data in real time.

[0594] If the user smiles, it is recorded as "positive" emotion data.

[0595] The recognized emotion data is sent to the server.

[0596] Step 5:

[0597] After viewing an advertisement, if a user feels it is inappropriate, he or she can provide his or her opinion using a feedback form. As input, the user's opinion or thoughts are entered into the feedback form, including emotional data. As output, the feedback data is sent to the server.

[0598] Specific behavior:

[0599] A user types "This ad is violent" into a feedback form on their device.

[0600] Pressing the Send Feedback button will create and send a request that also includes emotion data.

[0601] Step 6:

[0602] The device sends user feedback to the server. The server receives the feedback data as input and adds it to the retraining dataset for the generative model. The output is a retrained generative model.

[0603] Specific behavior:

[0604] The device posts the feedback data to "https: / / server / api / feedback".

[0605] The server adds the feedback to the dataset and retrains the AI ​​model.

[0606] Step 7:

[0607] The server interacts with other services via API to provide external analysis of advertising data. As input, there is an API request from the other service, the server analyzes the advertising data, and provides the results. As output, the analysis results are returned to the other service.

[0608] Specific behavior:

[0609] The video streaming service sends an API request to the server.

[0610] The server analyzes the advertising data using a generative AI model and sends the results back to the video streaming service.

[0611] (Application example 2)

[0612] 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."

[0613] Conventional ad filtering systems simply analyze the content of text and images to filter out inappropriate ads, but do not fully utilize user sentiment and feedback, resulting in a suboptimal ad experience for users. In particular, users may be presented with ads that they find annoying, resulting in a poor ad experience. Furthermore, the lack of a mechanism for imposing appropriate penalties on advertisers who place inappropriate ads also poses a problem of reduced reliability for the system as a whole.

[0614] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0615] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis results, means for recognizing user emotions, means for determining whether to display an advertisement based on the filtering results and user emotion data, means for collecting user feedback, and means for updating the generative model using the collected feedback. This makes it possible to utilize user emotion data to improve the accuracy of advertising filtering and provide a more appropriate advertising experience for users. Furthermore, by collecting feedback data and updating the generative model, the system is always kept up to date, minimizing the display of inappropriate advertisements.

[0616] "Advertising Content" means promotional material displayed within a website or application, including in the form of text, images, audio, video, or other content.

[0617] A "generative model" is an algorithm that uses artificial intelligence techniques to learn and generate, particularly for use in analyzing and filtering advertising content.

[0618] "Analysis" is the process of evaluating advertising content to determine whether it contains inappropriate information.

[0619] "Filtering" is the process of filtering out or blocking inappropriate advertising content based on analysis results.

[0620] "User" means an individual or organization that uses the system to view advertising content.

[0621] "Emotion recognition" is a technology that uses a camera and microphone to analyze a user's facial expressions and tone of voice to identify their emotions.

[0622] "Feedback" refers to information such as opinions, impressions, and evaluations that users input regarding advertising content.

[0623] "Model updating" is the process of retraining a generative model with newly collected data to improve its accuracy and functionality.

[0624] "API integration" is a technology that provides an interface for exchanging data and functions between different software systems.

[0625] A "penalty" is a sanction imposed on advertisers who post inappropriate advertisements, and may include restrictions on advertisement distribution.

[0626] A system for implementing this invention retrieves advertising content, analyzes and filters it using a generative model, recognizes user emotions, collects feedback, and updates the generative model to provide a more effective advertising experience.

[0627] As a specific embodiment, consider the following system: First, a server obtains advertising content from a website or application. The advertising content is downloaded from a specified URL using an HTTP request, and may include formats such as text, images, audio, and video.

[0628] The server then sends the acquired advertising content to a generative model, which analyzes the advertising content. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the generative model's analysis engine and begins analysis.

[0629] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[0630] Furthermore, while users are viewing ads, the emotion engine recognizes their emotions in real time. The emotion engine analyzes data acquired from the smartphone's camera and microphone to recognize emotions from the user's facial expressions and tone of voice. This emotion data is used in the filtering process and for future ad analysis.

[0631] If a user feels that an ad is inappropriate after viewing it, they can provide feedback through a feedback form on their smartphone. Users enter their opinions or concerns about the ad and press the submit button. This feedback also includes the user's emotional data.

[0632] The device sends the user's feedback to the server. The device then formats the feedback data appropriately and creates and sends an HTTP request to the server. The server receives the feedback and adds it to a dataset for updating the model. This dataset is then used to retrain the generative model and improve its accuracy.

[0633] The server also provides ad checking functionality through API integration with other services. For example, it will expose the ad checking API to platforms such as video streaming services and blogs, allowing external analysis of ad data. Penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0634] As an example of this system, consider the following prompt:

[0635] Get Ad: "Get ad content from URL"

[0636] Content analysis: "Analyze using generative AI models to determine appropriateness"

[0637] Emotion Recognition: "Recognize user emotions in real time from camera footage"

[0638] Feedback collection: "User opinions and sentiment data are sent as feedback."

[0639] In practicing this invention, it is recommended to use hardware and software such as smartphone cameras, microphones, generative AI models, API interfaces, EmotionRecognizers, and AdAnalyzers, as appropriate, to provide users with a safe and reliable advertising experience.

[0640] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0641] Step 1:

[0642] The server retrieves advertising content from a website or application by using an HTTP request to download data from a specified URL. The input is a URL, and the output is the advertising data, which can be text, images, audio, or video.

[0643] Step 2:

[0644] The server sends the acquired advertising content to the generative model, which analyzes the advertising content. The input is the advertising data, and the generative model analyzes it to determine whether it contains inappropriate content. The output is a label indicating whether the advertisement is appropriate or inappropriate.

[0645] Step 3:

[0646] The server filters advertising content based on the analysis results of the generative model. If an ad is deemed inappropriate, it is blocked. If an ad is deemed appropriate, it is displayed. The input is the result of the judgment in step 2, and the output is an instruction to block or display the ad.

[0647] Step 4:

[0648] The device uses an emotion engine to recognize the user's emotions while they are viewing an advertisement. The emotion engine analyzes data acquired from the smartphone's camera and microphone. The input is camera video and audio data, and the output is the user's emotions (e.g., joy, anger, sadness, etc.).

[0649] Step 5:

[0650] After watching an advertisement, a user provides feedback using a feedback form on the device. The user enters their opinion or impression about the advertisement into the form and presses the submit button. The input is the user's text feedback and emotion data, and the output is the feedback data sent to the server.

[0651] Step 6:

[0652] The server receives the feedback data sent from the terminal and adds the collected data to the update dataset for the generative model. The input is the feedback data, and the output is the training dataset for the generative model.

[0653] Step 7:

[0654] The server uses the collected feedback data to retrain the generative model and improve its accuracy: the input is an updated training dataset, and the output is an updated generative model.

[0655] Step 8:

[0656] The server provides ad checking functionality through API integration with other services. The platform sends an API request, and the server performs external analysis of the ad data. The input is ad data from the external platform, and the output is the analysis results. If inappropriate ad delivery is confirmed, the server imposes penalties on the advertiser.

[0657] 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.

[0658] 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.

[0659] 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.

[0660] [Third embodiment]

[0661] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0662] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0663] 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).

[0664] 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.

[0665] 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.

[0666] 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).

[0667] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0668] 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.

[0669] 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.

[0670] 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.

[0671] 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.

[0672] 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."

[0673] The present invention provides a system for acquiring advertising content, analyzing it using a generative model, filtering it, collecting feedback, and updating the model. Specific embodiments of each element are described below.

[0674] The server retrieves advertising content from websites or video streaming services. For example, it specifies a specific URL and downloads the advertising data from there. The retrieved data, which may be in the form of text, images, audio, or video, is then prepared for analysis by the generative model.

[0675] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The analysis results are used as an indicator to determine whether the advertisement is appropriate.

[0676] Based on the analysis of the ad content, the server decides whether to display it or not. If it is determined to contain inappropriate information, the ad is blocked and not shown to the user. On the other hand, if it is determined to be appropriate, the ad is displayed.

[0677] A dedicated UI is provided on the device to allow users to provide feedback on incorrectly displayed ads. Users fill out a feedback form, and this information is sent from the device to the server. This feedback data is used to update the generative model.

[0678] The server uses the collected feedback as part of model updates. By adding feedback data and retraining the generative model, it can detect fake information with high accuracy even on new data.

[0679] Furthermore, the server will provide other services with the ability to check advertising content through API integration. For example, it will provide ad checking functionality to platforms such as video streaming services and blogs, and analyze advertising data from outside. It will also improve the reliability of the entire system by imposing penalties on advertisers who deliver inappropriate ads.

[0680] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it will not be displayed; conversely, if it is determined to be appropriate, it will be displayed. Furthermore, appropriate feedback is provided to the advertiser to improve the model's performance.

[0681] As a result, the system of the present invention can effectively detect and block advertisements containing fake information, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0682] The processing flow will be explained below.

[0683] Step 1:

[0684] The server retrieves advertising content from websites and video streaming services. It sends an HTTP request from a specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[0685] Step 2:

[0686] The server sends the acquired advertising content to a generative model. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, which begins analysis.

[0687] Step 3:

[0688] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate, and uses this information to decide what to do next.

