System

Abstract

To provide a system.SOLUTION: A system includes: examination means that, by using a generative AI model and on the basis of a specific examination standard, conducts an examination for an advertisement material that is digital contents including any one of a text, an image, and moving images that are provided from a person in charge in an advertising agency and used as an advertisement; means that, by using the generative AI model, outputs a result of the examination and transmits the result of the examination to the person in charge; an emotion engine that recognizes the emotion of a user who browses the advertisement material; and adjustment means that inputs a specific prompt sentence to the generative AI model to adjust the examination standard on the basis of the emotion.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] Currently, the screening of advertising materials is largely done visually, and the cost of screening is high due to the huge number of submissions. In particular, the screening process, which involves subjective judgments such as "offensive / not offensive" and "adult / not adult," takes a long time because the criteria for such judgments are unclear. [Means for solving the problem]

[0005] The system disclosed herein includes an evaluation means that uses a generative AI model to evaluate advertising materials, which are digital content including any of text, images, and videos used as advertisements, provided by an advertising agency representative based on specific evaluation criteria; a means that uses a generative AI model to output the results of the evaluation and transmit the results of the evaluation to the representative; an emotion engine that recognizes the emotions of a user viewing the advertising material; and an adjustment means that adjusts the evaluation criteria based on the emotions by inputting specific prompt sentences into the generative AI model. [Brief explanation of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [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. 2 is a sequence diagram showing a flow of processing in the data processing system according to the first embodiment of the first form example. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1. [Figure 13] FIG. 10 is a sequence diagram showing a processing flow of a data processing system in a second embodiment of the second form example. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Embodiment Example 2. [Figure 15] FIG. 10 is a sequence diagram showing the flow of processing in a data processing system according to a third embodiment of the third embodiment. [Figure 16] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Embodiment 3. [Figure 17] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Embodiment 1 when an emotion engine is combined. [Figure 19] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the second embodiment of the second form example when an emotion engine is combined. [Figure 20] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 of Form Example 2 when an emotion engine is combined. [Figure 21] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in the third embodiment of the third form example when an emotion engine is combined. [Figure 22] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 3 of Form Example 3 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0007] 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.

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

[0009] 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, the 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), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] 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.

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

[0012] In the following embodiments, a communication I / F (Interface) with a 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.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. 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."

[0014] [First embodiment]

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

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

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

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

[0019] The reception device 38 includes a touch panel 38A, 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.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (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.

[0021] 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.

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

[0023] 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.

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

[0025] In the smart device 14, the processor 46 performs the reception output 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.

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

[0027] "Example 1"

[0028] The present invention is a system that incorporates a generation AI for reviewing advertising materials. Specifically, a web-based submission form or API is used as a means for accepting submissions of advertising materials. The submitted advertising materials are reviewed by the generation AI, and the review results are communicated to the user via an output means, such as a web interface or email notification.

[0029] "Example 2"

[0030] The generative AI makes subjective judgments such as whether the ad material is "offensive / not offensive" or "adult / not adult" by analyzing the content of the ad material using technologies such as image recognition and natural language processing, and makes a judgment based on the results.

[0031] "Example 3"

[0032] When the generative AI outputs the review results, it will include a means to clearly indicate the criteria used. Specifically, along with the review results, it will output information indicating what factors were used as the basis for the judgment. For example, if a content is judged to be "offensive," it will output information such as "contains violent language" as the reason.

[0033] The processing flow of each embodiment will be described below.

[0034] "Example 1"

[0035] Step 1: Advertisers submit their ad materials to the system via a web-based submission form or API.

[0036] Step 2: The submitted ad material is reviewed by the AI ​​generator, which analyzes the content of the ad material using techniques such as image recognition and natural language processing.

[0037] Step 3: The generating AI outputs the review results and communicates them to the advertiser via a web interface or email notification.

[0038] "Example 2"

[0039] Step 1: When advertising materials are submitted to the system, the generation AI makes a subjective judgment on the advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult."

[0040] Step 2: The generative AI analyzes the content of the advertising material using technologies such as image recognition and natural language processing, and makes a decision based on the results.

[0041] Step 3: The generating AI outputs the review results and their criteria, and notifies the advertiser of the review results. "Example 3"

[0042] Step 1: Once the ad material is submitted to the system, the generation AI reviews it.

[0043] Step 2: The generative AI outputs the results of the review and the criteria for its decision. Specifically, it outputs information indicating the factors that formed the basis for the decision along with the review results.

[0044] Step 3: The AI's review results and criteria are communicated to the advertiser. For example, if the content is deemed "offensive," the AI ​​outputs information such as "contains violent content" as the reason.

[0045] Example 1

[0046] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0047] The traditional review process for advertising materials is often done manually, which is time-consuming and labor-intensive, and prone to subjective judgment. Furthermore, there are also delays in notifying the review results, which creates the risk of missing the timing of advertising campaigns. To solve these problems, a system that provides an efficient and objective review process is needed.

[0048] 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.

[0049] In this invention, the server includes means for converting the content of the advertising material into text format, generating and inputting a prompt sentence to the generative AI model, means for acquiring the review results from the generative AI model, and means for notifying the user of the review results via a web interface or email notification, thereby automating the advertising material review process and enabling efficient and objective review.

[0050] "Advertising materials" are digital content such as text, images, and videos used as advertisements.

[0051] "Generative AI" is a system that uses artificial intelligence technology to automatically perform specific tasks, in this case reviewing advertising materials.

[0052] "Means for accepting submissions" refers to an interface through which users submit advertising materials to the system, and includes web forms, APIs, etc.

[0053] The "means for outputting the review results" refers to an interface that the generation AI uses to notify the user of the results of its review of the advertising material, and includes a web interface and email notification.

[0054] "Means for converting into text format" refers to the process of analyzing the content of advertising materials and converting them into a format that can be input as text data into a generative AI model.

[0055] A "prompt" is an instruction given to a generative AI model to perform a specific task.

[0056] "Means for obtaining review results from the generative AI model" refers to the process for receiving the review results output by the generative AI model and processing them within the system.

[0057] A "web interface" is an interface that allows a user to access the system through a web browser to perform operations and check results.

[0058] "Email notification" is a function that allows the system to send review results and other information to the user's email address.

[0059] MODE FOR CARRYING OUT THE INVENTION

[0060] The present invention provides a system for efficiently and objectively reviewing advertising materials. This system automatically reviews advertising materials using a generative AI model and notifies the user of the results. Specific embodiments are described below.

[0061] 1. Introduction

[0062] The system aims to automate the entire process, from submitting advertising materials to reviewing and notifying the results. Users submit advertising materials to the system using a web-based submission form or API, and the server reviews them using a generative AI model. The review results are communicated to users via a web interface or email notification.

[0063] 2. Hardware and Software Used

[0064] Server: Receives advertising materials, runs the generative AI model, and outputs the review results.

[0065] Generative AI model: An artificial intelligence model used to review advertising materials, such as OpenAI's GPT-4.

[0066] Web interface: An interface for users to submit advertising materials and check the review results.

[0067] Email notification system: A system for notifying users of the review results by email.

[0068] 3. Submit advertising materials

[0069] Users submit advertising materials to the system using a web-based submission form or API. Specifically, users open a web browser, access a specified URL, enter the required information in the form, and upload the advertising material files. When using the API, users submit the advertising material data as a POST request.

[0070] 4. Receiving and storing advertising materials

[0071] The server receives the ad material sent by the user. The server temporarily stores the data received through the submission form or API endpoint. For example, the server stores the ad material file in a specific directory and records the associated metadata in a database.

[0072] 5. Review Process

[0073] The server passes the received advertising material to the generative AI model and begins the review. Specifically, the server converts the content of the advertising material into text format and generates a prompt for the generative AI model. An example of the prompt is shown below.

[0074] Example prompt sentence:

[0075] The contents of the advertising material are shown below. Please review this advertising material to determine whether it is appropriate.

[0076] Advertisement: "New smartphone now available at a special price!"

[0077] This prompt sentence is input into the generative AI model.

[0078] 6. Obtaining and notifying the results of the review

[0079] The server receives the review results from the generative AI model. The generative AI model analyzes the content of the advertising material and returns an evaluation of its appropriateness and quality. For example, the generative AI model may return an evaluation such as "This advertising material is appropriate."

[0080] Once the results are obtained, the server notifies the user of the results. The output method is a web interface or email notification. Specifically, the server displays the results on a web page so that the user can log in and check the results. If email notification is set up, the results are also sent to the user's email address.

[0081] In this way, the system can conduct the advertising material review process efficiently and objectively.

[0082] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0083] Step 1:

[0084] Users submit advertising materials to the system using a web-based submission form or API. Specifically, users open a web browser, access a specified URL, enter the required information in the form, and upload the advertising material files. When using the API, users submit the advertising material data as a POST request.

[0085] Input: Advertising material files and related information

[0086] Output: Sending the data

[0087] Step 2:

[0088] The server receives the ad material sent by the user. The server temporarily stores the data received through the submission form or API endpoint. For example, the server stores the ad material file in a specific directory and records the associated metadata in a database.

[0089] Input: Advertising material data sent by the user

[0090] Output: Saved ad material files and metadata

[0091] Step 3:

[0092] The server passes the received advertising material to the generative AI model and begins the review. Specifically, the server converts the content of the advertising material into text format and generates a prompt for the generative AI model. An example of the prompt is shown below.

[0093] Example prompt sentence:

[0094] The contents of the advertising material are shown below. Please review this advertising material to determine whether it is appropriate.

[0095] Advertisement: "New smartphone now available at a special price!"

[0096] This prompt sentence is input into the generative AI model.

[0097] Input: Saved ad material file

[0098] Output: Generated prompt statement

[0099] Step 4:

[0100] The server receives the review results from the generative AI model. The generative AI model analyzes the content of the advertising material and returns an evaluation of its appropriateness and quality. For example, the generative AI model may return an evaluation such as "This advertising material is appropriate."

[0101] Input: Generated prompt statement

[0102] Output: Review results from the generative AI model

[0103] Step 5:

[0104] The server notifies the user of the acquired screening results. The output method is a web interface or email notification. Specifically, the server displays the screening results on a web page so that the user can log in and check the results. If email notification is set up, the screening results will also be sent to the user's email address.

[0105] Input: Review results from a generative AI model

[0106] Output: Display of review results on web page and email notification

[0107] In this way, the advertising material review process is efficient and objective.

[0108] (Application example 1)

[0109] Next, a description will be given of Application Example 1 of Embodiment 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."

[0110] In the conventional advertising material review system, it took a long time from submitting advertising materials to confirming the review results, making it difficult to manage advertising efficiently. In addition, because the review criteria for advertising materials were not clear, there was insufficient understanding of the review results, and revision work was often cumbersome.

[0111] 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.

[0112] In this invention, the server includes a generating AI means for reviewing advertising materials, a means for accepting submission of advertising materials, a means for outputting the review results by the generating AI, and a means for submitting advertising materials via an application installed on a smartphone and checking the review results in real time. This shortens the time from submitting advertising materials to checking the review results, enabling efficient advertising management. In addition, because the review criteria are clearly stated, users can better understand the review results and make corrections easier.

[0113] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[0114] "Veteration" is the process of evaluating advertising materials to determine whether they meet certain criteria.

[0115] "Generative AI" is a system that uses artificial intelligence technology to automatically review advertising materials.

[0116] "Submission" refers to the act of submitting advertising materials to the system.

[0117] A "means" is a method or device used to achieve a particular purpose.

[0118] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[0119] An "application" is software installed on a smartphone to provide specific functions or services.

[0120] "Real-time" refers to immediate processing or response without delay.

[0121] A "system" is a structure in which multiple elements work together to achieve a specific function.

[0122] The "review results" are the evaluation results that the generation AI outputs after evaluating the advertising material.

[0123] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, and a means for submitting advertising materials via an application installed on a smartphone and checking the review results in real time.

[0124] Explanation of program processing

[0125] The server runs a web server using Flask and provides an API endpoint that accepts the submission of advertising materials. When a user submits advertising materials through a smartphone application, the materials are sent to the server. The server reviews the advertising materials using OpenAI's generative AI model (e.g., text-davinci-003). The review results are returned from the server to the smartphone application in JSON format, allowing the user to check the review results in real time.

[0126] Hardware and software used

[0127] Hardware: Smartphone

[0128] Software: Flask (Python framework), OpenAI API

[0129] Specific examples

[0130] Consider a scenario where an ad agency employee submits a banner image for a new ad campaign for review. The employee opens a smartphone app and uploads the new banner image. The app sends the image to a server, where a generative AI model reviews it. The resulting feedback is, "This banner is fine, but please make the text font size a little larger."

[0131] Prompt Sentence Examples

[0132] Please review the following advertising materials:

[0133] [Banner image URL]

[0134] Judging results:

[0135] Using this prompt, the generative AI model is asked to review the ad material. The review results are displayed in real time on a smartphone app, allowing the person in charge to immediately check for any corrections.

[0136] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0137] Step 1:

[0138] A user launches a smartphone application, selects and uploads advertising materials (images, videos, text, etc.). The input is the advertising materials selected by the user, and the output is the data of the uploaded advertising materials. Specifically, the user taps the "Upload" button on the application and selects advertising materials from the device's storage.

[0139] Step 2:

[0140] The device (smartphone application) sends the uploaded ad material to the server. The input is the ad material data, and the output is the ad material data sent to the server. Specifically, the application uses an HTTP POST request to send the ad material to the server's API endpoint.

[0141] Step 3:

[0142] The server passes the received ad material to the generative AI model for review. The input is the ad material data sent to the server, and the output is the review result by the generative AI model. Specifically, the server calls the OpenAI API and sends the prompt text and ad material data to the generative AI model.

[0143] Step 4:

[0144] The generative AI model reviews the ad material and generates a review result. The input is the prompt and ad material data, and the output is a text of the review result. Specifically, the generative AI model analyzes the ad material and generates feedback on its appropriateness and corrections.

[0145] Step 5:

[0146] The server receives the review results from the generative AI model and returns them to the smartphone application. The input is the review results from the generative AI model, and the output is the review results sent to the smartphone application. Specifically, the server converts the review results into JSON format and sends them to the application as an HTTP response.

[0147] Step 6:

[0148] The terminal (smartphone application) displays the review results received from the server to the user. The input is the review results from the server, and the output is the review results displayed to the user. Specifically, the application analyzes the received JSON data and displays the review results on the screen.

[0149] Step 7:

[0150] The user checks the review results and modifies the advertising material as necessary. The input is the displayed review results, and the output is the modified advertising material. Specifically, the user modifies the advertising material based on the review results and uploads it again.

[0151] Example 2

[0152] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0153] In conventional advertising material review systems, when making subjective judgments such as whether an advertising material is "offensive / not offensive" or "adult / not," the criteria for judgment are vague, resulting in low reliability of the review results. Furthermore, there are few systems that can analyze both image data and text data, making it difficult to make a comprehensive judgment.

[0154] 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.

[0155] In this invention, the server includes means including a generation AI for screening advertising materials, means for accepting submission of advertising materials, means for outputting the screening results by the generation AI, image recognition means for analyzing image data, natural language processing means for analyzing text data, means for integrating the results of the image recognition means and the natural language processing means, and means for notifying the user of the screening results. This enables subjective judgments such as whether advertising materials are "offensive / not offensive" or "adult / not" to be made with high accuracy and reliability.

[0156] "Generative AI" refers to an artificial intelligence model used to review advertising materials, including techniques for making subjective judgments.

[0157] "Advertising materials" is a general term for digital content including images and text data used as advertisements.

[0158] "Means for accepting submissions" refers to an interface or protocol for users to upload advertising materials to the system.

[0159] "Means for outputting the review results" refers to a method or device for notifying the user of the results of the review conducted by the generation AI.

[0160] "Image recognition means" refers to a technology or device for analyzing image data of advertising materials and classifying them into specific categories.

[0161] "Natural language processing means" refers to technology or devices for analyzing text data of advertising materials and classifying them into specific categories.

[0162] "Means for integrating results" refers to a method or device for integrating the analytical results of the image recognition means and the natural language processing means to arrive at a final review result.

[0163] "Means of notifying the user" refers to the interface or protocol used to communicate the review results to the user.

[0164] This invention relates to a system for reviewing advertising materials. Specifically, it is a system that uses a generative AI model to make subjective judgments about advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult." This system has the function of accepting submissions of advertising materials and outputting the review results of the generative AI.

[0165] Hardware and software used

[0166] Hardware:

[0167] Server: A server with a powerful processor and sufficient memory is required, for example, a server with an Intel Xeon processor is suitable.

[0168] Device: Users need a device such as a computer or smartphone to upload advertising materials.

[0169] software:

[0170] Generative AI models: Use deep learning frameworks such as TENSORFLOW® and PyTorch to screen advertising materials.

[0171] Image Recognition: A pre-trained image recognition model is used to analyze the image data of advertising materials.

[0172] Natural Language Processing: We use natural language processing models such as BERT and GPT-3 (registered trademark) to analyze the text data of advertising materials.

[0173] Data processing and calculation

[0174] The server receives the advertising material uploaded by the user and analyzes the data in the following manner.

[0175] 1. Image Recognition:

[0176] The server uses TensorFlow and PyTorch to analyze the image data of the advertising materials, for example, classifying the advertising images into categories such as "offensive / not offensive" and "adult / not offensive."

[0177] 2. Natural Language Processing:

[0178] The server uses BERT or GPT-3 to analyze the text data of the advertising material. For example, the text data is classified into categories such as "offensive / not offensive" and "adult / not offensive."

[0179] 3. Synthesis of results:

[0180] The server combines the results of image recognition and natural language processing to make the final decision. For example, if the image is judged to be "offensive" and the text is also judged to be "offensive," the final result will be "offensive."

[0181] Specific examples

[0182] The user uploads the following image and text as advertising material.

[0183] Image: A photo of a person engaging in inappropriate behavior

[0184] Text: "Use this product and you'll be instantly popular!"

[0185] The server first analyzes the image using TensorFlow and determines that the image is "offensive." Next, it analyzes the text using BERT and determines that the text is "offensive." Finally, the server determines that the entire ad material is "offensive" and notifies the user of the result.

[0186] Prompt Sentence Examples

[0187] "Please decide whether the following advertising materials are offensive. Analyze both the image and the text and let us know your findings."

[0188] Image: [Image URL]

[0189] Text: "Use this product and you'll be instantly popular!"

[0190] In this way, the server uses the generative AI model to make a subjective judgment on the advertising material.

[0191] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0192] Step 1:

[0193] The user uploads advertising materials.

[0194] The user uses the system's upload form to select image files and text data, and clicks the submit button. The input is the image file and text data of the advertising material. The output is the advertising material data sent to the server.

[0195] Step 2:

[0196] The server receives the image data.

[0197] The server receives image data uploaded by the user via an HTTP request and temporarily stores it. The input is the image file sent by the user. The output is the image file saved on the server. Specifically, it is saved as / tmp / uploads / image.jpg.

[0198] Step 3:

[0199] The server receives the text data.

[0200] The server receives text data uploaded by the user via an HTTP request and temporarily stores it. The input is the text data sent by the user. The output is the text data stored on the server. Specifically, it is stored in / tmp / uploads / text.txt.

[0201] Step 4:

[0202] The server performs image recognition.

[0203] The server uses TensorFlow and PyTorch to analyze the received image data. The input is the image file stored on the server. The output is the classification of the image into categories such as "offensive / not offensive" and "adult / not." Specifically, it executes model.predict(' / tmp / uploads / image.jpg') and obtains the results.

[0204] Step 5:

[0205] The server performs natural language processing.

[0206] The server uses BERT or GPT-3 to analyze the received text data. The input is the text data stored on the server. The output is the classification of the text into categories such as "offensive / not offensive" and "adult / not adult." Specifically, execute bert_model.predict(' / tmp / uploads / text.txt') to obtain the results.

[0207] Step 6:

[0208] The server aggregates the results.

[0209] The server integrates the results of image recognition and natural language processing. The input is the result of image recognition and the result of natural language processing. The output is the integrated final judgment result. For example, if both results match "unpleasant," the final result is "unpleasant."

[0210] Step 7:

[0211] The server notifies the user of the results.

[0212] The server notifies the user of the integrated results. The input is the integrated final review result. The output is the review result notified to the user. Specifically, the HTTP response is "This advertising material is offensive."

[0213] (Application example 2)

[0214] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0215] When reviewing advertising materials, it is necessary to quickly and accurately make subjective judgments, such as whether the content of images and text is offensive or adult. However, traditional methods require humans to make these judgments, which is time-consuming and has inconsistent standards. Furthermore, there is a lack of an efficient system for checking the quality of advertising materials in advance, making it difficult to control the quality of advertising.

[0216] 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.

[0217] In this invention, the server includes a generating AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generating AI, a means for analyzing images and text of the advertising materials, and a means for integrating the analysis results to make a final decision. This enables the quality of advertising materials to be screened quickly and accurately, and advertising quality control to be achieved.

[0218] "Generative AI" is an artificial intelligence technology that analyzes the content of advertising materials and makes subjective judgments.

[0219] "Advertising materials" are digital content such as images and text used as advertisements.

[0220] "Means for accepting submissions" refers to the interface or function for uploading advertising materials to the system.

[0221] "Means for outputting review results" refers to a function for displaying or reporting the analysis results of the generating AI.

[0222] "Means for analyzing images and text" refers to techniques and algorithms for separately analyzing images and text in advertising materials.

[0223] The "means of integrating the analysis results and making the final decision" is a function that integrates the analysis results of images and text to make a final decision on the quality and appropriateness of advertising materials.

[0224] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, a means for analyzing images and text in the advertising materials, and a means for integrating the analysis results to make a final decision.

[0225] The server runs a generation AI to review the advertising materials. The generation AI analyzes the images and text of the advertising materials and makes subjective judgments such as "offensive / not offensive" and "adult / not." Specifically, it uses OpenCV as its image recognition technology and Hugging Face's Transformers as its natural language processing technology.

[0226] Users submit advertising materials through an interface to upload them to the system. The submitted advertising materials are sent to the server and analyzed by the generation AI. The analysis results are integrated by the server and output as the final review results.

[0227] For example, suppose a user uploads the following image and text as advertising material:

[0228] Image: path / to / ad_image.jpg

[0229] Text: "This is an example ad text."

[0230] The server first analyzes the image using an image recognition model, then analyzes the text using a natural language processing model, and combines the results to determine whether the ad material is offensive or not, or whether it is adult content or not.

