system

Abstract

We provide the system. [Solution] A means for receiving questions from users, A means of sending questions to multiple generative artificial intelligence systems, A means of receiving responses from each generative artificial intelligence system, A means of evaluating multiple received responses and selecting the most appropriate response, A means to automatically correct the selected answer, A means of displaying the corrected final answer and related advertisements, A system that includes this.

Description

Technical Field

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

Background Art

[0002] Patent Document 1 discloses a persona chatbot control method performed by at least one processor, including steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a 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

Summary of the Invention

Problems to be Solved by the Invention

[0004] Current generative artificial intelligence learns a large number of texts and outputs words with a high probability of coming next, but may generate incorrect texts due to insufficient learning or data bias. Therefore, it is inappropriate for real-time investigations and fact-checking, which is a major drawback. This problem is not easily improved. The present invention aims to provide a system that improves the accuracy of answers by combining multiple generative artificial intelligences.

Means for Solving the Problems

[0005] The present invention is a system comprising means for receiving questions from a user, means for sending questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, and means for displaying the corrected final answer and related advertisements. Furthermore, the system includes means for performing grammatical checks and fact-checking on the selected answer. It also includes means for setting evaluation criteria for selecting the optimal answer, calculating an evaluation score for each answer, and selecting the answer with the highest evaluation based on that score. In this way, it is possible to utilize multiple generative artificial intelligence systems to provide more accurate and reliable answers.

[0006] A "user" is an individual or group that uses the system to input questions and receive answers.

[0007] "Means for receiving questions" refers to a device or software that acquires questions entered by a user and stores them in a format that can be processed within the system.

[0008] "Generative artificial intelligence" is a technology that learns from large amounts of text data, predicts the next word or phrase, and generates text.

[0009] "Means for sending questions" refers to a device or software that converts received questions into an appropriate format and sends them to a generative artificial intelligence system.

[0010] "Means for receiving responses" refers to a device or software that receives responses sent back from a generative artificial intelligence, and stores and processes them within the system.

[0011] "Evaluation criteria" are standards such as fact-checking, grammatical accuracy, and reliability that are used to compare and evaluate multiple responses returned by generative artificial intelligence.

[0012] "Means for evaluating responses" refers to a device or software that compares multiple responses received based on multiple evaluation criteria and selects the most appropriate response.

[0013] "A means of selecting the most appropriate answer" refers to a device or software that selects the answer with the highest rating based on a score calculated according to evaluation criteria.

[0014] "Means for automatically correcting answers" refers to a device or software that performs grammatical checks and fact-checks on selected answers and corrects them to be accurate and easy to read.

[0015] "Means for displaying relevant advertisements" refers to a device or software that retrieves relevant advertisements based on the user's responses and displays them to the user.

[0016] "Grammar checking" is the process of detecting and correcting grammatical errors in generated text.

[0017] "Fact-checking" is the process of verifying whether the generated text contains accurate information using reliable external sources.

[0018] "Means for calculating evaluation scores" refers to a device or software that calculates a score based on evaluation criteria set for each response. [Brief explanation of the drawing]

[0019] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] It is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] It is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] It is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] It is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] It shows an emotion map to which a plurality of emotions are mapped. [Figure 10] It shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Embodiments for Carrying Out the Invention

[0020] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0021] First, the terms used in the following description will be explained.

[0022] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), and APU (Accelerated Processing Unit).

[0023] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0024] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0025] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0026] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0027] [First Embodiment]

[0028] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0029] As shown in Figure 1, the 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.

[0030] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0031] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0032] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0033] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0034] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0035] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0036] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0037] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0038] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0039] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0040] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[0041] Receive user questions

[0042] The terminal accepts questions from the user via an input form. These questions are entered by the user, for example, through a web browser or mobile application. Let's say the user enters, "When are the next Olympics?" This question is transmitted to the server in real time.

[0043] Send a question to multiple AIs

[0044] The server converts the received question into the appropriate format and sends it to multiple generative artificial intelligence (AI) systems. For example, it makes asynchronous requests to the API endpoints of each AI service. This causes each AI to generate an answer to the question using its own algorithm.

[0045] Receive responses from each AI

[0046] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[0047] Select the most appropriate answer.

[0048] The server evaluates multiple received responses and selects the most appropriate one based on criteria such as fact-checking, grammatical accuracy, and reliability. A scoring function is used for evaluation; for example, if "July 2024" receives the highest score, it will be selected.

[0049] Automatically correct the selected answer.

[0050] The server performs grammatical checks and fact-checks on the selected answers. For example, it checks the answer "July 2024" and makes any necessary corrections. Furthermore, it adds more detailed information, such as "from July 26, 2024," using reliable sources.

[0051] Display the final answer and related ads.

[0052] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0053] Specific example

[0054] If a user asks, "When is the next Olympics?":

[0055] 1. The terminal receives the question and forwards it to the server.

[0056] 2. The server sends questions to multiple generative artificial intelligence systems.

[0057] 3. Each AI will respond with answers such as "2024," "next year," or "July 2024."

[0058] 4. The server selects the most appropriate answer, "July 2024".

[0059] 5. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[0060] 6. The device displays the user the final answer, "The next Olympics will be held from July 26, 2024," along with related advertisements.

[0061] This system allows users to quickly obtain accurate and reliable answers.

[0062] The following describes the processing flow.

[0063] Step 1:

[0064] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[0065] Step 2:

[0066] The terminal sends the question to the server. The entered question is transferred to the server in real time. During this process, the question is transmitted securely using network communication.

[0067] Step 3:

[0068] The server sends a question to multiple generative artificial intelligence (AI) services. Upon receiving the question, the server converts its content into an appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it might send a question to "AI Service 1," "AI Service 2," and "AI Service 3."

[0069] Step 4:

[0070] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, it might receive responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3".

[0071] Step 5:

[0072] The server evaluates the responses and selects the most appropriate one. A scoring function is applied to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. The response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[0073] Step 6:

[0074] The selected answer will be automatically corrected. The server will perform grammatical checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as necessary. Furthermore, it will add detailed information such as "from July 26, 2024" by referring to reliable sources.

[0075] Step 7:

[0076] The system generates the final answer and related ads. The server retrieves the relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the answer.

[0077] Step 8:

[0078] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[0079] Step 9:

[0080] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0081] In this way, users can quickly obtain accurate and reliable answers simply by entering their questions.

[0082] (Example 1)

[0083] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0084] Conventional information retrieval systems often provide inaccurate information in response to user inquiries. Even with the use of numerous generative artificial intelligence systems, responses may vary or contain unreliable information. In such situations, it is difficult for users to obtain accurate information quickly. Furthermore, users bear the burden of having to judge the reliability of the information they receive themselves.

[0085] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0086] In this invention, the server includes means for receiving questions from users, means for forwarding received questions to the server, means for sending questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the multiple received answers and selecting the most appropriate answer, means for automatically correcting the selected answer, and means for displaying the corrected final answer and related advertisements. This enables users to quickly obtain accurate and reliable answers.

[0087] A "user" refers to an individual or group that inputs questions into a system and receives answers.

[0088] A "terminal" is a device used by a user to input questions and send them to a server, and specifically includes personal computers and smartphones.

[0089] A "server" refers to a computer system that receives questions from users, sends the questions to multiple generative artificial intelligence systems, evaluates and corrects the received answers, and then sends the final answer to the terminal.

[0090] "Generative artificial intelligence" refers to systems that utilize machine learning and natural language processing technologies to automatically generate answers to user questions.

[0091] "Means of receiving questions" refers to a combination of software and hardware for inputting, storing, and transferring user questions to a server in a digital format.

[0092] "Means for sending questions" refers to communication protocols or API interfaces that allow a server to send questions received from a user to multiple generative artificial intelligence systems.

[0093] "Means of receiving responses" refers to communication protocols and databases used by servers to receive and store responses from generative artificial intelligence.

[0094] "Means of evaluating responses" refers to algorithms or evaluation systems that analyze multiple received responses and select the most appropriate response based on evaluation criteria.

[0095] "Means for correcting answers" refers to programs or libraries used to perform grammatical checks and factual verification on selected answers, and then make corrections or revisions.

[0096] "Means for displaying corrected final answers and related advertisements" refers to software and interfaces for displaying the final answers and related advertisements on the user's device.

[0097] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[0098] This system includes a terminal that receives user questions, a server that processes the received questions and transmits them to multiple generative artificial intelligence systems, and means for evaluating and correcting the answers from the generative artificial intelligence systems and providing the user with the final answers and related advertisements.

[0099] Hardware and Software Overview

[0100] terminal

[0101] A device is a device on which the user enters a question, and primarily includes personal computers, smartphones, and tablets. The device provides an interface for entering the question using a web browser or mobile application.

[0102] server

[0103] This is a central computer system for receiving user questions and transmitting them to multiple generative artificial intelligence systems. The server has the following main functions:

[0104] API interface: A means of communication for sending questions to and receiving answers from generative artificial intelligence.

[0105] Database: A storage device for storing and evaluating received responses.

[0106] Evaluation system: An algorithm that evaluates each answer and selects the most appropriate answer.

[0107] Generative artificial intelligence

[0108] This refers to a system that automatically generates answers to questions sent from a server, and includes Google Cloud AI, OpenAI, IBM Watson, and others.

[0109] System operation example

[0110] The user types "When is the next Olympics?" into their web browser. At this point, the terminal receives the user's question and forwards it to the server.

[0111] The server converts the received question into an appropriate format and sends an API request to a generative artificial intelligence (AI) such as Google Cloud AI, OpenAI, or IBM Watson. Multiple generative AIs answer the question based on their own algorithms and return the answer to the server.

[0112] The server receives responses from each generative artificial intelligence system and stores them in a database. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[0113] The server evaluates the multiple responses it receives. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. A scoring algorithm is used to assign points to each response, and the response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[0114] The server performs grammar checks and fact-checks on the selected answer. Grammar checks utilize language processing libraries, and fact-checking relies on reliable sources. As a result, detailed information such as "from July 26, 2024" is added.

[0115] The corrected final answer is sent to the device, which then displays it to the user. Related advertisements are also displayed simultaneously. For example, an advertisement for sporting goods might appear along with the answer, "The next Olympics are from July 26, 2024."

[0116] Example of a prompt

[0117] Question: "When is the next Olympics?"

[0118] Question: "What is the date of the next Olympic Games?"

[0119] Question: "I would like to know the dates of the Olympics."

[0120] As described above, this system allows users to quickly obtain accurate and reliable answers.

[0121] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0122] Step 1:

[0123] The user enters a question.

[0124] Input: The user enters the question via a web browser or mobile application.

[0125] Action: The user types "When are the next Olympics?" into the browser's search form.

[0126] Data processing: Format the input text data into JSON format.

[0127] Output: The formatted question data is temporarily stored on the device.

[0128] Step 2:

[0129] The terminal forwards the question to the server.

[0130] Input: Question data entered by the user (in JSON format).

[0131] Operation: The device uses the JavaScript (registered trademark) fetch API to send the question data to the server as an HTTP request.

[0132] Data processing: Include the question data in the HTTP request body.

[0133] Output: Question data arrives at the server and the server receives it.

[0134] Step 3:

[0135] The server sends questions to multiple generative artificial intelligence systems.

[0136] Input: Question data received from the terminal.

[0137] Operation: The server processes the received question data into an API request format.

[0138] Data processing: Convert the question data into a format compatible with the API of each generative artificial intelligence system. For example, convert from JSON to URL-encoded format.

[0139] Output: The converted question data is sent as an HTTP request to each generative artificial intelligence system.

[0140] Step 4:

[0141] The server receives responses from each generative artificial intelligence system.

[0142] Input: Response data from each generative artificial intelligence.

[0143] Operation: The server asynchronously waits for and receives responses from each generative artificial intelligence system.

[0144] Data processing: Store the received response data in a list.

[0145] Output: A list containing the answers of each generative artificial intelligence.

[0146] Step 5:

[0147] The server evaluates the received responses and selects the most appropriate one.

[0148] Input: Multiple response data (lists) from various generative artificial intelligence systems.

[0149] Operation: The server uses an evaluation algorithm to score each response.

[0150] Data processing: A scoring function is applied to each response to calculate an evaluation score based on reliability and grammatical accuracy.

[0151] Output: Select the answer with the highest rating.

[0152] Step 6:

[0153] The server automatically corrects the selected answer.

[0154] Input: The most appropriate response data.

[0155] Operation: The server performs grammar checks and fact-checks on the selected answer. For grammar checks, it uses, for example, the Grammarly API, and for fact-checks, it refers to reliable sources.

[0156] Data processing: This involves correcting grammar and adding detailed information.

[0157] Output: Corrected final response data.

[0158] Step 7:

[0159] The device displays the final answer and related ads.

[0160] Input: Final response data and related advertising data sent from the server.

[0161] Function: Displays data received by the device on web pages and applications. Dynamically updates information using HTML and JavaScript.

[0162] Data processing: Convert the final response data and advertising data into a format that is compatible with the user interface.

[0163] Output: The user is shown the information "The next Olympic Games will be held from July 26, 2024" and advertisements for related sporting goods.

[0164] Through the steps outlined above, users can quickly obtain accurate and reliable answers.

[0165] (Application Example 1)

[0166] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0167] Conventional generative artificial intelligence systems have a problem in that they do not provide accurate and appropriate answers to user questions with sufficient precision. Furthermore, they lack the ability to appropriately recommend related information and content related to the generated answers, resulting in a situation where user needs and convenience are not fully met.

[0168] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0169] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related information, and means for providing related content based on the displayed information. This enables the user to obtain accurate and reliable answers and simultaneously acquire related content.

[0170] A "user" refers to each individual who uses this system.

[0171] "Means for receiving questions" refers to a device or software that receives input from a user and processes its contents.

[0172] "Generative artificial intelligence" refers to AI models that can generate answers to given questions or inputs.

[0173] "Means for sending questions" refers to a device or software that transmits questions received from a user to multiple generative artificial intelligence systems.

[0174] "Means for receiving responses" refers to a device or software for receiving responses generated by generative artificial intelligence.

[0175] "Means for evaluating responses" refers to a device or software for evaluating multiple received responses and determining which is the most appropriate.

[0176] "Means for selecting the most appropriate answer" refers to a device or software for selecting the optimal answer based on evaluation.

[0177] "Means for automatically correcting answers" refers to a device or software for automatically correcting selected answers based on grammar or facts.

[0178] "Means for displaying the final answer and related information" refers to a device or software for displaying the corrected final answer and related information to the user.

[0179] "Means for providing related content" refers to a device or software for providing additional content to the user based on the displayed information.

[0180] This invention is a system that processes user questions in real time, provides the most appropriate answers, and recommends relevant content. Specific embodiments of this system are described below.

[0181] The system primarily consists of a server, terminals, and users. Users submit questions via their terminals (such as smartphones and tablets). A dedicated application installed on the terminal receives questions from users and forwards them to the server. This application also provides an interface for users to input their questions.

[0182] The server sends questions received from users to multiple generative artificial intelligence (AI) models and receives responses from each AI model. In this process, the server uses OpenAI's GPT-3® API and other AI service technologies to send questions. For example, if a user asks, "What are some recommended new movies?", the server sends this question to the AI ​​models and receives responses from each.

[0183] The server evaluates the multiple responses received and selects the most appropriate one. This evaluation considers factors such as accuracy, reliability, and grammatical correctness. The server calculates a score for each response based on the set evaluation criteria and selects the response with the highest score. The server then automatically performs grammatical checks and fact-checks on the selected response and corrects it to its final form.

[0184] The final corrected answer and related information are sent to the device. The device displays related content to the user along with the final answer. This related content might include, for example, a link to watch the movie or a list of related movies in the case of a question about movie recommendations. The device uses web scraping technology to retrieve and provide the related content to the user.

[0185] Hardware and software to be used

[0186] Hardware: Servers, smartphones, tablets

[0187] Software: OpenAI GPT-3 API, requests, BeautifulSoup, dedicated application (for smartphones and tablets)

[0188] Specific example

[0189] The following shows the specific processing that occurs when a user asks, "What are some of the latest movies you recommend?"

[0190] 1. The user enters "What are some recommended new movies?" into their device.

[0191] 2. The terminal forwards this question to the server.

[0192] 3. The server sends questions to multiple generative AI models and receives responses.

[0193] 4. The server selects the best response from those received and performs grammatical checks and fact-checks.

[0194] 5. The server sends the corrected response and related content to the device.

[0195] 6. The device displays to the user a link to watch the relevant movie and reviews along with the final answer.

[0196] Examples of prompt statements

[0197] "What are some of the latest movies you recommend?"

[0198] "What are your recommended travel destinations this year?"

[0199] "When is the next Olympics?"

[0200] In this way, users can quickly obtain accurate and reliable answers, along with useful content related to those answers. This system can meet user needs and significantly improve convenience.

[0201] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0202] Step 1:

[0203] The user enters a question.

[0204] The user uses a device (such as a smartphone or tablet) to enter a question into the application's input form. For example, they might enter, "What are some recommended new movies?" This input is then accepted by the device.

[0205] Input: A question entered by the user (e.g., "What are some recommended new movies?")

[0206] Output: Question text data

[0207] Step 2:

[0208] The terminal forwards the question to the server.

[0209] The application installed on the device transfers the received question text data to the server.

[0210] Input: Question text data

[0211] Output: Question data transferred to the server

[0212] Step 3:

[0213] The server sends questions to each generative artificial intelligence system.

[0214] The server converts the received question into the appropriate format and sends the question to multiple generative artificial intelligence (AI) models. Specifically, it makes asynchronous requests to the API endpoints of each AI service.

[0215] Input: Question data

[0216] Output: Question requests sent to multiple AI models

[0217] Step 4:

[0218] Receive an answer from a generative artificial intelligence.

[0219] The server receives responses from each generative artificial intelligence system and stores each response in a list.