[0689] Step 4:

[0690] The server filters the ads based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the information is appropriate, the ad is displayed. The server records the filtering results in a log.

[0691] Step 5:

[0692] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button.

[0693] Step 6:

[0694] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[0695] Step 7:

[0696] The server receives the feedback and adds the feedback data to the dataset for model updates, which are then used as new training data to retrain the generative model.

[0697] Step 8:

[0698] The server retrains the generative model to detect fake information and inappropriate ads with higher accuracy. The retrained model is then used for the next ad analysis.

[0699] Step 9:

[0700] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[0701] Step 10:

[0702] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. In addition, advertisers who display appropriate ads are charged a service fee.

[0703] Through these steps, the system effectively analyzes and filters advertising content, providing users with a safe and reliable advertising experience.

[0704] Example 1

[0705] 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."

[0706] In today's digital advertising market, there are many advertisements containing fake information or inappropriate content, which not only impair the user experience but also reduce brand credibility. Conventional systems have had difficulty efficiently detecting and blocking these inappropriate advertisements. Furthermore, they lack mechanisms for effectively utilizing user feedback to improve ad quality. The present invention aims to solve these issues and provide a safe and reliable advertising experience.

[0707] 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.

[0708] In this invention, the server includes a device for acquiring advertising content, a device for analyzing the advertising content using a generative model, a device for filtering the advertising content based on the analysis results from the generative model, a device for collecting user feedback, and a device for updating the generative model using the collected feedback. This makes it possible to effectively detect and block advertisements containing fake information or inappropriate content. Furthermore, by reflecting user feedback, the accuracy of the generative model can be continuously improved, thereby providing a safer and more reliable advertising experience.

[0709] "Advertising content" refers to promotional information expressed in the form of text, images, audio, video, etc.

[0710] A "generative model" is an artificial intelligence model that uses an algorithm trained in advance based on official information to analyze and judge advertising content.

[0711] "Filtering" is the process of determining the appropriateness of advertising content and removing inappropriate advertisements based on the analysis results from the generative model.

[0712] "Feedback" refers to opinions and evaluation information provided by users regarding the display results of advertising content.

[0713] A "device" is a mechanism including hardware and software for performing each function of the system.

[0714] "API integration" is an interface that allows functions and data to be shared and linked between different software systems.

[0715] "Penalty" means a sanction or penalty imposed on an advertiser who places an inappropriate advertisement.

[0716] MODE FOR CARRYING OUT THE INVENTION

[0717] This system acquires advertising content, analyzes it using a generative model, filters it, collects feedback, and updates the model. This system consists of three main entities: a server, a terminal, and a user.

[0718] Server Features

[0719] The server acts as a device for acquiring advertising content, downloading advertising data from websites and video streaming services. Specifically, it uses the Python requests library to send HTTP requests to acquire advertising content. For example, it downloads advertising data from the specified URL https: / / example.com / ad and saves it in the form of text, images, audio, and video.

[0720] The server then sends the acquired advertising content to a generative model for analysis. This generative model is a generative AI model such as OpenAI's GPT, which determines the ad appropriateness by inputting specific prompts. The model is highly reliable because it is trained based on official information.

[0721] Specifically, the following prompt sentences are input to the generative AI model:

[0722] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[0723] Based on the analysis results of the generative model, the system decides whether to display an ad. If the ad contains inappropriate information, it is blocked, and if it is deemed appropriate, it is displayed.

[0724] User operations

[0725] The device is provided with a dedicated UI for users to provide feedback on advertisements. Users enter their opinions and ratings in a feedback form, and this information is sent from the device to the server. For example, a user may send feedback such as "This advertisement is misleading."

[0726] Server model update function

[0727] The server retrains the generative model based on the feedback collected from users. By adding this feedback data, the generative model improves its detection accuracy for new data. Retraining is the process of adding new feedback information to an existing dataset to improve the performance of the generative model.

[0728] API integration function

[0729] Furthermore, the server provides an API integration function, allowing it to obtain advertising content in cooperation with other services. This function allows it to link with external systems such as video streaming services and blog platforms to optimize advertising. It can also impose penalties on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0730] Specific examples

[0731] For example, a video streaming platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If an ad is determined to be inappropriate, it will not be displayed; only if it is appropriate will the ad be displayed. Appropriate feedback is also provided to the advertiser, contributing to improving the model's performance.

[0732] As a result, the system of the present invention can effectively detect and block advertisements containing fake information or inappropriate content. By using this system, it is possible to provide users with a safe and reliable advertising experience and improve the overall quality of services.

[0733] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0734] Processing Steps

[0735] Step 1:

[0736] The server retrieves advertising content from websites and video streaming services. Specifically, it uses the Python requests library to send an HTTP request and download the advertising data.

[0737] Input: Specific URL https: / / example.com / ad

[0738] Data processing: Sending HTTP requests, downloading advertising data

[0739] Output: Advertising data stored in the form of text, images, audio, and video

[0740] Specific operation: Specify a URL, send a request, and save the response data in local storage.

[0741] Step 2:

[0742] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained on public information and includes algorithms for detecting inappropriate content.

[0743] Input: Ad data obtained and saved in Step 1

[0744] Data processing: Formatting advertising data, creating prompts, and inputting them into the generative model

[0745] Output: Analysis results of advertisement content by generative model

[0746] Specific operation: Convert the ad data into text format, generate a prompt sentence like the following, and input it into the generative model:

[0747] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[0748] Step 3:

[0749] The server determines whether to display advertising content based on the analysis results from the generative model. If the content is deemed inappropriate, the ad is blocked, and if it is deemed appropriate, the ad is displayed to the user.

[0750] Input: Analysis results of the generative model

[0751] Data processing: Analysis of analysis results, filtering of advertisements deemed inappropriate

[0752] Output: Filtered ad content (displayable or not)

[0753] Specific operations: Analyze the analysis results, add inappropriate ads to a block list, and display appropriate ads on the user's screen.

[0754] Step 4:

[0755] Users provide feedback on advertising content through a dedicated UI, which is then sent from the device to the server.

[0756] Input: User feedback (e.g., "This ad is misleading")

[0757] Data Processing: Feedback data collection and transmission

[0758] Output: Feedback data is saved on the server

[0759] Specific operation: The user fills in the feedback form and presses the submit button. The data is sent from the device to the server.

[0760] Step 5:

[0761] The server retrains the generative model based on the collected feedback.

[0762] Input: Collected feedback data

[0763] Data processing: Integrating feedback data and retraining the generative model

[0764] Output: Updated generative model

[0765] What it does: Add the feedback data to the existing training dataset and retrain the generative model.

[0766] Step 6:

[0767] The server provides API integration functionality, connects with other services to acquire and analyze advertising content, and penalizes advertisers who deliver inappropriate ads.

[0768] Input: API request from external service

[0769] Data processing: Acquire external advertising data and analyze it with a generative model

[0770] Output: Sending analysis results back to external services, notifying inappropriate advertisers of penalties

[0771] Specific operation: Receives API requests from video streaming services, analyzes advertising data, returns analysis results to external services, and notifies advertisers of penalties if necessary.

[0772] (Application example 1)

[0773] 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."

[0774] In today's digital advertising environment, displaying inappropriate ads to users is a frequent problem. Displaying ads containing inappropriate content or fake information can ruin the user experience and reduce the overall reliability of the service. Another challenge is the inefficient collection of feedback and subsequent model updates. Furthermore, ads are presented in a variety of formats (text, image, audio, video), requiring appropriate filtering for each.

[0775] 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.

[0776] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis result, means for collecting feedback from users, means for updating the generative model using the collected feedback, and means for executing an advertising filter application that uses the collected feedback to update the model, thereby making it possible to effectively filter inappropriate advertisements and update the generative model in real time based on user feedback.

[0777] "Advertising Content" means advertising information provided in the form of text, images, audio, or video.

[0778] A "generative model" is an algorithm trained based on official information, and is an artificial intelligence model used to detect inappropriate information.

[0779] "Feedback" refers to the opinions and ratings users provide about ads, which are used to update the model.

[0780] "Format" refers to the text, image, audio, or video form of the advertising content.

[0781] An "advertising filter application" is an application that has the functionality to capture, analyze, and filter advertising content and collect user feedback.

[0782] "Server" refers to a computer system that performs analysis and filtering of advertising content on a network and operates and updates generative models.

[0783] "API integration" means providing and using an application program interface to obtain and share advertising content with external services.

[0784] A "penalty" is a punishment imposed on advertisers who post inappropriate advertisements, and is a measure taken to maintain the reliability of the service.

[0785] "Smartphone application" means software that runs on a smartphone and acquires, analyzes, and filters advertising content.

[0786] The present invention is a system that acquires advertising content, analyzes it using a generative model, filters it, collects feedback from users, and updates the model. Specific embodiments of each element are described below.

[0787] System Configuration

[0788] server:

[0789] The server plays a central role in retrieving, analyzing, and filtering advertising content, as well as updating the generative model. The software uses Flask as the API server, and the Hugging Face transformers library is used to run the generative model. A computer system with sufficient computing power is recommended.

[0790] Device:

[0791] The user's smartphone runs an ad filter application, which sends API requests to the server and receives analysis results of ad content. It also collects user feedback via a feedback form and sends it to the server.

[0792] Acquiring and analyzing advertising content

[0793] 1. The server retrieves advertising content from the web page or app the user is viewing, which can be in the form of text, images, audio, or video.