[0231] An example of a prompt for the generative AI model is as follows:

[0232] Prompt for image recognition model: "Analyze the image for offensive or adult content."

[0233] Prompt the natural language processing model: "Analyze the text for offensive or adult content."

[0234] This allows for quick and accurate review of the quality of advertising materials, enabling quality control of advertising.

[0235] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0236] Step 1:

[0237] Users upload advertising materials to the system.

[0238] Input: Advertising materials (images and text)

[0239] Output: Advertising material sent to the server

[0240] Specific operation: The user selects the image file and text of the advertising material through the system interface and clicks the upload button, which sends the advertising material to the server.

[0241] Step 2:

[0242] The server receives images of the advertising material and inputs them into an image recognition model.

[0243] Input: Ad image sent to the server

[0244] Output: Input data to the image recognition model

[0245] Specific operation: The server preprocesses the received advertising image using OpenCV and converts it into a format suitable for the image recognition model, then inputs it into the image recognition model.

[0246] Step 3:

[0247] The server analyzes the advertisement image using an image recognition model.

[0248] Input: Input data to the image recognition model

[0249] Output: Image analysis results (offensive / not offensive, adult / not)

[0250] Specific operation: The server runs an analysis on the image recognition model using the prompt "Analyze the image for offensive or adult content." and obtains the results.

[0251] Step 4:

[0252] The server receives the text of the advertising material and inputs it into a natural language processing model.

[0253] Input: Ad text sent to the server

[0254] Output: Input data to the natural language processing model

[0255] Specific operation: The server preprocesses the received advertisement text and converts it into a format suitable for the natural language processing model, then inputs it into the natural language processing model.

[0256] Step 5:

[0257] The server analyzes the advertisement text using a natural language processing model.

[0258] Input: Input data to the natural language processing model

[0259] Output: Text analysis results (offensive / not offensive, adult / not)

[0260] Specific operation: The server runs analysis against the natural language processing model using the prompt sentence "Analyze the text for offensive or adult content." and obtains the results.

[0261] Step 6:

[0262] The server integrates the image analysis results and text analysis results and makes the final decision.

[0263] Input: Image analysis results, text analysis results

[0264] Output: Final judgment result (offensive / not offensive, adult / not)

[0265] Specific operation: The server integrates the image analysis results and text analysis results to evaluate the overall quality of the advertising material, and makes a final decision based on the integrated results.

[0266] Step 7:

[0267] The server outputs the final decision result to the user.

[0268] Input: Final decision result

[0269] Output: Notification of review results to the user

[0270] Specific operation: The server displays the final judgment result through an interface to notify the user. The user can check the review result and modify the advertising material as necessary.

[0271] Example 3

[0272] Next, a description will be given of a third embodiment of the third embodiment. 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."

[0273] In conventional advertising material review systems, only the review results are provided, and the criteria for the review are not clearly stated, making it difficult for users to understand the reasons for the review results. In addition, because the review results include subjective judgments, the reliability and transparency of the results are lacking.

[0274] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0275] In this invention, the server includes a generative AI model for reviewing advertising materials, a means for accepting submission of advertising materials, a means for outputting the review results by the generative AI model, and a means for clearly displaying the review results along with the judgment criteria, thereby enabling users to clearly understand the reasons for the review results and improving the reliability and transparency of the review results.

[0276] A "generative AI model" is an artificial intelligence algorithm that analyzes data entered by the user and generates review results and their criteria.

[0277] "Advertising materials" are digital content such as text, images, and videos used as advertisements.

[0278] The "means for accepting submissions" refers to an interface or protocol that allows users to submit advertising materials to the system.

[0279] The "means for outputting the review results" is a function for providing the review results generated by the generative AI model to the user.

[0280] "Means to clearly indicate the criteria for judgment" is a function that shows the user the factors that form the basis of the results when the generative AI model outputs the review results.

[0281] This invention is a system for reviewing advertising materials, and uses a generative AI model to clearly show the review results and the criteria for judgment. Specific embodiments of this system are described below.

[0282] The server analyzes the advertising material entered by the user using a generative AI model (e.g., a general natural language processing model). This generative AI model can make subjective judgments such as whether the advertising material is "offensive / not offensive" or "adult / not adult."

[0283] The user inputs advertising material into the input field of the device. For example, the user inputs the text "This movie is very violent and unbearable to watch." The device sends this input text data to the server. The server inputs the received text data into the generative AI model for analysis.

[0284] The generative AI model analyzes the input text data and generates a review result and its criteria. For example, for the text "This movie is extremely violent and unbearable to watch," it generates the review result "unpleasant" and the criterion "contains violent content."

[0285] The server sends the review results and criteria obtained from the generative AI model to the user's device. The device then displays the received review results and criteria to the user. For example, it might display "Review result: Offensive" or "Criteria: Contains violent content" on the screen.

[0286] As a concrete example, by inputting the following prompt sentence into the generative AI model, the review results and judgment criteria can be obtained:

[0287] Please review the following text and state your findings and criteria:

[0288] "This movie is extremely violent and unbearable to watch."

[0289] In this way, the user can clearly understand the screening results and the criteria for their judgment. This system improves the reliability and transparency of the screening results, making it easier for the user to understand the reasons for the screening results. The flow of the identification process in the third embodiment will be described with reference to FIG. 15.

[0290] Step 1:

[0291] The user inputs text data.

[0292] The user inputs the text data of the advertising material that he / she wishes to review into an input field on the terminal. For example, he / she inputs the text "This movie is extremely violent and unbearable to watch."

[0293] Input: Text data of advertising materials

[0294] Output: Text data entered into the terminal

[0295] Step 2:

[0296] The terminal transmits the input text data to the server.

[0297] The terminal packages the entered text data in JSON format and sends it to the server as an HTTP request.

[0298] Input: Text data entered into the terminal

[0299] Output: Text data sent to the server

[0300] Step 3:

[0301] The server receives the text data and inputs it into the generative AI model.

[0302] The server receives the text data sent from the device and inputs it into a generative AI model, which uses, for example, a natural language processing model.

[0303] Input: Text data sent to the server

[0304] Output: Text data input to the generative AI model

[0305] Step 4:

[0306] The generative AI model analyzes the text data and generates review results and judgment criteria.

[0307] The generative AI model analyzes the input text data and generates a review result and its criteria. For example, it generates a review result such as "offensive" and a criterion such as "contains violent content."

[0308] Input: Text data fed into the generative AI model

[0309] Output: Review results and criteria

[0310] Step 5:

[0311] The server sends the review results and criteria to the user's terminal.

[0312] The server packages the review results and judgment criteria obtained from the generative AI model in JSON format and sends them to the user's device as an HTTP response.

[0313] Input: Review results and criteria

[0314] Output: Review results and criteria sent to the user's device

[0315] Step 6:

[0316] The terminal displays the review results and criteria to the user.

[0317] The device analyzes the review results and criteria received from the server and displays them to the user. For example, it might display on the screen, "Review result: Offensive" or "Criteria: Contains violent content."

[0318] Input: Review results and criteria sent to the user's device

[0319] Output: The review results and criteria displayed to the user

[0320] (Application example 3)

[0321] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[0322] In conventional advertising material review systems, the review results are output by the AI ​​generator, but the criteria for the review are not clearly stated, making it difficult for users to understand the reasons for the review results. Furthermore, content distribution services lacked a means for users to check the appropriateness of content before viewing it. This increased the risk of users encountering unpleasant content.

[0323] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0324] In this invention, the server includes a generating AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generating AI, a means for clearly indicating the judgment criteria along with the screening results, a means for generating a prompt sentence for screening content, a means for obtaining the screening results using the generating AI model, and a means for providing the screening results to the user, which makes it easier for the user to understand the reasons for the screening results and enables the user to confirm the appropriateness of the content before viewing it.

[0325] "Generative AI" is a system that uses artificial intelligence techniques to generate data and perform specific tasks.

[0326] "Advertising materials" are content such as text, images, and videos used as advertisements.

[0327] The "means for accepting submissions" refers to an interface or system for receiving advertising materials from users.

[0328] The "means for outputting the review results" is a system for displaying the results of the review conducted by the generation AI to the user.

[0329] The "means of clearly indicating the criteria for judgment" is a system that shows the user the basis for the judgment when the generation AI outputs the review results.

[0330] The "means for generating prompt sentences" is a system that creates instruction sentences to request review from the generation AI.

[0331] "Means for obtaining review results using a generative AI model" refers to a system that uses a generative AI model to review advertising materials and content and obtain the results.

[0332] "Means for providing the review results to the user" refers to a system for notifying the user of the review results conducted by the generation AI.

[0333] A system for implementing this invention includes a generating AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generating AI, a means for clearly indicating the judgment criteria along with the review results, a means for generating prompt sentences for reviewing content, a means for obtaining the review results using a generating AI model, and a means for providing the review results to a user.

[0334] The server provides an interface for receiving advertising materials from users. When a user submits advertising materials, the server uses a generation AI to review the materials. The generation AI makes subjective judgments such as whether the advertising materials are "offensive / not offensive" or "adult / not" and returns the results to the server.

[0335] The server then displays the results of the review along with the criteria for its decision. For example, if the content is deemed "offensive," the server will provide information such as "contains violent content" as the reason. This makes it easier for users to understand the reasons for the review.

[0336] Furthermore, the server generates a prompt to review the content. The generated prompt is an instruction to the generative AI model to request a review. The generative AI model reviews the advertising materials and content based on the prompt and returns the results to the server.

[0337] The server provides the user with the review results obtained using the generative AI model, allowing the user to confirm the appropriateness of the content before viewing it.

[0338] The hardware used includes an internet-connected server and the user's smartphone, and the software uses Python and the OpenAI API.

[0339] As a concrete example, consider a case where a part of a movie a user wants to watch "contains violent scenes." In this case, the generative AI receives the prompt "This movie contains violent scenes," judges the movie to be "unpleasant," and returns the reason as "It contains violent content."

[0340] Example prompt sentence:

[0341] Please review the following content:

[0342] This film contains violent scenes.

[0343] Please specify whether or not you find this content offensive and why.

[0344] In this way, the user can check the results of the content review and the reasons for it before viewing.

[0345] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0346] Step 1:

[0347] The user submits advertising materials.

[0348] Users upload advertising materials (text, images, videos, etc.) to the server. The server receives this advertising material and stores it in a database. The input is the advertising material provided by the user, and the output is the advertising material stored on the server.

[0349] Step 2:

[0350] The server reviews the advertising material using generated AI.

[0351] The server passes the stored ad material to the generation AI for review. The generation AI analyzes the ad material and makes a subjective judgment such as "offensive / not offensive" or "adult / not." The input is the ad material stored on the server, and the output is the review results by the generation AI.

[0352] Step 3:

[0353] The server obtains the review results and the criteria.

[0354] The generation AI returns the review results along with the criteria for its judgment to the server. For example, if it judges something to be "offensive," it will provide information such as "it contains violent content" as the reason. The input is the review results by the generation AI, and the output is the review results and the criteria for that judgment.

[0355] Step 4:

[0356] The server generates a prompt.

[0357] The server generates a prompt to request the AI ​​to review the ad. The prompt includes information about the ad material to be reviewed. The input is the information about the ad material to be reviewed, and the output is the generated prompt.

[0358] Step 5:

[0359] The server uses the generative AI model to obtain the review results.

[0360] The server passes the generated prompt to the generative AI model for review. The generative AI model reviews the advertising material based on the prompt and returns the results to the server. The input is the generated prompt, and the output is the review result by the generative AI model.

[0361] Step 6:

[0362] The server provides the review results to the user.

[0363] The server notifies the user of the review results obtained using the generative AI model. The user can check the review results and their judgment criteria. The input is the review results obtained by the generative AI model, and the output is the review results and their judgment criteria provided to the user.

[0364] 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.

[0365] "Example 1"

[0366] In one embodiment of the present invention, a system including a generation AI that reviews advertising materials, a means for accepting submissions of advertising materials, and a means for outputting the review results of the generation AI is combined with an emotion engine that recognizes user emotions. Specifically, the emotion engine analyzes the user's reaction (e.g., facial expression, tone of voice, etc.) when viewing the advertising materials and feeds the results back to the generation AI. This allows the generation AI to review the advertising materials based on the user's emotions.

[0367] "Example 2"

[0368] In another embodiment of the present invention, the emotion engine further includes a means for analyzing a user's reaction and feeding the result back to the generation AI's review process. Specifically, the emotion engine analyzes the user's reaction when viewing the advertising material and feeds the result back to the generation AI. This allows the generation AI to review the advertising material based on the user's emotion.

[0369] "Example 3"

[0370] In yet another embodiment of the present invention, the emotion engine further includes a means for detecting changes in the user's emotions and adjusting the screening criteria of the generation AI in response to those changes. Specifically, the emotion engine analyzes the user's reaction when viewing advertising material and feeds back the results to the generation AI. This allows the generation AI to adjust the screening criteria of the advertising material in response to changes in the user's emotions.

[0371] The processing flow of each embodiment will be described below.

[0372] "Example 1"

[0373] Step 1: User views advertising material.

[0374] Step 2: The emotion engine analyzes the user's reactions (e.g., facial expressions, tone of voice, etc.).

[0375] Step 3: The analysis results are fed back to the generation AI.

[0376] Step 4: Generative AI reviews advertising materials based on user sentiment.

[0377] "Example 2"

[0378] Step 1: User views advertising material.

[0379] Step 2: The emotion engine analyzes the user's reaction.

[0380] Step 3: The analysis results are fed back to the generation AI.

[0381] Step 4: Generative AI reviews advertising materials based on user sentiment.

[0382] "Example 3"

[0383] Step 1: User views advertising material.

[0384] Step 2: The emotion engine analyzes the user's reaction.

[0385] Step 3: The analysis results are fed back to the generation AI.

[0386] Step 4: The generative AI adjusts the review criteria for advertising materials according to changes in user emotions.

[0387] Example 1

[0388] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0389] In conventional advertising material review systems, the evaluation of the appropriateness and quality of advertising materials involves subjective judgment, which can lead to inconsistent review results. Furthermore, because the review is conducted without taking into account the user's feelings, it is difficult to accurately evaluate the actual impact of advertising materials on users. Furthermore, because the criteria for the review results are not clearly stated, users find it difficult to accept the review results.

[0390] The identification process by the identification processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means. In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, a sentiment analysis engine means for recognizing the user's emotions, and a means for feeding back the analysis results of the sentiment analysis engine to the generation AI. This allows for consistent screening of advertising materials, enables evaluation taking into account the user's emotions, and makes it easier for users to accept the screening results because the criteria for the screening results are clearly stated.

[0391] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[0392] "Advertising materials" are digital content such as images, text, and videos used as advertisements.

[0393] "Means for accepting submissions" refers to an interface such as a web form or API that allows users to upload advertising materials to the system.

[0394] "Means for outputting the review results" refers to methods such as a web interface or email notification for notifying the user of the review results generated by the generation AI.

[0395] An "emotion analysis engine" is software or hardware that analyzes a user's facial expressions, tone of voice, etc. to recognize the user's emotions.

[0396] The "means of feedback" refers to a method for sending the analysis results of the sentiment analysis engine to the generation AI and reflecting them in the review process.

[0397] The present invention is a system that uses a generative AI to review advertising materials. A specific embodiment of this system will be described below.

[0398] System Configuration

[0399] This system consists of the following main components:

[0400] 1. A means of accepting ad material submissions: A web form or API that allows users to upload ad material to the system.

[0401] 2. Generative AI: Artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[0402] 3. Means of outputting the review results: A web interface or email notification to notify the user of the review results generated by the generative AI.

[0403] 4. Emotion Analysis Engine: Software or hardware that analyzes a user's facial expressions and tone of voice to recognize their emotions.

[0404] 5. Feedback method: The results of the sentiment analysis engine's analysis are sent to the generation AI and reflected in the review process.

[0405] Submitting advertising materials

[0406] Users submit advertising materials to the system using a web-based submission form or API. Specifically, when uploading advertising images through a web browser, users enter the necessary information into the submission form, attach the image file, and submit it.

[0407] Receipt and analysis of advertising materials

[0408] The server receives the advertising material sent by the user and temporarily stores it in a database. The server then passes the received advertising material to the generation AI, which analyzes the content of the advertising material and evaluates its appropriateness and quality. For example, the generation AI checks whether the advertising text contains inappropriate language and whether the images comply with regulations.

[0409] User sentiment analysis

[0410] When a user views advertising materials, the emotion analysis engine analyzes the user's reactions. Specifically, the user's facial expressions and tone of voice are collected through a camera and microphone, and the emotion analysis engine analyzes the data. For example, it can recognize emotions such as whether the user smiled or was surprised when they saw the advertisement.

[0411] Feedback and review updates

[0412] The server feeds back the results of the emotion analysis engine to the generation AI, which then uses this feedback to make a final review of the ad material. For example, if the user response is positive, the ad material will receive a high rating.

[0413] Notification of review results

[0414] The server notifies the user of the final review results by the generation AI. Notification methods include a web interface and email notification. For example, the user can check the review results through the web interface.

[0415] Examples and prompts

[0416] As a concrete example, consider the case where a user submits an advertising image for a new cosmetic product. The user accesses a web form, enters the required information, and uploads the image file. The server receives this image and passes it to the generation AI. The generation AI analyzes the content of the image and evaluates its appropriateness. At the same time, an emotion analysis engine analyzes the user's facial expressions and tone of voice as they view the advertising image, and feeds the results back to the generation AI. Finally, the generation AI notifies the user of the review results via a web interface.

[0417] Example prompt sentence:

[0418] "Please review the advertisement images for a new cosmetic product. Analyze the user's facial expressions and tone of voice, and give a thumbs up if the user responds positively."

[0419] In this way, this system combines generative AI and an emotion analysis engine to efficiently and accurately review advertising materials.

[0420] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0421] Step 1:

[0422] Users submit advertising materials to the system using a web-based submission form or API. Specifically, when uploading advertising images through a web browser, users enter the necessary information into the submission form, attach the image file, and submit it.

[0423] Input: Ad material (image file), ad title, description, category, and other information

[0424] Output: Advertising material data sent to the server

[0425] Specific behavior:

[0426] A user accesses a web form.

[0427] The user inputs information such as the title, description, and category of the advertising material.

[0428] The user selects an advertisement image file and clicks the upload button.

[0429] Step 2:

[0430] The server receives the advertising material sent by the user and temporarily stores it in a database.

[0431] Input: Advertising material data sent by the user

[0432] Output: Advertising material data stored in the database

[0433] Specific behavior:

[0434] The server receives an HTTP request.

[0435] The server stores the advertising material data in a database.

[0436] Step 3:

[0437] The server passes the received advertising material to the generation AI, which analyzes the content of the advertising material and evaluates its appropriateness and quality.

[0438] Input: Advertising material data stored in the database

[0439] Output: Review results by the generated AI

[0440] Specific behavior:

[0441] The server inputs advertising material data into the generated AI model.

[0442] Generative AI analyzes the text and images of the ad.

[0443] The generation AI generates the review results and returns them to the server.

[0444] Step 4:

[0445] When a user views advertising materials, the emotion analysis engine analyzes the user's reactions. Specifically, the user's facial expressions and tone of voice are collected through a camera and microphone, and the emotion analysis engine analyzes the data.

[0446] Input: User's facial expression data, tone of voice data

[0447] Output: Sentiment analysis results from the sentiment analysis engine

[0448] Specific behavior:

[0449] A user accesses a web page for viewing advertising material.

[0450] The user's device activates the camera and microphone.

[0451] The emotion analysis engine analyzes the user's facial expressions and tone of voice in real time.

[0452] Step 5:

[0453] The server feeds back the results of the emotion analysis engine to the generation AI, which then performs a final review of the advertising materials based on this feedback.

[0454] Input: Emotion analysis results from the emotion analysis engine

[0455] Output: Final review results from the updated generative AI

[0456] Specific behavior:

[0457] The server receives the analysis results from the emotion analysis engine.

[0458] The server sends the analysis results to the generation AI.

[0459] The generative AI updates the review results based on the feedback.

[0460] Step 6:

[0461] The server notifies the user of the final review results by the generation AI via a web interface or email.

[0462] Input: Final review results from the updated generation AI

[0463] Output: Review results notified to the user

[0464] Specific behavior:

[0465] The server stores the review results in a database.

[0466] The server sends an email notification to the user.

[0467] The user checks the results of the review through a web interface.

[0468] (Application example 1)

[0469] Next, a description will be given of Application Example 1 of Embodiment 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."

[0470] Conventional advertising material review systems were unable to take into account subjective factors or users' emotional reactions when determining whether the content of advertising material was appropriate. This made it difficult to accurately evaluate the impact of advertising material on viewers, resulting in the risk of inappropriate advertising being delivered. Furthermore, because the criteria for review results were not clearly stated, advertisers were sometimes dissatisfied with the review results.

[0471] 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.

[0472] In this invention, the server includes a generation AI means for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine means for recognizing user emotions, and a means for analyzing the user's emotional responses and providing feedback to the generation AI. This makes it possible to take the user's emotional responses into consideration when reviewing advertising materials, thereby providing more accurate review results. Furthermore, by clearly indicating the criteria for the review results, advertisers can more easily accept the review results.

[0473] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[0474] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[0475] "Submission methods" are interfaces such as web-based forms or APIs for uploading advertising materials into the system.

[0476] "Means for outputting the review results" refers to methods such as a web interface or email notification for notifying the user of the review results generated by the generation AI.

[0477] An "emotion engine" is software or hardware that analyzes a user's facial expressions, tone of voice, etc. to recognize emotions.

[0478] The "means for analyzing emotional responses and providing feedback to the generation AI" is a system that collects emotional data when users view advertising materials and provides it to the generation AI.

[0479] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for recognizing user emotions, and a means for analyzing the user's emotional reactions and providing feedback to the generation AI. Specific embodiments of the system are described below.

[0480] The server runs a generative AI model to review ad materials. The generative AI model analyzes the content of the ad materials and evaluates their appropriateness and quality. Ad materials are submitted via a web-based form or API. When a user uploads an ad material, the server passes the material to the generative AI model for review.

[0481] The results of the review are output to the user via a web interface or email notification. The review results include subjective judgments such as whether the advertising material is "offensive / not offensive" or "adult / not." The review criteria are also clearly stated, making it easy for users to understand the results.