[0220] Input: Request for response from AI model

[0221] Output: List of answers (Example: ["Movie A", "Movie B", "Best Movie C"])

[0222] Step 5:

[0223] The server selects the best answer.

[0224] The server evaluates multiple received responses and selects the most appropriate response based on predefined evaluation criteria. These criteria include accuracy, reliability, and grammatical correctness. A scoring function is used to calculate evaluation scores, and the response with the highest score is selected.

[0225] Input: Answer list

[0226] Output: Selected best answer (e.g., "Best movie C")

[0227] Step 6:

[0228] Correct the selected answer.

[0229] The server performs grammatical checks and fact-checks on the selected answer and makes corrections as needed. For example, it uses an online grammar checking service to check the grammatical structure and consults relevant databases for fact-checking.

[0230] Input: Selected answer

[0231] Output: Corrected final answer (e.g., "The best movie recommendation is the latest movie C")

[0232] Step 7:

[0233] The server retrieves the relevant content.

[0234] Based on the finalized response, the server retrieves relevant content from the web. Specifically, it uses web scraping techniques to collect links to watch related movies and reviews.

[0235] Input: Corrected final answer

[0236] Output: Related content (e.g., viewing links, reviews)

[0237] Step 8:

[0238] The device displays the final answer and related content.

[0239] The terminal receives the corrected final answer and related content sent from the server and displays it to the user.

[0240] Input: Final answer and related content

[0241] Output: The final answer and related content displayed to the user (e.g., "The best movie recommendation is the latest movie C. Here is the viewing link and review.")

[0242] As a result, users can quickly obtain accurate and reliable answers, along with relevant and useful content.

[0243] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0244] This invention provides a system that uses multiple generative artificial intelligence systems to provide the most appropriate answer to a user's question, and improves the quality of the answer by combining it with an emotion engine that recognizes the user's emotions. Specific examples of this system are described below.

[0245] Recognizing user emotions

[0246] As the device receives a question from the user via an input form, the emotion engine recognizes the user's emotions. For example, if a user types "When are the next Olympics?", the emotion engine analyzes the user's tone of voice and facial expressions to determine whether the user is feeling curiosity, anticipation, or excitement. This data is analyzed in real time in response to text input, and as a result, an emotion status is generated.

[0247] Receive a question and add sentiment data.

[0248] The device sends the sentiment status generated by the sentiment engine to the server along with the user's question. This status is processed together with the question.

[0249] Send a question to multiple AIs

[0250] The server converts the question and associated sentiment data into an appropriate format and sends it to multiple generative artificial intelligence (AI) services. Asynchronous requests are made to the API endpoint of each AI service, and each AI uses its own algorithm to generate an answer to the question.

[0251] Receive and evaluate the responses from each AI.

[0252] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, responses such as "2024" from "AI Service 1," "next year" from "AI Service 2," and "July 2024" from "AI Service 3" might be obtained. During this process, the emotional status is also considered and influences the evaluation criteria.

[0253] Select the most appropriate answer.

[0254] The server evaluates multiple responses received and selects the most appropriate one. The evaluation also takes into account the user's emotional status; for example, if the user is excited, they are more likely to choose a more detailed and future-oriented response (such as "July 2024").

[0255] Automatically correct the selected answer.

[0256] The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the server performs a grammar check on this answer and corrects it as needed. Furthermore, it adds detailed information such as "from July 26, 2024" by referring to reliable sources and adjusts the tone of the sentence to match the sentiment status.

[0257] Generate related ads based on the final answer

[0258] The server retrieves relevant ads based on the corrected final response. It uses the ad delivery service's API to search for relevant ads and provide them along with the response. For example, a user with an excited emotional status might see ads for sports equipment or event tickets.

[0259] Display the final answer and related ads.

[0260] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0261] Specific example

[0262] If a user asks, "When is the next Olympics?":

[0263] 1. The device receives the question, and the emotion engine analyzes the user's emotions. It recognizes that the user is agitated.

[0264] 2. The device sends the question and sentiment status to the server.

[0265] 3. The server sends questions to multiple generative artificial intelligence systems.

[0266] 4. Each AI will respond with answers such as "2024," "next year," or "July 2024." Emotional status will also be included in the evaluation.

[0267] 5. The server selects the most appropriate answer, "July 2024". Additional date details are provided based on the sentiment status.

[0268] 6. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[0269] 7. The server retrieves relevant ads based on the sentiment status.

[0270] 8. The device displays the final answer, "The next Olympics are from July 26, 2024," along with related advertisements. For example, an excited user might see an advertisement for sports equipment.

[0271] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[0272] The following describes the processing flow.

[0273] Step 1:

[0274] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[0275] Step 2:

[0276] The device recognizes the user's emotions. The emotion engine analyzes the user's voice tone, facial expressions, keyboard typing style, etc., to generate an emotional status. For example, it recognizes when the user is asking a question with anticipation or is excited. This emotional information is also temporarily stored.

[0277] Step 3:

[0278] The device sends the question and sentiment status to the server. Sentiment information is securely transmitted to the server along with the question content using network communication.

[0279] Step 4:

[0280] The server receives the question and sentiment status, and sends the question to multiple generative artificial intelligence (AI) services. It converts the received question into the appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it sends the question to "AI Service 1," "AI Service 2," and "AI Service 3."

[0281] Step 5:

[0282] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3" might be obtained.

[0283] Step 6:

[0284] The server evaluates the responses and selects the most appropriate one. It applies a scoring function to evaluate the multiple received responses. The evaluation criteria include fact-checking, grammatical accuracy, reliability, and also the sentiment status. For example, if the user is excited, a detailed and future-oriented response (such as "July 2024") will receive a high score.

[0285] Step 7:

[0286] Automatically correct the selected response. The server performs a grammar check and fact-check on the selected response. For example, if "July 2024" is selected, a grammar check is performed on this response and corrected if necessary. Furthermore, refer to reliable information sources and add detailed information such as "starting from July 26, 2024". The text is also corrected in terms of tone and expression according to the sentiment status.

[0287] Step 8:

[0288] Generate the final response and relevant advertisements. The server obtains relevant advertisements based on the corrected final response. Using the API of the advertising delivery service, relevant advertisements are searched and provided together with the final answer. For example, for a user with an excited sentiment status, advertisements such as sports supplies and event tickets are selected.

[0289] Step 9:

[0290] The server sends the final response and relevant advertisements to the terminal. The corrected response and related advertisements are bundled and transferred to the terminal.

[0291] Step 10:

[0292] The terminal displays the final response and relevant advertisements to the user. The terminal displays the received information on the screen. For example, the response "The next Olympics will start from July 26, 2024" is displayed, and at the same time, an advertisement for sports supplies related to it is displayed.

[0293] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[0294] (Example 2)

[0295] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0296] Conventional question-answering systems have the problem of not providing answers that take user emotions into consideration, making it difficult for users to obtain satisfactory answers. Furthermore, when selecting the most appropriate answer from answers obtained from multiple generative AI systems, the evaluation does not include emotional data, which can lead to a decrease in the quality of the answers. In addition, when providing related information or advertisements, there is a problem in that the content displayed is not appropriate based on the user's emotional state.

[0297] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0298] In this invention, the server includes means for recognizing the user's emotions, means for attaching emotion data to a question and transmitting it, means for evaluating multiple received answers and selecting the most appropriate answer, and means for performing grammatical checks and fact-checking on the selected answer. This makes it possible to provide the most appropriate answer obtained from multiple generative artificial intelligence systems while taking the user's emotions into consideration.

[0299] A "user" is someone who uses the system to input questions and receive answers.

[0300] A "question" refers to the information or points of concern that a user enters into the system.

[0301] "Emotions" refer to the user's psychological state and include curiosity, anticipation, excitement, and so on.

[0302] "Emotional data" refers to information generated by analyzing the emotions of users and is added to questions.

[0303] "Generative artificial intelligence" refers to algorithms and systems for natural language generation like humans.

[0304] "Grammar check" is a process of verifying whether a text uses correct grammar.

[0305] "Fact-checking" is a process of verifying whether information is accurate and reliable.

[0306] "Evaluation criteria" are criteria for evaluating multiple answers or information, taking into account the emotional data of users.

[0307] "Evaluation score" is a quantification of how appropriate each answer is based on the evaluation criteria.

[0308] "Advertisement" provides information on products and services related to the final answer.

[0309] "Asynchronous request" is a method of sending requests to multiple AI services simultaneously and processing them independently.

[0310] The present invention is a system that uses multiple generative artificial intelligence for questions from users and provides the most appropriate answer. This system includes functions of recognizing the emotions of users, generating answers while considering the emotional data, and providing related advertisements. Hereinafter, specific embodiments for implementing this system will be described.

[0311] System configuration:

[0312] This system is composed of a terminal, a server, and multiple generative artificial intelligence (AI) services. The main hardware and software used are as follows:

[0313] 1. Device (PC or smartphone): This is the device used by the user to input questions and view the final answers. It is desirable that it be equipped with a camera and microphone for analyzing voice tone and facial expressions.

[0314] 2. Server: This is the central computing unit that performs data processing, AI integration, question and answer evaluation, grammar checking, fact-checking, and relevant advertisement retrieval.

[0315] 3. Emotion Engine: This is a software tool for analyzing user emotions. Specifically, it uses speech recognition software or facial recognition tools (e.g., Google Speech-to-Text or Microsoft® Azure® Face API).

[0316] 4. Generative artificial intelligence services: Algorithms for generating answers to questions. Specifically, this includes OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[0317] Operation details:

[0318] The system operates as follows:

[0319] 1. The user enters the question:

[0320] Enter the question into the input form on the terminal. For example, a question like, "When is the next Olympics?"

[0321] 2. The device analyzes emotions:

[0322] The system analyzes the user's voice tone and facial expressions using an emotion engine, and the results are obtained as emotion data. For example, if the user is asked "When is the next Olympics?", the system will analyze that the user is excited.

[0323] 3. The device sends the question and sentiment data to the server:

[0324] The device sends the analysis results (question and emotion status) to the server as a data packet. Specifically, data such as "Question: When are the next Olympics?" and "Emotion Status: Excited" is sent in JSON format.

[0325] 4. The server sends questions to multiple generative AI systems:

[0326] The server converts the received question and sentiment data into an appropriate format and sends asynchronous requests to multiple generative AI services. For example, it sends the question "When are the next Olympics?" to OpenAI's GPT-3, Google's BERT, Microsoft's Turing, and others.

[0327] 5. Evaluate the answers obtained from each AI:

[0328] The server collects the responses returned by each generative AI and stores them in a list. Each response is something like "2024," "next year," or "July 2024." During this process, sentiment data is taken into consideration to select the most appropriate response.

[0329] 6. Grammatical check and fact-checking of selected answers:

[0330] The server grammatically checks the selected answer and fact-checks it by referring to reliable sources. For example, if "July 2024" is selected, it will refer to the official Olympic website and add the specific date "July 26, 2024".

[0331] 7. Obtaining related advertisements:

[0332] The server retrieves relevant ads based on the user's emotions. For example, excited users are shown ads for sports equipment or event tickets.

[0333] 8. Display of the final answer and related ads:

[0334] The corrected final answer is sent from the server to the device, which then displays it to the user. Relevant advertisements are also displayed along with the answer.

[0335] Specific example:

[0336] When a user asks "When is the next Olympics?", the specific actions taken are as follows:

[0337] 1. The user enters "When is the next Olympics?" into their device.

[0338] 2. The device uses an emotion engine to analyze the user's emotions (e.g., excited).

[0339] 3. The device sends the question and sentiment data to the server.

[0340] 4. The server sends questions to multiple generative AI systems (e.g., OpenAI's GPT-3, Google's BERT, Microsoft's Turing).

[0341] 5. Collect responses from each AI and select the most appropriate response (e.g., "July 2024").

[0342] 6. Check the grammar and facts of the answer (e.g., add any supplementary information to "July 26, 2024").

[0343] 7. Obtain relevant ads based on emotions (e.g., ads for sporting goods).

[0344] 8. The device displays an advertisement with the final answer: "The next Olympics will be held from July 26, 2024."

[0345] Example of a prompt:

[0346] "When is the next Olympics? A question from an excited user."

[0347] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0348] Step 1:

[0349] The user enters a question.

[0350] Operation: The user enters a question through the input form on their device. For example, they might enter, "When are the next Olympics?"

[0351] Input: The user enters a question into the input form.

[0352] Output: The terminal retrieves the user's question data.

[0353] Step 2:

[0354] The device analyzes the question, voice tone, and facial expressions.

[0355] Operation: The device uses an emotion engine to analyze the user's voice tone and facial expressions in real time. Specifically, it captures voice with a microphone and analyzes facial expressions with a camera. Software used includes speech recognition tools (e.g., Google Speech-to-Text) and face recognition APIs (e.g., Microsoft Azure Face API).

[0356] Input: User's questions and data on voice tone and facial expressions.

[0357] Output: Emotional status analyzed by the emotion engine (e.g., excitement, anticipation, etc.).

[0358] Step 3:

[0359] The device sends the question and sentiment status to the server.

[0360] Operation: The terminal sends the question and the obtained sentiment status to the server as data packets. Possible transmission formats include JSON.

[0361] Input: User's question and analyzed sentiment status.

[0362] Output: Question and sentiment data sent to the server in JSON format, etc.

[0363] Step 4:

[0364] The server sends questions to multiple generative AIs.

[0365] Operation: The server sends received question and sentiment data as asynchronous requests to multiple generative artificial intelligence systems. Recipients include OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[0366] Input: User question and sentiment data received by the server.

[0367] Output: Question and sentiment data sent to a generative AI service.

[0368] Step 5:

[0369] Receive and evaluate responses from each generative AI.

[0370] Operation: The server receives responses from each generative AI and stores each response in a list. It sets evaluation criteria considering the emotional status and evaluates the best response.

[0371] Input: Responses and emotional status from each generative AI.

[0372] Output: Store the most appropriate answer in a list and calculate the evaluation score.

[0373] Step 6:

[0374] Select the most appropriate answer and perform grammatical checks and fact-checks.

[0375] Operation: The server selects the most appropriate answer based on the evaluation score and checks it using a grammar checking tool (e.g., Grammarly API) and fact-checking sources (e.g., the official Olympic website).

[0376] Input: The answer with the highest rating score.

[0377] Output: Corrected answer after grammar check and fact-checking (e.g., "from July 26, 2024").

[0378] Step 7:

[0379] Get relevant ads based on your emotional status

[0380] Operation: The server considers the user's emotional state and retrieves relevant ads from the ad delivery system (e.g., Google AdSense API).

[0381] Input: Emotional status and optimal response.

[0382] Output: Relevant ads selected based on emotional status.

[0383] Step 8:

[0384] Send the final answer and related advertisements to your device and display them.

[0385] Operation: The corrected final answer and related advertisements are sent to the device, which then displays them to the user. For example, an advertisement for sports equipment might be displayed along with the answer, "The next Olympics are from July 26, 2024."

[0386] Input: Corrected final answer and related ads.

[0387] Output: The final answer and related advertisements displayed on the user's device.

[0388] (Application Example 2)

[0389] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0390] Conventional generative artificial intelligence systems often lacked sufficient accuracy and quality in their responses to user questions. A particular challenge was the decrease in user satisfaction due to responses generated without considering the user's emotions. Furthermore, the process of appropriately evaluating responses from multiple generative AI systems and selecting the optimal answer was complex, and efficiency improvements were needed. Additionally, there was the difficulty in effectively presenting relevant products that responded to the user's emotions when suggesting products in virtual stores.

[0391] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0392] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related advertisements, means for recognizing the user's emotions, means for transmitting questions containing emotion data to the generative artificial intelligence systems, and means for selecting the optimal answer based on the emotion data. This enables the generation of optimal answers that take the user's emotions into consideration and the effective suggestion of related products based on those answers.

[0393] "Means for receiving user questions" refers to a device or program that has the function of receiving question information entered by a user.

[0394] "Means for sending questions to multiple generative artificial intelligence systems" refers to a device or program that has the function of sending a user's questions to multiple generative AI services.

[0395] "Means for receiving responses from each generative artificial intelligence" refers to a device or program that has the function of receiving response information returned from a generative AI service.

[0396] "A means of evaluating multiple received responses and selecting the most appropriate response" refers to a device or program that has the function of analyzing and evaluating responses from multiple generative AI systems and selecting the most appropriate one.

[0397] "Means for automatically correcting selected answers" refers to a device or program that has the function of automatically performing grammatical checks and factual verification on selected answers and correcting them as necessary.

[0398] "Means for displaying corrected final answers and related advertisements" refers to a device or program that has the function of presenting corrected answers and related advertisements to a user.

[0399] "Means for recognizing user emotions" refers to a device or program that analyzes a user's voice and facial expressions to determine their emotions.

[0400] "Means for sending questions containing emotional data to a generative artificial intelligence system" refers to a device or program that has the function of adding emotional data to a user's question and sending it to a generative AI service.

[0401] "Means for selecting the optimal response based on emotional data" refers to a device or program that has the function of evaluating responses from multiple generative AI systems, taking emotional data into consideration, and selecting the best one.

[0402] "Means for performing grammatical checks" refers to a device or program that has the function of verifying and correcting the grammatical consistency of a text.

[0403] "Means of fact-checking" refers to a device or program that has the function of verifying the content of a received response based on reliable sources.

[0404] "Means for setting evaluation criteria for selecting the optimal answer" refers to a device or program that has the function of setting criteria (e.g., accuracy, relevance, etc.) for selecting the best answer from among multiple answers.

[0405] "Means for calculating the evaluation score for each response" refers to a device or program that has the function of assigning a score to each response based on established evaluation criteria.

[0406] "Means for selecting the highest-rated response based on scores" refers to a device or program that has the function of selecting the highest-rated response based on evaluation scores.

[0407] "Means for setting evaluation criteria that take emotional data into consideration" refers to a device or program that has the function of setting criteria for evaluating responses while taking into consideration the user's emotional data.