[0794] 2. The acquired advertising content is sent to a generative model, which is trained in advance on official information and detects whether the content contains inappropriate content.

[0795] Filtering and Feedback

[0796] 1. The server filters advertising content based on the analysis results of the generative model. Advertising content deemed inappropriate is blocked, and only advertising content deemed appropriate is displayed to the user.

[0797] 2. Users can provide feedback on the ads they see. They can easily enter their opinions and ratings using the feedback form provided in the smartphone application.

[0798] Model Update

[0799] 1. The server stores the feedback collected from users and uses it to update the generative model. The feedback data is used to retrain the model, enabling it to analyze new data with high accuracy.

[0800] Specific examples

[0801] For example, a user browsing a news app can submit their opinion about an ad they see using a feedback form asking, "Do you think this ad is inappropriate?" This information is sent to the server and used as data for retraining the generative model.

[0802] Prompt Sentence Examples

[0803] "Analyze the text of this ad to determine if it contains inappropriate information. Ad text: 'Download your favorite characters now! 100% authentic!'"

[0804] This allows users to view more appropriate advertisements than before, improving the reliability of the service as a whole and user satisfaction.

[0805] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0806] Step 1:

[0807] The server obtains advertising content from web pages and apps viewed by users. Specifically, it downloads advertising content from specific URLs and data sources, converts it into text, images, audio, and video formats, and saves it. The input is the URL and data source information, and the output is the raw advertising content data.

[0808] Step 2:

[0809] The server sends the acquired advertising content to the generative model for analysis. The generative model uses a specific algorithm to detect whether the advertising content is inappropriate. The input is the raw data of the advertising content, and the output is the analysis result (appropriate / inappropriate).

[0810] Step 3:

[0811] The server filters the ad content based on the analysis results of the generative model. If inappropriate content is detected, the ad content is removed from the display list. The input is the analysis results of the generative model and the ad content itself, and the output is the filtered list.

[0812] Step 4:

[0813] The server sends the filtered advertising content to the terminal, which displays it to the user and saves the display results. The input is the filtered advertising content list, and the output is the advertisements displayed to the user.

[0814] Step 5:

[0815] Users provide feedback on the displayed advertisements by entering their opinions and ratings using a feedback form on their device and sending the data to the server. The input is the feedback content, and the output is the feedback data stored on the server.

[0816] Step 6:

[0817] The server stores the collected feedback data and uses it to retrain the generative model. New feedback data is used to update the generative model, improving the accuracy of the next analysis. The input is the feedback data, and the output is the updated generative model.

[0818] Step 7:

[0819] The server also connects with external services via API to provide advertising content filtering functionality. It retrieves external advertising content in response to API requests and analyzes it using a generative model. The input is advertising content information from external services, and the output is the analysis results and filtered advertising content.

[0820] Step 8:

[0821] The server imposes penalties on advertisers who have posted inappropriate ads. Based on the analysis results, it notifies the advertiser of a warning or penalty. The input is a list of ad content that is judged to be inappropriate, and the output is a warning or penalty notice to the advertiser.

[0822] 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.

[0823] The present invention combines a system that acquires advertising content, analyzes it using a generative model, filters it, collects user feedback, and updates the model with an emotion engine that recognizes user emotions. Specific embodiments of each element are described below.

[0824] The server retrieves advertising content from websites and video streaming services. The server sends an HTTP request from the specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[0825] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, where the analysis begins.

[0826] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[0827] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The emotion engine analyzes data acquired from the camera and microphone to recognize emotions from the user's facial expressions and tone of voice. The recognized emotion data is used in the filtering process of advertising content.

[0828] If a user feels that an ad is inappropriate, they can provide feedback using the device's feedback form. The user enters their opinion or problem with the ad and presses the submit button. This feedback also includes the user's emotional data.

[0829] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[0830] The server receives the feedback and adds the feedback data to the dataset for model update. The feedback data and the user's sentiment data are used as new training data to retrain the generative model. The retrained generative model is used for the next advertising analysis.

[0831] Furthermore, the server will provide ad checking functionality through API integration with other services. For example, the ad checking API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. Furthermore, penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0832] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it is not displayed; if it is determined to be appropriate, it is displayed. In addition, the accuracy of ad filtering is improved based on user emotion data, providing a more appropriate advertising experience for users.

[0833] In this way, the system of the present invention can effectively analyze and filter advertising content, and improve accuracy by incorporating user sentiment feedback, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0834] The processing flow will be explained below.

[0835] Step 1:

[0836] The server retrieves advertising content from websites and video streaming services via URLs, sending HTTP requests and downloading text, images, audio, and video data.

[0837] Step 2:

[0838] The server sends the acquired advertising content to a generative model for analysis. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The advertising data is passed to the model's analysis engine, and analysis begins.

[0839] Step 3:

[0840] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate. The server uses this information to decide what to do next.

[0841] Step 4:

[0842] The server filters the ad content based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it is displayed. The filtering results are recorded in a log.

[0843] Step 5:

[0844] While the ad is being displayed, the emotion engine recognizes the user's emotions in real time by analyzing data acquired from the device's camera and microphone, and recognizing emotions from the user's facial expressions and tone of voice.

[0845] Step 6:

[0846] The emotion engine recognizes the user's emotion data and sends it to the server. The emotion data is used for feedback and to evaluate the effectiveness of advertising displays.

[0847] Step 7:

[0848] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button. The feedback also includes the user's emotional data.

[0849] Step 8:

[0850] The device sends the user's feedback to the server. The device compiles the feedback data into an appropriate format, creates an HTTP request to the server, and sends it.

[0851] Step 9:

[0852] The server receives the feedback and adds it to the dataset for updating the generative model. The feedback data and the user's emotion data are used as new training data to retrain the generative model.

[0853] Step 10:

[0854] The server retrains the generative model to detect fake information and inappropriate ads with high accuracy. The retrained model is then used for the next ad analysis.

[0855] Step 11:

[0856] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[0857] Step 12:

[0858] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. Service fees are collected from companies that display appropriate ads.

[0859] As described above, this system acquires, analyzes, and filters advertising content, and utilizes user sentiment data to provide more accurate feedback and update the model, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[0860] Example 2

[0861] 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."

[0862] Conventional ad delivery systems lacked sufficient means to prevent the display of inappropriate ads, resulting in a poor quality ad experience for users. Furthermore, they lacked a mechanism for collecting feedback that took user sentiment into account and improving the accuracy of ad filtering based on that feedback. As a result, users were unable to enjoy a safe and reliable ad experience. This led to issues such as a decline in the reliability of the service and user satisfaction.

[0863] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0864] In this invention, the server includes a means for acquiring advertising content, a means for analyzing the advertising content using a generative model, and a means for filtering the advertising content based on the analysis result, thereby enabling analysis of advertising content and blocking of inappropriate advertisements.

[0865] The server further includes means for recognizing user emotions, means for collecting user feedback, and means for updating the generative model using the collected feedback and emotion data, thereby improving the accuracy of advertisement filtering based on the user emotion data and providing a safer and more appropriate advertising experience for the user.

[0866] "Advertising content" refers to advertising information provided in the form of text, images, audio, video, etc.

[0867] A "generative model" is an artificial intelligence model that is trained based on official information and has an algorithm for analyzing the content of advertising content.

[0868] "Filtering" is a process of classifying advertising content into appropriate and inappropriate content based on the analysis results, and preventing inappropriate content from being displayed to the user.

[0869] The "emotion engine" is a system that analyzes the user's facial expressions, tone of voice, etc. to recognize the user's emotions in real time.

[0870] "Feedback" refers to the opinions and impressions provided by users regarding advertising content, as well as the emotional data recognized at that time.

[0871] "Model updating" is the process of retraining a generative model using collected feedback and emotion data to improve the model's accuracy.

[0872] "API integration" is a method of exchanging data with other services through an application programming interface.

[0873] A "penalty" is a sanction imposed on advertisers who distribute inappropriate ads.

[0874] The present invention provides a system for effectively analyzing and filtering advertising content, and improving accuracy by reflecting feedback based on user sentiment. Specific embodiments of each element are described below.

[0875] Acquiring advertising content

[0876] The server collects advertising content from websites and video streaming services. To do this, the server sends an HTTP request to a specified URL and downloads advertising data such as text, images, audio, and video. This data is then stored in the server's database.

[0877] Ad content analysis

[0878] The server sends the acquired advertising content to a generative AI model for content analysis. This generative AI model is trained based on official information and has algorithms to detect inappropriate content and fake information. The server then sends the advertising data to the model's API endpoint and receives the analysis results.

[0879] Ad content filtering

[0880] Based on the analysis results, the server filters the ad content. Specifically, inappropriate ads are blocked and only appropriate ads are displayed. The filtering results are also recorded.

[0881] Emotion recognition by emotion engine

[0882] While a user is viewing an ad, the emotion engine recognizes the user's emotions in real time. The emotion engine obtains data from the camera and microphone and analyzes the user's facial expressions and tone of voice to recognize emotions. This recognition data is used in the ad filtering process.

[0883] User feedback

[0884] If a user feels that an advertisement is inappropriate after viewing it, they can provide their opinion using a feedback form. The feedback includes the user's emotional data and is sent from the device to the server.