[0482] Furthermore, it is equipped with an emotion engine that recognizes the user's emotions. The emotion engine analyzes the user's facial expressions and tone of voice when viewing advertising materials and feeds this emotional data back to the generation AI. This allows the generation AI to conduct screening that takes the user's emotional reactions into account.

[0483] For example, when a user uploads a new advertising video, the generative AI reviews the video, analyzes the user's facial expressions while watching the video, and provides feedback along with the review results. For example, the following prompt sentence can be input into the generative AI model:

[0484] "To review advertising materials, please analyze the following advertising video. Please output the review results, taking into account the user's emotional response.

[0485] Advertisement video: [Advertisement video URL]

[0486] User's emotional response: [user's facial expression data]

[0487] In this way, it becomes possible to take into account the emotional response of the user when screening advertising materials, thereby providing more accurate screening results.

[0488] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0489] Step 1:

[0490] The user uploads advertising materials.

[0491] Input: Advertising materials (images, videos, text, etc.)

[0492] How it works: Users upload advertising materials to the server using a web-based form or API.

[0493] Output: Ad material file saved on the server

[0494] Step 2:

[0495] The server passes the advertising material to a generative AI model for review.

[0496] Input: Ad material file stored on the server

[0497] How it works: The server feeds the uploaded ad material into a generative AI model to evaluate its appropriateness and quality.

[0498] Output: Review results from the generative AI model

[0499] Step 3:

[0500] The server notifies the user of the review results.

[0501] Input: Review results from the generative AI model

[0502] How it works: The server sends the results of the review to the user via a web interface or email notification.

[0503] Output: Review results notified to the user

[0504] Step 4:

[0505] When a user views advertising material, the emotion engine analyzes their emotional response.

[0506] Input: advertising materials, user facial expressions and tone of voice

[0507] How it works: When a user views advertising materials, the emotion engine uses the camera and microphone to capture the user's facial expressions and tone of voice and analyzes the emotional data.

[0508] Output: Parsed emotion data

[0509] Step 5:

[0510] The server feeds the emotion data back to the generative AI model.

[0511] Input: Parsed emotion data

[0512] Operation: The server inputs the emotion data obtained from the emotion engine into the generative AI model and reflects it in the review results.

[0513] Output: Final review results taking into account emotional data

[0514] Step 6:

[0515] The server notifies the user of the final review result.

[0516] Input: Final review results taking into account emotional data

[0517] How it works: The server sends the final results to the user via a web interface or email notification.

[0518] Output: Final review results notified to the user

[0519] Example 2

[0520] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0521] Conventional advertising material review systems have difficulty accurately evaluating the subjective impact of advertising materials on users. Furthermore, there was a lack of a means to collect user reactions in real time and reflect them in the review process, making it impossible to accurately determine the appropriateness of advertising materials. This created a risk of displaying advertising materials that were unpleasant to users.

[0522] 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.

[0523] In this invention, the server includes means including a generation AI for reviewing advertising materials, means for accepting submission of advertising materials, means for outputting the review results by the generation AI, means including an emotion engine for collecting user reactions, and means for feeding back the analysis results of the emotion engine to the generation AI. This makes it possible to evaluate with high accuracy the subjective impact of advertising materials on users, collect user reactions in real time, and reflect them in the review process.

[0524] "Generative AI" is an artificial intelligence model that analyzes the content of advertising materials and makes subjective judgments.

[0525] "Advertising materials" are digital content such as images and text used as advertisements.

[0526] The "means for accepting submissions" is an interface that allows users to upload advertising materials to the system.

[0527] The "means of outputting the review results" is a function for notifying the user of the results of the analysis performed by the generation AI.

[0528] The "emotion engine" is a system for collecting and analyzing user reactions.

[0529] The "means of feedback" is a function that sends the analysis results of the emotion engine to the generation AI to improve the accuracy of the model.

[0530] "User response" refers to behavioral data such as the user's facial expression, gaze, and clicking behavior when viewing advertising material.

[0531] MODE FOR CARRYING OUT THE INVENTION

[0532] The present invention relates to a system for screening advertising materials. A specific embodiment of this system will be described below.

[0533] System configuration

[0534] This system consists of three main components: a server, a terminal, and a user. The server analyzes and reviews advertising materials using a generative AI model. The terminal provides an interface for users to upload advertising materials and check the review results. The terminal also has an emotion engine for collecting user responses. The user is responsible for uploading advertising materials and receiving the review results.

[0535] Hardware and software used

[0536] Server: Use a server equipped with a high-performance processor and large memory capacity. Deep learning frameworks such as TensorFlow and PyTorch are used to run generative AI models.

[0537] Device: A personal computer or smartphone equipped with a camera or sensor is used. Image processing libraries such as OpenCV and Dlib are used to run the emotion engine.

[0538] User: Uses a device with internet connectivity to upload advertising materials.

[0539] Data processing and calculation

[0540] The server receives the advertising materials uploaded by the user and analyzes them using the generative AI model. Specifically, it processes and calculates the data as follows:

[0541] Image recognition: The server uses TensorFlow to analyze images of advertising materials and identify elements within the images, thereby determining whether the advertising materials are offensive or not, or whether they are adult content or not.

[0542] Natural language processing: The server analyzes the text contained in the advertising material using BERT and GPT-3 to determine whether the content is appropriate.

[0543] The terminal collects the user's reactions when viewing advertising materials. Specifically, it processes and calculates the data as follows:

[0544] User reaction collection: The device camera captures the user's facial expressions in real time, which are analyzed by the emotion engine. Mouse movements and clicks are also recorded.

[0545] Specific examples

[0546] For example, suppose a user uploads a new advertising banner. The server uses image recognition technology to analyze the elements in the banner, while simultaneously analyzing the text content using a natural language processing model. When the user views the banner, the device's camera captures the user's facial expressions, which are analyzed by the emotion engine. The analysis results are sent to the server and fed back to the generative AI model. Finally, the server determines whether the advertising banner is appropriate and notifies the user of the result.

[0547] Prompt Sentence Examples

[0548] Below is an example of a prompt sentence to input to the generative AI model.

[0549] "Please decide whether this advertising material is offensive."

[0550] "Please analyze this advertising material to see if it contains adult content."

[0551] In this way, a system is constructed in which the server, terminal, and user work together to determine the appropriateness of advertising materials.

[0552] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0553] Step 1: Upload your creative

[0554] Input: The user uploads advertising materials (images and text) using the terminal.

[0555] Specific operation: The user opens a browser on the device, accesses the web page provided by the server, selects the advertising material file, and clicks the upload button.

[0556] Output: The server receives the uploaded advertising material.

[0557] Step 2: Analyze the advertising material

[0558] Input: The server receives the uploaded advertising material.

[0559] Specific operation: The server uses image recognition technology to analyze the image of the advertising material. It uses TensorFlow to identify elements within the image and determine whether the advertising material is "offensive / not offensive" or "adult / not." At the same time, it analyzes the text data using BERT and GPT-3 to determine whether the content is appropriate.

[0560] Output: The server obtains the analysis results from the generative AI model.

[0561] Step 3: Collect user responses

[0562] Input: User views advertising material.

[0563] How it works: While a user is viewing advertising materials, the device camera captures the user's facial expressions in real time, as well as mouse movements and clicks. This data is then sent to the emotion engine.

[0564] Output: The device collects the user's reaction data.

[0565] Step 4: Analysis by the Emotion Engine

[0566] Input: User response data sent from the device.

[0567] Specific operation: The server inputs the user's facial expression data into the emotion engine and determines whether the user felt "uncomfortable." Mouse movements and clicks are also analyzed.

[0568] Output: The server receives the analysis results of the emotion engine.

[0569] Step 5: Processing feedback

[0570] Input: Sentiment engine analysis results.

[0571] Specific operation: The server feeds back the analysis results of the emotion engine to the generative AI model and adjusts the model parameters, thereby improving the accuracy of the generative AI model.

[0572] Output: The server gets an updated generative AI model.

[0573] Step 6: Final decision and output of results

[0574] Input: Analysis results from the updated generative AI model.

[0575] Specific operation: If the server determines that the advertising material is "offensive" or "adult," it will notify the user of the result by email or notification.

[0576] Output: The user can check the results of the review on the terminal.

[0577] (Application example 2)

[0578] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0579] Conventional advertising material review systems have had the problem of making it difficult to accurately determine whether an advertisement is offensive to users or contains adult content. Furthermore, because the review does not take into account user emotions and reactions, there is also the problem of not being able to improve the user experience.

[0580] 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.

[0581] In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, an emotion engine means for analyzing user reactions, and a means for feeding back the analysis results of the emotion engine to the generation AI. This makes it possible to accurately determine whether advertising materials are offensive to users or contain adult content, and to perform screening taking into account the user's emotions and reactions.

[0582] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and makes subjective judgments.

[0583] "Advertising materials" are digital content such as images, text, and videos used as advertisements.

[0584] "Means for accepting submissions" refers to the interface or process for uploading advertising materials to the system.

[0585] "Means for outputting review results" refers to the interface or process for displaying the results of the review conducted by the generation AI to the user.

[0586] An "emotion engine" is software or hardware that analyzes a user's reactions and determines their emotions.

[0587] "Feedback means" refers to the interface or process for transmitting the analysis results of the emotion engine to the generative AI and reflecting them in the review process.

[0588] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for analyzing user reactions, and a means for feeding back the analysis results of the emotion engine to the generation AI.

[0589] The server receives the image and text of the ad material as input and analyzes its content using generative AI. Specifically, it uses image recognition and natural language processing models to determine whether the ad material is offensive or adult. OpenCV is used for image recognition, and Hugging Face Transformers is used for natural language processing.

[0590] The emotion engine analyzes the reaction data (video and audio) when a user views advertising materials. The emotion engine uses the Emotion Engine to determine the user's emotions. The analysis results are fed back to the generation AI, which then reviews the ad based on the user's emotions.

[0591] The device displays the results of the AI's review to the user, including the criteria for determining whether the ad material is offensive or adult content.

[0592] For example, if an ad image contains a violent scene, the image recognition model will judge it as "offensive" and the ad will be deemed unpleasant. If the ad text contains discriminatory language, the natural language processing model will judge it as "negative" and the ad will be deemed unpleasant. Furthermore, if a user shows a negative reaction, such as grimacing, when viewing the ad, the emotion engine will judge it as "negative" and the ad will be deemed unpleasant.

[0593] An example of a prompt to input to a generative AI model is as follows:

[0594] "Analyze ad images and text to detect offensive or adult content. Analyze user response data and determine the ad as offensive if there is a negative response."

[0595] In this way, it is possible to accurately determine whether advertising material is offensive to users or contains adult content, and furthermore, to conduct screening that takes into account the user's feelings and reactions.

[0596] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0597] Step 1:

[0598] The server receives the image and text of the advertising material as input. When the user uploads the advertising material, the server acquires the data. The input data consists of image files and text data.

[0599] Step 2:

[0600] The server analyzes advertising images using an image recognition model. Specifically, it preprocesses the images using OpenCV and classifies the image content using Hugging Face Transformers. Data processing includes image resizing and noise removal. The output is a classification of the image with labels such as "offensive" or "adult."

[0601] Step 3:

[0602] The server analyzes the ad text using a natural language processing model. Specifically, it uses Hugging Face's Transformers to perform sentiment analysis of the text. Data processing involves tokenizing the text and calculating a sentiment score. The output is a classification of the text with labels such as "negative" or "neutral."

[0603] Step 4:

[0604] The server combines the results of steps 2 and 3 to determine whether the advertising material is offensive or adult. Specifically, it combines image and text labels to make a comprehensive assessment. The output is a judgment result such as whether the advertising material is "offensive" or "not offensive," or "adult" or "not adult."

[0605] Step 5:

[0606] The device collects reaction data (video and audio) when the user views advertising materials and sends it to the server. The input data is real-time data of the user's facial expressions and voice.

[0607] Step 6:

[0608] The server analyzes the user's reaction data using an emotion engine. Specifically, it uses the Emotion Engine to determine the user's emotions. Data processing involves extracting video frames and analyzing audio spectra. The output is a classification of the user's emotions with labels such as "positive" or "negative."

[0609] Step 7:

[0610] The server feeds back the emotion engine's analysis results to the generation AI and updates the final review result of the advertising material. Specifically, if the user's emotion is negative, the advertising material is judged to be unpleasant. The output is the updated review result.

[0611] Step 8:

[0612] The terminal displays the final review results received from the server to the user. Specifically, the criteria for determining whether the advertising material is offensive or adult are also clearly indicated. The output is the review results displayed to the user.

[0613] Example 3

[0614] Next, a description will be given of a third embodiment of the third embodiment. 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."

[0615] In conventional advertising material review systems, it was difficult to provide a satisfactory explanation to users because the criteria used to determine the review results were unclear. Additionally, the review criteria could not be flexibly adjusted in response to changes in user emotions, making it difficult to respond to diverse user reactions.

[0616] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0617] In this invention, the server includes a generative AI model for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results of the generative AI model, a means for clearly displaying the screening results along with the judgment criteria, an emotion engine for detecting changes in the user's emotions, and a means for adjusting the screening criteria of the generative AI model based on feedback from the emotion engine. This increases the transparency of the screening results and enables flexible adjustment of the screening criteria according to the user's emotions.

[0618] "Generative AI Model" means an artificial intelligence model used to screen advertising materials.

[0619] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[0620] "Means for accepting submissions" refers to the interface or protocol for uploading or transmitting advertising materials to the system.

[0621] The "means for outputting the review results" is a function for displaying or notifying the user or administrator of the results of the generative AI model's review of the advertising material.

[0622] "Means to clearly indicate the criteria for judgment" is a function that clearly indicates the factors and reasons that form the basis for the judgment when the generative AI model outputs the review results.

[0623] An "emotion engine" is software or hardware for detecting and analyzing changes in a user's emotions.

[0624] "Means for adjusting the evaluation criteria of the generative AI model based on feedback" is a function for dynamically changing the evaluation criteria of the generative AI model based on the user's emotional data obtained from the emotion engine.

[0625] The present invention is a system for reviewing advertising materials, which includes a generative AI model, an emotion engine, and various means for linking them. Specific embodiments of this system are described below.

[0626] Hardware and software used

[0627] Hardware: Servers, user devices (PCs, smartphones, etc.)

[0628] Software: Generative AI model, emotion engine

[0629] Server Operation

[0630] The server runs a generative AI model to review advertising materials. When advertising materials are submitted by advertising providers or users, the server receives the materials and inputs them into the generative AI model. The generative AI model makes subjective judgments about the advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult," and outputs the results.

[0631] In addition, the server outputs information that clearly explains the criteria used to determine the ad's content, along with the results of the review. For example, if the ad is deemed "offensive," the server will provide information such as "contains violent content" as the reason.

[0632] Device behavior

[0633] The device displays the advertising material received from the server to the user. When the user views the advertising material, the device uses an emotion engine to analyze the user's reaction. The emotion engine uses a camera and microphone to analyze the user's facial expressions and tone of voice and generate emotion data.

[0634] User Actions

[0635] The user views the advertising material displayed on their device. When a change in the user's emotion is detected, the information is sent from the device to the server. The server adjusts the evaluation criteria of the generative AI model based on the received emotion data.

[0636] Specific examples

[0637] Example 1: Advertising material review

[0638] The server receives a "new game ad video" from an ad provider. The generative AI model reviews this ad video and determines it to be "offensive." It outputs the information that the reason for this is that it "contains violent content."

[0639] Example prompt: "Please review this advertising material and print out your results and criteria."

[0640] Example 2: Feedback from an Emotion Engine

[0641] The device displays to the user an "advertising video for a new game" received from the server. While the user is watching the advertising video, the device analyzes facial expressions with a camera and detects "discomfort." The user's "discomfort" data is sent from the device to the server. The server adjusts the evaluation criteria of the generative AI model based on the received "discomfort" data and reflects it in future evaluations.

[0642] Example prompt: "Detect changes in the user's emotions and provide that information as feedback to the generative AI."

[0643] In this way, the system can provide flexible screening criteria that correspond to the user's feelings. The flow of the identification process in the third embodiment will be described with reference to FIG.

[0644] Step 1:

[0645] The server receives advertising material from advertising providers or users.

[0646] Specifically, the server receives an HTTP request and stores the advertising material data.

[0647] Input: Advertising materials (images, videos, text, etc.)

[0648] Output: Saved advertising material data

[0649] Step 2:

[0650] The server inputs the received advertising material into a generative AI model and reviews it.

[0651] Specifically, the server passes the advertising material data to the generative AI model and obtains the review results.

[0652] Input: Advertising material data

[0653] Output: Review result (e.g. "offensive", "adult", etc.)

[0654] Step 3:

[0655] The server outputs the review results and judgment criteria obtained from the generative AI model.

[0656] Specifically, the server generates the review results and criteria in JSON format and returns them to the user.

[0657] Input: Review results

[0658] Output: JSON data of the review results and criteria

[0659] Step 4:

[0660] The terminal displays the advertising material received from the server to the user.

[0661] Specifically, the terminal displays advertising material on the screen and waits for the user's response.

[0662] Input: Advertising material data

[0663] Output: Displayed advertising material

[0664] Step 5:

[0665] The terminal uses an emotion engine to analyze the user's reaction.

[0666] Specifically, the device uses a camera and microphone to analyze the user's facial expressions and tone of voice and generate emotional data.

[0667] Input: User facial and voice data

[0668] Output: Emotion data

[0669] Step 6:

[0670] The user's emotion data is transmitted from the terminal to the server.

[0671] Specifically, the terminal transmits emotion data generated by the emotion engine to the server.

[0672] Input: Emotion data

[0673] Output: Emotion data sent to the server

[0674] Step 7:

[0675] The server adjusts the evaluation criteria for the generative AI model based on the received emotional data.

[0676] Specifically, the server analyzes the emotion data and updates the parameters of the generative AI model.

[0677] Input: Emotion data

[0678] Output: Adjusted generative AI model parameters

[0679] (Application example 3)

[0680] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[0681] In conventional advertising material review systems, when the AI ​​generator outputs the review results, the criteria for its judgment are not clearly stated, resulting in a lack of transparency for users. Furthermore, because the review is conducted based on uniform criteria without taking into account the user's feelings, there is a problem in that flexible review based on the user's feelings cannot be performed. This increases the possibility that the advertising material will cause discomfort to users, potentially reducing the effectiveness of the advertisement.

[0682] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[0683] In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, an emotion engine means for analyzing user emotions, a means for feeding back the analysis results of the emotion engine to the generation AI, and a means for the generation AI to adjust the screening criteria based on the user emotions. This enables flexible screening according to user emotions, improves the transparency of the screening results, and makes it possible to reduce the discomfort that advertising materials cause to users.

[0684] "Generative AI" refers to artificial intelligence used to review advertising materials and determine whether they are suitable based on specific criteria.

[0685] "Advertising materials" refers to all content used as advertisements, including images, videos, and text.

[0686] "Means for accepting submissions" refers to the interface or process for uploading advertising materials to the system.

[0687] "Means for outputting the review results" refers to the interface or process for displaying the results of the review conducted by the generation AI to the user.

[0688] An "emotion engine" is software or hardware for analyzing a user's emotions, and detects emotions from the user's facial expressions, voice, etc.

[0689] "Means of feedback" refers to the interface or process for transmitting the analysis results of the emotion engine to the generative AI.

[0690] "Means for adjusting the review criteria" refers to an interface or process that allows the generative AI to dynamically change the review criteria based on user sentiment.

[0691] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for analyzing user emotions, a means for feeding back the analysis results of the emotion engine to the generation AI, and a means for the generation AI to adjust the review criteria based on the user's emotions.

[0692] Hardware and software used

[0693] Hardware: Smartphone camera, smartphone display

[0694] Software: OpenCV (image processing library), Keras (sentiment analysis model framework), Transformers (generative AI model framework)

[0695] Data processing and calculation

[0696] Display of advertising materials

[0697] The server uses OpenCV to display the advertising material on the smartphone display, allowing users to visually check the advertising material.

[0698] Emotion analysis

[0699] The device (smartphone) captures the user's face using a camera and analyzes emotions using a Keras model. Specifically, the captured image is converted to grayscale, face detection is performed, and then the image is input into the emotion analysis model. The model estimates emotions from the user's facial expressions and outputs the results.

[0700] Emotional Feedback

[0701] The server feeds back the analysis results of the emotion engine to the generation AI. Specifically, the analyzed emotional information is input to the generation AI as a prompt sentence, and the generation AI adjusts the screening criteria for advertising materials based on that information.

[0702] Adjustment of review criteria

[0703] The generative AI dynamically changes the judging criteria based on the emotional information fed back, enabling flexible judging based on the user's emotions.

[0704] Specific examples

[0705] For example, if a promotional video for a new smartphone is displayed as advertising material and the user responds with "Happy," that emotion is fed back to the generating AI, which then adjusts the review criteria for the next advertising material.

[0706] Prompt Sentence Examples

[0707] The user's emotion is Happy. Based on this emotion, please evaluate the following advertising material: Promotional video for a new smartphone

[0708] In this way, the screening criteria for advertising materials can be dynamically adjusted based on user sentiment, which improves the transparency of screening results and reduces the discomfort that advertising materials cause to users.

[0709] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[0710] Step 1:

[0711] The server displays the advertising material on the smartphone display. Specifically, it uses OpenCV to read the image or video file of the advertising material and displays it on the smartphone display. The input is the file path of the advertising material, and the output is the advertising material displayed on the display.

[0712] Step 2:

[0713] The device (smartphone) captures the user's face using a camera. Specifically, the camera is activated and captures the user's face in real time. The input is the video stream from the camera, and the output is the captured image of the user's face.

[0714] Step 3:

[0715] The device converts the captured face image to grayscale and performs face detection. Specifically, it applies a face detection algorithm using OpenCV to identify the face position. The input is the captured face image, and the output is face position information.

[0716] Step 4:

[0717] The device extracts the detected face area and inputs it into an emotion analysis model. Specifically, the face image is input into an emotion analysis model pre-trained using Keras, and emotions are estimated. The input is the detected face image, and the output is the estimated emotion label.

[0718] Step 5:

[0719] The server feeds back the analysis results of the emotion engine to the generation AI. Specifically, it inputs the estimated emotion label to the generation AI as a prompt sentence. The input is the estimated emotion label, and the output is the prompt sentence input to the generation AI.