[0408] This invention is a system that improves the quality of responses by using multiple generative artificial intelligence systems to provide the most appropriate answers to user questions, and by combining this with an emotion engine that recognizes the user's emotions. This system mainly consists of a server and terminals.

[0409] Hardware and software configuration

[0410] Hardware to use

[0411] Smart glasses, head-mounted displays, or smartphones

[0412] Software to use

[0413] Emotion recognition libraries (e.g., Microsoft Azure Emotion Recognition API)

[0414] Generative AI libraries (e.g., OpenAI, Google AI)

[0415] Programming language: Python

[0416] System operation

[0417] User question received

[0418] Users ask questions about products within the virtual store using a device (smart glasses, head-mounted display, or smartphone). The device receives the question via an input form, and an emotion engine analyzes the user's voice tone and facial expressions to generate an emotion status. This emotion status is then sent to the server.

[0419] Sending questions and sentiment data

[0420] The server receives the question and sentiment status, converts them into the appropriate format, and sends them to multiple generative artificial intelligences. Asynchronous requests are made to the API endpoint of each generative AI, and each AI generates an answer to the question using its own algorithm.

[0421] Receiving and evaluating responses

[0422] The server receives responses from each generative AI system and stores them in a list. The server evaluates each response, taking into account the user's emotional status, and selects the most appropriate response. The evaluation also takes emotional status into account; if the user is excited, a more detailed and future-oriented response is more likely to be selected.

[0423] Correcting responses and generating ads

[0424] The server performs grammar checks and fact-checks on the selected answers. It then references reliable sources to add further details. Additionally, it utilizes ad delivery service APIs to retrieve relevant ads based on the sentiment status.

[0425] Final answer and ad display

[0426] The corrected final answers and relevant advertisements are sent to the device and displayed to the user. This provides the user with accurate, reliable answers and relevant advertisements that are tailored to their emotions.

[0427] Specific example

[0428] If a user asks "Is this jacket waterproof?" in a virtual store, the following process will occur.

[0429] 1. Use the microphone and camera on the smart glasses to capture the user's questions and facial expressions.

[0430] 2. The emotion engine analyzes the user's voice tone and facial expressions and generates an emotion status of "excited".

[0431] 3. The question and emotion status are sent to the server, and requests are sent to multiple generative AI systems.

[0432] 4. Each generation AI generates a response such as "Yes, it's waterproof," "Yes, it's waterproof tested," or "Yes, it's waterproof and windproof."

[0433] 5. The server, considering its "excited" emotional status, selects the detailed answer "Yes, it is waterproof and windproof."

[0434] 6. As related advertisements, suggest items such as waterproof jackets and waterproof sprays.

[0435] 7. Display the final answer and related advertisements on the smart glasses' display.

[0436] Example input prompts for a generative AI model

[0437] User question: "Is this jacket waterproof?"

[0438] User's emotion: "Excited"

[0439] Please generate variations of the answer.

[0440] This invention makes it possible to generate optimal responses that take user emotions into consideration, and to effectively suggest related products based on those responses.

[0441] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0442] Step 1:

[0443] The user enters a question about the product. A device (smart glasses, head-mounted display, or smartphone) receives this question, and simultaneously, an emotion engine analyzes the user's voice tone and facial expressions. The device generates analyzed emotion data and sends it to the server.

[0444] Input: User's question, user's voice tone, facial expression

[0445] Data processing: The emotion engine analyzes voice tone and facial expressions to generate emotion data.

[0446] Output: Questionnaire data, sentiment data

[0447] Step 2:

[0448] The server receives user questions and sentiment data. It converts the received data into an appropriate format (e.g., JSON) and sends asynchronous requests to the API endpoints of each generative artificial intelligence system.

[0449] Input: Question data, sentiment data

[0450] Data processing: Convert question data and sentiment data to JSON format.

[0451] Output: Request to generative AI service

[0452] Step 3:

[0453] Each generative artificial intelligence generates an answer to a question and returns the answer data to the server. The server receives this answer data.

[0454] Input: Response from a generative AI service

[0455] Data processing: Each generative AI generates answers using its own unique algorithm.

[0456] Output: List of response data

[0457] Step 4:

[0458] The server evaluates multiple response data received. Considering user sentiment data, it calculates a score for each response based on evaluation criteria to select the best answer. The response with the highest score is then selected.

[0459] Input: List of response data, sentiment data

[0460] Data processing: Assign a score to each response based on evaluation criteria.

[0461] Output: Optimal response data

[0462] Step 5:

[0463] The server performs grammar checks and fact-checks on the selected optimal response data. Corrections are made as needed, and the final response is generated.

[0464] Input: Optimal response data

[0465] Data processing: grammar check, fact-checking, correction.

[0466] Output: Corrected final response data

[0467] Step 6:

[0468] The server retrieves relevant advertisements based on the final response data. Using the ad delivery service's API, it searches for advertisements that match the user's sentiment data and selects the most suitable advertisement.

[0469] Input: Final response data, sentiment data

[0470] Data processing: Obtain advertising data from advertising distribution services.

[0471] Output: Related ad data

[0472] Step 7:

[0473] The corrected final response data and related advertising data are sent from the server to the device. The device then displays them to the user, specifically on the display of smart glasses or a head-mounted display.

[0474] Input: Final response data, related ad data

[0475] Data processing: Conversion to display format

[0476] Output: User-viewable responses and advertisements

[0477] This process allows users to quickly receive detailed and reliable answers that address their emotions, and relevant advertisements are presented effectively.

[0478] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0479] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0480] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0481] [Second Embodiment]

[0482] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0483] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0484] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0485] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0486] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0487] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0488] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0489] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0490] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0491] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0492] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0493] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0494] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[0495] Receive user questions

[0496] The terminal accepts questions from the user via an input form. These questions are entered by the user, for example, through a web browser or mobile application. Let's say the user enters, "When are the next Olympics?" This question is transmitted to the server in real time.

[0497] Send a question to multiple AIs

[0498] The server converts the received question into the appropriate format and sends it to multiple generative artificial intelligence (AI) systems. For example, it makes asynchronous requests to the API endpoints of each AI service. This causes each AI to generate an answer to the question using its own algorithm.

[0499] Receive responses from each AI

[0500] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[0501] Select the most appropriate answer.

[0502] The server evaluates multiple received responses and selects the most appropriate one based on criteria such as fact-checking, grammatical accuracy, and reliability. A scoring function is used for evaluation; for example, if "July 2024" receives the highest score, it will be selected.

[0503] Automatically correct the selected answer.

[0504] The server performs grammatical checks and fact-checks on the selected answers. For example, it checks the answer "July 2024" and makes any necessary corrections. Furthermore, it adds more detailed information, such as "from July 26, 2024," using reliable sources.

[0505] Display the final answer and related ads.

[0506] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0507] Specific example

[0508] If a user asks, "When is the next Olympics?":

[0509] 1. The terminal receives the question and forwards it to the server.

[0510] 2. The server sends questions to multiple generative artificial intelligence systems.

[0511] 3. Each AI will respond with answers such as "2024," "next year," or "July 2024."

[0512] 4. The server selects the most appropriate answer, "July 2024".

[0513] 5. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[0514] 6. The device displays the user the final answer, "The next Olympics will be held from July 26, 2024," along with related advertisements.

[0515] This system allows users to quickly obtain accurate and reliable answers.

[0516] The following describes the processing flow.

[0517] Step 1:

[0518] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[0519] Step 2:

[0520] The terminal sends the question to the server. The entered question is transferred to the server in real time. During this process, the question is transmitted securely using network communication.

[0521] Step 3:

[0522] The server sends a question to multiple generative artificial intelligence (AI) services. Upon receiving the question, the server converts its content into an appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it might send a question to "AI Service 1," "AI Service 2," and "AI Service 3."

[0523] Step 4:

[0524] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, it might receive responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3".

[0525] Step 5:

[0526] The server evaluates the responses and selects the most appropriate one. A scoring function is applied to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. The response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[0527] Step 6:

[0528] The selected answer will be automatically corrected. The server will perform grammatical checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as necessary. Furthermore, it will add detailed information such as "from July 26, 2024" by referring to reliable sources.

[0529] Step 7:

[0530] The system generates the final answer and related ads. The server retrieves the relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the answer.

[0531] Step 8:

[0532] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[0533] Step 9:

[0534] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0535] In this way, users can quickly obtain accurate and reliable answers simply by entering their questions.

[0536] (Example 1)

[0537] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0538] Conventional information retrieval systems often provide inaccurate information in response to user inquiries. Even with the use of numerous generative artificial intelligence systems, responses may vary or contain unreliable information. In such situations, it is difficult for users to obtain accurate information quickly. Furthermore, users bear the burden of having to judge the reliability of the information they receive themselves.

[0539] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0540] In this invention, the server includes means for receiving questions from users, means for forwarding received questions to the server, means for sending questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the multiple received answers and selecting the most appropriate answer, means for automatically correcting the selected answer, and means for displaying the corrected final answer and related advertisements. This enables users to quickly obtain accurate and reliable answers.

[0541] A "user" refers to an individual or group that inputs questions into a system and receives answers.

[0542] A "terminal" is a device used by a user to input questions and send them to a server, and specifically includes personal computers and smartphones.

[0543] A "server" refers to a computer system that receives questions from users, sends the questions to multiple generative artificial intelligence systems, evaluates and corrects the received answers, and then sends the final answer to the terminal.

[0544] "Generative artificial intelligence" refers to systems that utilize machine learning and natural language processing technologies to automatically generate answers to user questions.

[0545] "Means of receiving questions" refers to a combination of software and hardware for inputting, storing, and transferring user questions to a server in a digital format.

[0546] "Means for sending questions" refers to communication protocols or API interfaces that allow a server to send questions received from a user to multiple generative artificial intelligence systems.

[0547] "Means of receiving responses" refers to communication protocols and databases used by servers to receive and store responses from generative artificial intelligence.

[0548] "Means of evaluating responses" refers to algorithms or evaluation systems that analyze multiple received responses and select the most appropriate response based on evaluation criteria.

[0549] "Means for correcting answers" refers to programs or libraries used to perform grammatical checks and factual verification on selected answers, and then make corrections or revisions.

[0550] "Means for displaying corrected final answers and related advertisements" refers to software and interfaces for displaying the final answers and related advertisements on the user's device.

[0551] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[0552] This system includes a terminal that receives user questions, a server that processes the received questions and transmits them to multiple generative artificial intelligence systems, and means for evaluating and correcting the answers from the generative artificial intelligence systems and providing the user with the final answers and related advertisements.

[0553] Hardware and Software Overview

[0554] terminal

[0555] A device is a device on which the user enters a question, and primarily includes personal computers, smartphones, and tablets. The device provides an interface for entering the question using a web browser or mobile application.

[0556] server

[0557] This is a central computer system for receiving user questions and transmitting them to multiple generative artificial intelligence systems. The server has the following main functions:

[0558] API interface: A means of communication for sending questions to and receiving answers from generative artificial intelligence.

[0559] Database: A storage device for storing and evaluating received responses.

[0560] Evaluation system: An algorithm that evaluates each answer and selects the most appropriate answer.

[0561] Generative artificial intelligence

[0562] This refers to a system that automatically generates answers to questions sent from a server, and includes Google Cloud AI, OpenAI, and IBM Watson.

[0563] System operation example

[0564] The user types "When is the next Olympics?" into their web browser. At this point, the terminal receives the user's question and forwards it to the server.

[0565] The server converts the received question into an appropriate format and sends an API request to a generative artificial intelligence (AI) such as Google Cloud AI, OpenAI, or IBM Watson. Multiple generative AIs answer the question based on their own algorithms and return the answer to the server.

[0566] The server receives responses from each generative artificial intelligence system and stores them in a database. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[0567] The server evaluates the multiple responses it receives. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. A scoring algorithm is used to assign points to each response, and the response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[0568] The server performs grammar checks and fact-checks on the selected answer. Grammar checks utilize language processing libraries, and fact-checking relies on reliable sources. As a result, detailed information such as "from July 26, 2024" is added.

[0569] The corrected final answer is sent to the device, which then displays it to the user. Related advertisements are also displayed simultaneously. For example, an advertisement for sporting goods might appear along with the answer, "The next Olympics are from July 26, 2024."

[0570] Example of a prompt

[0571] Question: "When is the next Olympics?"

[0572] Question: "What is the date of the next Olympic Games?"

[0573] Question: "I would like to know the dates of the Olympics."

[0574] As described above, this system allows users to quickly obtain accurate and reliable answers.

[0575] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0576] Step 1:

[0577] The user enters a question.

[0578] Input: The user enters the question via a web browser or mobile application.

[0579] Action: The user types "When are the next Olympics?" into the browser's search form.

[0580] Data processing: Format the input text data into JSON format.

[0581] Output: The formatted question data is temporarily stored on the device.

[0582] Step 2:

[0583] The terminal forwards the question to the server.

[0584] Input: Question data entered by the user (in JSON format).

[0585] Operation: The device uses the JavaScript fetch API to send the question data to the server as an HTTP request.

[0586] Data processing: Include the question data in the HTTP request body.

[0587] Output: Question data arrives at the server and the server receives it.

[0588] Step 3:

[0589] The server sends questions to multiple generative artificial intelligence systems.

[0590] Input: Question data received from the terminal.

[0591] Operation: The server processes the received question data into an API request format.

[0592] Data processing: Convert the question data into a format compatible with the API of each generative artificial intelligence system. For example, convert from JSON to URL-encoded format.

[0593] Output: The converted question data is sent as an HTTP request to each generative artificial intelligence system.

[0594] Step 4:

[0595] The server receives responses from each generative artificial intelligence system.

[0596] Input: Response data from each generative artificial intelligence.

[0597] Operation: The server asynchronously waits for and receives responses from each generative artificial intelligence system.

[0598] Data processing: Store the received response data in a list.

[0599] Output: A list containing the answers of each generative artificial intelligence.

[0600] Step 5:

[0601] The server evaluates the received responses and selects the most appropriate one.

[0602] Input: Multiple response data (lists) from various generative artificial intelligence systems.

[0603] Operation: The server uses an evaluation algorithm to score each response.

[0604] Data processing: A scoring function is applied to each response to calculate an evaluation score based on reliability and grammatical accuracy.

[0605] Output: Select the answer with the highest rating.

[0606] Step 6:

[0607] The server automatically corrects the selected answer.

[0608] Input: The most appropriate response data.

[0609] Operation: The server performs grammar checks and fact-checks on the selected answer. For grammar checks, it uses, for example, the Grammarly API, and for fact-checks, it refers to reliable sources.

[0610] Data processing: This involves correcting grammar and adding detailed information.

[0611] Output: Corrected final response data.

[0612] Step 7:

[0613] The device displays the final answer and related ads.

[0614] Input: Final response data and related advertising data sent from the server.

[0615] Function: Displays data received by the device on web pages and applications. Dynamically updates information using HTML and JavaScript.

[0616] Data processing: Convert the final response data and advertising data into a format that is compatible with the user interface.

[0617] Output: The user is shown the information "The next Olympic Games will be held from July 26, 2024" and advertisements for related sporting goods.

[0618] Through the steps outlined above, users can quickly obtain accurate and reliable answers.

[0619] (Application Example 1)

[0620] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0621] Conventional generative artificial intelligence systems have a problem in that they do not provide accurate and appropriate answers to user questions with sufficient precision. Furthermore, they lack the ability to appropriately recommend related information and content related to the generated answers, resulting in a situation where user needs and convenience are not fully met.

[0622] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0623] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related information, and means for providing related content based on the displayed information. This enables the user to obtain accurate and reliable answers and simultaneously acquire related content.

[0624] A "user" refers to each individual who uses this system.

[0625] "Means for receiving questions" refers to a device or software that receives input from a user and processes its contents.

[0626] "Generative artificial intelligence" refers to AI models that can generate answers to given questions or inputs.

[0627] "Means for sending questions" refers to a device or software that transmits questions received from a user to multiple generative artificial intelligence systems.

[0628] "Means for receiving responses" refers to a device or software for receiving responses generated by generative artificial intelligence.

[0629] "Means for evaluating responses" refers to a device or software for evaluating multiple received responses and determining which is the most appropriate.

[0630] "Means for selecting the most appropriate answer" refers to a device or software for selecting the optimal answer based on evaluation.

[0631] "Means for automatically correcting answers" refers to a device or software for automatically correcting selected answers based on grammar or facts.

[0632] "Means for displaying the final answer and related information" refers to a device or software for displaying the corrected final answer and related information to the user.

[0633] "Means for providing related content" refers to a device or software for providing additional content to the user based on the displayed information.

[0634] This invention is a system that processes user questions in real time, provides the most appropriate answers, and recommends relevant content. Specific embodiments of this system are described below.

[0635] The system primarily consists of a server, terminals, and users. Users submit questions via their terminals (such as smartphones and tablets). A dedicated application installed on the terminal receives questions from users and forwards them to the server. This application also provides an interface for users to input their questions.

[0636] The server sends questions received from users to multiple generative artificial intelligence (AI) models and receives responses from each AI model. In this process, the server uses OpenAI's GPT-3 API or other AI service technologies to send questions. For example, if a user asks, "What are some recommended new movies?", the server sends this question to the AI ​​models and receives responses from each.

[0637] The server evaluates the multiple responses received and selects the most appropriate one. This evaluation considers factors such as accuracy, reliability, and grammatical correctness. The server calculates a score for each response based on the set evaluation criteria and selects the response with the highest score. The server then automatically performs grammatical checks and fact-checks on the selected response and corrects it to its final form.

[0638] The final corrected answer and related information are sent to the device. The device displays related content to the user along with the final answer. This related content might include, for example, a link to watch the movie or a list of related movies in the case of a question about movie recommendations. The device uses web scraping technology to retrieve and provide the related content to the user.

[0639] Hardware and software to be used

[0640] Hardware: Servers, smartphones, tablets

[0641] Software: OpenAI GPT-3 API, requests, BeautifulSoup, dedicated application (for smartphones and tablets)

[0642] Specific example

[0643] The following shows the specific processing that occurs when a user asks, "What are some of the latest movies you recommend?"