[0885] Send feedback and update the model

[0886] The device sends user feedback to the server, which receives it and adds it to the retraining dataset for the generative model, which the server uses to retrain the model and improve its accuracy.

[0887] API integration with other services

[0888] Furthermore, the server will provide an ad checking function through API integration with other services. Specifically, the API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. The server will also improve reliability by imposing penalties on advertisers who display inappropriate ads.

[0889] Specific examples

[0890] As a concrete example, consider a video streaming platform that sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative AI model and returns the results to the platform. If the ad is deemed inappropriate, it will not be displayed; if it is deemed appropriate, it will be displayed. Furthermore, the accuracy of ad filtering can be improved based on user emotional data, providing a more appropriate advertising experience for users.

[0891] Prompt Sentence Examples

[0892] "Please use a generative model to analyze whether this ad contains inappropriate information and return the results."

[0893] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0894] Step 1:

[0895] The server collects advertising content from websites and video streaming services. As input, a specified URL is given, and the server sends an HTTP GET request to that URL. As output, advertising data such as text, images, audio, and video are stored in the server's database.

[0896] Specific behavior:

[0897] The server sends a GET request to "https: / / example.com / ad".

[0898] The server analyzes the JSON response and saves it as advertising data.

[0899] Step 2:

[0900] The server sends the acquired advertising content to the generative AI model for content analysis. As input, the advertising data is sent as a POST request to the API endpoint of the generative AI model. As output, the analysis results are returned and stored on the server.

[0901] Specific behavior:

[0902] The server sends the ad data as a POST request to "https: / / ai-model-service / api / analyze".

[0903] The server receives and stores the analysis results returned by the model.

[0904] Step 3:

[0905] The server filters the ad content based on the analysis results. The analysis results are the input, and the filtering blocks inappropriate ads. The output is that only appropriate ads are listed and recorded for display.

[0906] Specific behavior:

[0907] The server analyzes the analysis results it receives.

[0908] If it is determined to be fake news, the advertising data will be deleted.

[0909] Appropriate ad data is stored in the "approved_ads" table.

[0910] Step 4:

[0911] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The input is data acquired from the device's camera and microphone, and the output is analyzed emotion data sent to the server.

[0912] Specific behavior:

[0913] The emotion engine acquires and analyzes camera data in real time.

[0914] If the user smiles, it is recorded as "positive" emotion data.

[0915] The recognized emotion data is sent to the server.

[0916] Step 5:

[0917] After viewing an advertisement, if a user feels it is inappropriate, he or she can provide his or her opinion using a feedback form. As input, the user's opinion or thoughts are entered into the feedback form, including emotional data. As output, the feedback data is sent to the server.

[0918] Specific behavior:

[0919] A user types "This ad is violent" into a feedback form on their device.

[0920] Pressing the Send Feedback button will create and send a request that also includes emotion data.

[0921] Step 6:

[0922] The device sends user feedback to the server. The server receives the feedback data as input and adds it to the retraining dataset for the generative model. The output is a retrained generative model.

[0923] Specific behavior:

[0924] The device posts the feedback data to "https: / / server / api / feedback".

[0925] The server adds the feedback to the dataset and retrains the AI ​​model.

[0926] Step 7:

[0927] The server interacts with other services via API to provide external analysis of advertising data. As input, there is an API request from the other service, the server analyzes the advertising data, and provides the results. As output, the analysis results are returned to the other service.

[0928] Specific behavior:

[0929] The video streaming service sends an API request to the server.

[0930] The server analyzes the advertising data using a generative AI model and sends the results back to the video streaming service.

[0931] (Application example 2)

[0932] 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."

[0933] Conventional ad filtering systems simply analyze the content of text and images to filter out inappropriate ads, but do not fully utilize user sentiment and feedback, resulting in a suboptimal ad experience for users. In particular, users may be presented with ads that they find annoying, resulting in a poor ad experience. Furthermore, the lack of a mechanism for imposing appropriate penalties on advertisers who place inappropriate ads also poses a problem of reduced reliability for the system as a whole.

[0934] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0935] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis results, means for recognizing user emotions, means for determining whether to display an advertisement based on the filtering results and user emotion data, means for collecting user feedback, and means for updating the generative model using the collected feedback. This makes it possible to utilize user emotion data to improve the accuracy of advertising filtering and provide a more appropriate advertising experience for users. Furthermore, by collecting feedback data and updating the generative model, the system is always kept up to date, minimizing the display of inappropriate advertisements.

[0936] "Advertising Content" means promotional material displayed within a website or application, including in the form of text, images, audio, video, or other content.

[0937] A "generative model" is an algorithm that uses artificial intelligence techniques to learn and generate, particularly for use in analyzing and filtering advertising content.

[0938] "Analysis" is the process of evaluating advertising content to determine whether it contains inappropriate information.

[0939] "Filtering" is the process of filtering out or blocking inappropriate advertising content based on analysis results.

[0940] "User" means an individual or organization that uses the system to view advertising content.

[0941] "Emotion recognition" is a technology that uses a camera and microphone to analyze a user's facial expressions and tone of voice to identify their emotions.

[0942] "Feedback" refers to information such as opinions, impressions, and evaluations that users input regarding advertising content.

[0943] "Model updating" is the process of retraining a generative model with newly collected data to improve its accuracy and functionality.

[0944] "API integration" is a technology that provides an interface for exchanging data and functions between different software systems.

[0945] A "penalty" is a sanction imposed on advertisers who post inappropriate advertisements, and may include restrictions on advertisement distribution.

[0946] A system for implementing this invention retrieves advertising content, analyzes and filters it using a generative model, recognizes user emotions, collects feedback, and updates the generative model to provide a more effective advertising experience.

[0947] As a specific embodiment, consider the following system: First, a server obtains advertising content from a website or application. The advertising content is downloaded from a specified URL using an HTTP request, and may include formats such as text, images, audio, and video.

[0948] The server then sends the acquired advertising content to a generative model, which analyzes the advertising content. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the generative model's analysis engine and begins analysis.

[0949] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[0950] Furthermore, while users are viewing ads, the emotion engine recognizes their emotions in real time. The emotion engine analyzes data acquired from the smartphone's camera and microphone to recognize emotions from the user's facial expressions and tone of voice. This emotion data is used in the filtering process and for future ad analysis.

[0951] If a user feels that an ad is inappropriate after viewing it, they can provide feedback through a feedback form on their smartphone. Users enter their opinions or concerns about the ad and press the submit button. This feedback also includes the user's emotional data.

[0952] The device sends the user's feedback to the server. The device then formats the feedback data appropriately and creates and sends an HTTP request to the server. The server receives the feedback and adds it to a dataset for updating the model. This dataset is then used to retrain the generative model and improve its accuracy.

[0953] The server also provides ad checking functionality through API integration with other services. For example, it will expose the ad checking API to platforms such as video streaming services and blogs, allowing external analysis of ad data. Penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[0954] As an example of this system, consider the following prompt:

[0955] Get Ad: "Get ad content from URL"

[0956] Content analysis: "Analyze using generative AI models to determine appropriateness"

[0957] Emotion Recognition: "Recognize user emotions in real time from camera footage"

[0958] Feedback collection: "User opinions and sentiment data are sent as feedback."

[0959] In practicing this invention, it is recommended to use hardware and software such as smartphone cameras, microphones, generative AI models, API interfaces, EmotionRecognizers, and AdAnalyzers, as appropriate, to provide users with a safe and reliable advertising experience.

[0960] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0961] Step 1:

[0962] The server retrieves advertising content from a website or application by using an HTTP request to download data from a specified URL. The input is a URL, and the output is the advertising data, which can be text, images, audio, or video.

[0963] Step 2:

[0964] The server sends the acquired advertising content to the generative model, which analyzes the advertising content. The input is the advertising data, and the generative model analyzes it to determine whether it contains inappropriate content. The output is a label indicating whether the advertisement is appropriate or inappropriate.

[0965] Step 3:

[0966] The server filters advertising content based on the analysis results of the generative model. If an ad is deemed inappropriate, it is blocked. If an ad is deemed appropriate, it is displayed. The input is the result of the judgment in step 2, and the output is an instruction to block or display the ad.

[0967] Step 4:

[0968] The device uses an emotion engine to recognize the user's emotions while they are viewing an advertisement. The emotion engine analyzes data acquired from the smartphone's camera and microphone. The input is camera video and audio data, and the output is the user's emotions (e.g., joy, anger, sadness, etc.).

[0969] Step 5:

[0970] After watching an advertisement, a user provides feedback using a feedback form on the device. The user enters their opinion or impression about the advertisement into the form and presses the submit button. The input is the user's text feedback and emotion data, and the output is the feedback data sent to the server.

[0971] Step 6:

[0972] The server receives the feedback data sent from the terminal and adds the collected data to the update dataset for the generative model. The input is the feedback data, and the output is the training dataset for the generative model.

[0973] Step 7:

[0974] The server uses the collected feedback data to retrain the generative model and improve its accuracy: the input is an updated training dataset, and the output is an updated generative model.

[0975] Step 8:

[0976] The server provides ad checking functionality through API integration with other services. The platform sends an API request, and the server performs external analysis of the ad data. The input is ad data from the external platform, and the output is the analysis results. If inappropriate ad delivery is confirmed, the server imposes penalties on the advertiser.