[0720] Step 6:

[0721] The generative AI adjusts the ad material's review criteria based on the feedback emotional information. Specifically, it analyzes the prompt text and dynamically changes the review criteria. The input is the prompt text, and the output is the adjusted review criteria.

[0722] Step 7:

[0723] The server outputs the results of the generative AI's review. Specifically, it judges the suitability of the advertising material based on the adjusted review criteria and displays the results to the user. The input is the adjusted review criteria and the advertising material, and the output is the review results.

[0724] 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.

[0725] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ) 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.

[0726] Another example of generative AI is Gemini (registered trademark) (Internet search engine). <url: https: gemini.google.com ?hl="ja">) are listed.

[0727] 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.

[0728] [Second embodiment]

[0729] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0730] 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.

[0731] 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).

[0732] 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.

[0733] 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.

[0734] 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).

[0735] 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.

[0736] 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.

[0737] 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.

[0738] 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.

[0739] 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.

[0740] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0741] "Example 1"

[0742] The present invention is a system that incorporates a generation AI for reviewing advertising materials. Specifically, a web-based submission form or API is used as a means for accepting submissions of advertising materials. The submitted advertising materials are reviewed by the generation AI, and the review results are communicated to the user via an output means, such as a web interface or email notification.

[0743] "Example 2"

[0744] The generative AI makes subjective judgments such as whether the ad material is "offensive / not offensive" or "adult / not adult" by analyzing the content of the ad material using technologies such as image recognition and natural language processing, and makes a judgment based on the results.

[0745] "Example 3"

[0746] When the generative AI outputs the review results, it will include a means to clearly indicate the criteria used. Specifically, along with the review results, it will output information indicating what factors were used as the basis for the judgment. For example, if a content is judged to be "offensive," it will output information such as "contains violent language" as the reason.

[0747] The processing flow of each embodiment will be described below.

[0748] "Example 1"

[0749] Step 1: Advertisers submit their ad materials to the system via a web-based submission form or API.

[0750] Step 2: The submitted ad material is reviewed by the AI ​​generator, which analyzes the content of the ad material using techniques such as image recognition and natural language processing.

[0751] Step 3: The generating AI outputs the review results and communicates them to the advertiser via a web interface or email notification.

[0752] "Example 2"

[0753] Step 1: When advertising materials are submitted to the system, the generation AI makes a subjective judgment on the advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult."

[0754] Step 2: The generative AI analyzes the content of the advertising material using technologies such as image recognition and natural language processing, and makes a decision based on the results.

[0755] Step 3: The generating AI outputs the review results and their criteria, and notifies the advertiser of the review results. "Example 3"

[0756] Step 1: Once the ad material is submitted to the system, the generation AI reviews it.

[0757] Step 2: The generative AI outputs the results of the review and the criteria for its decision. Specifically, it outputs information indicating the factors that formed the basis for the decision along with the review results.

[0758] Step 3: The AI's review results and criteria are communicated to the advertiser. For example, if the content is deemed "offensive," the AI ​​outputs information such as "contains violent content" as the reason.

[0759] Example 1

[0760] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0761] The traditional advertising material review process is often done manually, which is time-consuming and labor-intensive, and prone to subjective judgment. Furthermore, there are sometimes delays in notifying review results, which creates the risk of missing the timing of advertising campaigns. To solve these problems, a system that provides an efficient and objective review process is needed.

[0762] 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.

[0763] In this invention, the server includes means for converting the content of the advertising material into text format, generating and inputting a prompt sentence to the generative AI model, means for acquiring the review results from the generative AI model, and means for notifying the user of the review results via a web interface or email notification, thereby automating the advertising material review process and enabling efficient and objective review.

[0764] "Advertising materials" are digital content such as text, images, and videos used as advertisements.

[0765] "Generative AI" is a system that uses artificial intelligence technology to automatically perform specific tasks, in this case reviewing advertising materials.

[0766] "Means for accepting submissions" refers to an interface through which users submit advertising materials to the system, and includes web forms, APIs, etc.

[0767] The "means for outputting the review results" refers to an interface that the generation AI uses to notify the user of the results of its review of the advertising material, and includes a web interface and email notification.

[0768] "Means for converting into text format" refers to the process of analyzing the content of advertising materials and converting them into a format that can be input as text data into a generative AI model.

[0769] A "prompt" is an instruction given to a generative AI model to perform a specific task.

[0770] "Means for obtaining review results from the generative AI model" refers to the process for receiving the review results output by the generative AI model and processing them within the system.

[0771] A "web interface" is an interface that allows a user to access the system through a web browser to perform operations and check results.

[0772] "Email notification" is a function that allows the system to send review results and other information to the user's email address.

[0773] MODE FOR CARRYING OUT THE INVENTION

[0774] The present invention provides a system for efficiently and objectively reviewing advertising materials. This system automatically reviews advertising materials using a generative AI model and notifies the user of the results. Specific embodiments are described below.

[0775] 1. Introduction

[0776] The system aims to automate the entire process, from submitting advertising materials to reviewing and notifying the results. Users submit advertising materials to the system using a web-based submission form or API, and the server reviews them using a generative AI model. The review results are communicated to users via a web interface or email notification.

[0777] 2. Hardware and Software Used

[0778] Server: Receives advertising materials, runs the generative AI model, and outputs the review results.

[0779] Generative AI model: An artificial intelligence model used to review advertising materials, such as OpenAI's GPT-4.

[0780] Web interface: An interface for users to submit advertising materials and check the review results.

[0781] Email notification system: A system for notifying users of the review results by email.

[0782] 3. Submit advertising materials

[0783] Users submit advertising materials to the system using a web-based submission form or API. Specifically, users open a web browser, access a specified URL, enter the required information in the form, and upload the advertising material files. When using the API, users submit the advertising material data as a POST request.

[0784] 4. Receiving and storing advertising materials

[0785] The server receives the ad material sent by the user. The server temporarily stores the data received through the submission form or API endpoint. For example, the server stores the ad material file in a specific directory and records the associated metadata in a database.

[0786] 5. Review Process

[0787] The server passes the received advertising material to the generative AI model and begins the review. Specifically, the server converts the content of the advertising material into text format and generates a prompt for the generative AI model. An example of the prompt is shown below.

[0788] Example prompt sentence:

[0789] The contents of the advertising material are shown below. Please review this advertising material to determine whether it is appropriate.

[0790] Advertisement: "New smartphone now available at a special price!"

[0791] This prompt sentence is input into the generative AI model.

[0792] 6. Obtaining and notifying the results of the review

[0793] The server receives the review results from the generative AI model. The generative AI model analyzes the content of the advertising material and returns an evaluation of its appropriateness and quality. For example, the generative AI model may return an evaluation such as "This advertising material is appropriate."

[0794] Once the results are obtained, the server notifies the user of the results. The output method is a web interface or email notification. Specifically, the server displays the results on a web page so that the user can log in and check the results. If email notification is set up, the results are also sent to the user's email address.

[0795] In this way, the system can conduct the advertising material review process efficiently and objectively.

[0796] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0797] Step 1:

[0798] Users submit advertising materials to the system using a web-based submission form or API. Specifically, users open a web browser, access a specified URL, enter the required information in the form, and upload the advertising material files. When using the API, users submit the advertising material data as a POST request.

[0799] Input: Advertising material files and related information

[0800] Output: Sending the data

[0801] Step 2:

[0802] The server receives the ad material sent by the user. The server temporarily stores the data received through the submission form or API endpoint. For example, the server stores the ad material file in a specific directory and records the associated metadata in a database.

[0803] Input: Advertising material data sent by the user

[0804] Output: Saved ad material files and metadata

[0805] Step 3:

[0806] The server passes the received advertising material to the generative AI model and begins the review. Specifically, the server converts the content of the advertising material into text format and generates a prompt for the generative AI model. An example of the prompt is shown below.

[0807] Example prompt sentence:

[0808] The contents of the advertising material are shown below. Please review this advertising material to determine whether it is appropriate.

[0809] Advertisement: "New smartphone now available at a special price!"

[0810] This prompt sentence is input into the generative AI model.

[0811] Input: Saved ad material file

[0812] Output: Generated prompt statement

[0813] Step 4:

[0814] The server receives the review results from the generative AI model. The generative AI model analyzes the content of the advertising material and returns an evaluation of its appropriateness and quality. For example, the generative AI model may return an evaluation such as "This advertising material is appropriate."

[0815] Input: Generated prompt statement

[0816] Output: Review results from the generative AI model

[0817] Step 5:

[0818] The server notifies the user of the acquired screening results. The output method is a web interface or email notification. Specifically, the server displays the screening results on a web page so that the user can log in and check the results. If email notification is set up, the screening results will also be sent to the user's email address.

[0819] Input: Review results from a generative AI model

[0820] Output: Display of review results on web page and email notification

[0821] In this way, the advertising material review process is efficient and objective.

[0822] (Application example 1)

[0823] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0824] In the conventional advertising material review system, it took a long time from submitting advertising materials to confirming the review results, making it difficult to manage advertising efficiently. In addition, because the review criteria for advertising materials were not clear, there was insufficient understanding of the review results, and revision work was often cumbersome.

[0825] 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.

[0826] In this invention, the server includes a generating AI means for reviewing advertising materials, a means for accepting submission of advertising materials, a means for outputting the review results by the generating AI, and a means for submitting advertising materials via an application installed on a smartphone and checking the review results in real time. This shortens the time from submitting advertising materials to checking the review results, enabling efficient advertising management. In addition, because the review criteria are clearly stated, users can better understand the review results and make corrections easier.

[0827] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[0828] "Veteration" is the process of evaluating advertising materials to determine whether they meet certain criteria.

[0829] "Generative AI" is a system that uses artificial intelligence technology to automatically review advertising materials.

[0830] "Submission" refers to the act of submitting advertising materials to the system.

[0831] A "means" is a method or device used to achieve a particular purpose.

[0832] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[0833] An "application" is software installed on a smartphone to provide specific functions or services.

[0834] "Real-time" refers to immediate processing or response without delay.

[0835] A "system" is a structure in which multiple elements work together to achieve a specific function.

[0836] The "review results" are the evaluation results that the generation AI outputs after evaluating the advertising material.

[0837] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, and a means for submitting advertising materials via an application installed on a smartphone and checking the review results in real time.

[0838] Explanation of program processing

[0839] The server runs a web server using Flask and provides an API endpoint that accepts the submission of advertising materials. When a user submits advertising materials through a smartphone application, the materials are sent to the server. The server reviews the advertising materials using OpenAI's generative AI model (e.g., text-davinci-003). The review results are returned from the server to the smartphone application in JSON format, allowing the user to check the review results in real time.

[0840] Hardware and software used

[0841] Hardware: Smartphone

[0842] Software: Flask (Python framework), OpenAI API

[0843] Specific examples

[0844] Consider a scenario where an ad agency employee submits a banner image for a new ad campaign for review. The employee opens a smartphone app and uploads the new banner image. The app sends the image to a server, where a generative AI model reviews it. The resulting feedback is, "This banner is fine, but please make the text font size a little larger."

[0845] Prompt Sentence Examples

[0846] Please review the following advertising materials:

[0847] [Banner image URL]

[0848] Judging results:

[0849] Using this prompt, the generative AI model is asked to review the ad material. The review results are displayed in real time on a smartphone app, allowing the person in charge to immediately check for any corrections.

[0850] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0851] Step 1:

[0852] A user launches a smartphone application, selects and uploads advertising materials (images, videos, text, etc.). The input is the advertising materials selected by the user, and the output is the data of the uploaded advertising materials. Specifically, the user taps the "Upload" button on the application and selects advertising materials from the device's storage.

[0853] Step 2:

[0854] The device (smartphone application) sends the uploaded ad material to the server. The input is the ad material data, and the output is the ad material data sent to the server. Specifically, the application uses an HTTP POST request to send the ad material to the server's API endpoint.

[0855] Step 3:

[0856] The server passes the received ad material to the generative AI model for review. The input is the ad material data sent to the server, and the output is the review result by the generative AI model. Specifically, the server calls the OpenAI API and sends the prompt text and ad material data to the generative AI model.

[0857] Step 4:

[0858] The generative AI model reviews the ad material and generates a review result. The input is the prompt and ad material data, and the output is a text of the review result. Specifically, the generative AI model analyzes the ad material and generates feedback on its appropriateness and corrections.

[0859] Step 5:

[0860] The server receives the review results from the generative AI model and returns them to the smartphone application. The input is the review results from the generative AI model, and the output is the review results sent to the smartphone application. Specifically, the server converts the review results into JSON format and sends them to the application as an HTTP response.

[0861] Step 6:

[0862] The terminal (smartphone application) displays the review results received from the server to the user. The input is the review results from the server, and the output is the review results displayed to the user. Specifically, the application analyzes the received JSON data and displays the review results on the screen.

[0863] Step 7:

[0864] The user checks the review results and modifies the advertising material as necessary. The input is the displayed review results, and the output is the modified advertising material. Specifically, the user modifies the advertising material based on the review results and uploads it again.

[0865] Example 2

[0866] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0867] In conventional advertising material review systems, when making subjective judgments such as whether an advertising material is "offensive / not offensive" or "adult / not," the criteria for judgment are vague, resulting in low reliability of the review results. Furthermore, there are few systems that can analyze both image data and text data, making it difficult to make a comprehensive judgment.

[0868] 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.

[0869] In this invention, the server includes means including a generation AI for screening advertising materials, means for accepting submission of advertising materials, means for outputting the screening results by the generation AI, image recognition means for analyzing image data, natural language processing means for analyzing text data, means for integrating the results of the image recognition means and the natural language processing means, and means for notifying the user of the screening results. This enables subjective judgments such as whether advertising materials are "offensive / not offensive" or "adult / not" to be made with high accuracy and reliability.

[0870] "Generative AI" refers to an artificial intelligence model used to review advertising materials, including techniques for making subjective judgments.

[0871] "Advertising materials" is a general term for digital content including images and text data used as advertisements.

[0872] "Means for accepting submissions" refers to an interface or protocol for users to upload advertising materials to the system.

[0873] "Means for outputting the review results" refers to a method or device for notifying the user of the results of the review conducted by the generation AI.

[0874] "Image recognition means" refers to a technology or device for analyzing image data of advertising materials and classifying them into specific categories.

[0875] "Natural language processing means" refers to technology or devices for analyzing text data of advertising materials and classifying them into specific categories.

[0876] "Means for integrating results" refers to a method or device for integrating the analytical results of the image recognition means and the natural language processing means to arrive at a final review result.

[0877] "Means of notifying the user" refers to the interface or protocol used to communicate the review results to the user.

[0878] This invention relates to a system for reviewing advertising materials. Specifically, it is a system that uses a generative AI model to make subjective judgments about advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult." This system has the function of accepting submissions of advertising materials and outputting the review results of the generative AI.

[0879] Hardware and software used

[0880] Hardware:

[0881] Server: A server with a powerful processor and sufficient memory is required, for example, a server with an Intel Xeon processor is suitable.

[0882] Device: Users need a device such as a computer or smartphone to upload advertising materials.

[0883] software:

[0884] Generative AI models: Use deep learning frameworks such as TensorFlow and PyTorch to screen advertising materials.

[0885] Image Recognition: A pre-trained image recognition model is used to analyze the image data of advertising materials.

[0886] Natural Language Processing: Use natural language processing models such as BERT and GPT-3 to analyze text data from advertising materials.

[0887] Data processing and calculation

[0888] The server receives the advertising material uploaded by the user and analyzes the data in the following manner.

[0889] 1. Image Recognition:

[0890] The server uses TensorFlow and PyTorch to analyze the image data of the advertising materials, for example, classifying the advertising images into categories such as "offensive / not offensive" and "adult / not offensive."

[0891] 2. Natural Language Processing:

[0892] The server uses BERT or GPT-3 to analyze the text data of the advertising material. For example, the text data is classified into categories such as "offensive / not offensive" and "adult / not offensive."

[0893] 3. Synthesis of results:

[0894] The server combines the results of image recognition and natural language processing to make the final decision. For example, if the image is judged to be "offensive" and the text is also judged to be "offensive," the final result will be "offensive."

[0895] Specific examples

[0896] The user uploads the following image and text as advertising material.

[0897] Image: A photo of a person engaging in inappropriate behavior

[0898] Text: "Use this product and you'll be instantly popular!"

[0899] The server first analyzes the image using TensorFlow and determines that the image is "offensive." Next, it analyzes the text using BERT and determines that the text is "offensive." Finally, the server determines that the entire ad material is "offensive" and notifies the user of the result.

[0900] Prompt Sentence Examples

[0901] "Please decide whether the following advertising materials are offensive. Analyze both the image and the text and let us know your findings."

[0902] Image: [Image URL]

[0903] Text: "Use this product and you'll be instantly popular!"

[0904] In this way, the server uses the generative AI model to make a subjective judgment on the advertising material.

[0905] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0906] Step 1:

[0907] The user uploads advertising materials.

[0908] The user uses the system's upload form to select image files and text data, and clicks the submit button. The input is the image file and text data of the advertising material. The output is the advertising material data sent to the server.

[0909] Step 2:

[0910] The server receives the image data.

[0911] The server receives image data uploaded by the user via an HTTP request and temporarily stores it. The input is the image file sent by the user. The output is the image file saved on the server. Specifically, it is saved as / tmp / uploads / image.jpg.

[0912] Step 3:

[0913] The server receives the text data.

[0914] The server receives text data uploaded by the user via an HTTP request and temporarily stores it. The input is the text data sent by the user. The output is the text data stored on the server. Specifically, it is stored in / tmp / uploads / text.txt.

[0915] Step 4:

[0916] The server performs image recognition.

[0917] The server uses TensorFlow and PyTorch to analyze the received image data. The input is the image file stored on the server. The output is the classification of the image into categories such as "offensive / not offensive" and "adult / not." Specifically, it executes model.predict(' / tmp / uploads / image.jpg') and obtains the results.

[0918] Step 5:

[0919] The server performs natural language processing.

[0920] The server uses BERT or GPT-3 to analyze the received text data. The input is the text data stored on the server. The output is the classification of the text into categories such as "offensive / not offensive" and "adult / not adult." Specifically, execute bert_model.predict(' / tmp / uploads / text.txt') to obtain the results.

[0921] Step 6:

[0922] The server aggregates the results.

[0923] The server integrates the results of image recognition and natural language processing. The input is the result of image recognition and the result of natural language processing. The output is the integrated final judgment result. For example, if both results match "unpleasant," the final result is "unpleasant."

[0924] Step 7:

[0925] The server notifies the user of the results.

[0926] The server notifies the user of the integrated results. The input is the integrated final review result. The output is the review result notified to the user. Specifically, the HTTP response is "This advertising material is offensive."

[0927] (Application example 2)

[0928] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0929] When reviewing advertising materials, it is necessary to quickly and accurately make subjective judgments, such as whether the content of images and text is offensive or adult. However, traditional methods require humans to make these judgments, which is time-consuming and has inconsistent standards. Furthermore, there is a lack of an efficient system for checking the quality of advertising materials in advance, making it difficult to control the quality of advertising.

[0930] 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.

[0931] In this invention, the server includes a generating AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generating AI, a means for analyzing images and text of the advertising materials, and a means for integrating the analysis results to make a final decision. This enables the quality of advertising materials to be screened quickly and accurately, and advertising quality control to be achieved.

[0932] "Generative AI" is an artificial intelligence technology that analyzes the content of advertising materials and makes subjective judgments.

[0933] "Advertising materials" are digital content such as images and text used as advertisements.

[0934] "Means for accepting submissions" refers to the interface or function for uploading advertising materials to the system.

[0935] "Means for outputting review results" refers to a function for displaying or reporting the analysis results of the generating AI.

[0936] "Means for analyzing images and text" refers to techniques and algorithms for separately analyzing images and text in advertising materials.

[0937] The "means of integrating the analysis results and making the final decision" is a function that integrates the analysis results of images and text to make a final decision on the quality and appropriateness of advertising materials.

[0938] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, a means for analyzing images and text in the advertising materials, and a means for integrating the analysis results to make a final decision.

[0939] The server runs a generation AI to review the advertising materials. The generation AI analyzes the images and text of the advertising materials and makes subjective judgments such as "offensive / not offensive" and "adult / not." Specifically, it uses OpenCV as its image recognition technology and Hugging Face's Transformers as its natural language processing technology.

[0940] Users submit advertising materials through an interface to upload them to the system. The submitted advertising materials are sent to the server and analyzed by the generation AI. The analysis results are integrated by the server and output as the final review results.

[0941] For example, suppose a user uploads the following image and text as advertising material:

[0942] Image: path / to / ad_image.jpg

[0943] Text: "This is an example ad text."

[0944] The server first analyzes the image using an image recognition model, then analyzes the text using a natural language processing model, and combines the results to determine whether the ad material is offensive or not, or whether it is adult content or not.

[0945] An example of a prompt for the generative AI model is as follows:

[0946] Prompt for image recognition model: "Analyze the image for offensive or adult content."

[0947] Prompt the natural language processing model: "Analyze the text for offensive or adult content."

[0948] This allows for quick and accurate review of the quality of advertising materials, enabling quality control of advertising.

[0949] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0950] Step 1:

[0951] Users upload advertising materials to the system.

[0952] Input: Advertising materials (images and text)

[0953] Output: Advertising material sent to the server

[0954] Specific operation: The user selects the image file and text of the advertising material through the system interface and clicks the upload button, which sends the advertising material to the server.

[0955] Step 2:

[0956] The server receives images of the advertising material and inputs them into an image recognition model.

[0957] Input: Ad image sent to the server

[0958] Output: Input data to the image recognition model

[0959] Specific operation: The server preprocesses the received advertising image using OpenCV and converts it into a format suitable for the image recognition model, then inputs it into the image recognition model.

[0960] Step 3:

[0961] The server analyzes the advertisement image using an image recognition model.

[0962] Input: Input data to the image recognition model

[0963] Output: Image analysis results (offensive / not offensive, adult / not)

[0964] Specific operation: The server runs an analysis on the image recognition model using the prompt "Analyze the image for offensive or adult content." and obtains the results.

[0965] Step 4:

[0966] The server receives the text of the advertising material and inputs it into a natural language processing model.

[0967] Input: Ad text sent to the server

[0968] Output: Input data to the natural language processing model

[0969] Specific operation: The server preprocesses the received advertisement text and converts it into a format suitable for the natural language processing model, then inputs it into the natural language processing model.