[0644] 1. The user enters "What are some recommended new movies?" into their device.

[0645] 2. The terminal forwards this question to the server.

[0646] 3. The server sends questions to multiple generative AI models and receives responses.

[0647] 4. The server selects the best response from those received and performs grammatical checks and fact-checks.

[0648] 5. The server sends the corrected response and related content to the device.

[0649] 6. The device displays to the user a link to watch the relevant movie and reviews along with the final answer.

[0650] Examples of prompt statements

[0651] "What are some of the latest movies you recommend?"

[0652] "What are your recommended travel destinations this year?"

[0653] "When is the next Olympics?"

[0654] In this way, users can quickly obtain accurate and reliable answers, along with useful content related to those answers. This system can meet user needs and significantly improve convenience.

[0655] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0656] Step 1:

[0657] The user enters a question.

[0658] The user uses a device (such as a smartphone or tablet) to enter a question into the application's input form. For example, they might enter, "What are some recommended new movies?" This input is then accepted by the device.

[0659] Input: A question entered by the user (e.g., "What are some recommended new movies?")

[0660] Output: Question text data

[0661] Step 2:

[0662] The terminal forwards the question to the server.

[0663] The application installed on the device transfers the received question text data to the server.

[0664] Input: Question text data

[0665] Output: Question data transferred to the server

[0666] Step 3:

[0667] The server sends questions to each generative artificial intelligence system.

[0668] The server converts the received question into the appropriate format and sends the question to multiple generative artificial intelligence (AI) models. Specifically, it makes asynchronous requests to the API endpoints of each AI service.

[0669] Input: Question data

[0670] Output: Question requests sent to multiple AI models

[0671] Step 4:

[0672] Receive an answer from a generative artificial intelligence.

[0673] The server receives responses from each generative artificial intelligence system and stores each response in a list.

[0674] Input: Request for response from AI model

[0675] Output: List of answers (Example: ["Movie A", "Movie B", "Best Movie C"])

[0676] Step 5:

[0677] The server selects the best answer.

[0678] The server evaluates multiple received responses and selects the most appropriate response based on predefined evaluation criteria. These criteria include accuracy, reliability, and grammatical correctness. A scoring function is used to calculate evaluation scores, and the response with the highest score is selected.

[0679] Input: Answer list

[0680] Output: Selected best answer (e.g., "Best movie C")

[0681] Step 6:

[0682] Correct the selected answer.

[0683] The server performs grammatical checks and fact-checks on the selected answer and makes corrections as needed. For example, it uses an online grammar checking service to check the grammatical structure and consults relevant databases for fact-checking.

[0684] Input: Selected answer

[0685] Output: Corrected final answer (e.g., "The best movie recommendation is the latest movie C")

[0686] Step 7:

[0687] The server retrieves the relevant content.

[0688] Based on the finalized response, the server retrieves relevant content from the web. Specifically, it uses web scraping techniques to collect links to watch related movies and reviews.

[0689] Input: Corrected final answer

[0690] Output: Related content (e.g., viewing links, reviews)

[0691] Step 8:

[0692] The device displays the final answer and related content.

[0693] The terminal receives the corrected final answer and related content sent from the server and displays it to the user.

[0694] Input: Final answer and related content

[0695] Output: The final answer and related content displayed to the user (e.g., "The best movie recommendation is the latest movie C. Here is the viewing link and review.")

[0696] As a result, users can quickly obtain accurate and reliable answers, along with relevant and useful content.

[0697] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0698] This invention provides a system that uses multiple generative artificial intelligence systems to provide the most appropriate answer to a user's question, and improves the quality of the answer by combining it with an emotion engine that recognizes the user's emotions. Specific examples of this system are described below.

[0699] Recognizing user emotions

[0700] As the device receives a question from the user via an input form, the emotion engine recognizes the user's emotions. For example, if a user types "When are the next Olympics?", the emotion engine analyzes the user's tone of voice and facial expressions to determine whether the user is feeling curiosity, anticipation, or excitement. This data is analyzed in real time in response to text input, and as a result, an emotion status is generated.

[0701] Receive a question and add sentiment data.

[0702] The device sends the sentiment status generated by the sentiment engine to the server along with the user's question. This status is processed together with the question.

[0703] Send a question to multiple AIs

[0704] The server converts the question and associated sentiment data into an appropriate format and sends it to multiple generative artificial intelligence (AI) services. Asynchronous requests are made to the API endpoint of each AI service, and each AI uses its own algorithm to generate an answer to the question.

[0705] Receive and evaluate the responses from each AI.

[0706] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, responses such as "2024" from "AI Service 1," "next year" from "AI Service 2," and "July 2024" from "AI Service 3" might be obtained. During this process, the emotional status is also considered and influences the evaluation criteria.

[0707] Select the most appropriate answer.

[0708] The server evaluates multiple responses received and selects the most appropriate one. The evaluation also takes into account the user's emotional status; for example, if the user is excited, they are more likely to choose a more detailed and future-oriented response (such as "July 2024").

[0709] Automatically correct the selected answer.

[0710] The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the server performs a grammar check on this answer and corrects it as needed. Furthermore, it adds detailed information such as "from July 26, 2024" by referring to reliable sources and adjusts the tone of the sentence to match the sentiment status.

[0711] Generate related ads based on the final answer

[0712] The server retrieves relevant ads based on the corrected final response. It uses the ad delivery service's API to search for relevant ads and provide them along with the response. For example, a user with an excited emotional status might see ads for sports equipment or event tickets.

[0713] Display the final answer and related ads.

[0714] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0715] Specific example

[0716] If a user asks, "When is the next Olympics?":

[0717] 1. The device receives the question, and the emotion engine analyzes the user's emotions. It recognizes that the user is agitated.

[0718] 2. The device sends the question and sentiment status to the server.

[0719] 3. The server sends questions to multiple generative artificial intelligence systems.

[0720] 4. Each AI will respond with answers such as "2024," "next year," or "July 2024." Emotional status will also be included in the evaluation.

[0721] 5. The server selects the most appropriate answer, "July 2024". Additional date details are provided based on the sentiment status.

[0722] 6. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[0723] 7. The server retrieves relevant ads based on the sentiment status.

[0724] 8. The device displays the final answer, "The next Olympics are from July 26, 2024," along with related advertisements. For example, an excited user might see an advertisement for sports equipment.

[0725] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[0726] The following describes the processing flow.

[0727] Step 1:

[0728] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[0729] Step 2:

[0730] The device recognizes the user's emotions. The emotion engine analyzes the user's voice tone, facial expressions, keyboard typing style, etc., to generate an emotional status. For example, it recognizes when the user is asking a question with anticipation or is excited. This emotional information is also temporarily stored.

[0731] Step 3:

[0732] The device sends the question and sentiment status to the server. Sentiment information is securely transmitted to the server along with the question content using network communication.

[0733] Step 4:

[0734] The server receives the question and sentiment status, and sends the question to multiple generative artificial intelligence (AI) services. It converts the received question into the appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it sends the question to "AI Service 1," "AI Service 2," and "AI Service 3."

[0735] Step 5:

[0736] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3" might be obtained.

[0737] Step 6:

[0738] The server evaluates the responses and selects the most appropriate one. It applies a scoring function to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, reliability, and even sentiment status. For example, if the user is excited, detailed, future-oriented responses (such as "July 2024") will receive higher scores.

[0739] Step 7:

[0740] The selected answer is automatically corrected. The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as needed. Furthermore, it will refer to reliable sources and add more detailed information such as "from July 26, 2024". The text may also be corrected to match the tone and expression of the sentiment status.

[0741] Step 8:

[0742] The system generates the final answer and relevant ads. The server retrieves relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the final answer. For example, users with an excited emotion status might see ads for sports goods or event tickets.

[0743] Step 9:

[0744] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[0745] Step 10:

[0746] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0747] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[0748] (Example 2)

[0749] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0750] Conventional question-answering systems have the problem of not providing answers that take user emotions into consideration, making it difficult for users to obtain satisfactory answers. Furthermore, when selecting the most appropriate answer from answers obtained from multiple generative AI systems, the evaluation does not include emotional data, which can lead to a decrease in the quality of the answers. In addition, when providing related information or advertisements, there is a problem in that the content displayed is not appropriate based on the user's emotional state.

[0751] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0752] In this invention, the server includes means for recognizing the user's emotions, means for attaching emotion data to a question and transmitting it, means for evaluating multiple received answers and selecting the most appropriate answer, and means for performing grammatical checks and fact-checking on the selected answer. This makes it possible to provide the most appropriate answer obtained from multiple generative artificial intelligence systems while taking the user's emotions into consideration.

[0753] A "user" is someone who uses the system to input questions and receive answers.

[0754] A "question" refers to the information or points of concern that a user enters into the system.

[0755] "Emotions" refer to the user's psychological state and include curiosity, anticipation, excitement, and so on.

[0756] "Emotional data" refers to information generated by analyzing the user's emotions and is added to the questions.

[0757] "Generative artificial intelligence" refers to algorithms and systems that generate natural language in a way that mimics human speech.

[0758] "Grammar check" is the process of verifying whether a text uses correct grammar.

[0759] "Fact-checking" is the process of verifying whether information is accurate and reliable.

[0760] "Evaluation criteria" are standards for evaluating multiple responses and pieces of information, and they take into account user sentiment data.

[0761] An "evaluation score" is a numerical representation of how appropriate each response is based on the evaluation criteria.

[0762] "Advertisements" are those that provide information about products or services related to the final answer.

[0763] An "asynchronous request" is a method of sending requests to multiple AI services simultaneously and processing them independently.

[0764] This invention is a system that uses multiple generative artificial intelligence systems to provide the most appropriate answers to user questions. This system includes functions to recognize the user's emotions, generate answers while considering that emotion data, and also provide relevant advertisements. The specific implementation of this system is described below.

[0765] System configuration:

[0766] This system consists of terminals, servers, and multiple generative artificial intelligence (AI) services. The main hardware and software used are as follows:

[0767] 1. Device (PC or smartphone): This is the device used by the user to input questions and view the final answers. It is desirable that it be equipped with a camera and microphone for analyzing voice tone and facial expressions.

[0768] 2. Server: This is the central computing unit that performs data processing, AI integration, question and answer evaluation, grammar checking, fact-checking, and relevant advertisement retrieval.

[0769] 3. Emotion Engine: This is a software tool for analyzing user emotions. Specifically, it uses speech recognition software or facial recognition tools (e.g., Google Speech-to-Text or Microsoft Azure Face API).

[0770] 4. Generative artificial intelligence services: Algorithms for generating answers to questions. Specifically, this includes OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[0771] Operation details:

[0772] The system operates as follows:

[0773] 1. The user enters the question:

[0774] Enter the question into the input form on the terminal. For example, a question like, "When is the next Olympics?"

[0775] 2. The device analyzes emotions:

[0776] The system analyzes the user's voice tone and facial expressions using an emotion engine, and the results are obtained as emotion data. For example, if the user is asked "When is the next Olympics?", the system will analyze that the user is excited.

[0777] 3. The device sends the question and sentiment data to the server:

[0778] The device sends the analysis results (question and emotion status) to the server as a data packet. Specifically, data such as "Question: When are the next Olympics?" and "Emotion Status: Excited" is sent in JSON format.

[0779] 4. The server sends questions to multiple generative AI systems:

[0780] The server converts the received question and sentiment data into an appropriate format and sends asynchronous requests to multiple generative AI services. For example, it sends the question "When are the next Olympics?" to OpenAI's GPT-3, Google's BERT, Microsoft's Turing, and others.

[0781] 5. Evaluate the answers obtained from each AI:

[0782] The server collects the responses returned by each generative AI and stores them in a list. Each response is something like "2024," "next year," or "July 2024." During this process, sentiment data is taken into consideration to select the most appropriate response.

[0783] 6. Grammatical check and fact-checking of selected answers:

[0784] The server grammatically checks the selected answer and fact-checks it by referring to reliable sources. For example, if "July 2024" is selected, it will refer to the official Olympic website and add the specific date "July 26, 2024".

[0785] 7. Obtaining related advertisements:

[0786] The server retrieves relevant ads based on the user's emotions. For example, excited users are shown ads for sports equipment or event tickets.

[0787] 8. Display of the final answer and related ads:

[0788] The corrected final answer is sent from the server to the device, which then displays it to the user. Relevant advertisements are also displayed along with the answer.

[0789] Specific example:

[0790] When a user asks "When is the next Olympics?", the specific actions taken are as follows:

[0791] 1. The user enters "When is the next Olympics?" into their device.

[0792] 2. The device uses an emotion engine to analyze the user's emotions (e.g., excited).

[0793] 3. The device sends the question and sentiment data to the server.

[0794] 4. The server sends questions to multiple generative AI systems (e.g., OpenAI's GPT-3, Google's BERT, Microsoft's Turing).

[0795] 5. Collect responses from each AI and select the most appropriate response (e.g., "July 2024").

[0796] 6. Check the grammar and facts of the answer (e.g., add any supplementary information to "July 26, 2024").

[0797] 7. Obtain relevant ads based on emotions (e.g., ads for sporting goods).

[0798] 8. The device displays an advertisement with the final answer: "The next Olympics will be held from July 26, 2024."

[0799] Example of a prompt:

[0800] "When is the next Olympics? A question from an excited user."

[0801] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0802] Step 1:

[0803] The user enters a question.

[0804] Operation: The user enters a question through the input form on their device. For example, they might enter, "When are the next Olympics?"

[0805] Input: The user enters a question into the input form.

[0806] Output: The terminal retrieves the user's question data.

[0807] Step 2:

[0808] The device analyzes the question, voice tone, and facial expressions.

[0809] Operation: The device uses an emotion engine to analyze the user's voice tone and facial expressions in real time. Specifically, it captures voice with a microphone and analyzes facial expressions with a camera. Software used includes speech recognition tools (e.g., Google Speech-to-Text) and face recognition APIs (e.g., Microsoft Azure Face API).

[0810] Input: User's questions and data on voice tone and facial expressions.

[0811] Output: Emotional status analyzed by the emotion engine (e.g., excitement, anticipation, etc.).

[0812] Step 3:

[0813] The device sends the question and sentiment status to the server.

[0814] Operation: The terminal sends the question and the obtained sentiment status to the server as data packets. Possible transmission formats include JSON.

[0815] Input: User's question and analyzed sentiment status.

[0816] Output: Question and sentiment data sent to the server in JSON format, etc.

[0817] Step 4:

[0818] The server sends questions to multiple generative AIs.

[0819] Operation: The server sends received question and sentiment data as asynchronous requests to multiple generative artificial intelligence systems. Recipients include OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[0820] Input: User question and sentiment data received by the server.

[0821] Output: Question and sentiment data sent to a generative AI service.

[0822] Step 5:

[0823] Receive and evaluate responses from each generative AI.

[0824] Operation: The server receives responses from each generative AI and stores each response in a list. It sets evaluation criteria considering the emotional status and evaluates the best response.

[0825] Input: Responses and emotional status from each generative AI.

[0826] Output: Store the most appropriate answer in a list and calculate the evaluation score.

[0827] Step 6:

[0828] Select the most appropriate answer and perform grammatical checks and fact-checks.

[0829] Operation: The server selects the most appropriate answer based on the evaluation score and checks it using a grammar checking tool (e.g., Grammarly API) and fact-checking sources (e.g., the official Olympic website).

[0830] Input: The answer with the highest rating score.

[0831] Output: Corrected answer after grammar check and fact-checking (e.g., "from July 26, 2024").

[0832] Step 7:

[0833] Get relevant ads based on your emotional status

[0834] Operation: The server considers the user's emotional state and retrieves relevant ads from the ad delivery system (e.g., Google AdSense API).

[0835] Input: Emotional status and optimal response.

[0836] Output: Relevant ads selected based on emotional status.

[0837] Step 8:

[0838] Send the final answer and related advertisements to your device and display them.

[0839] Operation: The corrected final answer and related advertisements are sent to the device, which then displays them to the user. For example, an advertisement for sports equipment might be displayed along with the answer, "The next Olympics are from July 26, 2024."

[0840] Input: Corrected final answer and related ads.

[0841] Output: The final answer and related advertisements displayed on the user's device.

[0842] (Application Example 2)

[0843] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0844] Conventional generative artificial intelligence systems often lacked sufficient accuracy and quality in their responses to user questions. A particular challenge was the decrease in user satisfaction due to responses generated without considering the user's emotions. Furthermore, the process of appropriately evaluating responses from multiple generative AI systems and selecting the optimal answer was complex, and efficiency improvements were needed. Additionally, there was the difficulty in effectively presenting relevant products that responded to the user's emotions when suggesting products in virtual stores.

[0845] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0846] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related advertisements, means for recognizing the user's emotions, means for transmitting questions containing emotion data to the generative artificial intelligence systems, and means for selecting the optimal answer based on the emotion data. This enables the generation of optimal answers that take the user's emotions into consideration and the effective suggestion of related products based on those answers.

[0847] "Means for receiving user questions" refers to a device or program that has the function of receiving question information entered by a user.

[0848] "Means for sending questions to multiple generative artificial intelligence systems" refers to a device or program that has the function of sending a user's questions to multiple generative AI services.

[0849] "Means for receiving responses from each generative artificial intelligence" refers to a device or program that has the function of receiving response information returned from a generative AI service.

[0850] "A means of evaluating multiple received responses and selecting the most appropriate response" refers to a device or program that has the function of analyzing and evaluating responses from multiple generative AI systems and selecting the most appropriate one.

[0851] "Means for automatically correcting selected answers" refers to a device or program that has the function of automatically performing grammatical checks and factual verification on selected answers and correcting them as necessary.

[0852] "Means for displaying corrected final answers and related advertisements" refers to a device or program that has the function of presenting corrected answers and related advertisements to a user.

[0853] "Means for recognizing user emotions" refers to a device or program that analyzes a user's voice and facial expressions to determine their emotions.