[0977] 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.

[0978] 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.

[0979] 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.

[0980] [Fourth embodiment]

[0981] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[0982] 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.

[0983] 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).

[0984] 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.

[0985] 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.

[0986] 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).

[0987] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0988] 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.

[0989] 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.

[0990] 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.

[0991] 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.

[0992] 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.

[0993] 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."

[0994] The present invention provides a system for acquiring advertising content, analyzing it using a generative model, filtering it, collecting feedback, and updating the model. Specific embodiments of each element are described below.

[0995] The server retrieves advertising content from websites or video streaming services. For example, it specifies a specific URL and downloads the advertising data from there. The retrieved data, which may be in the form of text, images, audio, or video, is then prepared for analysis by the generative model.

[0996] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The analysis results are used as an indicator to determine whether the advertisement is appropriate.

[0997] Based on the analysis of the ad content, the server decides whether to display it or not. If it is determined to contain inappropriate information, the ad is blocked and not shown to the user. On the other hand, if it is determined to be appropriate, the ad is displayed.

[0998] A dedicated UI is provided on the device to allow users to provide feedback on incorrectly displayed ads. Users fill out a feedback form, and this information is sent from the device to the server. This feedback data is used to update the generative model.

[0999] The server uses the collected feedback as part of model updates. By adding feedback data and retraining the generative model, it can detect fake information with high accuracy even on new data.

[1000] Furthermore, the server will provide other services with the ability to check advertising content through API integration. For example, it will provide ad checking functionality to platforms such as video streaming services and blogs, and analyze advertising data from outside. It will also improve the reliability of the entire system by imposing penalties on advertisers who deliver inappropriate ads.

[1001] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it will not be displayed; conversely, if it is determined to be appropriate, it will be displayed. Furthermore, appropriate feedback is provided to the advertiser to improve the model's performance.

[1002] As a result, the system of the present invention can effectively detect and block advertisements containing fake information, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[1003] The processing flow will be explained below.

[1004] Step 1:

[1005] The server retrieves advertising content from websites and video streaming services. It sends an HTTP request from a specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[1006] Step 2:

[1007] The server sends the acquired advertising content to a generative model. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, which begins analysis.

[1008] Step 3:

[1009] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate, and uses this information to decide what to do next.

[1010] Step 4:

[1011] The server filters the ads based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the information is appropriate, the ad is displayed. The server records the filtering results in a log.

[1012] Step 5:

[1013] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button.

[1014] Step 6:

[1015] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[1016] Step 7:

[1017] The server receives the feedback and adds the feedback data to the dataset for model updates, which are then used as new training data to retrain the generative model.

[1018] Step 8:

[1019] The server retrains the generative model to detect fake information and inappropriate ads with higher accuracy. The retrained model is then used for the next ad analysis.

[1020] Step 9:

[1021] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[1022] Step 10:

[1023] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. In addition, advertisers who display appropriate ads are charged a service fee.

[1024] Through these steps, the system effectively analyzes and filters advertising content, providing users with a safe and reliable advertising experience.

[1025] Example 1

[1026] 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."

[1027] In today's digital advertising market, there are many advertisements containing fake information or inappropriate content, which not only impair the user experience but also reduce brand credibility. Conventional systems have had difficulty efficiently detecting and blocking these inappropriate advertisements. Furthermore, they lack mechanisms for effectively utilizing user feedback to improve ad quality. The present invention aims to solve these issues and provide a safe and reliable advertising experience.

[1028] 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.

[1029] In this invention, the server includes a device for acquiring advertising content, a device for analyzing the advertising content using a generative model, a device for filtering the advertising content based on the analysis results from the generative model, a device for collecting user feedback, and a device for updating the generative model using the collected feedback. This makes it possible to effectively detect and block advertisements containing fake information or inappropriate content. Furthermore, by reflecting user feedback, the accuracy of the generative model can be continuously improved, thereby providing a safer and more reliable advertising experience.

[1030] "Advertising content" refers to promotional information expressed in the form of text, images, audio, video, etc.

[1031] A "generative model" is an artificial intelligence model that uses an algorithm trained in advance based on official information to analyze and judge advertising content.

[1032] "Filtering" is the process of determining the appropriateness of advertising content and removing inappropriate advertisements based on the analysis results from the generative model.

[1033] "Feedback" refers to opinions and evaluation information provided by users regarding the display results of advertising content.

[1034] A "device" is a mechanism including hardware and software for performing each function of the system.

[1035] "API integration" is an interface that allows functions and data to be shared and linked between different software systems.

[1036] "Penalty" means a sanction or penalty imposed on an advertiser who places an inappropriate advertisement.

[1037] MODE FOR CARRYING OUT THE INVENTION

[1038] This system acquires advertising content, analyzes it using a generative model, filters it, collects feedback, and updates the model. This system consists of three main entities: a server, a terminal, and a user.

[1039] Server Features

[1040] The server acts as a device for acquiring advertising content, downloading advertising data from websites and video streaming services. Specifically, it uses the Python requests library to send HTTP requests to acquire advertising content. For example, it downloads advertising data from the specified URL https: / / example.com / ad and saves it in the form of text, images, audio, and video.

[1041] The server then sends the acquired advertising content to a generative model for analysis. This generative model is a generative AI model such as OpenAI's GPT, which determines the ad appropriateness by inputting specific prompts. The model is highly reliable because it is trained based on official information.

[1042] Specifically, the following prompt sentences are input to the generative AI model:

[1043] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[1044] Based on the analysis results of the generative model, the system decides whether to display an ad. If the ad contains inappropriate information, it is blocked, and if it is deemed appropriate, it is displayed.

[1045] User operations

[1046] The device is provided with a dedicated UI for users to provide feedback on advertisements. Users enter their opinions and ratings in a feedback form, and this information is sent from the device to the server. For example, a user may send feedback such as "This advertisement is misleading."

[1047] Server model update function

[1048] The server retrains the generative model based on the feedback collected from users. By adding this feedback data, the generative model improves its detection accuracy for new data. Retraining is the process of adding new feedback information to an existing dataset to improve the performance of the generative model.

[1049] API integration function

[1050] Furthermore, the server provides an API integration function, allowing it to obtain advertising content in cooperation with other services. This function allows it to link with external systems such as video streaming services and blog platforms to optimize advertising. It can also impose penalties on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[1051] Specific examples

[1052] For example, a video streaming platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If an ad is determined to be inappropriate, it will not be displayed; only if it is appropriate will the ad be displayed. Appropriate feedback is also provided to the advertiser, contributing to improving the model's performance.

[1053] As a result, the system of the present invention can effectively detect and block advertisements containing fake information or inappropriate content. By using this system, it is possible to provide users with a safe and reliable advertising experience and improve the overall quality of services.

[1054] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1055] Processing Steps

[1056] Step 1:

[1057] The server retrieves advertising content from websites and video streaming services. Specifically, it uses the Python requests library to send an HTTP request and download the advertising data.

[1058] Input: Specific URL https: / / example.com / ad

[1059] Data processing: Sending HTTP requests, downloading advertising data

[1060] Output: Advertising data stored in the form of text, images, audio, and video

[1061] Specific operation: Specify a URL, send a request, and save the response data in local storage.

[1062] Step 2:

[1063] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained on public information and includes algorithms for detecting inappropriate content.

[1064] Input: Ad data obtained and saved in Step 1

[1065] Data processing: Formatting advertising data, creating prompts, and inputting them into the generative model

[1066] Output: Analysis results of advertisement content by generative model

[1067] Specific operation: Convert the ad data into text format, generate a prompt sentence like the following, and input it into the generative model:

[1068] "Please judge whether the following ad content contains inappropriate information. Example: 'Advertisement text: We'll teach you how to get rid of giant centipedes.'"

[1069] Step 3:

[1070] The server determines whether to display advertising content based on the analysis results from the generative model. If the content is deemed inappropriate, the ad is blocked, and if it is deemed appropriate, the ad is displayed to the user.

[1071] Input: Analysis results of the generative model

[1072] Data processing: Analysis of analysis results, filtering of advertisements deemed inappropriate

[1073] Output: Filtered ad content (displayable or not)

[1074] Specific operations: Analyze the analysis results, add inappropriate ads to a block list, and display appropriate ads on the user's screen.

[1075] Step 4:

[1076] Users provide feedback on advertising content through a dedicated UI, which is then sent from the device to the server.

[1077] Input: User feedback (e.g., "This ad is misleading")

[1078] Data Processing: Feedback data collection and transmission

[1079] Output: Feedback data is saved on the server

[1080] Specific operation: The user fills in the feedback form and presses the submit button. The data is sent from the device to the server.

[1081] Step 5:

[1082] The server retrains the generative model based on the collected feedback.

[1083] Input: Collected feedback data

[1084] Data processing: Integrating feedback data and retraining the generative model

[1085] Output: Updated generative model

[1086] What it does: Add the feedback data to the existing training dataset and retrain the generative model.

[1087] Step 6:

[1088] The server provides API integration functionality, connects with other services to acquire and analyze advertising content, and penalizes advertisers who deliver inappropriate ads.

[1089] Input: API request from external service

[1090] Data processing: Acquire external advertising data and analyze it with a generative model

[1091] Output: Sending analysis results back to external services, notifying inappropriate advertisers of penalties

[1092] Specific operation: Receives API requests from video streaming services, analyzes advertising data, returns analysis results to external services, and notifies advertisers of penalties if necessary.