[0970] Step 5:

[0971] The server analyzes the advertisement text using a natural language processing model.

[0972] Input: Input data to the natural language processing model

[0973] Output: Text analysis results (offensive / not offensive, adult / not)

[0974] Specific operation: The server runs analysis against the natural language processing model using the prompt sentence "Analyze the text for offensive or adult content." and obtains the results.

[0975] Step 6:

[0976] The server integrates the image analysis results and text analysis results and makes the final decision.

[0977] Input: Image analysis results, text analysis results

[0978] Output: Final judgment result (offensive / not offensive, adult / not)

[0979] Specific operation: The server integrates the image analysis results and text analysis results to evaluate the overall quality of the advertising material, and makes a final decision based on the integrated results.

[0980] Step 7:

[0981] The server outputs the final decision result to the user.

[0982] Input: Final decision result

[0983] Output: Notification of review results to the user

[0984] Specific operation: The server displays the final judgment result through an interface to notify the user. The user can check the review result and modify the advertising material as necessary.

[0985] Example 3

[0986] Next, a description will be given of Example 3 of Form Example 3. 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."

[0987] In conventional advertising material review systems, only the review results are provided, and the criteria for the review are not clearly stated, making it difficult for users to understand the reasons for the review results. In addition, because the review results include subjective judgments, the reliability and transparency of the results are lacking.

[0988] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[0989] In this invention, the server includes a generative AI model for reviewing advertising materials, a means for accepting submission of advertising materials, a means for outputting the review results by the generative AI model, and a means for clearly displaying the review results along with the judgment criteria, thereby enabling users to clearly understand the reasons for the review results and improving the reliability and transparency of the review results.

[0990] A "generative AI model" is an artificial intelligence algorithm that analyzes data entered by the user and generates review results and their criteria.

[0991] "Advertising materials" are digital content such as text, images, and videos used as advertisements.

[0992] The "means for accepting submissions" refers to an interface or protocol that allows users to submit advertising materials to the system.

[0993] The "means for outputting the review results" is a function for providing the review results generated by the generative AI model to the user.

[0994] "Means to clearly indicate the criteria for judgment" is a function that shows the user the factors that form the basis of the results when the generative AI model outputs the review results.

[0995] This invention is a system for reviewing advertising materials, and uses a generative AI model to clearly show the review results and the criteria for judgment. Specific embodiments of this system are described below.

[0996] The server analyzes the advertising material entered by the user using a generative AI model (e.g., a general natural language processing model). This generative AI model can make subjective judgments such as whether the advertising material is "offensive / not offensive" or "adult / not adult."

[0997] The user inputs advertising material into the input field of the device. For example, the user inputs the text "This movie is very violent and unbearable to watch." The device sends this input text data to the server. The server inputs the received text data into the generative AI model for analysis.

[0998] The generative AI model analyzes the input text data and generates a review result and its criteria. For example, for the text "This movie is extremely violent and unbearable to watch," it generates the review result "unpleasant" and the criterion "contains violent content."

[0999] The server sends the review results and criteria obtained from the generative AI model to the user's device. The device then displays the received review results and criteria to the user. For example, it might display "Review result: Offensive" or "Criteria: Contains violent content" on the screen.

[1000] As a concrete example, by inputting the following prompt sentence into the generative AI model, the review results and judgment criteria can be obtained:

[1001] Please review the following text and state your findings and criteria:

[1002] "This movie is extremely violent and unbearable to watch."

[1003] In this way, the user can clearly understand the screening results and the criteria for their judgment. This system improves the reliability and transparency of the screening results, making it easier for the user to understand the reasons for the screening results. The flow of the identification process in the third embodiment will be described with reference to FIG. 15.

[1004] Step 1:

[1005] The user inputs text data.

[1006] The user inputs the text data of the advertising material that he / she wishes to review into an input field on the terminal. For example, he / she inputs the text "This movie is extremely violent and unbearable to watch."

[1007] Input: Text data of advertising materials

[1008] Output: Text data entered into the terminal

[1009] Step 2:

[1010] The terminal transmits the input text data to the server.

[1011] The terminal packages the entered text data in JSON format and sends it to the server as an HTTP request.

[1012] Input: Text data entered into the terminal

[1013] Output: Text data sent to the server

[1014] Step 3:

[1015] The server receives the text data and inputs it into the generative AI model.

[1016] The server receives the text data sent from the device and inputs it into a generative AI model, which uses, for example, a natural language processing model.

[1017] Input: Text data sent to the server

[1018] Output: Text data input to the generative AI model

[1019] Step 4:

[1020] The generative AI model analyzes the text data and generates review results and judgment criteria.

[1021] The generative AI model analyzes the input text data and generates a review result and its criteria. For example, it generates a review result such as "offensive" and a criterion such as "contains violent content."

[1022] Input: Text data fed into the generative AI model

[1023] Output: Review results and criteria

[1024] Step 5:

[1025] The server sends the review results and criteria to the user's terminal.

[1026] The server packages the review results and judgment criteria obtained from the generative AI model in JSON format and sends them to the user's device as an HTTP response.

[1027] Input: Review results and criteria

[1028] Output: Review results and criteria sent to the user's device

[1029] Step 6:

[1030] The terminal displays the review results and criteria to the user.

[1031] The device analyzes the review results and criteria received from the server and displays them to the user. For example, it might display on the screen, "Review result: Offensive" or "Criteria: Contains violent content."

[1032] Input: Review results and criteria sent to the user's device

[1033] Output: The review results and criteria displayed to the user

[1034] (Application example 3)

[1035] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1036] In conventional advertising material review systems, the review results are output by the AI ​​generator, but the criteria for the review are not clearly stated, making it difficult for users to understand the reasons for the review results. Furthermore, content distribution services lacked a means for users to check the appropriateness of content before viewing it. This increased the risk of users encountering unpleasant content.

[1037] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1038] In this invention, the server includes a generating AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generating AI, a means for clearly indicating the judgment criteria along with the screening results, a means for generating a prompt sentence for screening content, a means for obtaining the screening results using the generating AI model, and a means for providing the screening results to the user, which makes it easier for the user to understand the reasons for the screening results and enables the user to confirm the appropriateness of the content before viewing it.

[1039] "Generative AI" is a system that uses artificial intelligence techniques to generate data and perform specific tasks.

[1040] "Advertising materials" are content such as text, images, and videos used as advertisements.

[1041] The "means for accepting submissions" refers to an interface or system for receiving advertising materials from users.

[1042] The "means for outputting the review results" is a system for displaying the results of the review conducted by the generation AI to the user.

[1043] The "means of clearly indicating the criteria for judgment" is a system that shows the user the basis for the judgment when the generation AI outputs the review results.

[1044] The "means for generating prompt sentences" is a system that creates instruction sentences to request review from the generation AI.

[1045] "Means for obtaining review results using a generative AI model" refers to a system that uses a generative AI model to review advertising materials and content and obtain the results.

[1046] "Means for providing the review results to the user" refers to a system for notifying the user of the review results conducted by the generation AI.

[1047] A system for implementing this invention includes a generating AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generating AI, a means for clearly indicating the judgment criteria along with the review results, a means for generating prompt sentences for reviewing content, a means for obtaining the review results using a generating AI model, and a means for providing the review results to a user.

[1048] The server provides an interface for receiving advertising materials from users. When a user submits advertising materials, the server uses a generation AI to review the materials. The generation AI makes subjective judgments such as whether the advertising materials are "offensive / not offensive" or "adult / not" and returns the results to the server.

[1049] The server then displays the results of the review along with the criteria for its decision. For example, if the content is deemed "offensive," the server will provide information such as "contains violent content" as the reason. This makes it easier for users to understand the reasons for the review.

[1050] Furthermore, the server generates a prompt to review the content. The generated prompt is an instruction to the generative AI model to request a review. The generative AI model reviews the advertising materials and content based on the prompt and returns the results to the server.

[1051] The server provides the user with the review results obtained using the generative AI model, allowing the user to confirm the appropriateness of the content before viewing it.

[1052] The hardware used includes an internet-connected server and the user's smartphone, and the software uses Python and the OpenAI API.

[1053] As a concrete example, consider a case where a part of a movie a user wants to watch "contains violent scenes." In this case, the generative AI receives the prompt "This movie contains violent scenes," judges the movie to be "unpleasant," and returns the reason as "It contains violent content."

[1054] Example prompt sentence:

[1055] Please review the following content:

[1056] This film contains violent scenes.

[1057] Please specify whether or not you find this content offensive and why.

[1058] In this way, the user can check the results of the content review and the reasons for it before viewing.

[1059] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1060] Step 1:

[1061] The user submits advertising materials.

[1062] Users upload advertising materials (text, images, videos, etc.) to the server. The server receives this advertising material and stores it in a database. The input is the advertising material provided by the user, and the output is the advertising material stored on the server.

[1063] Step 2:

[1064] The server reviews the advertising material using generated AI.

[1065] The server passes the stored ad material to the generation AI for review. The generation AI analyzes the ad material and makes a subjective judgment such as "offensive / not offensive" or "adult / not." The input is the ad material stored on the server, and the output is the review results by the generation AI.

[1066] Step 3:

[1067] The server obtains the review results and the criteria.

[1068] The generation AI returns the review results along with the criteria for its judgment to the server. For example, if it judges something to be "offensive," it will provide information such as "it contains violent content" as the reason. The input is the review results by the generation AI, and the output is the review results and the criteria for that judgment.

[1069] Step 4:

[1070] The server generates a prompt.

[1071] The server generates a prompt to request the AI ​​to review the ad. The prompt includes information about the ad material to be reviewed. The input is the information about the ad material to be reviewed, and the output is the generated prompt.

[1072] Step 5:

[1073] The server uses the generative AI model to obtain the review results.

[1074] The server passes the generated prompt to the generative AI model for review. The generative AI model reviews the advertising material based on the prompt and returns the results to the server. The input is the generated prompt, and the output is the review result by the generative AI model.

[1075] Step 6:

[1076] The server provides the review results to the user.

[1077] The server notifies the user of the review results obtained using the generative AI model. The user can check the review results and their judgment criteria. The input is the review results obtained by the generative AI model, and the output is the review results and their judgment criteria provided to the user.

[1078] 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.

[1079] "Example 1"

[1080] In one embodiment of the present invention, a system including a generation AI that reviews advertising materials, a means for accepting submissions of advertising materials, and a means for outputting the review results of the generation AI is combined with an emotion engine that recognizes user emotions. Specifically, the emotion engine analyzes the user's reaction (e.g., facial expression, tone of voice, etc.) when viewing the advertising materials and feeds the results back to the generation AI. This allows the generation AI to review the advertising materials based on the user's emotions.

[1081] "Example 2"

[1082] In another embodiment of the present invention, the emotion engine further includes a means for analyzing a user's reaction and feeding the result back to the generation AI's review process. Specifically, the emotion engine analyzes the user's reaction when viewing the advertising material and feeds the result back to the generation AI. This allows the generation AI to review the advertising material based on the user's emotion.

[1083] "Example 3"

[1084] In yet another embodiment of the present invention, the emotion engine further includes a means for detecting changes in the user's emotions and adjusting the screening criteria of the generation AI in response to those changes. Specifically, the emotion engine analyzes the user's reaction when viewing advertising material and feeds back the results to the generation AI. This allows the generation AI to adjust the screening criteria of the advertising material in response to changes in the user's emotions.

[1085] The processing flow of each embodiment will be described below.

[1086] "Example 1"

[1087] Step 1: User views advertising material.

[1088] Step 2: The emotion engine analyzes the user's reactions (e.g., facial expressions, tone of voice, etc.).

[1089] Step 3: The analysis results are fed back to the generation AI.

[1090] Step 4: Generative AI reviews advertising materials based on user sentiment.

[1091] "Example 2"

[1092] Step 1: User views advertising material.

[1093] Step 2: The emotion engine analyzes the user's reaction.

[1094] Step 3: The analysis results are fed back to the generation AI.

[1095] Step 4: Generative AI reviews advertising materials based on user sentiment.

[1096] "Example 3"

[1097] Step 1: User views advertising material.

[1098] Step 2: The emotion engine analyzes the user's reaction.

[1099] Step 3: The analysis results are fed back to the generation AI.

[1100] Step 4: The generative AI adjusts the review criteria for advertising materials according to changes in user emotions.

[1101] Example 1

[1102] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1103] In conventional advertising material review systems, the evaluation of the appropriateness and quality of advertising materials involves subjective judgment, which can lead to inconsistent review results. Furthermore, because the review is conducted without taking into account the user's feelings, it is difficult to accurately evaluate the actual impact of advertising materials on users. Furthermore, because the criteria for the review results are not clearly stated, users find it difficult to accept the review results.

[1104] The identification process by the identification processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means. In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, a sentiment analysis engine means for recognizing the user's emotions, and a means for feeding back the analysis results of the sentiment analysis engine to the generation AI. This allows for consistent screening of advertising materials, enables evaluation taking into account the user's emotions, and makes it easier for users to accept the screening results because the criteria for the screening results are clearly stated.

[1105] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[1106] "Advertising materials" are digital content such as images, text, and videos used as advertisements.

[1107] "Means for accepting submissions" refers to an interface such as a web form or API that allows users to upload advertising materials to the system.

[1108] "Means for outputting the review results" refers to methods such as a web interface or email notification for notifying the user of the review results generated by the generation AI.

[1109] An "emotion analysis engine" is software or hardware that analyzes a user's facial expressions, tone of voice, etc. to recognize the user's emotions.

[1110] The "means of feedback" refers to a method for sending the analysis results of the sentiment analysis engine to the generation AI and reflecting them in the review process.

[1111] The present invention is a system that uses a generative AI to review advertising materials. A specific embodiment of this system will be described below.

[1112] System Configuration

[1113] This system consists of the following main components:

[1114] 1. A means of accepting ad material submissions: A web form or API that allows users to upload ad material to the system.

[1115] 2. Generative AI: Artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[1116] 3. Means of outputting the review results: A web interface or email notification to notify the user of the review results generated by the generative AI.

[1117] 4. Emotion Analysis Engine: Software or hardware that analyzes a user's facial expressions and tone of voice to recognize their emotions.

[1118] 5. Feedback method: The results of the sentiment analysis engine's analysis are sent to the generation AI and reflected in the review process.

[1119] Submitting advertising materials

[1120] Users submit advertising materials to the system using a web-based submission form or API. Specifically, when uploading advertising images through a web browser, users enter the necessary information into the submission form, attach the image file, and submit it.

[1121] Receipt and analysis of advertising materials

[1122] The server receives the advertising material sent by the user and temporarily stores it in a database. The server then passes the received advertising material to the generation AI, which analyzes the content of the advertising material and evaluates its appropriateness and quality. For example, the generation AI checks whether the advertising text contains inappropriate language and whether the images comply with regulations.

[1123] User sentiment analysis

[1124] When a user views advertising materials, the emotion analysis engine analyzes the user's reactions. Specifically, the user's facial expressions and tone of voice are collected through a camera and microphone, and the emotion analysis engine analyzes the data. For example, it can recognize emotions such as whether the user smiled or was surprised when they saw the advertisement.

[1125] Feedback and review updates

[1126] The server feeds back the results of the emotion analysis engine to the generation AI, which then uses this feedback to make a final review of the ad material. For example, if the user response is positive, the ad material will receive a high rating.

[1127] Notification of review results

[1128] The server notifies the user of the final review results by the generation AI. Notification methods include a web interface and email notification. For example, the user can check the review results through the web interface.

[1129] Examples and prompts

[1130] As a concrete example, consider the case where a user submits an advertising image for a new cosmetic product. The user accesses a web form, enters the required information, and uploads the image file. The server receives this image and passes it to the generation AI. The generation AI analyzes the content of the image and evaluates its appropriateness. At the same time, an emotion analysis engine analyzes the user's facial expressions and tone of voice as they view the advertising image, and feeds the results back to the generation AI. Finally, the generation AI notifies the user of the review results via a web interface.

[1131] Example prompt sentence:

[1132] "Please review the advertisement images for a new cosmetic product. Analyze the user's facial expressions and tone of voice, and give a thumbs up if the user responds positively."

[1133] In this way, this system combines generative AI and an emotion analysis engine to efficiently and accurately review advertising materials.

[1134] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1135] Step 1:

[1136] Users submit advertising materials to the system using a web-based submission form or API. Specifically, when uploading advertising images through a web browser, users enter the necessary information into the submission form, attach the image file, and submit it.

[1137] Input: Ad material (image file), ad title, description, category, and other information

[1138] Output: Advertising material data sent to the server

[1139] Specific behavior:

[1140] A user accesses a web form.

[1141] The user inputs information such as the title, description, and category of the advertising material.

[1142] The user selects an advertisement image file and clicks the upload button.

[1143] Step 2:

[1144] The server receives the advertising material sent by the user and temporarily stores it in a database.

[1145] Input: Advertising material data sent by the user

[1146] Output: Advertising material data stored in the database

[1147] Specific behavior:

[1148] The server receives an HTTP request.

[1149] The server stores the advertising material data in a database.

[1150] Step 3:

[1151] The server passes the received advertising material to the generation AI, which analyzes the content of the advertising material and evaluates its appropriateness and quality.

[1152] Input: Advertising material data stored in the database

[1153] Output: Review results by the generated AI

[1154] Specific behavior:

[1155] The server inputs advertising material data into the generated AI model.

[1156] Generative AI analyzes the text and images of the ad.

[1157] The generation AI generates the review results and returns them to the server.

[1158] Step 4:

[1159] When a user views advertising materials, the emotion analysis engine analyzes the user's reactions. Specifically, the user's facial expressions and tone of voice are collected through a camera and microphone, and the emotion analysis engine analyzes the data.

[1160] Input: User's facial expression data, tone of voice data

[1161] Output: Sentiment analysis results from the sentiment analysis engine

[1162] Specific behavior:

[1163] A user accesses a web page for viewing advertising material.

[1164] The user's device activates the camera and microphone.

[1165] The emotion analysis engine analyzes the user's facial expressions and tone of voice in real time.

[1166] Step 5:

[1167] The server feeds back the results of the emotion analysis engine to the generation AI, which then performs a final review of the advertising materials based on this feedback.

[1168] Input: Emotion analysis results from the emotion analysis engine

[1169] Output: Final review results from the updated generative AI

[1170] Specific behavior:

[1171] The server receives the analysis results from the emotion analysis engine.

[1172] The server sends the analysis results to the generation AI.

[1173] The generative AI updates the review results based on the feedback.

[1174] Step 6:

[1175] The server notifies the user of the final review results by the generation AI via a web interface or email.

[1176] Input: Final review results from the updated generation AI

[1177] Output: Review results notified to the user

[1178] Specific behavior:

[1179] The server stores the review results in a database.

[1180] The server sends an email notification to the user.

[1181] The user checks the results of the review through a web interface.

[1182] (Application example 1)

[1183] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1184] Conventional advertising material review systems were unable to take into account subjective factors or users' emotional reactions when determining whether the content of advertising material was appropriate. This made it difficult to accurately evaluate the impact of advertising material on viewers, resulting in the risk of inappropriate advertising being delivered. Furthermore, because the criteria for review results were not clearly stated, advertisers were sometimes dissatisfied with the review results.

[1185] 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.

[1186] In this invention, the server includes a generation AI means for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine means for recognizing user emotions, and a means for analyzing the user's emotional responses and providing feedback to the generation AI. This makes it possible to take the user's emotional responses into consideration when reviewing advertising materials, thereby providing more accurate review results. Furthermore, by clearly indicating the criteria for the review results, advertisers can more easily accept the review results.

[1187] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[1188] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[1189] "Submission methods" are interfaces such as web-based forms or APIs for uploading advertising materials into the system.

[1190] "Means for outputting the review results" refers to methods such as a web interface or email notification for notifying the user of the review results generated by the generation AI.

[1191] An "emotion engine" is software or hardware that analyzes a user's facial expressions, tone of voice, etc. to recognize emotions.

[1192] The "means for analyzing emotional responses and providing feedback to the generation AI" is a system that collects emotional data when users view advertising materials and provides it to the generation AI.

[1193] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for recognizing user emotions, and a means for analyzing the user's emotional reactions and providing feedback to the generation AI. Specific embodiments of the system are described below.

[1194] The server runs a generative AI model to review ad materials. The generative AI model analyzes the content of the ad materials and evaluates their appropriateness and quality. Ad materials are submitted via a web-based form or API. When a user uploads an ad material, the server passes the material to the generative AI model for review.

[1195] The results of the review are output to the user via a web interface or email notification. The review results include subjective judgments such as whether the advertising material is "offensive / not offensive" or "adult / not." The review criteria are also clearly stated, making it easy for users to understand the results.

[1196] Furthermore, it is equipped with an emotion engine that recognizes the user's emotions. The emotion engine analyzes the user's facial expressions and tone of voice when viewing advertising materials and feeds this emotional data back to the generation AI. This allows the generation AI to conduct screening that takes the user's emotional reactions into account.

[1197] For example, when a user uploads a new advertising video, the generative AI reviews the video, analyzes the user's facial expressions while watching the video, and provides feedback along with the review results. For example, the following prompt sentence can be input into the generative AI model:

[1198] "To review advertising materials, please analyze the following advertising video. Please output the review results, taking into account the user's emotional response.

[1199] Advertisement video: [Advertisement video URL]

[1200] User's emotional response: [user's facial expression data]

[1201] In this way, it becomes possible to take into account the emotional response of the user when screening advertising materials, thereby providing more accurate screening results.

[1202] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1203] Step 1:

[1204] The user uploads advertising materials.

[1205] Input: Advertising materials (images, videos, text, etc.)

[1206] How it works: Users upload advertising materials to the server using a web-based form or API.

[1207] Output: Ad material file saved on the server

[1208] Step 2:

[1209] The server passes the advertising material to a generative AI model for review.

[1210] Input: Ad material file stored on the server

[1211] How it works: The server feeds the uploaded ad material into a generative AI model to evaluate its appropriateness and quality.

[1212] Output: Review results from the generative AI model

[1213] Step 3:

[1214] The server notifies the user of the review results.

[1215] Input: Review results from the generative AI model

[1216] How it works: The server sends the results of the review to the user via a web interface or email notification.

[1217] Output: Review results notified to the user

[1218] Step 4:

[1219] When a user views advertising material, the emotion engine analyzes their emotional response.