[0854] "Means for sending questions containing emotional data to a generative artificial intelligence system" refers to a device or program that has the function of adding emotional data to a user's question and sending it to a generative AI service.

[0855] "Means for selecting the optimal response based on emotional data" refers to a device or program that has the function of evaluating responses from multiple generative AI systems, taking emotional data into consideration, and selecting the best one.

[0856] "Means for performing grammatical checks" refers to a device or program that has the function of verifying and correcting the grammatical consistency of a text.

[0857] "Means of fact-checking" refers to a device or program that has the function of verifying the content of a received response based on reliable sources.

[0858] "Means for setting evaluation criteria for selecting the optimal answer" refers to a device or program that has the function of setting criteria (e.g., accuracy, relevance, etc.) for selecting the best answer from among multiple answers.

[0859] "Means for calculating the evaluation score for each response" refers to a device or program that has the function of assigning a score to each response based on established evaluation criteria.

[0860] "Means for selecting the highest-rated response based on scores" refers to a device or program that has the function of selecting the highest-rated response based on evaluation scores.

[0861] "Means for setting evaluation criteria that take emotional data into consideration" refers to a device or program that has the function of setting criteria for evaluating responses while taking into consideration the user's emotional data.

[0862] This invention is a system that improves the quality of responses by using multiple generative artificial intelligence systems to provide the most appropriate answers to user questions, and by combining this with an emotion engine that recognizes the user's emotions. This system mainly consists of a server and terminals.

[0863] Hardware and software configuration

[0864] Hardware to use

[0865] Smart glasses, head-mounted displays, or smartphones

[0866] Software to use

[0867] Emotion recognition libraries (e.g., Microsoft Azure Emotion Recognition API)

[0868] Generative AI libraries (e.g., OpenAI, Google AI)

[0869] Programming language: Python

[0870] System operation

[0871] User question received

[0872] Users ask questions about products within the virtual store using a device (smart glasses, head-mounted display, or smartphone). The device receives the question via an input form, and an emotion engine analyzes the user's voice tone and facial expressions to generate an emotion status. This emotion status is then sent to the server.

[0873] Sending questions and sentiment data

[0874] The server receives the question and sentiment status, converts them into the appropriate format, and sends them to multiple generative artificial intelligences. Asynchronous requests are made to the API endpoint of each generative AI, and each AI generates an answer to the question using its own algorithm.

[0875] Receiving and evaluating responses

[0876] The server receives responses from each generative AI system and stores them in a list. The server evaluates each response, taking into account the user's emotional status, and selects the most appropriate response. The evaluation also takes emotional status into account; if the user is excited, a more detailed and future-oriented response is more likely to be selected.

[0877] Correcting responses and generating ads

[0878] The server performs grammar checks and fact-checks on the selected answers. It then references reliable sources to add further details. Additionally, it utilizes ad delivery service APIs to retrieve relevant ads based on the sentiment status.

[0879] Final answer and ad display

[0880] The corrected final answers and relevant advertisements are sent to the device and displayed to the user. This provides the user with accurate, reliable answers and relevant advertisements that are tailored to their emotions.

[0881] Specific example

[0882] If a user asks "Is this jacket waterproof?" in a virtual store, the following process will occur.

[0883] 1. Use the microphone and camera on the smart glasses to capture the user's questions and facial expressions.

[0884] 2. The emotion engine analyzes the user's voice tone and facial expressions and generates an emotion status of "excited".

[0885] 3. The question and emotion status are sent to the server, and requests are sent to multiple generative AI systems.

[0886] 4. Each generation AI generates a response such as "Yes, it's waterproof," "Yes, it's waterproof tested," or "Yes, it's waterproof and windproof."

[0887] 5. The server, considering its "excited" emotional status, selects the detailed answer "Yes, it is waterproof and windproof."

[0888] 6. As related advertisements, suggest items such as waterproof jackets and waterproof sprays.

[0889] 7. Display the final answer and related advertisements on the smart glasses' display.

[0890] Example input prompts for a generative AI model

[0891] User question: "Is this jacket waterproof?"

[0892] User's emotion: "Excited"

[0893] Please generate variations of the answer.

[0894] This invention makes it possible to generate optimal responses that take user emotions into consideration, and to effectively suggest related products based on those responses.

[0895] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0896] Step 1:

[0897] The user enters a question about the product. A device (smart glasses, head-mounted display, or smartphone) receives this question, and simultaneously, an emotion engine analyzes the user's voice tone and facial expressions. The device generates analyzed emotion data and sends it to the server.

[0898] Input: User's question, user's voice tone, facial expression

[0899] Data processing: The emotion engine analyzes voice tone and facial expressions to generate emotion data.

[0900] Output: Questionnaire data, sentiment data

[0901] Step 2:

[0902] The server receives user questions and sentiment data. It converts the received data into an appropriate format (e.g., JSON) and sends asynchronous requests to the API endpoints of each generative artificial intelligence system.

[0903] Input: Question data, sentiment data

[0904] Data processing: Convert question data and sentiment data to JSON format.

[0905] Output: Request to generative AI service

[0906] Step 3:

[0907] Each generative artificial intelligence generates an answer to a question and returns the answer data to the server. The server receives this answer data.

[0908] Input: Response from a generative AI service

[0909] Data processing: Each generative AI generates answers using its own unique algorithm.

[0910] Output: List of response data

[0911] Step 4:

[0912] The server evaluates multiple response data received. Considering user sentiment data, it calculates a score for each response based on evaluation criteria to select the best answer. The response with the highest score is then selected.

[0913] Input: List of response data, sentiment data

[0914] Data processing: Assign a score to each response based on evaluation criteria.

[0915] Output: Optimal response data

[0916] Step 5:

[0917] The server performs grammar checks and fact-checks on the selected optimal response data. Corrections are made as needed, and the final response is generated.

[0918] Input: Optimal response data

[0919] Data processing: grammar check, fact-checking, correction.

[0920] Output: Corrected final response data

[0921] Step 6:

[0922] The server retrieves relevant advertisements based on the final response data. Using the ad delivery service's API, it searches for advertisements that match the user's sentiment data and selects the most suitable advertisement.

[0923] Input: Final response data, sentiment data

[0924] Data processing: Obtain advertising data from advertising distribution services.

[0925] Output: Related ad data

[0926] Step 7:

[0927] The corrected final response data and related advertising data are sent from the server to the device. The device then displays them to the user, specifically on the display of smart glasses or a head-mounted display.

[0928] Input: Final response data, related ad data

[0929] Data processing: Conversion to display format

[0930] Output: User-viewable responses and advertisements

[0931] This process allows users to quickly receive detailed and reliable answers that address their emotions, and relevant advertisements are presented effectively.

[0932] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0933] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0934] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0935] [Third Embodiment]

[0936] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0937] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0938] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0939] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0940] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0941] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0942] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0943] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0944] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0945] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0946] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0947] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0948] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[0949] Receive user questions

[0950] The terminal accepts questions from the user via an input form. These questions are entered by the user, for example, through a web browser or mobile application. Let's say the user enters, "When are the next Olympics?" This question is transmitted to the server in real time.

[0951] Send a question to multiple AIs

[0952] The server converts the received question into the appropriate format and sends it to multiple generative artificial intelligence (AI) systems. For example, it makes asynchronous requests to the API endpoints of each AI service. This causes each AI to generate an answer to the question using its own algorithm.

[0953] Receive responses from each AI

[0954] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[0955] Select the most appropriate answer.

[0956] The server evaluates multiple received responses and selects the most appropriate one based on criteria such as fact-checking, grammatical accuracy, and reliability. A scoring function is used for evaluation; for example, if "July 2024" receives the highest score, it will be selected.

[0957] Automatically correct the selected answer.

[0958] The server performs grammatical checks and fact-checks on the selected answers. For example, it checks the answer "July 2024" and makes any necessary corrections. Furthermore, it adds more detailed information, such as "from July 26, 2024," using reliable sources.

[0959] Display the final answer and related ads.

[0960] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0961] Specific example

[0962] If a user asks, "When is the next Olympics?":

[0963] 1. The terminal receives the question and forwards it to the server.

[0964] 2. The server sends questions to multiple generative artificial intelligence systems.

[0965] 3. Each AI will respond with answers such as "2024," "next year," or "July 2024."

[0966] 4. The server selects the most appropriate answer, "July 2024".

[0967] 5. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[0968] 6. The device displays the user the final answer, "The next Olympics will be held from July 26, 2024," along with related advertisements.

[0969] This system allows users to quickly obtain accurate and reliable answers.

[0970] The following describes the processing flow.

[0971] Step 1:

[0972] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[0973] Step 2:

[0974] The terminal sends the question to the server. The entered question is transferred to the server in real time. During this process, the question is transmitted securely using network communication.

[0975] Step 3:

[0976] The server sends a question to multiple generative artificial intelligence (AI) services. Upon receiving the question, the server converts its content into an appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it might send a question to "AI Service 1," "AI Service 2," and "AI Service 3."

[0977] Step 4:

[0978] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, it might receive responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3".

[0979] Step 5:

[0980] The server evaluates the responses and selects the most appropriate one. A scoring function is applied to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. The response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[0981] Step 6:

[0982] The selected answer will be automatically corrected. The server will perform grammatical checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as necessary. Furthermore, it will add detailed information such as "from July 26, 2024" by referring to reliable sources.

[0983] Step 7:

[0984] The system generates the final answer and related ads. The server retrieves the relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the answer.

[0985] Step 8:

[0986] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[0987] Step 9:

[0988] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[0989] In this way, users can quickly obtain accurate and reliable answers simply by entering their questions.

[0990] (Example 1)

[0991] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0992] Conventional information retrieval systems often provide inaccurate information in response to user inquiries. Even with the use of numerous generative artificial intelligence systems, responses may vary or contain unreliable information. In such situations, it is difficult for users to obtain accurate information quickly. Furthermore, users bear the burden of having to judge the reliability of the information they receive themselves.

[0993] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0994] In this invention, the server includes means for receiving questions from users, means for forwarding received questions to the server, means for sending questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the multiple received answers and selecting the most appropriate answer, means for automatically correcting the selected answer, and means for displaying the corrected final answer and related advertisements. This enables users to quickly obtain accurate and reliable answers.

[0995] A "user" refers to an individual or group that inputs questions into a system and receives answers.

[0996] A "terminal" is a device used by a user to input questions and send them to a server, and specifically includes personal computers and smartphones.

[0997] A "server" refers to a computer system that receives questions from users, sends the questions to multiple generative artificial intelligence systems, evaluates and corrects the received answers, and then sends the final answer to the terminal.

[0998] "Generative artificial intelligence" refers to systems that utilize machine learning and natural language processing technologies to automatically generate answers to user questions.

[0999] "Means of receiving questions" refers to a combination of software and hardware for inputting, storing, and transferring user questions to a server in a digital format.

[1000] "Means for sending questions" refers to communication protocols or API interfaces that allow a server to send questions received from a user to multiple generative artificial intelligence systems.

[1001] "Means of receiving responses" refers to communication protocols and databases used by servers to receive and store responses from generative artificial intelligence.

[1002] "Means of evaluating responses" refers to algorithms or evaluation systems that analyze multiple received responses and select the most appropriate response based on evaluation criteria.

[1003] "Means for correcting answers" refers to programs or libraries used to perform grammatical checks and factual verification on selected answers, and then make corrections or revisions.

[1004] "Means for displaying corrected final answers and related advertisements" refers to software and interfaces for displaying the final answers and related advertisements on the user's device.

[1005] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[1006] This system includes a terminal that receives user questions, a server that processes the received questions and transmits them to multiple generative artificial intelligence systems, and means for evaluating and correcting the answers from the generative artificial intelligence systems and providing the user with the final answers and related advertisements.

[1007] Hardware and Software Overview

[1008] terminal

[1009] A device is a device on which the user enters a question, and primarily includes personal computers, smartphones, and tablets. The device provides an interface for entering the question using a web browser or mobile application.

[1010] server

[1011] This is a central computer system for receiving user questions and transmitting them to multiple generative artificial intelligence systems. The server has the following main functions:

[1012] API interface: A means of communication for sending questions to and receiving answers from generative artificial intelligence.

[1013] Database: A storage device for storing and evaluating received responses.

[1014] Evaluation system: An algorithm that evaluates each answer and selects the most appropriate answer.

[1015] Generative artificial intelligence

[1016] This refers to a system that automatically generates answers to questions sent from a server, and includes Google Cloud AI, OpenAI, and IBM Watson.

[1017] System operation example

[1018] The user types "When is the next Olympics?" into their web browser. At this point, the terminal receives the user's question and forwards it to the server.

[1019] The server converts the received question into an appropriate format and sends an API request to a generative artificial intelligence (AI) such as Google Cloud AI, OpenAI, or IBM Watson. Multiple generative AIs answer the question based on their own algorithms and return the answer to the server.

[1020] The server receives responses from each generative artificial intelligence system and stores them in a database. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[1021] The server evaluates the multiple responses it receives. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. A scoring algorithm is used to assign points to each response, and the response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[1022] The server performs grammar checks and fact-checks on the selected answer. Grammar checks utilize language processing libraries, and fact-checking relies on reliable sources. As a result, detailed information such as "from July 26, 2024" is added.

[1023] The corrected final answer is sent to the device, which then displays it to the user. Related advertisements are also displayed simultaneously. For example, an advertisement for sporting goods might appear along with the answer, "The next Olympics are from July 26, 2024."

[1024] Example of a prompt

[1025] Question: "When is the next Olympics?"

[1026] Question: "What is the date of the next Olympic Games?"

[1027] Question: "I would like to know the dates of the Olympics."

[1028] As described above, this system allows users to quickly obtain accurate and reliable answers.

[1029] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1030] Step 1:

[1031] The user enters a question.

[1032] Input: The user enters the question via a web browser or mobile application.

[1033] Action: The user types "When are the next Olympics?" into the browser's search form.

[1034] Data processing: Format the input text data into JSON format.

[1035] Output: The formatted question data is temporarily stored on the device.

[1036] Step 2:

[1037] The terminal forwards the question to the server.

[1038] Input: Question data entered by the user (in JSON format).

[1039] Operation: The device uses the JavaScript fetch API to send the question data to the server as an HTTP request.

[1040] Data processing: Include the question data in the HTTP request body.

[1041] Output: Question data arrives at the server and the server receives it.

[1042] Step 3:

[1043] The server sends questions to multiple generative artificial intelligence systems.

[1044] Input: Question data received from the terminal.

[1045] Operation: The server processes the received question data into an API request format.

[1046] Data processing: Convert the question data into a format compatible with the API of each generative artificial intelligence system. For example, convert from JSON to URL-encoded format.

[1047] Output: The converted question data is sent as an HTTP request to each generative artificial intelligence system.

[1048] Step 4:

[1049] The server receives responses from each generative artificial intelligence system.

[1050] Input: Response data from each generative artificial intelligence.

[1051] Operation: The server asynchronously waits for and receives responses from each generative artificial intelligence system.

[1052] Data processing: Store the received response data in a list.

[1053] Output: A list containing the answers of each generative artificial intelligence.

[1054] Step 5:

[1055] The server evaluates the received responses and selects the most appropriate one.

[1056] Input: Multiple response data (lists) from various generative artificial intelligence systems.

[1057] Operation: The server uses an evaluation algorithm to score each response.

[1058] Data processing: A scoring function is applied to each response to calculate an evaluation score based on reliability and grammatical accuracy.

[1059] Output: Select the answer with the highest rating.

[1060] Step 6:

[1061] The server automatically corrects the selected answer.

[1062] Input: The most appropriate response data.

[1063] Operation: The server performs grammar checks and fact-checks on the selected answer. For grammar checks, it uses, for example, the Grammarly API, and for fact-checks, it refers to reliable sources.

[1064] Data processing: This involves correcting grammar and adding detailed information.

[1065] Output: Corrected final response data.

[1066] Step 7:

[1067] The device displays the final answer and related ads.

[1068] Input: Final response data and related advertising data sent from the server.

[1069] Function: Displays data received by the device on web pages and applications. Dynamically updates information using HTML and JavaScript.

[1070] Data processing: Convert the final response data and advertising data into a format that is compatible with the user interface.

[1071] Output: The user is shown the information "The next Olympic Games will be held from July 26, 2024" and advertisements for related sporting goods.

[1072] Through the steps outlined above, users can quickly obtain accurate and reliable answers.

[1073] (Application Example 1)

[1074] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1075] Conventional generative artificial intelligence systems have a problem in that they do not provide accurate and appropriate answers to user questions with sufficient precision. Furthermore, they lack the ability to appropriately recommend related information and content related to the generated answers, resulting in a situation where user needs and convenience are not fully met.

[1076] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1077] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related information, and means for providing related content based on the displayed information. This enables the user to obtain accurate and reliable answers and simultaneously acquire related content.

[1078] A "user" refers to each individual who uses this system.

[1079] "Means for receiving questions" refers to a device or software that receives input from a user and processes its contents.

[1080] "Generative artificial intelligence" refers to AI models that can generate answers to given questions or inputs.

[1081] "Means for sending questions" refers to a device or software that transmits questions received from a user to multiple generative artificial intelligence systems.

[1082] "Means for receiving responses" refers to a device or software for receiving responses generated by generative artificial intelligence.

[1083] "Means for evaluating responses" refers to a device or software for evaluating multiple received responses and determining which is the most appropriate.

[1084] "Means for selecting the most appropriate answer" refers to a device or software for selecting the optimal answer based on evaluation.

[1085] "Means for automatically correcting answers" refers to a device or software for automatically correcting selected answers based on grammar or facts.

[1086] "Means for displaying the final answer and related information" refers to a device or software for displaying the corrected final answer and related information to the user.

[1087] "Means for providing related content" refers to a device or software for providing additional content to the user based on the displayed information.

[1088] This invention is a system that processes user questions in real time, provides the most appropriate answers, and recommends relevant content. Specific embodiments of this system are described below.