[1093] (Application example 1)

[1094] 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."

[1095] In today's digital advertising environment, displaying inappropriate ads to users is a frequent problem. Displaying ads containing inappropriate content or fake information can ruin the user experience and reduce the overall reliability of the service. Another challenge is the inefficient collection of feedback and subsequent model updates. Furthermore, ads are presented in a variety of formats (text, image, audio, video), requiring appropriate filtering for each.

[1096] 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.

[1097] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis result, means for collecting feedback from users, means for updating the generative model using the collected feedback, and means for executing an advertising filter application that uses the collected feedback to update the model, thereby making it possible to effectively filter inappropriate advertisements and update the generative model in real time based on user feedback.

[1098] "Advertising Content" means advertising information provided in the form of text, images, audio, or video.

[1099] A "generative model" is an algorithm trained based on official information, and is an artificial intelligence model used to detect inappropriate information.

[1100] "Feedback" refers to the opinions and ratings users provide about ads, which are used to update the model.

[1101] "Format" refers to the text, image, audio, or video form of the advertising content.

[1102] An "advertising filter application" is an application that has the functionality to capture, analyze, and filter advertising content and collect user feedback.

[1103] "Server" refers to a computer system that performs analysis and filtering of advertising content on a network and operates and updates generative models.

[1104] "API integration" means providing and using an application program interface to obtain and share advertising content with external services.

[1105] A "penalty" is a punishment imposed on advertisers who post inappropriate advertisements, and is a measure taken to maintain the reliability of the service.

[1106] "Smartphone application" means software that runs on a smartphone and acquires, analyzes, and filters advertising content.

[1107] The present invention is a system that acquires advertising content, analyzes it using a generative model, filters it, collects feedback from users, and updates the model. Specific embodiments of each element are described below.

[1108] System Configuration

[1109] server:

[1110] The server plays a central role in retrieving, analyzing, and filtering advertising content, as well as updating the generative model. The software uses Flask as the API server, and the Hugging Face transformers library is used to run the generative model. A computer system with sufficient computing power is recommended.

[1111] Device:

[1112] The user's smartphone runs an ad filter application, which sends API requests to the server and receives analysis results of ad content. It also collects user feedback via a feedback form and sends it to the server.

[1113] Acquiring and analyzing advertising content

[1114] 1. The server retrieves advertising content from the web page or app the user is viewing, which can be in the form of text, images, audio, or video.

[1115] 2. The acquired advertising content is sent to a generative model, which is trained in advance on official information and detects whether the content contains inappropriate content.

[1116] Filtering and Feedback

[1117] 1. The server filters advertising content based on the analysis results of the generative model. Advertising content deemed inappropriate is blocked, and only advertising content deemed appropriate is displayed to the user.

[1118] 2. Users can provide feedback on the ads they see. They can easily enter their opinions and ratings using the feedback form provided in the smartphone application.

[1119] Model Update

[1120] 1. The server stores the feedback collected from users and uses it to update the generative model. The feedback data is used to retrain the model, enabling it to analyze new data with high accuracy.

[1121] Specific examples

[1122] For example, a user browsing a news app can submit their opinion about an ad they see using a feedback form asking, "Do you think this ad is inappropriate?" This information is sent to the server and used as data for retraining the generative model.

[1123] Prompt Sentence Examples

[1124] "Analyze the text of this ad to determine if it contains inappropriate information. Ad text: 'Download your favorite characters now! 100% authentic!'"

[1125] This allows users to view more appropriate advertisements than before, improving the reliability of the service as a whole and user satisfaction.

[1126] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1127] Step 1:

[1128] The server obtains advertising content from web pages and apps viewed by users. Specifically, it downloads advertising content from specific URLs and data sources, converts it into text, images, audio, and video formats, and saves it. The input is the URL and data source information, and the output is the raw advertising content data.

[1129] Step 2:

[1130] The server sends the acquired advertising content to the generative model for analysis. The generative model uses a specific algorithm to detect whether the advertising content is inappropriate. The input is the raw data of the advertising content, and the output is the analysis result (appropriate / inappropriate).

[1131] Step 3:

[1132] The server filters the ad content based on the analysis results of the generative model. If inappropriate content is detected, the ad content is removed from the display list. The input is the analysis results of the generative model and the ad content itself, and the output is the filtered list.

[1133] Step 4:

[1134] The server sends the filtered advertising content to the terminal, which displays it to the user and saves the display results. The input is the filtered advertising content list, and the output is the advertisements displayed to the user.

[1135] Step 5:

[1136] Users provide feedback on the displayed advertisements by entering their opinions and ratings using a feedback form on their device and sending the data to the server. The input is the feedback content, and the output is the feedback data stored on the server.

[1137] Step 6:

[1138] The server stores the collected feedback data and uses it to retrain the generative model. New feedback data is used to update the generative model, improving the accuracy of the next analysis. The input is the feedback data, and the output is the updated generative model.

[1139] Step 7:

[1140] The server also connects with external services via API to provide advertising content filtering functionality. It retrieves external advertising content in response to API requests and analyzes it using a generative model. The input is advertising content information from external services, and the output is the analysis results and filtered advertising content.

[1141] Step 8:

[1142] The server imposes penalties on advertisers who have posted inappropriate ads. Based on the analysis results, it notifies the advertiser of a warning or penalty. The input is a list of ad content that is judged to be inappropriate, and the output is a warning or penalty notice to the advertiser.

[1143] 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.

[1144] The present invention combines a system that acquires advertising content, analyzes it using a generative model, filters it, collects user feedback, and updates the model with an emotion engine that recognizes user emotions. Specific embodiments of each element are described below.

[1145] The server retrieves advertising content from websites and video streaming services. The server sends an HTTP request from the specified URL and downloads the advertising data. The retrieved data includes text, images, audio, and video.

[1146] The server then sends the acquired advertising content to a generative model for analysis. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the model's analysis engine, where the analysis begins.

[1147] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[1148] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The emotion engine analyzes data acquired from the camera and microphone to recognize emotions from the user's facial expressions and tone of voice. The recognized emotion data is used in the filtering process of advertising content.

[1149] If a user feels that an ad is inappropriate, they can provide feedback using the device's feedback form. The user enters their opinion or problem with the ad and presses the submit button. This feedback also includes the user's emotional data.

[1150] The device sends the user's feedback to the server. The device organizes the feedback data into an appropriate format and creates and sends an HTTP request to the server.

[1151] The server receives the feedback and adds the feedback data to the dataset for model update. The feedback data and the user's sentiment data are used as new training data to retrain the generative model. The retrained generative model is used for the next advertising analysis.

[1152] Furthermore, the server will provide ad checking functionality through API integration with other services. For example, the ad checking API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. Furthermore, penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[1153] As a concrete example, a video distribution platform sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative model and returns the results to the platform. If the ad is determined to be inappropriate, it is not displayed; if it is determined to be appropriate, it is displayed. In addition, the accuracy of ad filtering is improved based on user emotion data, providing a more appropriate advertising experience for users.

[1154] In this way, the system of the present invention can effectively analyze and filter advertising content, and improve accuracy by incorporating user sentiment feedback, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[1155] The processing flow will be explained below.

[1156] Step 1:

[1157] The server retrieves advertising content from websites and video streaming services via URLs, sending HTTP requests and downloading text, images, audio, and video data.

[1158] Step 2:

[1159] The server sends the acquired advertising content to a generative model for analysis. The generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The advertising data is passed to the model's analysis engine, and analysis begins.

[1160] Step 3:

[1161] The server receives the results of the generative model's analysis, which includes a judgment on whether the ad is appropriate or inappropriate. The server uses this information to decide what to do next.

[1162] Step 4:

[1163] The server filters the ad content based on the analysis results. If the ad contains inappropriate information, it is blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it is displayed. The filtering results are recorded in a log.

[1164] Step 5:

[1165] While the ad is being displayed, the emotion engine recognizes the user's emotions in real time by analyzing data acquired from the device's camera and microphone, and recognizing emotions from the user's facial expressions and tone of voice.

[1166] Step 6:

[1167] The emotion engine recognizes the user's emotion data and sends it to the server. The emotion data is used for feedback and to evaluate the effectiveness of advertising displays.

[1168] Step 7:

[1169] If a user feels that a displayed advertisement is inappropriate, they can provide feedback using the feedback form on their device. The user enters their opinion or problem with the advertisement and presses the submit button. The feedback also includes the user's emotional data.

[1170] Step 8:

[1171] The device sends the user's feedback to the server. The device compiles the feedback data into an appropriate format, creates an HTTP request to the server, and sends it.

[1172] Step 9:

[1173] The server receives the feedback and adds it to the dataset for updating the generative model. The feedback data and the user's emotion data are used as new training data to retrain the generative model.

[1174] Step 10:

[1175] The server retrains the generative model to detect fake information and inappropriate ads with high accuracy. The retrained model is then used for the next ad analysis.

[1176] Step 11:

[1177] The server provides ad checking functionality through API integration with other services. For example, the ad checking API can be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data.

[1178] Step 12:

[1179] The server imposes penalties on advertisers who display inappropriate ads. Penalties are implemented in the form of fines or suspension of ad delivery. Service fees are collected from companies that display appropriate ads.