[1220] Input: advertising materials, user facial expressions and tone of voice

[1221] How it works: When a user views advertising materials, the emotion engine uses the camera and microphone to capture the user's facial expressions and tone of voice and analyzes the emotional data.

[1222] Output: Parsed emotion data

[1223] Step 5:

[1224] The server feeds the emotion data back to the generative AI model.

[1225] Input: Parsed emotion data

[1226] Operation: The server inputs the emotion data obtained from the emotion engine into the generative AI model and reflects it in the review results.

[1227] Output: Final review results taking into account emotional data

[1228] Step 6:

[1229] The server notifies the user of the final review result.

[1230] Input: Final review results taking into account emotional data

[1231] How it works: The server sends the final results to the user via a web interface or email notification.

[1232] Output: Final review results notified to the user

[1233] Example 2

[1234] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1235] Conventional advertising material review systems have difficulty accurately evaluating the subjective impact of advertising materials on users. Furthermore, there was a lack of a means to collect user reactions in real time and reflect them in the review process, making it impossible to accurately determine the appropriateness of advertising materials. This created a risk of displaying advertising materials that were unpleasant to users.

[1236] 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.

[1237] In this invention, the server includes means including a generation AI for reviewing advertising materials, means for accepting submission of advertising materials, means for outputting the review results by the generation AI, means including an emotion engine for collecting user reactions, and means for feeding back the analysis results of the emotion engine to the generation AI. This makes it possible to evaluate with high accuracy the subjective impact of advertising materials on users, collect user reactions in real time, and reflect them in the review process.

[1238] "Generative AI" is an artificial intelligence model that analyzes the content of advertising materials and makes subjective judgments.

[1239] "Advertising materials" are digital content such as images and text used as advertisements.

[1240] The "means for accepting submissions" is an interface that allows users to upload advertising materials to the system.

[1241] The "means of outputting the review results" is a function for notifying the user of the results of the analysis performed by the generation AI.

[1242] The "emotion engine" is a system for collecting and analyzing user reactions.

[1243] The "means of feedback" is a function that sends the analysis results of the emotion engine to the generation AI to improve the accuracy of the model.

[1244] "User response" refers to behavioral data such as the user's facial expression, gaze, and clicking behavior when viewing advertising material.

[1245] MODE FOR CARRYING OUT THE INVENTION

[1246] The present invention relates to a system for screening advertising materials. A specific embodiment of this system will be described below.

[1247] System configuration

[1248] This system consists of three main components: a server, a terminal, and a user. The server analyzes and reviews advertising materials using a generative AI model. The terminal provides an interface for users to upload advertising materials and check the review results. The terminal also has an emotion engine for collecting user responses. The user is responsible for uploading advertising materials and receiving the review results.

[1249] Hardware and software used

[1250] Server: Use a server equipped with a high-performance processor and large memory capacity. Deep learning frameworks such as TensorFlow and PyTorch are used to run generative AI models.

[1251] Device: A personal computer or smartphone equipped with a camera or sensor is used. Image processing libraries such as OpenCV and Dlib are used to run the emotion engine.

[1252] User: Uses a device with internet connectivity to upload advertising materials.

[1253] Data processing and calculation

[1254] The server receives the advertising materials uploaded by the user and analyzes them using the generative AI model. Specifically, it processes and calculates the data as follows:

[1255] Image recognition: The server uses TensorFlow to analyze images of advertising materials and identify elements within the images, thereby determining whether the advertising materials are offensive or not, or whether they are adult content or not.

[1256] Natural language processing: The server analyzes the text contained in the advertising material using BERT and GPT-3 to determine whether the content is appropriate.

[1257] The terminal collects the user's reactions when viewing advertising materials. Specifically, it processes and calculates the data as follows:

[1258] User reaction collection: The device camera captures the user's facial expressions in real time, which are analyzed by the emotion engine. Mouse movements and clicks are also recorded.

[1259] Specific examples

[1260] For example, suppose a user uploads a new advertising banner. The server uses image recognition technology to analyze the elements in the banner, while simultaneously analyzing the text content using a natural language processing model. When the user views the banner, the device's camera captures the user's facial expressions, which are analyzed by the emotion engine. The analysis results are sent to the server and fed back to the generative AI model. Finally, the server determines whether the advertising banner is appropriate and notifies the user of the result.

[1261] Prompt Sentence Examples

[1262] Below is an example of a prompt sentence to input to the generative AI model.

[1263] "Please decide whether this advertising material is offensive."

[1264] "Please analyze this advertising material to see if it contains adult content."

[1265] In this way, a system is constructed in which the server, terminal, and user work together to determine the appropriateness of advertising materials.

[1266] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1267] Step 1: Upload your creative

[1268] Input: The user uploads advertising materials (images and text) using the terminal.

[1269] Specific operation: The user opens a browser on the device, accesses the web page provided by the server, selects the advertising material file, and clicks the upload button.

[1270] Output: The server receives the uploaded advertising material.

[1271] Step 2: Analyze the advertising material

[1272] Input: The server receives the uploaded advertising material.

[1273] Specific operation: The server uses image recognition technology to analyze the image of the advertising material. It uses TensorFlow to identify elements within the image and determine whether the advertising material is "offensive / not offensive" or "adult / not." At the same time, it analyzes the text data using BERT and GPT-3 to determine whether the content is appropriate.

[1274] Output: The server obtains the analysis results from the generative AI model.

[1275] Step 3: Collect user responses

[1276] Input: User views advertising material.

[1277] How it works: While a user is viewing advertising materials, the device camera captures the user's facial expressions in real time, as well as mouse movements and clicks. This data is then sent to the emotion engine.

[1278] Output: The device collects the user's reaction data.

[1279] Step 4: Analysis by the Emotion Engine

[1280] Input: User response data sent from the device.

[1281] Specific operation: The server inputs the user's facial expression data into the emotion engine and determines whether the user felt "uncomfortable." Mouse movements and clicks are also analyzed.

[1282] Output: The server receives the analysis results of the emotion engine.

[1283] Step 5: Processing feedback

[1284] Input: Sentiment engine analysis results.

[1285] Specific operation: The server feeds back the analysis results of the emotion engine to the generative AI model and adjusts the model parameters, thereby improving the accuracy of the generative AI model.

[1286] Output: The server gets an updated generative AI model.

[1287] Step 6: Final decision and output of results

[1288] Input: Analysis results from the updated generative AI model.

[1289] Specific operation: If the server determines that the advertising material is "offensive" or "adult," it will notify the user of the result by email or notification.

[1290] Output: The user can check the results of the review on the terminal.

[1291] (Application example 2)

[1292] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1293] Conventional advertising material review systems have had the problem of making it difficult to accurately determine whether an advertisement is offensive to users or contains adult content. Furthermore, because the review does not take into account user emotions and reactions, there is also the problem of not being able to improve the user experience.

[1294] 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.

[1295] In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, an emotion engine means for analyzing user reactions, and a means for feeding back the analysis results of the emotion engine to the generation AI. This makes it possible to accurately determine whether advertising materials are offensive to users or contain adult content, and to perform screening taking into account the user's emotions and reactions.

[1296] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and makes subjective judgments.

[1297] "Advertising materials" are digital content such as images, text, and videos used as advertisements.

[1298] "Means for accepting submissions" refers to the interface or process for uploading advertising materials to the system.

[1299] "Means for outputting review results" refers to the interface or process for displaying the results of the review conducted by the generation AI to the user.

[1300] An "emotion engine" is software or hardware that analyzes a user's reactions and determines their emotions.

[1301] "Feedback means" refers to the interface or process for transmitting the analysis results of the emotion engine to the generative AI and reflecting them in the review process.

[1302] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for analyzing user reactions, and a means for feeding back the analysis results of the emotion engine to the generation AI.

[1303] The server receives the image and text of the ad material as input and analyzes its content using generative AI. Specifically, it uses image recognition and natural language processing models to determine whether the ad material is offensive or adult. OpenCV is used for image recognition, and Hugging Face Transformers is used for natural language processing.

[1304] The emotion engine analyzes the reaction data (video and audio) when a user views advertising materials. The emotion engine uses the Emotion Engine to determine the user's emotions. The analysis results are fed back to the generation AI, which then reviews the ad based on the user's emotions.

[1305] The device displays the results of the AI's review to the user, including the criteria for determining whether the ad material is offensive or adult content.

[1306] For example, if an ad image contains a violent scene, the image recognition model will judge it as "offensive" and the ad will be deemed unpleasant. If the ad text contains discriminatory language, the natural language processing model will judge it as "negative" and the ad will be deemed unpleasant. Furthermore, if a user shows a negative reaction, such as grimacing, when viewing the ad, the emotion engine will judge it as "negative" and the ad will be deemed unpleasant.

[1307] An example of a prompt to input to a generative AI model is as follows:

[1308] "Analyze ad images and text to detect offensive or adult content. Analyze user response data and determine the ad as offensive if there is a negative response."

[1309] In this way, it is possible to accurately determine whether advertising material is offensive to users or contains adult content, and furthermore, to conduct screening that takes into account the user's feelings and reactions.

[1310] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1311] Step 1:

[1312] The server receives the image and text of the advertising material as input. When the user uploads the advertising material, the server acquires the data. The input data consists of image files and text data.

[1313] Step 2:

[1314] The server analyzes advertising images using an image recognition model. Specifically, it preprocesses the images using OpenCV and classifies the image content using Hugging Face Transformers. Data processing includes image resizing and noise removal. The output is a classification of the image with labels such as "offensive" or "adult."

[1315] Step 3:

[1316] The server analyzes the ad text using a natural language processing model. Specifically, it uses Hugging Face's Transformers to perform sentiment analysis of the text. Data processing involves tokenizing the text and calculating a sentiment score. The output is a classification of the text with labels such as "negative" or "neutral."

[1317] Step 4:

[1318] The server combines the results of steps 2 and 3 to determine whether the advertising material is offensive or adult. Specifically, it combines image and text labels to make a comprehensive assessment. The output is a judgment result such as whether the advertising material is "offensive" or "not offensive," or "adult" or "not adult."

[1319] Step 5:

[1320] The device collects reaction data (video and audio) when the user views advertising materials and sends it to the server. The input data is real-time data of the user's facial expressions and voice.

[1321] Step 6:

[1322] The server analyzes the user's reaction data using an emotion engine. Specifically, it uses the Emotion Engine to determine the user's emotions. Data processing involves extracting video frames and analyzing audio spectra. The output is a classification of the user's emotions with labels such as "positive" or "negative."

[1323] Step 7:

[1324] The server feeds back the emotion engine's analysis results to the generation AI and updates the final review result of the advertising material. Specifically, if the user's emotion is negative, the advertising material is judged to be unpleasant. The output is the updated review result.

[1325] Step 8:

[1326] The terminal displays the final review results received from the server to the user. Specifically, the criteria for determining whether the advertising material is offensive or adult are also clearly indicated. The output is the review results displayed to the user.

[1327] Example 3

[1328] Next, a description will be given of Example 3 of Form Example 3. 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."

[1329] In conventional advertising material review systems, it was difficult to provide a satisfactory explanation to users because the criteria used to determine the review results were unclear. Additionally, the review criteria could not be flexibly adjusted in response to changes in user emotions, making it difficult to respond to diverse user reactions.

[1330] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1331] In this invention, the server includes a generative AI model for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results of the generative AI model, a means for clearly displaying the screening results along with the judgment criteria, an emotion engine for detecting changes in the user's emotions, and a means for adjusting the screening criteria of the generative AI model based on feedback from the emotion engine. This increases the transparency of the screening results and enables flexible adjustment of the screening criteria according to the user's emotions.

[1332] "Generative AI Model" means an artificial intelligence model used to screen advertising materials.

[1333] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[1334] "Means for accepting submissions" refers to the interface or protocol for uploading or transmitting advertising materials to the system.

[1335] The "means for outputting the review results" is a function for displaying or notifying the user or administrator of the results of the generative AI model's review of the advertising material.

[1336] "Means to clearly indicate the criteria for judgment" is a function that clearly indicates the factors and reasons that form the basis for the judgment when the generative AI model outputs the review results.

[1337] An "emotion engine" is software or hardware for detecting and analyzing changes in a user's emotions.

[1338] "Means for adjusting the evaluation criteria of the generative AI model based on feedback" is a function for dynamically changing the evaluation criteria of the generative AI model based on the user's emotional data obtained from the emotion engine.

[1339] The present invention is a system for reviewing advertising materials, which includes a generative AI model, an emotion engine, and various means for linking them. Specific embodiments of this system are described below.

[1340] Hardware and software used

[1341] Hardware: Servers, user devices (PCs, smartphones, etc.)

[1342] Software: Generative AI model, emotion engine

[1343] Server Operation

[1344] The server runs a generative AI model to review advertising materials. When advertising materials are submitted by advertising providers or users, the server receives the materials and inputs them into the generative AI model. The generative AI model makes subjective judgments about the advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult," and outputs the results.

[1345] In addition, the server outputs information that clearly explains the criteria used to determine the ad's content, along with the results of the review. For example, if the ad is deemed "offensive," the server will provide information such as "contains violent content" as the reason.

[1346] Device behavior

[1347] The device displays the advertising material received from the server to the user. When the user views the advertising material, the device uses an emotion engine to analyze the user's reaction. The emotion engine uses a camera and microphone to analyze the user's facial expressions and tone of voice and generate emotion data.

[1348] User Actions

[1349] The user views the advertising material displayed on their device. When a change in the user's emotion is detected, the information is sent from the device to the server. The server adjusts the evaluation criteria of the generative AI model based on the received emotion data.

[1350] Specific examples

[1351] Example 1: Advertising material review

[1352] The server receives a "new game ad video" from an ad provider. The generative AI model reviews this ad video and determines it to be "offensive." It outputs the information that the reason for this is that it "contains violent content."

[1353] Example prompt: "Please review this advertising material and print out your results and criteria."

[1354] Example 2: Feedback from an Emotion Engine

[1355] The device displays to the user an "advertising video for a new game" received from the server. While the user is watching the advertising video, the device analyzes facial expressions with a camera and detects "discomfort." The user's "discomfort" data is sent from the device to the server. The server adjusts the evaluation criteria of the generative AI model based on the received "discomfort" data and reflects it in future evaluations.

[1356] Example prompt: "Detect changes in the user's emotions and provide that information as feedback to the generative AI."

[1357] In this way, the system can provide flexible screening criteria that correspond to the user's feelings. The flow of the identification process in the third embodiment will be described with reference to FIG.

[1358] Step 1:

[1359] The server receives advertising material from advertising providers or users.

[1360] Specifically, the server receives an HTTP request and stores the advertising material data.

[1361] Input: Advertising materials (images, videos, text, etc.)

[1362] Output: Saved advertising material data

[1363] Step 2:

[1364] The server inputs the received advertising material into a generative AI model and reviews it.

[1365] Specifically, the server passes the advertising material data to the generative AI model and obtains the review results.

[1366] Input: Advertising material data

[1367] Output: Review result (e.g. "offensive", "adult", etc.)

[1368] Step 3:

[1369] The server outputs the review results and judgment criteria obtained from the generative AI model.

[1370] Specifically, the server generates the review results and criteria in JSON format and returns them to the user.

[1371] Input: Review results

[1372] Output: JSON data of the review results and criteria

[1373] Step 4:

[1374] The terminal displays the advertising material received from the server to the user.

[1375] Specifically, the terminal displays advertising material on the screen and waits for the user's response.

[1376] Input: Advertising material data

[1377] Output: Displayed advertising material

[1378] Step 5:

[1379] The terminal uses an emotion engine to analyze the user's reaction.

[1380] Specifically, the device uses a camera and microphone to analyze the user's facial expressions and tone of voice and generate emotional data.

[1381] Input: User facial and voice data

[1382] Output: Emotion data

[1383] Step 6:

[1384] The user's emotion data is transmitted from the terminal to the server.

[1385] Specifically, the terminal transmits emotion data generated by the emotion engine to the server.

[1386] Input: Emotion data

[1387] Output: Emotion data sent to the server

[1388] Step 7:

[1389] The server adjusts the evaluation criteria for the generative AI model based on the received emotional data.

[1390] Specifically, the server analyzes the emotion data and updates the parameters of the generative AI model.

[1391] Input: Emotion data

[1392] Output: Adjusted generative AI model parameters

[1393] (Application example 3)

[1394] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1395] In conventional advertising material review systems, when the AI ​​generator outputs the review results, the criteria for its judgment are not clearly stated, resulting in a lack of transparency for users. Furthermore, because the review is conducted based on uniform criteria without taking into account the user's feelings, there is a problem in that flexible review based on the user's feelings cannot be performed. This increases the possibility that the advertising material will cause discomfort to users, potentially reducing the effectiveness of the advertisement.

[1396] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1397] In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, an emotion engine means for analyzing user emotions, a means for feeding back the analysis results of the emotion engine to the generation AI, and a means for the generation AI to adjust the screening criteria based on the user emotions. This enables flexible screening according to user emotions, improves the transparency of the screening results, and makes it possible to reduce the discomfort that advertising materials cause to users.

[1398] "Generative AI" refers to artificial intelligence used to review advertising materials and determine whether they are suitable based on specific criteria.

[1399] "Advertising materials" refers to all content used as advertisements, including images, videos, and text.

[1400] "Means for accepting submissions" refers to the interface or process for uploading advertising materials to the system.

[1401] "Means for outputting the review results" refers to the interface or process for displaying the results of the review conducted by the generation AI to the user.

[1402] An "emotion engine" is software or hardware for analyzing a user's emotions, and detects emotions from the user's facial expressions, voice, etc.

[1403] "Means of feedback" refers to the interface or process for transmitting the analysis results of the emotion engine to the generative AI.

[1404] "Means for adjusting the review criteria" refers to an interface or process that allows the generative AI to dynamically change the review criteria based on user sentiment.

[1405] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for analyzing user emotions, a means for feeding back the analysis results of the emotion engine to the generation AI, and a means for the generation AI to adjust the review criteria based on the user's emotions.

[1406] Hardware and software used

[1407] Hardware: Smartphone camera, smartphone display

[1408] Software: OpenCV (image processing library), Keras (sentiment analysis model framework), Transformers (generative AI model framework)

[1409] Data processing and calculation

[1410] Display of advertising materials

[1411] The server uses OpenCV to display the advertising material on the smartphone display, allowing users to visually check the advertising material.

[1412] Emotion analysis

[1413] The device (smartphone) captures the user's face using a camera and analyzes emotions using a Keras model. Specifically, the captured image is converted to grayscale, face detection is performed, and then the image is input into the emotion analysis model. The model estimates emotions from the user's facial expressions and outputs the results.

[1414] Emotional Feedback

[1415] The server feeds back the analysis results of the emotion engine to the generation AI. Specifically, the analyzed emotional information is input to the generation AI as a prompt sentence, and the generation AI adjusts the screening criteria for advertising materials based on that information.

[1416] Adjustment of review criteria

[1417] The generative AI dynamically changes the judging criteria based on the emotional information fed back, enabling flexible judging based on the user's emotions.

[1418] Specific examples

[1419] For example, if a promotional video for a new smartphone is displayed as advertising material and the user responds with "Happy," that emotion is fed back to the generating AI, which then adjusts the review criteria for the next advertising material.

[1420] Prompt Sentence Examples

[1421] The user's emotion is Happy. Based on this emotion, please evaluate the following advertising material: Promotional video for a new smartphone

[1422] In this way, the screening criteria for advertising materials can be dynamically adjusted based on user sentiment, which improves the transparency of screening results and reduces the discomfort that advertising materials cause to users.

[1423] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1424] Step 1:

[1425] The server displays the advertising material on the smartphone display. Specifically, it uses OpenCV to read the image or video file of the advertising material and displays it on the smartphone display. The input is the file path of the advertising material, and the output is the advertising material displayed on the display.

[1426] Step 2:

[1427] The device (smartphone) captures the user's face using a camera. Specifically, the camera is activated and captures the user's face in real time. The input is the video stream from the camera, and the output is the captured image of the user's face.

[1428] Step 3:

[1429] The device converts the captured face image to grayscale and performs face detection. Specifically, it applies a face detection algorithm using OpenCV to identify the face position. The input is the captured face image, and the output is face position information.

[1430] Step 4:

[1431] The device extracts the detected face area and inputs it into an emotion analysis model. Specifically, the face image is input into an emotion analysis model pre-trained using Keras, and emotions are estimated. The input is the detected face image, and the output is the estimated emotion label.

[1432] Step 5:

[1433] The server feeds back the analysis results of the emotion engine to the generation AI. Specifically, it inputs the estimated emotion label to the generation AI as a prompt sentence. The input is the estimated emotion label, and the output is the prompt sentence input to the generation AI.

[1434] Step 6:

[1435] The generative AI adjusts the ad material's review criteria based on the feedback emotional information. Specifically, it analyzes the prompt text and dynamically changes the review criteria. The input is the prompt text, and the output is the adjusted review criteria.

[1436] Step 7:

[1437] The server outputs the results of the generative AI's review. Specifically, it judges the suitability of the advertising material based on the adjusted review criteria and displays the results to the user. The input is the adjusted review criteria and the advertising material, and the output is the review results.

[1438] 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.

[1439] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by 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.

[1440] Another example of generative AI is Gemini (internet search engine). <url: https: gemini.google.com ?hl="ja">) are listed.

[1441] 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.

[1442] [Third embodiment]

[1443] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1444] 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.

[1445] 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).

[1446] 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.

[1447] 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.

[1448] 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).

[1449] 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.

[1450] 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.

[1451] 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.

[1452] 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.

[1453] 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.

[1454] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1455] "Example 1"

[1456] The present invention is a system that incorporates a generation AI for reviewing advertising materials. Specifically, a web-based submission form or API is used as a means for accepting submissions of advertising materials. The submitted advertising materials are reviewed by the generation AI, and the review results are communicated to the user via an output means, such as a web interface or email notification.

[1457] "Example 2"

[1458] The generative AI makes subjective judgments such as whether the ad material is "offensive / not offensive" or "adult / not adult" by analyzing the content of the ad material using technologies such as image recognition and natural language processing, and makes a judgment based on the results.

[1459] "Example 3"

[1460] When the generative AI outputs the review results, it will include a means to clearly indicate the criteria used. Specifically, along with the review results, it will output information indicating what factors were used as the basis for the judgment. For example, if a content is judged to be "offensive," it will output information such as "contains violent language" as the reason.