[1089] The system primarily consists of a server, terminals, and users. Users submit questions via their terminals (such as smartphones and tablets). A dedicated application installed on the terminal receives questions from users and forwards them to the server. This application also provides an interface for users to input their questions.

[1090] The server sends questions received from users to multiple generative artificial intelligence (AI) models and receives responses from each AI model. In this process, the server uses OpenAI's GPT-3 API or other AI service technologies to send questions. For example, if a user asks, "What are some recommended new movies?", the server sends this question to the AI ​​models and receives responses from each.

[1091] The server evaluates the multiple responses received and selects the most appropriate one. This evaluation considers factors such as accuracy, reliability, and grammatical correctness. The server calculates a score for each response based on the set evaluation criteria and selects the response with the highest score. The server then automatically performs grammatical checks and fact-checks on the selected response and corrects it to its final form.

[1092] The final corrected answer and related information are sent to the device. The device displays related content to the user along with the final answer. This related content might include, for example, a link to watch the movie or a list of related movies in the case of a question about movie recommendations. The device uses web scraping technology to retrieve and provide the related content to the user.

[1093] Hardware and software to be used

[1094] Hardware: Servers, smartphones, tablets

[1095] Software: OpenAI GPT-3 API, requests, BeautifulSoup, dedicated application (for smartphones and tablets)

[1096] Specific example

[1097] The following shows the specific processing that occurs when a user asks, "What are some of the latest movies you recommend?"

[1098] 1. The user enters "What are some recommended new movies?" into their device.

[1099] 2. The terminal forwards this question to the server.

[1100] 3. The server sends questions to multiple generative AI models and receives responses.

[1101] 4. The server selects the best response from those received and performs grammatical checks and fact-checks.

[1102] 5. The server sends the corrected response and related content to the device.

[1103] 6. The device displays to the user a link to watch the relevant movie and reviews along with the final answer.

[1104] Examples of prompt statements

[1105] "What are some of the latest movies you recommend?"

[1106] "What are your recommended travel destinations this year?"

[1107] "When is the next Olympics?"

[1108] In this way, users can quickly obtain accurate and reliable answers, along with useful content related to those answers. This system can meet user needs and significantly improve convenience.

[1109] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1110] Step 1:

[1111] The user enters a question.

[1112] The user uses a device (such as a smartphone or tablet) to enter a question into the application's input form. For example, they might enter, "What are some recommended new movies?" This input is then accepted by the device.

[1113] Input: A question entered by the user (e.g., "What are some recommended new movies?")

[1114] Output: Question text data

[1115] Step 2:

[1116] The terminal forwards the question to the server.

[1117] The application installed on the device transfers the received question text data to the server.

[1118] Input: Question text data

[1119] Output: Question data transferred to the server

[1120] Step 3:

[1121] The server sends questions to each generative artificial intelligence system.

[1122] The server converts the received question into the appropriate format and sends the question to multiple generative artificial intelligence (AI) models. Specifically, it makes asynchronous requests to the API endpoints of each AI service.

[1123] Input: Question data

[1124] Output: Question requests sent to multiple AI models

[1125] Step 4:

[1126] Receive an answer from a generative artificial intelligence.

[1127] The server receives responses from each generative artificial intelligence system and stores each response in a list.

[1128] Input: Request for response from AI model

[1129] Output: List of answers (Example: ["Movie A", "Movie B", "Best Movie C"])

[1130] Step 5:

[1131] The server selects the best answer.

[1132] The server evaluates multiple received responses and selects the most appropriate response based on predefined evaluation criteria. These criteria include accuracy, reliability, and grammatical correctness. A scoring function is used to calculate evaluation scores, and the response with the highest score is selected.

[1133] Input: Answer list

[1134] Output: Selected best answer (e.g., "Best movie C")

[1135] Step 6:

[1136] Correct the selected answer.

[1137] The server performs grammatical checks and fact-checks on the selected answer and makes corrections as needed. For example, it uses an online grammar checking service to check the grammatical structure and consults relevant databases for fact-checking.

[1138] Input: Selected answer

[1139] Output: Corrected final answer (e.g., "The best movie recommendation is the latest movie C")

[1140] Step 7:

[1141] The server retrieves the relevant content.

[1142] Based on the finalized response, the server retrieves relevant content from the web. Specifically, it uses web scraping techniques to collect links to watch related movies and reviews.

[1143] Input: Corrected final answer

[1144] Output: Related content (e.g., viewing links, reviews)

[1145] Step 8:

[1146] The device displays the final answer and related content.

[1147] The terminal receives the corrected final answer and related content sent from the server and displays it to the user.

[1148] Input: Final answer and related content

[1149] Output: The final answer and related content displayed to the user (e.g., "The best movie recommendation is the latest movie C. Here is the viewing link and review.")

[1150] As a result, users can quickly obtain accurate and reliable answers, along with relevant and useful content.

[1151] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1152] This invention provides a system that uses multiple generative artificial intelligence systems to provide the most appropriate answer to a user's question, and improves the quality of the answer by combining it with an emotion engine that recognizes the user's emotions. Specific examples of this system are described below.

[1153] Recognizing user emotions

[1154] As the device receives a question from the user via an input form, the emotion engine recognizes the user's emotions. For example, if a user types "When are the next Olympics?", the emotion engine analyzes the user's tone of voice and facial expressions to determine whether the user is feeling curiosity, anticipation, or excitement. This data is analyzed in real time in response to text input, and as a result, an emotion status is generated.

[1155] Receive a question and add sentiment data.

[1156] The device sends the sentiment status generated by the sentiment engine to the server along with the user's question. This status is processed together with the question.

[1157] Send a question to multiple AIs

[1158] The server converts the question and associated sentiment data into an appropriate format and sends it to multiple generative artificial intelligence (AI) services. Asynchronous requests are made to the API endpoint of each AI service, and each AI uses its own algorithm to generate an answer to the question.

[1159] Receive and evaluate the responses from each AI.

[1160] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, responses such as "2024" from "AI Service 1," "next year" from "AI Service 2," and "July 2024" from "AI Service 3" might be obtained. During this process, the emotional status is also considered and influences the evaluation criteria.

[1161] Select the most appropriate answer.

[1162] The server evaluates multiple responses received and selects the most appropriate one. The evaluation also takes into account the user's emotional status; for example, if the user is excited, they are more likely to choose a more detailed and future-oriented response (such as "July 2024").

[1163] Automatically correct the selected answer.

[1164] The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the server performs a grammar check on this answer and corrects it as needed. Furthermore, it adds detailed information such as "from July 26, 2024" by referring to reliable sources and adjusts the tone of the sentence to match the sentiment status.

[1165] Generate related ads based on the final answer

[1166] The server retrieves relevant ads based on the corrected final response. It uses the ad delivery service's API to search for relevant ads and provide them along with the response. For example, a user with an excited emotional status might see ads for sports equipment or event tickets.

[1167] Display the final answer and related ads.

[1168] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[1169] Specific example

[1170] If a user asks, "When is the next Olympics?":

[1171] 1. The device receives the question, and the emotion engine analyzes the user's emotions. It recognizes that the user is agitated.

[1172] 2. The device sends the question and sentiment status to the server.

[1173] 3. The server sends questions to multiple generative artificial intelligence systems.

[1174] 4. Each AI will respond with answers such as "2024," "next year," or "July 2024." Emotional status will also be included in the evaluation.

[1175] 5. The server selects the most appropriate answer, "July 2024". Additional date details are provided based on the sentiment status.

[1176] 6. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[1177] 7. The server retrieves relevant ads based on the sentiment status.

[1178] 8. The device displays the final answer, "The next Olympics are from July 26, 2024," along with related advertisements. For example, an excited user might see an advertisement for sports equipment.

[1179] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[1180] The following describes the processing flow.

[1181] Step 1:

[1182] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[1183] Step 2:

[1184] The device recognizes the user's emotions. The emotion engine analyzes the user's voice tone, facial expressions, keyboard typing style, etc., to generate an emotional status. For example, it recognizes when the user is asking a question with anticipation or is excited. This emotional information is also temporarily stored.

[1185] Step 3:

[1186] The device sends the question and sentiment status to the server. Sentiment information is securely transmitted to the server along with the question content using network communication.

[1187] Step 4:

[1188] The server receives the question and sentiment status, and sends the question to multiple generative artificial intelligence (AI) services. It converts the received question into the appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it sends the question to "AI Service 1," "AI Service 2," and "AI Service 3."

[1189] Step 5:

[1190] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3" might be obtained.

[1191] Step 6:

[1192] The server evaluates the responses and selects the most appropriate one. It applies a scoring function to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, reliability, and even sentiment status. For example, if the user is excited, detailed, future-oriented responses (such as "July 2024") will receive higher scores.

[1193] Step 7:

[1194] The selected answer is automatically corrected. The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as needed. Furthermore, it will refer to reliable sources and add more detailed information such as "from July 26, 2024". The text may also be corrected to match the tone and expression of the sentiment status.

[1195] Step 8:

[1196] The system generates the final answer and relevant ads. The server retrieves relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the final answer. For example, users with an excited emotion status might see ads for sports goods or event tickets.

[1197] Step 9:

[1198] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[1199] Step 10:

[1200] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[1201] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[1202] (Example 2)

[1203] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1204] Conventional question-answering systems have the problem of not providing answers that take user emotions into consideration, making it difficult for users to obtain satisfactory answers. Furthermore, when selecting the most appropriate answer from answers obtained from multiple generative AI systems, the evaluation does not include emotional data, which can lead to a decrease in the quality of the answers. In addition, when providing related information or advertisements, there is a problem in that the content displayed is not appropriate based on the user's emotional state.

[1205] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1206] In this invention, the server includes means for recognizing the user's emotions, means for attaching emotion data to a question and transmitting it, means for evaluating multiple received answers and selecting the most appropriate answer, and means for performing grammatical checks and fact-checking on the selected answer. This makes it possible to provide the most appropriate answer obtained from multiple generative artificial intelligence systems while taking the user's emotions into consideration.

[1207] A "user" is someone who uses the system to input questions and receive answers.

[1208] A "question" refers to the information or points of concern that a user enters into the system.

[1209] "Emotions" refer to the user's psychological state and include curiosity, anticipation, excitement, and so on.

[1210] "Emotional data" refers to information generated by analyzing the user's emotions and is added to the questions.

[1211] "Generative artificial intelligence" refers to algorithms and systems that generate natural language in a way that mimics human speech.

[1212] "Grammar check" is the process of verifying whether a text uses correct grammar.

[1213] "Fact-checking" is the process of verifying whether information is accurate and reliable.

[1214] "Evaluation criteria" are standards for evaluating multiple responses and pieces of information, and they take into account user sentiment data.

[1215] An "evaluation score" is a numerical representation of how appropriate each response is based on the evaluation criteria.

[1216] "Advertisements" are those that provide information about products or services related to the final answer.

[1217] An "asynchronous request" is a method of sending requests to multiple AI services simultaneously and processing them independently.

[1218] This invention is a system that uses multiple generative artificial intelligence systems to provide the most appropriate answers to user questions. This system includes functions to recognize the user's emotions, generate answers while considering that emotion data, and also provide relevant advertisements. The specific implementation of this system is described below.

[1219] System configuration:

[1220] This system consists of terminals, servers, and multiple generative artificial intelligence (AI) services. The main hardware and software used are as follows:

[1221] 1. Device (PC or smartphone): This is the device used by the user to input questions and view the final answers. It is desirable that it be equipped with a camera and microphone for analyzing voice tone and facial expressions.

[1222] 2. Server: This is the central computing unit that performs data processing, AI integration, question and answer evaluation, grammar checking, fact-checking, and relevant advertisement retrieval.

[1223] 3. Emotion Engine: This is a software tool for analyzing user emotions. Specifically, it uses speech recognition software or facial recognition tools (e.g., Google Speech-to-Text or Microsoft Azure Face API).

[1224] 4. Generative artificial intelligence services: Algorithms for generating answers to questions. Specifically, this includes OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[1225] Operation details:

[1226] The system operates as follows:

[1227] 1. The user enters the question:

[1228] Enter the question into the input form on the terminal. For example, a question like, "When is the next Olympics?"

[1229] 2. The device analyzes emotions:

[1230] The system analyzes the user's voice tone and facial expressions using an emotion engine, and the results are obtained as emotion data. For example, if the user is asked "When is the next Olympics?", the system will analyze that the user is excited.

[1231] 3. The device sends the question and sentiment data to the server:

[1232] The device sends the analysis results (question and emotion status) to the server as a data packet. Specifically, data such as "Question: When are the next Olympics?" and "Emotion Status: Excited" is sent in JSON format.

[1233] 4. The server sends questions to multiple generative AI systems:

[1234] The server converts the received question and sentiment data into an appropriate format and sends asynchronous requests to multiple generative AI services. For example, it sends the question "When are the next Olympics?" to OpenAI's GPT-3, Google's BERT, Microsoft's Turing, and others.

[1235] 5. Evaluate the answers obtained from each AI:

[1236] The server collects the responses returned by each generative AI and stores them in a list. Each response is something like "2024," "next year," or "July 2024." During this process, sentiment data is taken into consideration to select the most appropriate response.

[1237] 6. Grammatical check and fact-checking of selected answers:

[1238] The server grammatically checks the selected answer and fact-checks it by referring to reliable sources. For example, if "July 2024" is selected, it will refer to the official Olympic website and add the specific date "July 26, 2024".

[1239] 7. Obtaining related advertisements:

[1240] The server retrieves relevant ads based on the user's emotions. For example, excited users are shown ads for sports equipment or event tickets.

[1241] 8. Display of the final answer and related ads:

[1242] The corrected final answer is sent from the server to the device, which then displays it to the user. Relevant advertisements are also displayed along with the answer.

[1243] Specific example:

[1244] When a user asks "When is the next Olympics?", the specific actions taken are as follows:

[1245] 1. The user enters "When is the next Olympics?" into their device.

[1246] 2. The device uses an emotion engine to analyze the user's emotions (e.g., excited).

[1247] 3. The device sends the question and sentiment data to the server.

[1248] 4. The server sends questions to multiple generative AI systems (e.g., OpenAI's GPT-3, Google's BERT, Microsoft's Turing).

[1249] 5. Collect responses from each AI and select the most appropriate response (e.g., "July 2024").

[1250] 6. Check the grammar and facts of the answer (e.g., add any supplementary information to "July 26, 2024").

[1251] 7. Obtain relevant ads based on emotions (e.g., ads for sporting goods).

[1252] 8. The device displays an advertisement with the final answer: "The next Olympics will be held from July 26, 2024."

[1253] Example of a prompt:

[1254] "When is the next Olympics? A question from an excited user."

[1255] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1256] Step 1:

[1257] The user enters a question.

[1258] Operation: The user enters a question through the input form on their device. For example, they might enter, "When are the next Olympics?"

[1259] Input: The user enters a question into the input form.

[1260] Output: The terminal retrieves the user's question data.

[1261] Step 2:

[1262] The device analyzes the question, voice tone, and facial expressions.

[1263] Operation: The device uses an emotion engine to analyze the user's voice tone and facial expressions in real time. Specifically, it captures voice with a microphone and analyzes facial expressions with a camera. Software used includes speech recognition tools (e.g., Google Speech-to-Text) and face recognition APIs (e.g., Microsoft Azure Face API).

[1264] Input: User's questions and data on voice tone and facial expressions.

[1265] Output: Emotional status analyzed by the emotion engine (e.g., excitement, anticipation, etc.).

[1266] Step 3:

[1267] The device sends the question and sentiment status to the server.

[1268] Operation: The terminal sends the question and the obtained sentiment status to the server as data packets. Possible transmission formats include JSON.

[1269] Input: User's question and analyzed sentiment status.

[1270] Output: Question and sentiment data sent to the server in JSON format, etc.

[1271] Step 4:

[1272] The server sends questions to multiple generative AIs.

[1273] Operation: The server sends received question and sentiment data as asynchronous requests to multiple generative artificial intelligence systems. Recipients include OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[1274] Input: User question and sentiment data received by the server.

[1275] Output: Question and sentiment data sent to a generative AI service.

[1276] Step 5:

[1277] Receive and evaluate responses from each generative AI.

[1278] Operation: The server receives responses from each generative AI and stores each response in a list. It sets evaluation criteria considering the emotional status and evaluates the best response.

[1279] Input: Responses and emotional status from each generative AI.

[1280] Output: Store the most appropriate answer in a list and calculate the evaluation score.

[1281] Step 6:

[1282] Select the most appropriate answer and perform grammatical checks and fact-checks.

[1283] Operation: The server selects the most appropriate answer based on the evaluation score and checks it using a grammar checking tool (e.g., Grammarly API) and fact-checking sources (e.g., the official Olympic website).

[1284] Input: The answer with the highest rating score.

[1285] Output: Corrected answer after grammar check and fact-checking (e.g., "from July 26, 2024").

[1286] Step 7:

[1287] Get relevant ads based on your emotional status

[1288] Operation: The server considers the user's emotional state and retrieves relevant ads from the ad delivery system (e.g., Google AdSense API).

[1289] Input: Emotional status and optimal response.

[1290] Output: Relevant ads selected based on emotional status.

[1291] Step 8:

[1292] Send the final answer and related advertisements to your device and display them.

[1293] Operation: The corrected final answer and related advertisements are sent to the device, which then displays them to the user. For example, an advertisement for sports equipment might be displayed along with the answer, "The next Olympics are from July 26, 2024."

[1294] Input: Corrected final answer and related ads.

[1295] Output: The final answer and related advertisements displayed on the user's device.

[1296] (Application Example 2)

[1297] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1298] Conventional generative artificial intelligence systems often lacked sufficient accuracy and quality in their responses to user questions. A particular challenge was the decrease in user satisfaction due to responses generated without considering the user's emotions. Furthermore, the process of appropriately evaluating responses from multiple generative AI systems and selecting the optimal answer was complex, and efficiency improvements were needed. Additionally, there was the difficulty in effectively presenting relevant products that responded to the user's emotions when suggesting products in virtual stores.