[1180] As described above, this system acquires, analyzes, and filters advertising content, and utilizes user sentiment data to provide more accurate feedback and update the model, thereby providing users with a safe and reliable advertising experience and improving the quality of service.

[1181] Example 2

[1182] 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."

[1183] Conventional ad delivery systems lacked sufficient means to prevent the display of inappropriate ads, resulting in a poor quality ad experience for users. Furthermore, they lacked a mechanism for collecting feedback that took user sentiment into account and improving the accuracy of ad filtering based on that feedback. As a result, users were unable to enjoy a safe and reliable ad experience. This led to issues such as a decline in the reliability of the service and user satisfaction.

[1184] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1185] In this invention, the server includes a means for acquiring advertising content, a means for analyzing the advertising content using a generative model, and a means for filtering the advertising content based on the analysis result, thereby enabling analysis of advertising content and blocking of inappropriate advertisements.

[1186] The server further includes means for recognizing user emotions, means for collecting user feedback, and means for updating the generative model using the collected feedback and emotion data, thereby improving the accuracy of advertisement filtering based on the user emotion data and providing a safer and more appropriate advertising experience for the user.

[1187] "Advertising content" refers to advertising information provided in the form of text, images, audio, video, etc.

[1188] A "generative model" is an artificial intelligence model that is trained based on official information and has an algorithm for analyzing the content of advertising content.

[1189] "Filtering" is a process of classifying advertising content into appropriate and inappropriate content based on the analysis results, and preventing inappropriate content from being displayed to the user.

[1190] The "emotion engine" is a system that analyzes the user's facial expressions, tone of voice, etc. to recognize the user's emotions in real time.

[1191] "Feedback" refers to the opinions and impressions provided by users regarding advertising content, as well as the emotional data recognized at that time.

[1192] "Model updating" is the process of retraining a generative model using collected feedback and emotion data to improve the model's accuracy.

[1193] "API integration" is a method of exchanging data with other services through an application programming interface.

[1194] A "penalty" is a sanction imposed on advertisers who distribute inappropriate ads.

[1195] The present invention provides a system for effectively analyzing and filtering advertising content, and improving accuracy by reflecting feedback based on user sentiment. Specific embodiments of each element are described below.

[1196] Acquiring advertising content

[1197] The server collects advertising content from websites and video streaming services. To do this, the server sends an HTTP request to a specified URL and downloads advertising data such as text, images, audio, and video. This data is then stored in the server's database.

[1198] Ad content analysis

[1199] The server sends the acquired advertising content to a generative AI model for content analysis. This generative AI model is trained based on official information and has algorithms to detect inappropriate content and fake information. The server then sends the advertising data to the model's API endpoint and receives the analysis results.

[1200] Ad content filtering

[1201] Based on the analysis results, the server filters the ad content. Specifically, inappropriate ads are blocked and only appropriate ads are displayed. The filtering results are also recorded.

[1202] Emotion recognition by emotion engine

[1203] While a user is viewing an ad, the emotion engine recognizes the user's emotions in real time. The emotion engine obtains data from the camera and microphone and analyzes the user's facial expressions and tone of voice to recognize emotions. This recognition data is used in the ad filtering process.

[1204] User feedback

[1205] If a user feels that an advertisement is inappropriate after viewing it, they can provide their opinion using a feedback form. The feedback includes the user's emotional data and is sent from the device to the server.

[1206] Send feedback and update the model

[1207] The device sends user feedback to the server, which receives it and adds it to the retraining dataset for the generative model, which the server uses to retrain the model and improve its accuracy.

[1208] API integration with other services

[1209] Furthermore, the server will provide an ad checking function through API integration with other services. Specifically, the API will be made available to platforms such as video streaming services and blogs, allowing external analysis of ad data. The server will also improve reliability by imposing penalties on advertisers who display inappropriate ads.

[1210] Specific examples

[1211] As a concrete example, consider a video streaming platform that sends an API request to a server to analyze advertising data. The server analyzes the advertising data using a generative AI model and returns the results to the platform. If the ad is deemed inappropriate, it will not be displayed; if it is deemed appropriate, it will be displayed. Furthermore, the accuracy of ad filtering can be improved based on user emotional data, providing a more appropriate advertising experience for users.

[1212] Prompt Sentence Examples

[1213] "Please use a generative model to analyze whether this ad contains inappropriate information and return the results."

[1214] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1215] Step 1:

[1216] The server collects advertising content from websites and video streaming services. As input, a specified URL is given, and the server sends an HTTP GET request to that URL. As output, advertising data such as text, images, audio, and video are stored in the server's database.

[1217] Specific behavior:

[1218] The server sends a GET request to "https: / / example.com / ad".

[1219] The server analyzes the JSON response and saves it as advertising data.

[1220] Step 2:

[1221] The server sends the acquired advertising content to the generative AI model for content analysis. As input, the advertising data is sent as a POST request to the API endpoint of the generative AI model. As output, the analysis results are returned and stored on the server.

[1222] Specific behavior:

[1223] The server sends the ad data as a POST request to "https: / / ai-model-service / api / analyze".

[1224] The server receives and stores the analysis results returned by the model.

[1225] Step 3:

[1226] The server filters the ad content based on the analysis results. The analysis results are the input, and the filtering blocks inappropriate ads. The output is that only appropriate ads are listed and recorded for display.

[1227] Specific behavior:

[1228] The server analyzes the analysis results it receives.

[1229] If it is determined to be fake news, the advertising data will be deleted.

[1230] Appropriate ad data is stored in the "approved_ads" table.

[1231] Step 4:

[1232] While a user is viewing an advertisement, the emotion engine recognizes the user's emotions in real time. The input is data acquired from the device's camera and microphone, and the output is analyzed emotion data sent to the server.

[1233] Specific behavior:

[1234] The emotion engine acquires and analyzes camera data in real time.

[1235] If the user smiles, it is recorded as "positive" emotion data.

[1236] The recognized emotion data is sent to the server.

[1237] Step 5:

[1238] After viewing an advertisement, if a user feels it is inappropriate, he or she can provide his or her opinion using a feedback form. As input, the user's opinion or thoughts are entered into the feedback form, including emotional data. As output, the feedback data is sent to the server.

[1239] Specific behavior:

[1240] A user types "This ad is violent" into a feedback form on their device.

[1241] Pressing the Send Feedback button will create and send a request that also includes emotion data.

[1242] Step 6:

[1243] The device sends user feedback to the server. The server receives the feedback data as input and adds it to the retraining dataset for the generative model. The output is a retrained generative model.

[1244] Specific behavior:

[1245] The device posts the feedback data to "https: / / server / api / feedback".

[1246] The server adds the feedback to the dataset and retrains the AI ​​model.

[1247] Step 7:

[1248] The server interacts with other services via API to provide external analysis of advertising data. As input, there is an API request from the other service, the server analyzes the advertising data, and provides the results. As output, the analysis results are returned to the other service.

[1249] Specific behavior:

[1250] The video streaming service sends an API request to the server.

[1251] The server analyzes the advertising data using a generative AI model and sends the results back to the video streaming service.

[1252] (Application example 2)

[1253] 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."

[1254] Conventional ad filtering systems simply analyze the content of text and images to filter out inappropriate ads, but do not fully utilize user sentiment and feedback, resulting in a suboptimal ad experience for users. In particular, users may be presented with ads that they find annoying, resulting in a poor ad experience. Furthermore, the lack of a mechanism for imposing appropriate penalties on advertisers who place inappropriate ads also poses a problem of reduced reliability for the system as a whole.

[1255] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1256] In this invention, the server includes means for acquiring advertising content, means for analyzing the advertising content using a generative model, means for filtering the advertising content based on the analysis results, means for recognizing user emotions, means for determining whether to display an advertisement based on the filtering results and user emotion data, means for collecting user feedback, and means for updating the generative model using the collected feedback. This makes it possible to utilize user emotion data to improve the accuracy of advertising filtering and provide a more appropriate advertising experience for users. Furthermore, by collecting feedback data and updating the generative model, the system is always kept up to date, minimizing the display of inappropriate advertisements.

[1257] "Advertising Content" means promotional material displayed within a website or application, including in the form of text, images, audio, video, or other content.

[1258] A "generative model" is an algorithm that uses artificial intelligence techniques to learn and generate, particularly for use in analyzing and filtering advertising content.

[1259] "Analysis" is the process of evaluating advertising content to determine whether it contains inappropriate information.

[1260] "Filtering" is the process of filtering out or blocking inappropriate advertising content based on analysis results.

[1261] "User" means an individual or organization that uses the system to view advertising content.

[1262] "Emotion recognition" is a technology that uses a camera and microphone to analyze a user's facial expressions and tone of voice to identify their emotions.

[1263] "Feedback" refers to information such as opinions, impressions, and evaluations that users input regarding advertising content.

[1264] "Model updating" is the process of retraining a generative model with newly collected data to improve its accuracy and functionality.

[1265] "API integration" is a technology that provides an interface for exchanging data and functions between different software systems.

[1266] A "penalty" is a sanction imposed on advertisers who post inappropriate advertisements, and may include restrictions on advertisement distribution.

[1267] A system for implementing this invention retrieves advertising content, analyzes and filters it using a generative model, recognizes user emotions, collects feedback, and updates the generative model to provide a more effective advertising experience.