[1461] The processing flow of each embodiment will be described below.

[1462] "Example 1"

[1463] Step 1: Advertisers submit their ad materials to the system via a web-based submission form or API.

[1464] Step 2: The submitted ad material is reviewed by the AI ​​generator, which analyzes the content of the ad material using techniques such as image recognition and natural language processing.

[1465] Step 3: The generating AI outputs the review results and communicates them to the advertiser via a web interface or email notification.

[1466] "Example 2"

[1467] Step 1: When advertising materials are submitted to the system, the generation AI makes a subjective judgment on the advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult."

[1468] Step 2: The generative AI analyzes the content of the advertising material using technologies such as image recognition and natural language processing, and makes a decision based on the results.

[1469] Step 3: The generating AI outputs the review results and their criteria, and notifies the advertiser of the review results. "Example 3"

[1470] Step 1: Once the ad material is submitted to the system, the generation AI reviews it.

[1471] Step 2: The generative AI outputs the results of the review and the criteria for its decision. Specifically, it outputs information indicating the factors that formed the basis for the decision along with the review results.

[1472] Step 3: The AI's review results and criteria are communicated to the advertiser. For example, if the content is deemed "offensive," the AI ​​outputs information such as "contains violent content" as the reason.

[1473] Example 1

[1474] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1475] The traditional review process for advertising materials is often done manually, which is time-consuming and labor-intensive, and prone to subjective judgment. Furthermore, there are also delays in notifying the review results, which creates the risk of missing the timing of advertising campaigns. To solve these problems, a system that provides an efficient and objective review process is needed.

[1476] 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.

[1477] In this invention, the server includes means for converting the content of the advertising material into text format, generating and inputting a prompt sentence to the generative AI model, means for acquiring the review results from the generative AI model, and means for notifying the user of the review results via a web interface or email notification, thereby automating the advertising material review process and enabling efficient and objective review.

[1478] "Advertising materials" are digital content such as text, images, and videos used as advertisements.

[1479] "Generative AI" is a system that uses artificial intelligence technology to automatically perform specific tasks, in this case reviewing advertising materials.

[1480] "Means for accepting submissions" refers to an interface through which users submit advertising materials to the system, and includes web forms, APIs, etc.

[1481] The "means for outputting the review results" refers to an interface that the generation AI uses to notify the user of the results of its review of the advertising material, and includes a web interface and email notification.

[1482] "Means for converting into text format" refers to the process of analyzing the content of advertising materials and converting them into a format that can be input as text data into a generative AI model.

[1483] A "prompt" is an instruction given to a generative AI model to perform a specific task.

[1484] "Means for obtaining review results from the generative AI model" refers to the process for receiving the review results output by the generative AI model and processing them within the system.

[1485] A "web interface" is an interface that allows a user to access the system through a web browser to perform operations and check results.

[1486] "Email notification" is a function that allows the system to send review results and other information to the user's email address.

[1487] MODE FOR CARRYING OUT THE INVENTION

[1488] The present invention provides a system for efficiently and objectively reviewing advertising materials. This system automatically reviews advertising materials using a generative AI model and notifies the user of the results. Specific embodiments are described below.

[1489] 1. Introduction

[1490] The system aims to automate the entire process, from submitting advertising materials to reviewing and notifying the results. Users submit advertising materials to the system using a web-based submission form or API, and the server reviews them using a generative AI model. The review results are communicated to users via a web interface or email notification.

[1491] 2. Hardware and Software Used

[1492] Server: Receives advertising materials, runs the generative AI model, and outputs the review results.

[1493] Generative AI model: An artificial intelligence model used to review advertising materials, such as OpenAI's GPT-4.

[1494] Web interface: An interface for users to submit advertising materials and check the review results.

[1495] Email notification system: A system for notifying users of the review results by email.

[1496] 3. Submit advertising materials

[1497] Users submit advertising materials to the system using a web-based submission form or API. Specifically, users open a web browser, access a specified URL, enter the required information in the form, and upload the advertising material files. When using the API, users submit the advertising material data as a POST request.

[1498] 4. Receiving and storing advertising materials

[1499] The server receives the ad material sent by the user. The server temporarily stores the data received through the submission form or API endpoint. For example, the server stores the ad material file in a specific directory and records the associated metadata in a database.

[1500] 5. Review Process

[1501] The server passes the received advertising material to the generative AI model and begins the review. Specifically, the server converts the content of the advertising material into text format and generates a prompt for the generative AI model. An example of the prompt is shown below.

[1502] Example prompt sentence:

[1503] The contents of the advertising material are shown below. Please review this advertising material to determine whether it is appropriate.

[1504] Advertisement: "New smartphone now available at a special price!"

[1505] This prompt sentence is input into the generative AI model.

[1506] 6. Obtaining and notifying the results of the review

[1507] The server receives the review results from the generative AI model. The generative AI model analyzes the content of the advertising material and returns an evaluation of its appropriateness and quality. For example, the generative AI model may return an evaluation such as "This advertising material is appropriate."

[1508] Once the results are obtained, the server notifies the user of the results. The output method is a web interface or email notification. Specifically, the server displays the results on a web page so that the user can log in and check the results. If email notification is set up, the results are also sent to the user's email address.

[1509] In this way, the system can conduct the advertising material review process efficiently and objectively.

[1510] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1511] Step 1:

[1512] Users submit advertising materials to the system using a web-based submission form or API. Specifically, users open a web browser, access a specified URL, enter the required information in the form, and upload the advertising material files. When using the API, users submit the advertising material data as a POST request.

[1513] Input: Advertising material files and related information

[1514] Output: Sending the data

[1515] Step 2:

[1516] The server receives the ad material sent by the user. The server temporarily stores the data received through the submission form or API endpoint. For example, the server stores the ad material file in a specific directory and records the associated metadata in a database.

[1517] Input: Advertising material data sent by the user

[1518] Output: Saved ad material files and metadata

[1519] Step 3:

[1520] The server passes the received advertising material to the generative AI model and begins the review. Specifically, the server converts the content of the advertising material into text format and generates a prompt for the generative AI model. An example of the prompt is shown below.

[1521] Example prompt sentence:

[1522] The contents of the advertising material are shown below. Please review this advertising material to determine whether it is appropriate.

[1523] Advertisement: "New smartphone now available at a special price!"

[1524] This prompt sentence is input into the generative AI model.

[1525] Input: Saved ad material file

[1526] Output: Generated prompt statement

[1527] Step 4:

[1528] The server receives the review results from the generative AI model. The generative AI model analyzes the content of the advertising material and returns an evaluation of its appropriateness and quality. For example, the generative AI model may return an evaluation such as "This advertising material is appropriate."

[1529] Input: Generated prompt statement

[1530] Output: Review results from the generative AI model

[1531] Step 5:

[1532] The server notifies the user of the acquired screening results. The output method is a web interface or email notification. Specifically, the server displays the screening results on a web page so that the user can log in and check the results. If email notification is set up, the screening results will also be sent to the user's email address.

[1533] Input: Review results from a generative AI model

[1534] Output: Display of review results on web page and email notification

[1535] In this way, the advertising material review process is efficient and objective.

[1536] (Application example 1)

[1537] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1538] In the conventional advertising material review system, it took a long time from submitting advertising materials to confirming the review results, making it difficult to manage advertising efficiently. In addition, because the review criteria for advertising materials were not clear, there was insufficient understanding of the review results, and revision work was often cumbersome.

[1539] 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.

[1540] In this invention, the server includes a generating AI means for reviewing advertising materials, a means for accepting submission of advertising materials, a means for outputting the review results by the generating AI, and a means for submitting advertising materials via an application installed on a smartphone and checking the review results in real time. This shortens the time from submitting advertising materials to checking the review results, enabling efficient advertising management. In addition, because the review criteria are clearly stated, users can better understand the review results and make corrections easier.

[1541] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[1542] "Veteration" is the process of evaluating advertising materials to determine whether they meet certain criteria.

[1543] "Generative AI" is a system that uses artificial intelligence technology to automatically review advertising materials.

[1544] "Submission" refers to the act of submitting advertising materials to the system.

[1545] A "means" is a method or device used to achieve a particular purpose.

[1546] A "smartphone" is a mobile phone that can connect to the Internet and install applications.

[1547] An "application" is software installed on a smartphone to provide specific functions or services.

[1548] "Real-time" refers to immediate processing or response without delay.

[1549] A "system" is a structure in which multiple elements work together to achieve a specific function.

[1550] The "review results" are the evaluation results that the generation AI outputs after evaluating the advertising material.

[1551] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, and a means for submitting advertising materials via an application installed on a smartphone and checking the review results in real time.

[1552] Explanation of program processing

[1553] The server runs a web server using Flask and provides an API endpoint that accepts the submission of advertising materials. When a user submits advertising materials through a smartphone application, the materials are sent to the server. The server reviews the advertising materials using OpenAI's generative AI model (e.g., text-davinci-003). The review results are returned from the server to the smartphone application in JSON format, allowing the user to check the review results in real time.

[1554] Hardware and software used

[1555] Hardware: Smartphone

[1556] Software: Flask (Python framework), OpenAI API

[1557] Specific examples

[1558] Consider a scenario where an ad agency employee submits a banner image for a new ad campaign for review. The employee opens a smartphone app and uploads the new banner image. The app sends the image to a server, where a generative AI model reviews it. The resulting feedback is, "This banner is fine, but please make the text font size a little larger."

[1559] Prompt Sentence Examples

[1560] Please review the following advertising materials:

[1561] [Banner image URL]

[1562] Judging results:

[1563] Using this prompt, the generative AI model is asked to review the ad material. The review results are displayed in real time on a smartphone app, allowing the person in charge to immediately check for any corrections.

[1564] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1565] Step 1:

[1566] A user launches a smartphone application, selects and uploads advertising materials (images, videos, text, etc.). The input is the advertising materials selected by the user, and the output is the data of the uploaded advertising materials. Specifically, the user taps the "Upload" button on the application and selects advertising materials from the device's storage.

[1567] Step 2:

[1568] The device (smartphone application) sends the uploaded ad material to the server. The input is the ad material data, and the output is the ad material data sent to the server. Specifically, the application uses an HTTP POST request to send the ad material to the server's API endpoint.

[1569] Step 3:

[1570] The server passes the received ad material to the generative AI model for review. The input is the ad material data sent to the server, and the output is the review result by the generative AI model. Specifically, the server calls the OpenAI API and sends the prompt text and ad material data to the generative AI model.

[1571] Step 4:

[1572] The generative AI model reviews the ad material and generates a review result. The input is the prompt and ad material data, and the output is a text of the review result. Specifically, the generative AI model analyzes the ad material and generates feedback on its appropriateness and corrections.

[1573] Step 5:

[1574] The server receives the review results from the generative AI model and returns them to the smartphone application. The input is the review results from the generative AI model, and the output is the review results sent to the smartphone application. Specifically, the server converts the review results into JSON format and sends them to the application as an HTTP response.

[1575] Step 6:

[1576] The terminal (smartphone application) displays the review results received from the server to the user. The input is the review results from the server, and the output is the review results displayed to the user. Specifically, the application analyzes the received JSON data and displays the review results on the screen.

[1577] Step 7:

[1578] The user checks the review results and modifies the advertising material as necessary. The input is the displayed review results, and the output is the modified advertising material. Specifically, the user modifies the advertising material based on the review results and uploads it again.

[1579] Example 2

[1580] Next, a description will be given of Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1581] In conventional advertising material review systems, when making subjective judgments such as whether an advertising material is "offensive / not offensive" or "adult / not," the criteria for judgment are vague, resulting in low reliability of the review results. Furthermore, there are few systems that can analyze both image data and text data, making it difficult to make a comprehensive judgment.

[1582] 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.

[1583] In this invention, the server includes means including a generation AI for screening advertising materials, means for accepting submission of advertising materials, means for outputting the screening results by the generation AI, image recognition means for analyzing image data, natural language processing means for analyzing text data, means for integrating the results of the image recognition means and the natural language processing means, and means for notifying the user of the screening results. This enables subjective judgments such as whether advertising materials are "offensive / not offensive" or "adult / not" to be made with high accuracy and reliability.

[1584] "Generative AI" refers to an artificial intelligence model used to review advertising materials, including techniques for making subjective judgments.

[1585] "Advertising materials" is a general term for digital content including images and text data used as advertisements.

[1586] "Means for accepting submissions" refers to an interface or protocol for users to upload advertising materials to the system.

[1587] "Means for outputting the review results" refers to a method or device for notifying the user of the results of the review conducted by the generation AI.

[1588] "Image recognition means" refers to a technology or device for analyzing image data of advertising materials and classifying them into specific categories.

[1589] "Natural language processing means" refers to technology or devices for analyzing text data of advertising materials and classifying them into specific categories.

[1590] "Means for integrating results" refers to a method or device for integrating the analytical results of the image recognition means and the natural language processing means to arrive at a final review result.

[1591] "Means of notifying the user" refers to the interface or protocol used to communicate the review results to the user.

[1592] This invention relates to a system for reviewing advertising materials. Specifically, it is a system that uses a generative AI model to make subjective judgments about advertising materials, such as whether they are "offensive / not offensive" or "adult / not adult." This system has the function of accepting submissions of advertising materials and outputting the review results of the generative AI.

[1593] Hardware and software used

[1594] Hardware:

[1595] Server: A server with a powerful processor and sufficient memory is required, for example, a server with an Intel Xeon processor is suitable.

[1596] Device: Users need a device such as a computer or smartphone to upload advertising materials.

[1597] software:

[1598] Generative AI models: Use deep learning frameworks such as TensorFlow and PyTorch to screen advertising materials.

[1599] Image Recognition: A pre-trained image recognition model is used to analyze the image data of advertising materials.

[1600] Natural Language Processing: Use natural language processing models such as BERT and GPT-3 to analyze text data from advertising materials.

[1601] Data processing and calculation

[1602] The server receives the advertising material uploaded by the user and analyzes the data in the following manner.

[1603] 1. Image Recognition:

[1604] The server uses TensorFlow and PyTorch to analyze the image data of the advertising materials, for example, classifying the advertising images into categories such as "offensive / not offensive" and "adult / not offensive."

[1605] 2. Natural Language Processing:

[1606] The server uses BERT or GPT-3 to analyze the text data of the advertising material. For example, the text data is classified into categories such as "offensive / not offensive" and "adult / not offensive."

[1607] 3. Synthesis of results:

[1608] The server combines the results of image recognition and natural language processing to make the final decision. For example, if the image is judged to be "offensive" and the text is also judged to be "offensive," the final result will be "offensive."

[1609] Specific examples

[1610] The user uploads the following image and text as advertising material.

[1611] Image: A photo of a person engaging in inappropriate behavior

[1612] Text: "Use this product and you'll be instantly popular!"

[1613] The server first analyzes the image using TensorFlow and determines that the image is "offensive." Next, it analyzes the text using BERT and determines that the text is "offensive." Finally, the server determines that the entire ad material is "offensive" and notifies the user of the result.

[1614] Prompt Sentence Examples

[1615] "Please decide whether the following advertising materials are offensive. Analyze both the image and the text and let us know your findings."

[1616] Image: [Image URL]

[1617] Text: "Use this product and you'll be instantly popular!"

[1618] In this way, the server uses the generative AI model to make a subjective judgment on the advertising material.

[1619] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1620] Step 1:

[1621] The user uploads advertising materials.

[1622] The user uses the system's upload form to select image files and text data, and clicks the submit button. The input is the image file and text data of the advertising material. The output is the advertising material data sent to the server.

[1623] Step 2:

[1624] The server receives the image data.

[1625] The server receives image data uploaded by the user via an HTTP request and temporarily stores it. The input is the image file sent by the user. The output is the image file saved on the server. Specifically, it is saved as / tmp / uploads / image.jpg.

[1626] Step 3:

[1627] The server receives the text data.

[1628] The server receives text data uploaded by the user via an HTTP request and temporarily stores it. The input is the text data sent by the user. The output is the text data stored on the server. Specifically, it is stored in / tmp / uploads / text.txt.

[1629] Step 4:

[1630] The server performs image recognition.

[1631] The server uses TensorFlow and PyTorch to analyze the received image data. The input is the image file stored on the server. The output is the classification of the image into categories such as "offensive / not offensive" and "adult / not." Specifically, it executes model.predict(' / tmp / uploads / image.jpg') and obtains the results.

[1632] Step 5:

[1633] The server performs natural language processing.

[1634] The server uses BERT or GPT-3 to analyze the received text data. The input is the text data stored on the server. The output is the classification of the text into categories such as "offensive / not offensive" and "adult / not adult." Specifically, execute bert_model.predict(' / tmp / uploads / text.txt') to obtain the results.

[1635] Step 6:

[1636] The server aggregates the results.

[1637] The server integrates the results of image recognition and natural language processing. The input is the result of image recognition and the result of natural language processing. The output is the integrated final judgment result. For example, if both results match "unpleasant," the final result is "unpleasant."

[1638] Step 7:

[1639] The server notifies the user of the results.

[1640] The server notifies the user of the integrated results. The input is the integrated final review result. The output is the review result notified to the user. Specifically, the HTTP response is "This advertising material is offensive."

[1641] (Application example 2)

[1642] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1643] When reviewing advertising materials, it is necessary to quickly and accurately make subjective judgments, such as whether the content of images and text is offensive or adult. However, traditional methods require humans to make these judgments, which is time-consuming and has inconsistent standards. Furthermore, there is a lack of an efficient system for checking the quality of advertising materials in advance, making it difficult to control the quality of advertising.

[1644] 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.

[1645] In this invention, the server includes a generating AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generating AI, a means for analyzing images and text of the advertising materials, and a means for integrating the analysis results to make a final decision. This enables the quality of advertising materials to be screened quickly and accurately, and advertising quality control to be achieved.

[1646] "Generative AI" is an artificial intelligence technology that analyzes the content of advertising materials and makes subjective judgments.

[1647] "Advertising materials" are digital content such as images and text used as advertisements.

[1648] "Means for accepting submissions" refers to the interface or function for uploading advertising materials to the system.

[1649] "Means for outputting review results" refers to a function for displaying or reporting the analysis results of the generating AI.

[1650] "Means for analyzing images and text" refers to techniques and algorithms for separately analyzing images and text in advertising materials.

[1651] The "means of integrating the analysis results and making the final decision" is a function that integrates the analysis results of images and text to make a final decision on the quality and appropriateness of advertising materials.

[1652] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, a means for analyzing images and text in the advertising materials, and a means for integrating the analysis results to make a final decision.

[1653] The server runs a generation AI to review the advertising materials. The generation AI analyzes the images and text of the advertising materials and makes subjective judgments such as "offensive / not offensive" and "adult / not." Specifically, it uses OpenCV as its image recognition technology and Hugging Face's Transformers as its natural language processing technology.

[1654] Users submit advertising materials through an interface to upload them to the system. The submitted advertising materials are sent to the server and analyzed by the generation AI. The analysis results are integrated by the server and output as the final review results.

[1655] For example, suppose a user uploads the following image and text as advertising material:

[1656] Image: path / to / ad_image.jpg

[1657] Text: "This is an example ad text."

[1658] The server first analyzes the image using an image recognition model, then analyzes the text using a natural language processing model, and combines the results to determine whether the ad material is offensive or not, or whether it is adult content or not.

[1659] An example of a prompt for the generative AI model is as follows:

[1660] Prompt for image recognition model: "Analyze the image for offensive or adult content."

[1661] Prompt the natural language processing model: "Analyze the text for offensive or adult content."

[1662] This allows for quick and accurate review of the quality of advertising materials, enabling quality control of advertising.

[1663] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1664] Step 1:

[1665] Users upload advertising materials to the system.

[1666] Input: Advertising materials (images and text)

[1667] Output: Advertising material sent to the server

[1668] Specific operation: The user selects the image file and text of the advertising material through the system interface and clicks the upload button, which sends the advertising material to the server.

[1669] Step 2:

[1670] The server receives images of the advertising material and inputs them into an image recognition model.

[1671] Input: Ad image sent to the server

[1672] Output: Input data to the image recognition model

[1673] Specific operation: The server preprocesses the received advertising image using OpenCV and converts it into a format suitable for the image recognition model, then inputs it into the image recognition model.

[1674] Step 3:

[1675] The server analyzes the advertisement image using an image recognition model.

[1676] Input: Input data to the image recognition model

[1677] Output: Image analysis results (offensive / not offensive, adult / not)

[1678] Specific operation: The server runs an analysis on the image recognition model using the prompt "Analyze the image for offensive or adult content." and obtains the results.

[1679] Step 4:

[1680] The server receives the text of the advertising material and inputs it into a natural language processing model.

[1681] Input: Ad text sent to the server

[1682] Output: Input data to the natural language processing model

[1683] Specific operation: The server preprocesses the received advertisement text and converts it into a format suitable for the natural language processing model, then inputs it into the natural language processing model.

[1684] Step 5:

[1685] The server analyzes the advertisement text using a natural language processing model.

[1686] Input: Input data to the natural language processing model

[1687] Output: Text analysis results (offensive / not offensive, adult / not)

[1688] Specific operation: The server runs analysis against the natural language processing model using the prompt sentence "Analyze the text for offensive or adult content." and obtains the results.

[1689] Step 6:

[1690] The server integrates the image analysis results and text analysis results and makes the final decision.

[1691] Input: Image analysis results, text analysis results

[1692] Output: Final judgment result (offensive / not offensive, adult / not)

[1693] Specific operation: The server integrates the image analysis results and text analysis results to evaluate the overall quality of the advertising material, and makes a final decision based on the integrated results.

[1694] Step 7:

[1695] The server outputs the final decision result to the user.

[1696] Input: Final decision result

[1697] Output: Notification of review results to the user

[1698] Specific operation: The server displays the final judgment result through an interface to notify the user. The user can check the review result and modify the advertising material as necessary.

[1699] Example 3

[1700] Next, a third embodiment of the third embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1701] In conventional advertising material review systems, only the review results are provided, and the criteria for the review are not clearly stated, making it difficult for users to understand the reasons for the review results. In addition, because the review results include subjective judgments, the reliability and transparency of the results are lacking.

[1702] The specific processing by the specific processing unit 290 of the data processing device 12 in the third embodiment is realized by the following means.

[1703] In this invention, the server includes a generative AI model for reviewing advertising materials, a means for accepting submission of advertising materials, a means for outputting the review results by the generative AI model, and a means for clearly displaying the review results along with the judgment criteria, thereby enabling users to clearly understand the reasons for the review results and improving the reliability and transparency of the review results.