[1299] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1300] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related advertisements, means for recognizing the user's emotions, means for transmitting questions containing emotion data to the generative artificial intelligence systems, and means for selecting the optimal answer based on the emotion data. This enables the generation of optimal answers that take the user's emotions into consideration and the effective suggestion of related products based on those answers.

[1301] "Means for receiving user questions" refers to a device or program that has the function of receiving question information entered by a user.

[1302] "Means for sending questions to multiple generative artificial intelligence systems" refers to a device or program that has the function of sending a user's questions to multiple generative AI services.

[1303] "Means for receiving responses from each generative artificial intelligence" refers to a device or program that has the function of receiving response information returned from a generative AI service.

[1304] "A means of evaluating multiple received responses and selecting the most appropriate response" refers to a device or program that has the function of analyzing and evaluating responses from multiple generative AI systems and selecting the most appropriate one.

[1305] "Means for automatically correcting selected answers" refers to a device or program that has the function of automatically performing grammatical checks and factual verification on selected answers and correcting them as necessary.

[1306] "Means for displaying corrected final answers and related advertisements" refers to a device or program that has the function of presenting corrected answers and related advertisements to a user.

[1307] "Means for recognizing user emotions" refers to a device or program that analyzes a user's voice and facial expressions to determine their emotions.

[1308] "Means for sending questions containing emotional data to a generative artificial intelligence system" refers to a device or program that has the function of adding emotional data to a user's question and sending it to a generative AI service.

[1309] "Means for selecting the optimal response based on emotional data" refers to a device or program that has the function of evaluating responses from multiple generative AI systems, taking emotional data into consideration, and selecting the best one.

[1310] "Means for performing grammatical checks" refers to a device or program that has the function of verifying and correcting the grammatical consistency of a text.

[1311] "Means of fact-checking" refers to a device or program that has the function of verifying the content of a received response based on reliable sources.

[1312] "Means for setting evaluation criteria for selecting the optimal answer" refers to a device or program that has the function of setting criteria (e.g., accuracy, relevance, etc.) for selecting the best answer from among multiple answers.

[1313] "Means for calculating the evaluation score for each response" refers to a device or program that has the function of assigning a score to each response based on established evaluation criteria.

[1314] "Means for selecting the highest-rated response based on scores" refers to a device or program that has the function of selecting the highest-rated response based on evaluation scores.

[1315] "Means for setting evaluation criteria that take emotional data into consideration" refers to a device or program that has the function of setting criteria for evaluating responses while taking into consideration the user's emotional data.

[1316] This invention is a system that improves the quality of responses by using multiple generative artificial intelligence systems to provide the most appropriate answers to user questions, and by combining this with an emotion engine that recognizes the user's emotions. This system mainly consists of a server and terminals.

[1317] Hardware and software configuration

[1318] Hardware to use

[1319] Smart glasses, head-mounted displays, or smartphones

[1320] Software to use

[1321] Emotion recognition libraries (e.g., Microsoft Azure Emotion Recognition API)

[1322] Generative AI libraries (e.g., OpenAI, Google AI)

[1323] Programming language: Python

[1324] System operation

[1325] User question received

[1326] Users ask questions about products within the virtual store using a device (smart glasses, head-mounted display, or smartphone). The device receives the question via an input form, and an emotion engine analyzes the user's voice tone and facial expressions to generate an emotion status. This emotion status is then sent to the server.

[1327] Sending questions and sentiment data

[1328] The server receives the question and sentiment status, converts them into the appropriate format, and sends them to multiple generative artificial intelligences. Asynchronous requests are made to the API endpoint of each generative AI, and each AI generates an answer to the question using its own algorithm.

[1329] Receiving and evaluating responses

[1330] The server receives responses from each generative AI system and stores them in a list. The server evaluates each response, taking into account the user's emotional status, and selects the most appropriate response. The evaluation also takes emotional status into account; if the user is excited, a more detailed and future-oriented response is more likely to be selected.

[1331] Correcting responses and generating ads

[1332] The server performs grammar checks and fact-checks on the selected answers. It then references reliable sources to add further details. Additionally, it utilizes ad delivery service APIs to retrieve relevant ads based on the sentiment status.

[1333] Final answer and ad display

[1334] The corrected final answers and relevant advertisements are sent to the device and displayed to the user. This provides the user with accurate, reliable answers and relevant advertisements that are tailored to their emotions.

[1335] Specific example

[1336] If a user asks "Is this jacket waterproof?" in a virtual store, the following process will occur.

[1337] 1. Use the microphone and camera on the smart glasses to capture the user's questions and facial expressions.

[1338] 2. The emotion engine analyzes the user's voice tone and facial expressions and generates an emotion status of "excited".

[1339] 3. The question and emotion status are sent to the server, and requests are sent to multiple generative AI systems.

[1340] 4. Each generation AI generates a response such as "Yes, it's waterproof," "Yes, it's waterproof tested," or "Yes, it's waterproof and windproof."

[1341] 5. The server, considering its "excited" emotional status, selects the detailed answer "Yes, it is waterproof and windproof."

[1342] 6. As related advertisements, suggest items such as waterproof jackets and waterproof sprays.

[1343] 7. Display the final answer and related advertisements on the smart glasses' display.

[1344] Example input prompts for a generative AI model

[1345] User question: "Is this jacket waterproof?"

[1346] User's emotion: "Excited"

[1347] Please generate variations of the answer.

[1348] This invention makes it possible to generate optimal responses that take user emotions into consideration, and to effectively suggest related products based on those responses.

[1349] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1350] Step 1:

[1351] The user enters a question about the product. A device (smart glasses, head-mounted display, or smartphone) receives this question, and simultaneously, an emotion engine analyzes the user's voice tone and facial expressions. The device generates analyzed emotion data and sends it to the server.

[1352] Input: User's question, user's voice tone, facial expression

[1353] Data processing: The emotion engine analyzes voice tone and facial expressions to generate emotion data.

[1354] Output: Questionnaire data, sentiment data

[1355] Step 2:

[1356] The server receives user questions and sentiment data. It converts the received data into an appropriate format (e.g., JSON) and sends asynchronous requests to the API endpoints of each generative artificial intelligence system.

[1357] Input: Question data, sentiment data

[1358] Data processing: Convert question data and sentiment data to JSON format.

[1359] Output: Request to generative AI service

[1360] Step 3:

[1361] Each generative artificial intelligence generates an answer to a question and returns the answer data to the server. The server receives this answer data.

[1362] Input: Response from a generative AI service

[1363] Data processing: Each generative AI generates answers using its own unique algorithm.

[1364] Output: List of response data

[1365] Step 4:

[1366] The server evaluates multiple response data received. Considering user sentiment data, it calculates a score for each response based on evaluation criteria to select the best answer. The response with the highest score is then selected.

[1367] Input: List of response data, sentiment data

[1368] Data processing: Assign a score to each response based on evaluation criteria.

[1369] Output: Optimal response data

[1370] Step 5:

[1371] The server performs grammar checks and fact-checks on the selected optimal response data. Corrections are made as needed, and the final response is generated.

[1372] Input: Optimal response data

[1373] Data processing: grammar check, fact-checking, correction.

[1374] Output: Corrected final response data

[1375] Step 6:

[1376] The server retrieves relevant advertisements based on the final response data. Using the ad delivery service's API, it searches for advertisements that match the user's sentiment data and selects the most suitable advertisement.

[1377] Input: Final response data, sentiment data

[1378] Data processing: Obtain advertising data from advertising distribution services.

[1379] Output: Related ad data

[1380] Step 7:

[1381] The corrected final response data and related advertising data are sent from the server to the device. The device then displays them to the user, specifically on the display of smart glasses or a head-mounted display.

[1382] Input: Final response data, related ad data

[1383] Data processing: Conversion to display format

[1384] Output: User-viewable responses and advertisements

[1385] This process allows users to quickly receive detailed and reliable answers that address their emotions, and relevant advertisements are presented effectively.

[1386] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1387] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1388] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1389] [Fourth Embodiment]

[1390] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1391] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1392] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1393] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1394] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1395] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1396] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1397] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1398] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1399] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[1400] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1401] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1402] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1403] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[1404] Receive user questions

[1405] The terminal accepts questions from the user via an input form. These questions are entered by the user, for example, through a web browser or mobile application. Let's say the user enters, "When are the next Olympics?" This question is transmitted to the server in real time.

[1406] Send a question to multiple AIs

[1407] The server converts the received question into the appropriate format and sends it to multiple generative artificial intelligence (AI) systems. For example, it makes asynchronous requests to the API endpoints of each AI service. This causes each AI to generate an answer to the question using its own algorithm.

[1408] Receive responses from each AI

[1409] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[1410] Select the most appropriate answer.

[1411] The server evaluates multiple received responses and selects the most appropriate one based on criteria such as fact-checking, grammatical accuracy, and reliability. A scoring function is used for evaluation; for example, if "July 2024" receives the highest score, it will be selected.

[1412] Automatically correct the selected answer.

[1413] The server performs grammatical checks and fact-checks on the selected answers. For example, it checks the answer "July 2024" and makes any necessary corrections. Furthermore, it adds more detailed information, such as "from July 26, 2024," using reliable sources.

[1414] Display the final answer and related ads.

[1415] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[1416] Specific example

[1417] If a user asks, "When is the next Olympics?":

[1418] 1. The terminal receives the question and forwards it to the server.

[1419] 2. The server sends questions to multiple generative artificial intelligence systems.

[1420] 3. Each AI will respond with answers such as "2024," "next year," or "July 2024."

[1421] 4. The server selects the most appropriate answer, "July 2024".

[1422] 5. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[1423] 6. The device displays the user the final answer, "The next Olympics will be held from July 26, 2024," along with related advertisements.

[1424] This system allows users to quickly obtain accurate and reliable answers.

[1425] The following describes the processing flow.

[1426] Step 1:

[1427] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[1428] Step 2:

[1429] The terminal sends the question to the server. The entered question is transferred to the server in real time. During this process, the question is transmitted securely using network communication.

[1430] Step 3:

[1431] The server sends a question to multiple generative artificial intelligence (AI) services. Upon receiving the question, the server converts its content into an appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it might send a question to "AI Service 1," "AI Service 2," and "AI Service 3."

[1432] Step 4:

[1433] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, it might receive responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3".

[1434] Step 5:

[1435] The server evaluates the responses and selects the most appropriate one. A scoring function is applied to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. The response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[1436] Step 6:

[1437] The selected answer will be automatically corrected. The server will perform grammatical checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as necessary. Furthermore, it will add detailed information such as "from July 26, 2024" by referring to reliable sources.

[1438] Step 7:

[1439] The system generates the final answer and related ads. The server retrieves the relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the answer.

[1440] Step 8:

[1441] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[1442] Step 9:

[1443] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[1444] In this way, users can quickly obtain accurate and reliable answers simply by entering their questions.

[1445] (Example 1)

[1446] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1447] Conventional information retrieval systems often provide inaccurate information in response to user inquiries. Even with the use of numerous generative artificial intelligence systems, responses may vary or contain unreliable information. In such situations, it is difficult for users to obtain accurate information quickly. Furthermore, users bear the burden of having to judge the reliability of the information they receive themselves.

[1448] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1449] In this invention, the server includes means for receiving questions from users, means for forwarding received questions to the server, means for sending questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the multiple received answers and selecting the most appropriate answer, means for automatically correcting the selected answer, and means for displaying the corrected final answer and related advertisements. This enables users to quickly obtain accurate and reliable answers.

[1450] A "user" refers to an individual or group that inputs questions into a system and receives answers.

[1451] A "terminal" is a device used by a user to input questions and send them to a server, and specifically includes personal computers and smartphones.

[1452] A "server" refers to a computer system that receives questions from users, sends the questions to multiple generative artificial intelligence systems, evaluates and corrects the received answers, and then sends the final answer to the terminal.

[1453] "Generative artificial intelligence" refers to systems that utilize machine learning and natural language processing technologies to automatically generate answers to user questions.

[1454] "Means of receiving questions" refers to a combination of software and hardware for inputting, storing, and transferring user questions to a server in a digital format.

[1455] "Means for sending questions" refers to communication protocols or API interfaces that allow a server to send questions received from a user to multiple generative artificial intelligence systems.

[1456] "Means of receiving responses" refers to communication protocols and databases used by servers to receive and store responses from generative artificial intelligence.

[1457] "Means of evaluating responses" refers to algorithms or evaluation systems that analyze multiple received responses and select the most appropriate response based on evaluation criteria.

[1458] "Means for correcting answers" refers to programs or libraries used to perform grammatical checks and factual verification on selected answers, and then make corrections or revisions.

[1459] "Means for displaying corrected final answers and related advertisements" refers to software and interfaces for displaying the final answers and related advertisements on the user's device.

[1460] This invention is a system that sends user questions to multiple generative artificial intelligence systems, selects the most accurate answer, automatically corrects it, and provides the user with the final answer and related advertisements. Specific examples of this system are described below.

[1461] This system includes a terminal that receives user questions, a server that processes the received questions and transmits them to multiple generative artificial intelligence systems, and means for evaluating and correcting the answers from the generative artificial intelligence systems and providing the user with the final answers and related advertisements.

[1462] Hardware and Software Overview

[1463] terminal

[1464] A device is a device on which the user enters a question, and primarily includes personal computers, smartphones, and tablets. The device provides an interface for entering the question using a web browser or mobile application.

[1465] server

[1466] This is a central computer system for receiving user questions and transmitting them to multiple generative artificial intelligence systems. The server has the following main functions:

[1467] API interface: A means of communication for sending questions to and receiving answers from generative artificial intelligence.

[1468] Database: A storage device for storing and evaluating received responses.

[1469] Evaluation system: An algorithm that evaluates each answer and selects the most appropriate answer.

[1470] Generative artificial intelligence

[1471] This refers to a system that automatically generates answers to questions sent from a server, and includes Google Cloud AI, OpenAI, and IBM Watson.

[1472] System operation example

[1473] The user types "When is the next Olympics?" into their web browser. At this point, the terminal receives the user's question and forwards it to the server.

[1474] The server converts the received question into an appropriate format and sends an API request to a generative artificial intelligence (AI) such as Google Cloud AI, OpenAI, or IBM Watson. Multiple generative AIs answer the question based on their own algorithms and return the answer to the server.

[1475] The server receives responses from each generative artificial intelligence system and stores them in a database. For example, "AI Service 1" might provide "2024," "AI Service 2" might provide "next year," and "AI Service 3" might provide "July 2024."

[1476] The server evaluates the multiple responses it receives. Evaluation criteria include fact-checking, grammatical accuracy, and reliability. A scoring algorithm is used to assign points to each response, and the response with the highest score is selected. For example, if "July 2024" receives the highest score, it will be selected.

[1477] The server performs grammar checks and fact-checks on the selected answer. Grammar checks utilize language processing libraries, and fact-checking relies on reliable sources. As a result, detailed information such as "from July 26, 2024" is added.

[1478] The corrected final answer is sent to the device, which then displays it to the user. Related advertisements are also displayed simultaneously. For example, an advertisement for sporting goods might appear along with the answer, "The next Olympics are from July 26, 2024."

[1479] Example of a prompt

[1480] Question: "When is the next Olympics?"

[1481] Question: "What is the date of the next Olympic Games?"

[1482] Question: "I would like to know the dates of the Olympics."

[1483] As described above, this system allows users to quickly obtain accurate and reliable answers.

[1484] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1485] Step 1:

[1486] The user enters a question.

[1487] Input: The user enters the question via a web browser or mobile application.

[1488] Action: The user types "When are the next Olympics?" into the browser's search form.

[1489] Data processing: Format the input text data into JSON format.

[1490] Output: The formatted question data is temporarily stored on the device.

[1491] Step 2:

[1492] The terminal forwards the question to the server.

[1493] Input: Question data entered by the user (in JSON format).

[1494] Operation: The device uses the JavaScript fetch API to send the question data to the server as an HTTP request.

[1495] Data processing: Include the question data in the HTTP request body.

[1496] Output: Question data arrives at the server and the server receives it.

[1497] Step 3:

[1498] The server sends questions to multiple generative artificial intelligence systems.

[1499] Input: Question data received from the terminal.

[1500] Operation: The server processes the received question data into an API request format.

[1501] Data processing: Convert the question data into a format compatible with the API of each generative artificial intelligence system. For example, convert from JSON to URL-encoded format.

[1502] Output: The converted question data is sent as an HTTP request to each generative artificial intelligence system.

[1503] Step 4:

[1504] The server receives responses from each generative artificial intelligence system.

[1505] Input: Response data from each generative artificial intelligence.

[1506] Operation: The server asynchronously waits for and receives responses from each generative artificial intelligence system.

[1507] Data processing: Store the received response data in a list.

[1508] Output: A list containing the answers of each generative artificial intelligence.

[1509] Step 5:

[1510] The server evaluates the received responses and selects the most appropriate one.

[1511] Input: Multiple response data (lists) from various generative artificial intelligence systems.

[1512] Operation: The server uses an evaluation algorithm to score each response.

[1513] Data processing: A scoring function is applied to each response to calculate an evaluation score based on reliability and grammatical accuracy.

[1514] Output: Select the answer with the highest rating.

[1515] Step 6:

[1516] The server automatically corrects the selected answer.

[1517] Input: The most appropriate response data.

[1518] Operation: The server performs grammar checks and fact-checks on the selected answer. For grammar checks, it uses, for example, the Grammarly API, and for fact-checks, it refers to reliable sources.

[1519] Data processing: This involves correcting grammar and adding detailed information.

[1520] Output: Corrected final response data.

[1521] Step 7:

[1522] The device displays the final answer and related ads.

[1523] Input: Final response data and related advertising data sent from the server.

[1524] Function: Displays data received by the device on web pages and applications. Dynamically updates information using HTML and JavaScript.