[1268] As a specific embodiment, consider the following system: First, a server obtains advertising content from a website or application. The advertising content is downloaded from a specified URL using an HTTP request, and may include formats such as text, images, audio, and video.

[1269] The server then sends the acquired advertising content to a generative model, which analyzes the advertising content. This generative model is trained based on official information and incorporates algorithms to detect fake information and inappropriate content. The server then passes the advertising data to the generative model's analysis engine and begins analysis.

[1270] Based on the analysis results, the server filters the ad content. If the ad contains inappropriate information, it will be blocked and not displayed to the user. On the other hand, if the ad is deemed appropriate, it will be displayed. The server records the filtering results.

[1271] Furthermore, while users are viewing ads, the emotion engine recognizes their emotions in real time. The emotion engine analyzes data acquired from the smartphone's camera and microphone to recognize emotions from the user's facial expressions and tone of voice. This emotion data is used in the filtering process and for future ad analysis.

[1272] If a user feels that an ad is inappropriate after viewing it, they can provide feedback through a feedback form on their smartphone. Users enter their opinions or concerns about the ad and press the submit button. This feedback also includes the user's emotional data.

[1273] The device sends the user's feedback to the server. The device then formats the feedback data appropriately and creates and sends an HTTP request to the server. The server receives the feedback and adds it to a dataset for updating the model. This dataset is then used to retrain the generative model and improve its accuracy.

[1274] The server also provides ad checking functionality through API integration with other services. For example, it will expose the ad checking API to platforms such as video streaming services and blogs, allowing external analysis of ad data. Penalties will be imposed on advertisers who deliver inappropriate ads, improving the reliability of the entire system.

[1275] As an example of this system, consider the following prompt:

[1276] Get Ad: "Get ad content from URL"

[1277] Content analysis: "Analyze using generative AI models to determine appropriateness"

[1278] Emotion Recognition: "Recognize user emotions in real time from camera footage"

[1279] Feedback collection: "User opinions and sentiment data are sent as feedback."

[1280] In practicing this invention, it is recommended to use hardware and software such as smartphone cameras, microphones, generative AI models, API interfaces, EmotionRecognizers, and AdAnalyzers, as appropriate, to provide users with a safe and reliable advertising experience.

[1281] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1282] Step 1:

[1283] The server retrieves advertising content from a website or application by using an HTTP request to download data from a specified URL. The input is a URL, and the output is the advertising data, which can be text, images, audio, or video.

[1284] Step 2:

[1285] The server sends the acquired advertising content to the generative model, which analyzes the advertising content. The input is the advertising data, and the generative model analyzes it to determine whether it contains inappropriate content. The output is a label indicating whether the advertisement is appropriate or inappropriate.

[1286] Step 3:

[1287] The server filters advertising content based on the analysis results of the generative model. If an ad is deemed inappropriate, it is blocked. If an ad is deemed appropriate, it is displayed. The input is the result of the judgment in step 2, and the output is an instruction to block or display the ad.

[1288] Step 4:

[1289] The device uses an emotion engine to recognize the user's emotions while they are viewing an advertisement. The emotion engine analyzes data acquired from the smartphone's camera and microphone. The input is camera video and audio data, and the output is the user's emotions (e.g., joy, anger, sadness, etc.).

[1290] Step 5:

[1291] After watching an advertisement, a user provides feedback using a feedback form on the device. The user enters their opinion or impression about the advertisement into the form and presses the submit button. The input is the user's text feedback and emotion data, and the output is the feedback data sent to the server.

[1292] Step 6:

[1293] The server receives the feedback data sent from the terminal and adds the collected data to the update dataset for the generative model. The input is the feedback data, and the output is the training dataset for the generative model.

[1294] Step 7:

[1295] The server uses the collected feedback data to retrain the generative model and improve its accuracy: the input is an updated training dataset, and the output is an updated generative model.

[1296] Step 8:

[1297] The server provides ad checking functionality through API integration with other services. The platform sends an API request, and the server performs external analysis of the ad data. The input is ad data from the external platform, and the output is the analysis results. If inappropriate ad delivery is confirmed, the server imposes penalties on the advertiser.

[1298] 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.

[1299] 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.

[1300] 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.

[1301] 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.

[1302] 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.

[1303] 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.

[1304] 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).

[1305] 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.

[1306] 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."

[1307] 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.

[1308] 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).

[1309] 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.

[1310] 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.

[1311] 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.

[1312] 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.

[1313] 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.

[1314] 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.

[1315] 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.

[1316] 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.

[1317] 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.

[1318] 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.

[1319] The following is further disclosed regarding the above embodiment.

[1320] (Claim 1)

[1321] a means for obtaining advertising content;

[1322] means for analyzing advertising content with a generative model;

[1323] means for filtering advertising content based on the analysis results;

[1324] a means for collecting feedback from users;

[1325] a means for updating the generative model using the collected feedback;

[1326] A system including:

[1327] (Claim 2)

[1328] A means of obtaining advertising content through API integration with other services,

[1329] A means of providing API integration,

[1330] measures to penalize advertisers who place inappropriate ads;

[1331] The system of claim 1 further comprising:

[1332] (Claim 3)

[1333] The advertising content can be text, images, audio, or video;

[1334] The generative model is trained based on formal information.

[1335] The system of claim 1 further comprising:

[1336] "Example 1"

[1337] (Claim 1)

[1338] a device for acquiring advertising content;

[1339] a device for analyzing advertising content using a generative model;

[1340] an apparatus for filtering advertising content based on analysis results from the generative model;

[1341] a device for collecting feedback from users;

[1342] a device for updating the generative model using the collected feedback;

[1343] A system including:

[1344] (Claim 2)

[1345] A device that acquires advertising content through API collaboration with other services;

[1346] A device that provides API integration,

[1347] a device for imposing penalties on advertisers who have posted inappropriate advertisements;

[1348] The system of claim 1 further comprising:

[1349] (Claim 3)

[1350] A device in which the advertising content is text, image, audio, or video;

[1351] A device in which the generative model is trained based on formal information;

[1352] The system of claim 1 further comprising:

[1353] "Application Example 1"

[1354] (Claim 1)

[1355] a means for obtaining advertising content;

[1356] means for analyzing advertising content with a generative model;

[1357] means for filtering advertising content based on the analysis results;

[1358] a means for collecting feedback from users;

[1359] a means for updating the generative model using the collected feedback;

[1360] means for running an advertisement filter application that utilizes the collected feedback to update the model;

[1361] A system including:

[1362] (Claim 2)

[1363] A means of obtaining advertising content through API integration with other services,

[1364] A means of providing API integration,

[1365] measures to penalize advertisers who place inappropriate ads;

[1366] means for analyzing the text, image, audio, and video formats of advertising content;

[1367] The system of claim 1 further comprising:

[1368] (Claim 3)

[1369] The advertising content can be text, images, audio, or video;

[1370] The generative model is trained based on formal information.

[1371] a means for providing a smartphone application that blocks inappropriate advertisements based on the analysis results;

[1372] The system of claim 1 further comprising:

[1373] "Example 2: Combining Emotion Engines"

[1374] (Claim 1)

[1375] a means for obtaining advertising content;

[1376] means for analyzing advertising content with a generative model;

[1377] means for filtering advertising content based on the analysis results;

[1378] means for recognizing a user's emotion;

[1379] a means for collecting feedback from users;

[1380] a means for updating the generative model using the collected feedback and emotion data;

[1381] A system including:

[1382] (Claim 2)

[1383] A means of obtaining advertising content through API integration with other services,

[1384] A means of providing API integration,

[1385] measures to penalize advertisers who place inappropriate ads;

[1386] The system of claim 1 further comprising:

[1387] (Claim 3)

[1388] The advertising content can be text, images, audio, or video;

[1389] The generative model is trained based on formal information.

[1390] The system of claim 1 further comprising:

[1391] "Application example 2 when combining emotion engines"

[1392] (Claim 1)

[1393] a means for obtaining advertising content;

[1394] means for analyzing advertising content with a generative model;

[1395] means for filtering advertising content based on the analysis results;

[1396] means for recognizing a user's emotion;

[1397] A means for determining whether to display an advertisement based on the filtering result and the user's emotion data;

[1398] a means for collecting feedback from users;

[1399] a means for updating the generative model using the collected feedback;

[1400] A system including:

[1401] (Claim 2)

[1402] A means of obtaining advertising content through API integration with other services,

[1403] A means of providing API integration,

[1404] measures to penalize advertisers who place inappropriate ads;

[1405] The system of claim 1 further comprising:

[1406] (Claim 3)

[1407] The advertising content can be text, images, audio, or video;

[1408] The generative model is trained based on formal information.

[1409] a means for real-time user emotion recognition;

[1410] The system of claim 1 further comprising: [Explanation of symbols]

[1411] 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 obtaining advertising content; means for analyzing advertising content with a generative model; means for filtering advertising content based on the analysis results; a means for collecting feedback from users; a means for updating the generative model using the collected feedback; and A system including:

2. A means of obtaining advertising content through API integration with other services, A means of providing API integration, measures to penalize advertisers who place inappropriate ads; The system of claim 1 further comprising:

3. The advertising content can be text, images, audio, or video; The generative model is trained based on formal information. The system of claim 1 further comprising:

Images