[1704] A "generative AI model" is an artificial intelligence algorithm that analyzes data entered by the user and generates review results and their criteria.

[1705] "Advertising materials" are digital content such as text, images, and videos used as advertisements.

[1706] The "means for accepting submissions" refers to an interface or protocol that allows users to submit advertising materials to the system.

[1707] The "means for outputting the review results" is a function for providing the review results generated by the generative AI model to the user.

[1708] "Means to clearly indicate the criteria for judgment" is a function that shows the user the factors that form the basis of the results when the generative AI model outputs the review results.

[1709] This invention is a system for reviewing advertising materials, and uses a generative AI model to clearly show the review results and the criteria for judgment. Specific embodiments of this system are described below.

[1710] The server analyzes the advertising material entered by the user using a generative AI model (e.g., a general natural language processing model). This generative AI model can make subjective judgments such as whether the advertising material is "offensive / not offensive" or "adult / not adult."

[1711] The user inputs advertising material into the input field of the device. For example, the user inputs the text "This movie is very violent and unbearable to watch." The device sends this input text data to the server. The server inputs the received text data into the generative AI model for analysis.

[1712] The generative AI model analyzes the input text data and generates a review result and its criteria. For example, for the text "This movie is extremely violent and unbearable to watch," it generates the review result "unpleasant" and the criterion "contains violent content."

[1713] The server sends the review results and criteria obtained from the generative AI model to the user's device. The device then displays the received review results and criteria to the user. For example, it might display "Review result: Offensive" or "Criteria: Contains violent content" on the screen.

[1714] As a concrete example, by inputting the following prompt sentence into the generative AI model, the review results and judgment criteria can be obtained:

[1715] Please review the following text and state your findings and criteria:

[1716] "This movie is extremely violent and unbearable to watch."

[1717] In this way, the user can clearly understand the screening results and the criteria for their judgment. This system improves the reliability and transparency of the screening results, making it easier for the user to understand the reasons for the screening results. The flow of the identification process in the third embodiment will be described with reference to FIG. 15.

[1718] Step 1:

[1719] The user inputs text data.

[1720] The user inputs the text data of the advertising material that he / she wishes to review into an input field on the terminal. For example, he / she inputs the text "This movie is extremely violent and unbearable to watch."

[1721] Input: Text data of advertising materials

[1722] Output: Text data entered into the terminal

[1723] Step 2:

[1724] The terminal transmits the input text data to the server.

[1725] The terminal packages the entered text data in JSON format and sends it to the server as an HTTP request.

[1726] Input: Text data entered into the terminal

[1727] Output: Text data sent to the server

[1728] Step 3:

[1729] The server receives the text data and inputs it into the generative AI model.

[1730] The server receives the text data sent from the device and inputs it into a generative AI model, which uses, for example, a natural language processing model.

[1731] Input: Text data sent to the server

[1732] Output: Text data input to the generative AI model

[1733] Step 4:

[1734] The generative AI model analyzes the text data and generates review results and judgment criteria.

[1735] The generative AI model analyzes the input text data and generates a review result and its criteria. For example, it generates a review result such as "offensive" and a criterion such as "contains violent content."

[1736] Input: Text data fed into the generative AI model

[1737] Output: Review results and criteria

[1738] Step 5:

[1739] The server sends the review results and criteria to the user's terminal.

[1740] The server packages the review results and judgment criteria obtained from the generative AI model in JSON format and sends them to the user's device as an HTTP response.

[1741] Input: Review results and criteria

[1742] Output: Review results and criteria sent to the user's device

[1743] Step 6:

[1744] The terminal displays the review results and criteria to the user.

[1745] The device analyzes the review results and criteria received from the server and displays them to the user. For example, it might display on the screen, "Review result: Offensive" or "Criteria: Contains violent content."

[1746] Input: Review results and criteria sent to the user's device

[1747] Output: The review results and criteria displayed to the user

[1748] (Application example 3)

[1749] Next, a description will be given of Application Example 3 of Form Example 3. 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."

[1750] In conventional advertising material review systems, the review results are output by the AI ​​generator, but the criteria for the review are not clearly stated, making it difficult for users to understand the reasons for the review results. Furthermore, content distribution services lacked a means for users to check the appropriateness of content before viewing it. This increased the risk of users encountering unpleasant content.

[1751] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 3 is realized by the following means.

[1752] In this invention, the server includes a generating AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generating AI, a means for clearly indicating the judgment criteria along with the screening results, a means for generating a prompt sentence for screening content, a means for obtaining the screening results using the generating AI model, and a means for providing the screening results to the user, which makes it easier for the user to understand the reasons for the screening results and enables the user to confirm the appropriateness of the content before viewing it.

[1753] "Generative AI" is a system that uses artificial intelligence techniques to generate data and perform specific tasks.

[1754] "Advertising materials" are content such as text, images, and videos used as advertisements.

[1755] The "means for accepting submissions" refers to an interface or system for receiving advertising materials from users.

[1756] The "means for outputting the review results" is a system for displaying the results of the review conducted by the generation AI to the user.

[1757] The "means of clearly indicating the criteria for judgment" is a system that shows the user the basis for the judgment when the generation AI outputs the review results.

[1758] The "means for generating prompt sentences" is a system that creates instruction sentences to request review from the generation AI.

[1759] "Means for obtaining review results using a generative AI model" refers to a system that uses a generative AI model to review advertising materials and content and obtain the results.

[1760] "Means for providing the review results to the user" refers to a system for notifying the user of the review results conducted by the generation AI.

[1761] A system for implementing this invention includes a generating AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generating AI, a means for clearly indicating the judgment criteria along with the review results, a means for generating prompt sentences for reviewing content, a means for obtaining the review results using a generating AI model, and a means for providing the review results to a user.

[1762] The server provides an interface for receiving advertising materials from users. When a user submits advertising materials, the server uses a generation AI to review the materials. The generation AI makes subjective judgments such as whether the advertising materials are "offensive / not offensive" or "adult / not" and returns the results to the server.

[1763] The server then displays the results of the review along with the criteria for its decision. For example, if the content is deemed "offensive," the server will provide information such as "contains violent content" as the reason. This makes it easier for users to understand the reasons for the review.

[1764] Furthermore, the server generates a prompt to review the content. The generated prompt is an instruction to the generative AI model to request a review. The generative AI model reviews the advertising materials and content based on the prompt and returns the results to the server.

[1765] The server provides the user with the review results obtained using the generative AI model, allowing the user to confirm the appropriateness of the content before viewing it.

[1766] The hardware used includes an internet-connected server and the user's smartphone, and the software uses Python and the OpenAI API.

[1767] As a concrete example, consider a case where a part of a movie a user wants to watch "contains violent scenes." In this case, the generative AI receives the prompt "This movie contains violent scenes," judges the movie to be "unpleasant," and returns the reason as "It contains violent content."

[1768] Example prompt sentence:

[1769] Please review the following content:

[1770] This film contains violent scenes.

[1771] Please specify whether or not you find this content offensive and why.

[1772] In this way, the user can check the results of the content review and the reasons for it before viewing.

[1773] The flow of the specific processing in Application Example 3 will be described with reference to FIG.

[1774] Step 1:

[1775] The user submits advertising materials.

[1776] Users upload advertising materials (text, images, videos, etc.) to the server. The server receives this advertising material and stores it in a database. The input is the advertising material provided by the user, and the output is the advertising material stored on the server.

[1777] Step 2:

[1778] The server reviews the advertising material using generated AI.

[1779] The server passes the stored ad material to the generation AI for review. The generation AI analyzes the ad material and makes a subjective judgment such as "offensive / not offensive" or "adult / not." The input is the ad material stored on the server, and the output is the review results by the generation AI.

[1780] Step 3:

[1781] The server obtains the review results and the criteria.

[1782] The generation AI returns the review results along with the criteria for its judgment to the server. For example, if it judges something to be "offensive," it will provide information such as "it contains violent content" as the reason. The input is the review results by the generation AI, and the output is the review results and the criteria for that judgment.

[1783] Step 4:

[1784] The server generates a prompt.

[1785] The server generates a prompt to request the AI ​​to review the ad. The prompt includes information about the ad material to be reviewed. The input is the information about the ad material to be reviewed, and the output is the generated prompt.

[1786] Step 5:

[1787] The server uses the generative AI model to obtain the review results.

[1788] The server passes the generated prompt to the generative AI model for review. The generative AI model reviews the advertising material based on the prompt and returns the results to the server. The input is the generated prompt, and the output is the review result by the generative AI model.

[1789] Step 6:

[1790] The server provides the review results to the user.

[1791] The server notifies the user of the review results obtained using the generative AI model. The user can check the review results and their judgment criteria. The input is the review results obtained by the generative AI model, and the output is the review results and their judgment criteria provided to the user.

[1792] 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.

[1793] "Example 1"

[1794] In one embodiment of the present invention, a system including a generation AI that reviews advertising materials, a means for accepting submissions of advertising materials, and a means for outputting the review results of the generation AI is combined with an emotion engine that recognizes user emotions. Specifically, the emotion engine analyzes the user's reaction (e.g., facial expression, tone of voice, etc.) when viewing the advertising materials and feeds the results back to the generation AI. This allows the generation AI to review the advertising materials based on the user's emotions.

[1795] "Example 2"

[1796] In another embodiment of the present invention, the emotion engine further includes a means for analyzing a user's reaction and feeding the result back to the generation AI's review process. Specifically, the emotion engine analyzes the user's reaction when viewing the advertising material and feeds the result back to the generation AI. This allows the generation AI to review the advertising material based on the user's emotion.

[1797] "Example 3"

[1798] In yet another embodiment of the present invention, the emotion engine further includes a means for detecting changes in the user's emotions and adjusting the screening criteria of the generation AI in response to those changes. Specifically, the emotion engine analyzes the user's reaction when viewing advertising material and feeds back the results to the generation AI. This allows the generation AI to adjust the screening criteria of the advertising material in response to changes in the user's emotions.

[1799] The processing flow of each embodiment will be described below.

[1800] "Example 1"

[1801] Step 1: User views advertising material.

[1802] Step 2: The emotion engine analyzes the user's reactions (e.g., facial expressions, tone of voice, etc.).

[1803] Step 3: The analysis results are fed back to the generation AI.

[1804] Step 4: Generative AI reviews advertising materials based on user sentiment.

[1805] "Example 2"

[1806] Step 1: User views advertising material.

[1807] Step 2: The emotion engine analyzes the user's reaction.

[1808] Step 3: The analysis results are fed back to the generation AI.

[1809] Step 4: Generative AI reviews advertising materials based on user sentiment.

[1810] "Example 3"

[1811] Step 1: User views advertising material.

[1812] Step 2: The emotion engine analyzes the user's reaction.

[1813] Step 3: The analysis results are fed back to the generation AI.

[1814] Step 4: The generative AI adjusts the review criteria for advertising materials according to changes in user emotions.

[1815] Example 1

[1816] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1817] In conventional advertising material review systems, the evaluation of the appropriateness and quality of advertising materials involves subjective judgment, which can lead to inconsistent review results. Furthermore, because the review is conducted without taking into account the user's feelings, it is difficult to accurately evaluate the actual impact of advertising materials on users. Furthermore, because the criteria for the review results are not clearly stated, users find it difficult to accept the review results.

[1818] The identification process by the identification processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means. In this invention, the server includes a generation AI means for screening advertising materials, a means for accepting submission of advertising materials, a means for outputting the screening results by the generation AI, a sentiment analysis engine means for recognizing the user's emotions, and a means for feeding back the analysis results of the sentiment analysis engine to the generation AI. This allows for consistent screening of advertising materials, enables evaluation taking into account the user's emotions, and makes it easier for users to accept the screening results because the criteria for the screening results are clearly stated.

[1819] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[1820] "Advertising materials" are digital content such as images, text, and videos used as advertisements.

[1821] "Means for accepting submissions" refers to an interface such as a web form or API that allows users to upload advertising materials to the system.

[1822] "Means for outputting the review results" refers to methods such as a web interface or email notification for notifying the user of the review results generated by the generation AI.

[1823] An "emotion analysis engine" is software or hardware that analyzes a user's facial expressions, tone of voice, etc. to recognize the user's emotions.

[1824] The "means of feedback" refers to a method for sending the analysis results of the sentiment analysis engine to the generation AI and reflecting them in the review process.

[1825] The present invention is a system that uses a generative AI to review advertising materials. A specific embodiment of this system will be described below.

[1826] System Configuration

[1827] This system consists of the following main components:

[1828] 1. A means of accepting ad material submissions: A web form or API that allows users to upload ad material to the system.

[1829] 2. Generative AI: Artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[1830] 3. Means of outputting the review results: A web interface or email notification to notify the user of the review results generated by the generative AI.

[1831] 4. Emotion Analysis Engine: Software or hardware that analyzes a user's facial expressions and tone of voice to recognize their emotions.

[1832] 5. Feedback method: The results of the sentiment analysis engine's analysis are sent to the generation AI and reflected in the review process.

[1833] Submitting advertising materials

[1834] Users submit advertising materials to the system using a web-based submission form or API. Specifically, when uploading advertising images through a web browser, users enter the necessary information into the submission form, attach the image file, and submit it.

[1835] Receipt and analysis of advertising materials

[1836] The server receives the advertising material sent by the user and temporarily stores it in a database. The server then passes the received advertising material to the generation AI, which analyzes the content of the advertising material and evaluates its appropriateness and quality. For example, the generation AI checks whether the advertising text contains inappropriate language and whether the images comply with regulations.

[1837] User sentiment analysis

[1838] When a user views advertising materials, the emotion analysis engine analyzes the user's reactions. Specifically, the user's facial expressions and tone of voice are collected through a camera and microphone, and the emotion analysis engine analyzes the data. For example, it can recognize emotions such as whether the user smiled or was surprised when they saw the advertisement.

[1839] Feedback and review updates

[1840] The server feeds back the results of the emotion analysis engine to the generation AI, which then uses this feedback to make a final review of the ad material. For example, if the user response is positive, the ad material will receive a high rating.

[1841] Notification of review results

[1842] The server notifies the user of the final review results by the generation AI. Notification methods include a web interface and email notification. For example, the user can check the review results through the web interface.

[1843] Examples and prompts

[1844] As a concrete example, consider the case where a user submits an advertising image for a new cosmetic product. The user accesses a web form, enters the required information, and uploads the image file. The server receives this image and passes it to the generation AI. The generation AI analyzes the content of the image and evaluates its appropriateness. At the same time, an emotion analysis engine analyzes the user's facial expressions and tone of voice as they view the advertising image, and feeds the results back to the generation AI. Finally, the generation AI notifies the user of the review results via a web interface.

[1845] Example prompt sentence:

[1846] "Please review the advertisement images for a new cosmetic product. Analyze the user's facial expressions and tone of voice, and give a thumbs up if the user responds positively."

[1847] In this way, this system combines generative AI and an emotion analysis engine to efficiently and accurately review advertising materials.

[1848] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1849] Step 1:

[1850] Users submit advertising materials to the system using a web-based submission form or API. Specifically, when uploading advertising images through a web browser, users enter the necessary information into the submission form, attach the image file, and submit it.

[1851] Input: Ad material (image file), ad title, description, category, and other information

[1852] Output: Advertising material data sent to the server

[1853] Specific behavior:

[1854] A user accesses a web form.

[1855] The user inputs information such as the title, description, and category of the advertising material.

[1856] The user selects an advertisement image file and clicks the upload button.

[1857] Step 2:

[1858] The server receives the advertising material sent by the user and temporarily stores it in a database.

[1859] Input: Advertising material data sent by the user

[1860] Output: Advertising material data stored in the database

[1861] Specific behavior:

[1862] The server receives an HTTP request.

[1863] The server stores the advertising material data in a database.

[1864] Step 3:

[1865] The server passes the received advertising material to the generation AI, which analyzes the content of the advertising material and evaluates its appropriateness and quality.

[1866] Input: Advertising material data stored in the database

[1867] Output: Review results by the generated AI

[1868] Specific behavior:

[1869] The server inputs advertising material data into the generated AI model.

[1870] Generative AI analyzes the text and images of the ad.

[1871] The generation AI generates the review results and returns them to the server.

[1872] Step 4:

[1873] When a user views advertising materials, the emotion analysis engine analyzes the user's reactions. Specifically, the user's facial expressions and tone of voice are collected through a camera and microphone, and the emotion analysis engine analyzes the data.

[1874] Input: User's facial expression data, tone of voice data

[1875] Output: Sentiment analysis results from the sentiment analysis engine

[1876] Specific behavior:

[1877] A user accesses a web page for viewing advertising material.

[1878] The user's device activates the camera and microphone.

[1879] The emotion analysis engine analyzes the user's facial expressions and tone of voice in real time.

[1880] Step 5:

[1881] The server feeds back the results of the emotion analysis engine to the generation AI, which then performs a final review of the advertising materials based on this feedback.

[1882] Input: Emotion analysis results from the emotion analysis engine

[1883] Output: Final review results from the updated generative AI

[1884] Specific behavior:

[1885] The server receives the analysis results from the emotion analysis engine.

[1886] The server sends the analysis results to the generation AI.

[1887] The generative AI updates the review results based on the feedback.

[1888] Step 6:

[1889] The server notifies the user of the final review results by the generation AI via a web interface or email.

[1890] Input: Final review results from the updated generation AI

[1891] Output: Review results notified to the user

[1892] Specific behavior:

[1893] The server stores the review results in a database.

[1894] The server sends an email notification to the user.

[1895] The user checks the results of the review through a web interface.

[1896] (Application example 1)

[1897] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1898] Conventional advertising material review systems were unable to take into account subjective factors or users' emotional reactions when determining whether the content of advertising material was appropriate. This made it difficult to accurately evaluate the impact of advertising material on viewers, resulting in the risk of inappropriate advertising being delivered. Furthermore, because the criteria for review results were not clearly stated, advertisers were sometimes dissatisfied with the review results.

[1899] 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.

[1900] In this invention, the server includes a generation AI means for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine means for recognizing user emotions, and a means for analyzing the user's emotional responses and providing feedback to the generation AI. This makes it possible to take the user's emotional responses into consideration when reviewing advertising materials, thereby providing more accurate review results. Furthermore, by clearly indicating the criteria for the review results, advertisers can more easily accept the review results.

[1901] "Generative AI" is artificial intelligence that analyzes the content of advertising materials and evaluates their appropriateness and quality.

[1902] "Advertising materials" are digital content such as images, videos, and text used as advertisements.

[1903] "Submission methods" are interfaces such as web-based forms or APIs for uploading advertising materials into the system.

[1904] "Means for outputting the review results" refers to methods such as a web interface or email notification for notifying the user of the review results generated by the generation AI.

[1905] An "emotion engine" is software or hardware that analyzes a user's facial expressions, tone of voice, etc. to recognize emotions.

[1906] The "means for analyzing emotional responses and providing feedback to the generation AI" is a system that collects emotional data when users view advertising materials and provides it to the generation AI.

[1907] A system for implementing this invention includes a generation AI for reviewing advertising materials, a means for accepting submissions of advertising materials, a means for outputting the review results by the generation AI, an emotion engine for recognizing user emotions, and a means for analyzing the user's emotional reactions and providing feedback to the generation AI. Specific embodiments of the system are described below.

[1908] The server runs a generative AI model to review ad materials. The generative AI model analyzes the content of the ad materials and evaluates their appropriateness and quality. Ad materials are submitted via a web-based form or API. When a user uploads an ad material, the server passes the material to the generative AI model for review.

[1909] The results of the review are output to the user via a web interface or email notification. The review results include subjective judgments such as whether the advertising material is "offensive / not offensive" or "adult / not." The review criteria are also clearly stated, making it easy for users to understand the results.

[1910] Furthermore, it is equipped with an emotion engine that recognizes the user's emotions. The emotion engine analyzes the user's facial expressions and tone of voice when viewing advertising materials and feeds this emotional data back to the generation AI. This allows the generation AI to conduct screening that takes the user's emotional reactions into account.

[1911] For example, when a user uploads a new advertising video, the generative AI reviews the video, analyzes the user's facial expressions while watching the video, and provides feedback along with the review results. For example, the following prompt sentence can be input into the generative AI model:

[1912] "To review advertising materials, please analyze the following advertising video. Please output the review results, taking into account the user's emotional response.

[1913] Advertisement video: [Advertisement video URL]

[1914] User's emotional response: [user's facial expression data]

[1915] In this way, it becomes possible to take into account the emotional response of the user when screening advertising materials, thereby providing more accurate screening results.

[1916] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1917] Step 1:

[1918] The user uploads advertising materials.

[1919] Input: Advertising materials (images, videos, text, etc.)

[1920] How it works: Users upload advertising materials to the server using a web-based form or API.

[1921] Output: Ad material file saved on the server

[1922] Step 2:

[1923] The server passes the advertising material to a generative AI model for review.

[1924] Input: Ad material file stored on the server

[1925] How it works: The server feeds the uploaded ad material into a generative AI model to evaluate its appropriateness and quality.

[1926] Output: Review results from the generative AI model

[1927] Step 3:

[1928] The server notifies the user of the review results.

[1929] Input: Review results from the generative AI model

[1930] How it works: The server sends the results of the review to the user via a web interface or email notification.

[1931] Output: Review results notified to the user

[1932] Step 4:

[1933] When a user views advertising material, the emotion engine analyzes their emotional response.

[1934] Input: advertising materials, user facial expressions and tone of voice

[1935] How it works: When a user views advertising mat...

Claims

1. A review means for reviewing advertising materials, which are digital content including any of text, images, and videos used as advertisements, provided by a person in charge of an advertising agency based on specific review criteria using a generative AI model; A means for outputting the results of the screening using a generative AI model and transmitting the results of the screening to the person in charge; an emotion engine that recognizes the emotion of a user viewing the advertising material; An adjustment means for adjusting the screening criteria based on the emotion by inputting a specific prompt sentence into a generative AI model; A system including:

2. The system according to claim 1 further comprises a notification means for determining whether the advertising material contains offensive elements, and if it is determined that the advertising material contains offensive elements, for notifying the user or the person in charge of a message indicating the reason why the advertising material contains offensive elements.

3. The system according to claim 2 , wherein the notification means notifies the user or the person in charge of the message when the advertising material contains a violent scene or a discriminatory expression.

Images