[1525] Data processing: Convert the final response data and advertising data into a format that is compatible with the user interface.

[1526] Output: The user is shown the information "The next Olympic Games will be held from July 26, 2024" and advertisements for related sporting goods.

[1527] Through the steps outlined above, users can quickly obtain accurate and reliable answers.

[1528] (Application Example 1)

[1529] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1530] Conventional generative artificial intelligence systems have a problem in that they do not provide accurate and appropriate answers to user questions with sufficient precision. Furthermore, they lack the ability to appropriately recommend related information and content related to the generated answers, resulting in a situation where user needs and convenience are not fully met.

[1531] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1532] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related information, and means for providing related content based on the displayed information. This enables the user to obtain accurate and reliable answers and simultaneously acquire related content.

[1533] A "user" refers to each individual who uses this system.

[1534] "Means for receiving questions" refers to a device or software that receives input from a user and processes its contents.

[1535] "Generative artificial intelligence" refers to AI models that can generate answers to given questions or inputs.

[1536] "Means for sending questions" refers to a device or software that transmits questions received from a user to multiple generative artificial intelligence systems.

[1537] "Means for receiving responses" refers to a device or software for receiving responses generated by generative artificial intelligence.

[1538] "Means for evaluating responses" refers to a device or software for evaluating multiple received responses and determining which is the most appropriate.

[1539] "Means for selecting the most appropriate answer" refers to a device or software for selecting the optimal answer based on evaluation.

[1540] "Means for automatically correcting answers" refers to a device or software for automatically correcting selected answers based on grammar or facts.

[1541] "Means for displaying the final answer and related information" refers to a device or software for displaying the corrected final answer and related information to the user.

[1542] "Means for providing related content" refers to a device or software for providing additional content to the user based on the displayed information.

[1543] This invention is a system that processes user questions in real time, provides the most appropriate answers, and recommends relevant content. Specific embodiments of this system are described below.

[1544] The system primarily consists of a server, terminals, and users. Users submit questions via their terminals (such as smartphones and tablets). A dedicated application installed on the terminal receives questions from users and forwards them to the server. This application also provides an interface for users to input their questions.

[1545] The server sends questions received from users to multiple generative artificial intelligence (AI) models and receives responses from each AI model. In this process, the server uses OpenAI's GPT-3 API or other AI service technologies to send questions. For example, if a user asks, "What are some recommended new movies?", the server sends this question to the AI ​​models and receives responses from each.

[1546] The server evaluates the multiple responses received and selects the most appropriate one. This evaluation considers factors such as accuracy, reliability, and grammatical correctness. The server calculates a score for each response based on the set evaluation criteria and selects the response with the highest score. The server then automatically performs grammatical checks and fact-checks on the selected response and corrects it to its final form.

[1547] The final corrected answer and related information are sent to the device. The device displays related content to the user along with the final answer. This related content might include, for example, a link to watch the movie or a list of related movies in the case of a question about movie recommendations. The device uses web scraping technology to retrieve and provide the related content to the user.

[1548] Hardware and software to be used

[1549] Hardware: Servers, smartphones, tablets

[1550] Software: OpenAI GPT-3 API, requests, BeautifulSoup, dedicated application (for smartphones and tablets)

[1551] Specific example

[1552] The following shows the specific processing that occurs when a user asks, "What are some of the latest movies you recommend?"

[1553] 1. The user enters "What are some recommended new movies?" into their device.

[1554] 2. The terminal forwards this question to the server.

[1555] 3. The server sends questions to multiple generative AI models and receives responses.

[1556] 4. The server selects the best response from those received and performs grammatical checks and fact-checks.

[1557] 5. The server sends the corrected response and related content to the device.

[1558] 6. The device displays to the user a link to watch the relevant movie and reviews along with the final answer.

[1559] Examples of prompt statements

[1560] "What are some of the latest movies you recommend?"

[1561] "What are your recommended travel destinations this year?"

[1562] "When is the next Olympics?"

[1563] In this way, users can quickly obtain accurate and reliable answers, along with useful content related to those answers. This system can meet user needs and significantly improve convenience.

[1564] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1565] Step 1:

[1566] The user enters a question.

[1567] The user uses a device (such as a smartphone or tablet) to enter a question into the application's input form. For example, they might enter, "What are some recommended new movies?" This input is then accepted by the device.

[1568] Input: A question entered by the user (e.g., "What are some recommended new movies?")

[1569] Output: Question text data

[1570] Step 2:

[1571] The terminal forwards the question to the server.

[1572] The application installed on the device transfers the received question text data to the server.

[1573] Input: Question text data

[1574] Output: Question data transferred to the server

[1575] Step 3:

[1576] The server sends questions to each generative artificial intelligence system.

[1577] The server converts the received question into the appropriate format and sends the question to multiple generative artificial intelligence (AI) models. Specifically, it makes asynchronous requests to the API endpoints of each AI service.

[1578] Input: Question data

[1579] Output: Question requests sent to multiple AI models

[1580] Step 4:

[1581] Receive an answer from a generative artificial intelligence.

[1582] The server receives responses from each generative artificial intelligence system and stores each response in a list.

[1583] Input: Request for response from AI model

[1584] Output: List of answers (Example: ["Movie A", "Movie B", "Best Movie C"])

[1585] Step 5:

[1586] The server selects the best answer.

[1587] The server evaluates multiple received responses and selects the most appropriate response based on predefined evaluation criteria. These criteria include accuracy, reliability, and grammatical correctness. A scoring function is used to calculate evaluation scores, and the response with the highest score is selected.

[1588] Input: Answer list

[1589] Output: Selected best answer (e.g., "Best movie C")

[1590] Step 6:

[1591] Correct the selected answer.

[1592] The server performs grammatical checks and fact-checks on the selected answer and makes corrections as needed. For example, it uses an online grammar checking service to check the grammatical structure and consults relevant databases for fact-checking.

[1593] Input: Selected answer

[1594] Output: Corrected final answer (e.g., "The best movie recommendation is the latest movie C")

[1595] Step 7:

[1596] The server retrieves the relevant content.

[1597] Based on the finalized response, the server retrieves relevant content from the web. Specifically, it uses web scraping techniques to collect links to watch related movies and reviews.

[1598] Input: Corrected final answer

[1599] Output: Related content (e.g., viewing links, reviews)

[1600] Step 8:

[1601] The device displays the final answer and related content.

[1602] The terminal receives the corrected final answer and related content sent from the server and displays it to the user.

[1603] Input: Final answer and related content

[1604] Output: The final answer and related content displayed to the user (e.g., "The best movie recommendation is the latest movie C. Here is the viewing link and review.")

[1605] As a result, users can quickly obtain accurate and reliable answers, along with relevant and useful content.

[1606] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1607] This invention provides a system that uses multiple generative artificial intelligence systems to provide the most appropriate answer to a user's question, and improves the quality of the answer by combining it with an emotion engine that recognizes the user's emotions. Specific examples of this system are described below.

[1608] Recognizing user emotions

[1609] As the device receives a question from the user via an input form, the emotion engine recognizes the user's emotions. For example, if a user types "When are the next Olympics?", the emotion engine analyzes the user's tone of voice and facial expressions to determine whether the user is feeling curiosity, anticipation, or excitement. This data is analyzed in real time in response to text input, and as a result, an emotion status is generated.

[1610] Receive a question and add sentiment data.

[1611] The device sends the sentiment status generated by the sentiment engine to the server along with the user's question. This status is processed together with the question.

[1612] Send a question to multiple AIs

[1613] The server converts the question and associated sentiment data into an appropriate format and sends it to multiple generative artificial intelligence (AI) services. Asynchronous requests are made to the API endpoint of each AI service, and each AI uses its own algorithm to generate an answer to the question.

[1614] Receive and evaluate the responses from each AI.

[1615] The server receives responses from each generative artificial intelligence system and stores them in a list. For example, responses such as "2024" from "AI Service 1," "next year" from "AI Service 2," and "July 2024" from "AI Service 3" might be obtained. During this process, the emotional status is also considered and influences the evaluation criteria.

[1616] Select the most appropriate answer.

[1617] The server evaluates multiple responses received and selects the most appropriate one. The evaluation also takes into account the user's emotional status; for example, if the user is excited, they are more likely to choose a more detailed and future-oriented response (such as "July 2024").

[1618] Automatically correct the selected answer.

[1619] The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the server performs a grammar check on this answer and corrects it as needed. Furthermore, it adds detailed information such as "from July 26, 2024" by referring to reliable sources and adjusts the tone of the sentence to match the sentiment status.

[1620] Generate related ads based on the final answer

[1621] The server retrieves relevant ads based on the corrected final response. It uses the ad delivery service's API to search for relevant ads and provide them along with the response. For example, a user with an excited emotional status might see ads for sports equipment or event tickets.

[1622] Display the final answer and related ads.

[1623] The corrected final answer is sent from the server to the device, and the device displays the final answer and related advertisements to the user. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[1624] Specific example

[1625] If a user asks, "When is the next Olympics?":

[1626] 1. The device receives the question, and the emotion engine analyzes the user's emotions. It recognizes that the user is agitated.

[1627] 2. The device sends the question and sentiment status to the server.

[1628] 3. The server sends questions to multiple generative artificial intelligence systems.

[1629] 4. Each AI will respond with answers such as "2024," "next year," or "July 2024." Emotional status will also be included in the evaluation.

[1630] 5. The server selects the most appropriate answer, "July 2024". Additional date details are provided based on the sentiment status.

[1631] 6. Perform grammatical checks and fact-checks on the selected answers, and add the detailed information, "from July 26, 2024."

[1632] 7. The server retrieves relevant ads based on the sentiment status.

[1633] 8. The device displays the final answer, "The next Olympics are from July 26, 2024," along with related advertisements. For example, an excited user might see an advertisement for sports equipment.

[1634] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[1635] The following describes the processing flow.

[1636] Step 1:

[1637] The user enters a question. The user enters the question into an input form in a web browser or mobile application. For example, they might enter, "When are the next Olympics?" The device receives this question and temporarily saves the input.

[1638] Step 2:

[1639] The device recognizes the user's emotions. The emotion engine analyzes the user's voice tone, facial expressions, keyboard typing style, etc., to generate an emotional status. For example, it recognizes when the user is asking a question with anticipation or is excited. This emotional information is also temporarily stored.

[1640] Step 3:

[1641] The device sends the question and sentiment status to the server. Sentiment information is securely transmitted to the server along with the question content using network communication.

[1642] Step 4:

[1643] The server receives the question and sentiment status, and sends the question to multiple generative artificial intelligence (AI) services. It converts the received question into the appropriate format and makes asynchronous requests to the API endpoints of multiple generative artificial intelligence (AI) services. For example, it sends the question to "AI Service 1," "AI Service 2," and "AI Service 3."

[1644] Step 5:

[1645] The server receives responses from each generative artificial intelligence. The server receives the responses from each AI and stores them in a list. For example, responses such as "2024" from "AI Service 1", "next year" from "AI Service 2", and "July 2024" from "AI Service 3" might be obtained.

[1646] Step 6:

[1647] The server evaluates the responses and selects the most appropriate one. It applies a scoring function to evaluate multiple received responses. Evaluation criteria include fact-checking, grammatical accuracy, reliability, and even sentiment status. For example, if the user is excited, detailed, future-oriented responses (such as "July 2024") will receive higher scores.

[1648] Step 7:

[1649] The selected answer is automatically corrected. The server performs grammar checks and fact-checks on the selected answer. For example, if "July 2024" is selected, the answer will be grammatically checked and corrected as needed. Furthermore, it will refer to reliable sources and add more detailed information such as "from July 26, 2024". The text may also be corrected to match the tone and expression of the sentiment status.

[1650] Step 8:

[1651] The system generates the final answer and relevant ads. The server retrieves relevant ads based on the corrected final answer. It uses the ad delivery service's API to search for relevant ads and provide them along with the final answer. For example, users with an excited emotion status might see ads for sports goods or event tickets.

[1652] Step 9:

[1653] The server sends the final answer and related advertisements to the device. The corrected answer and related advertisements are then transferred to the device together.

[1654] Step 10:

[1655] The device displays the final answer and related advertisements to the user. The device displays the received information on the screen. For example, the answer "The next Olympics are from July 26, 2024" might be displayed, along with advertisements for related sports equipment.

[1656] In this way, users can simply enter a question and quickly receive accurate, reliable answers that are tailored to their emotions.

[1657] (Example 2)

[1658] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1659] Conventional question-answering systems have the problem of not providing answers that take user emotions into consideration, making it difficult for users to obtain satisfactory answers. Furthermore, when selecting the most appropriate answer from answers obtained from multiple generative AI systems, the evaluation does not include emotional data, which can lead to a decrease in the quality of the answers. In addition, when providing related information or advertisements, there is a problem in that the content displayed is not appropriate based on the user's emotional state.

[1660] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1661] In this invention, the server includes means for recognizing the user's emotions, means for attaching emotion data to a question and transmitting it, means for evaluating multiple received answers and selecting the most appropriate answer, and means for performing grammatical checks and fact-checking on the selected answer. This makes it possible to provide the most appropriate answer obtained from multiple generative artificial intelligence systems while taking the user's emotions into consideration.

[1662] A "user" is someone who uses the system to input questions and receive answers.

[1663] A "question" refers to the information or points of concern that a user enters into the system.

[1664] "Emotions" refer to the user's psychological state and include curiosity, anticipation, excitement, and so on.

[1665] "Emotional data" refers to information generated by analyzing the user's emotions and is added to the questions.

[1666] "Generative artificial intelligence" refers to algorithms and systems that generate natural language in a way that mimics human speech.

[1667] "Grammar check" is the process of verifying whether a text uses correct grammar.

[1668] "Fact-checking" is the process of verifying whether information is accurate and reliable.

[1669] "Evaluation criteria" are standards for evaluating multiple responses and pieces of information, and they take into account user sentiment data.

[1670] An "evaluation score" is a numerical representation of how appropriate each response is based on the evaluation criteria.

[1671] "Advertisements" are those that provide information about products or services related to the final answer.

[1672] An "asynchronous request" is a method of sending requests to multiple AI services simultaneously and processing them independently.

[1673] This invention is a system that uses multiple generative artificial intelligence systems to provide the most appropriate answers to user questions. This system includes functions to recognize the user's emotions, generate answers while considering that emotion data, and also provide relevant advertisements. The specific implementation of this system is described below.

[1674] System configuration:

[1675] This system consists of terminals, servers, and multiple generative artificial intelligence (AI) services. The main hardware and software used are as follows:

[1676] 1. Device (PC or smartphone): This is the device used by the user to input questions and view the final answers. It is desirable that it be equipped with a camera and microphone for analyzing voice tone and facial expressions.

[1677] 2. Server: This is the central computing unit that performs data processing, AI integration, question and answer evaluation, grammar checking, fact-checking, and relevant advertisement retrieval.

[1678] 3. Emotion Engine: This is a software tool for analyzing user emotions. Specifically, it uses speech recognition software or facial recognition tools (e.g., Google Speech-to-Text or Microsoft Azure Face API).

[1679] 4. Generative artificial intelligence services: Algorithms for generating answers to questions. Specifically, this includes OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[1680] Operation details:

[1681] The system operates as follows:

[1682] 1. The user enters the question:

[1683] Enter the question into the input form on the terminal. For example, a question like, "When is the next Olympics?"

[1684] 2. The device analyzes emotions:

[1685] The system analyzes the user's voice tone and facial expressions using an emotion engine, and the results are obtained as emotion data. For example, if the user is asked "When is the next Olympics?", the system will analyze that the user is excited.

[1686] 3. The device sends the question and sentiment data to the server:

[1687] The device sends the analysis results (question and emotion status) to the server as a data packet. Specifically, data such as "Question: When are the next Olympics?" and "Emotion Status: Excited" is sent in JSON format.

[1688] 4. The server sends questions to multiple generative AI systems:

[1689] The server converts the received question and sentiment data into an appropriate format and sends asynchronous requests to multiple generative AI services. For example, it sends the question "When are the next Olympics?" to OpenAI's GPT-3, Google's BERT, Microsoft's Turing, and others.

[1690] 5. Evaluate the answers obtained from each AI:

[1691] The server collects the responses returned by each generative AI and stores them in a list. Each response is something like "2024," "next year," or "July 2024." During this process, sentiment data is taken into consideration to select the most appropriate response.

[1692] 6. Grammatical check and fact-checking of selected answers:

[1693] The server grammatically checks the selected answer and fact-checks it by referring to reliable sources. For example, if "July 2024" is selected, it will refer to the official Olympic website and add the specific date "July 26, 2024".

[1694] 7. Obtaining related advertisements:

[1695] The server retrieves relevant ads based on the user's emotions. For example, excited users are shown ads for sports equipment or event tickets.

[1696] 8. Display of the final answer and related ads:

[1697] The corrected final answer is sent from the server to the device, which then displays it to the user. Relevant advertisements are also displayed along with the answer.

[1698] Specific example:

[1699] When a user asks "When is the next Olympics?", the specific actions taken are as follows:

[1700] 1. The user enters "When is the next Olympics?" into their device.

[1701] 2. The device uses an emotion engine to analyze the user's emotions (e.g., excited).

[1702] 3. The device sends the question and sentiment data to the server.

[1703] 4. The server sends questions to multiple generative AI systems (e.g., OpenAI's GPT-3, Google's BERT, Microsoft's Turing).

[1704] 5. Collect responses from each AI and select the most appropriate response (e.g., "July 2024").

[1705] 6. Check the grammar and facts of the answer (e.g., add any supplementary information to "July 26, 2024").

[1706] 7. Obtain relevant ads based on emotions (e.g., ads for sporting goods).

[1707] 8. The device displays an advertisement with the final answer: "The next Olympics will be held from July 26, 2024."

[1708] Example of a prompt:

[1709] "When is the next Olympics? A question from an excited user."

[1710] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1711] Step 1:

[1712] The user enters a question.

[1713] Operation: The user enters a question through the input form on their device. For example, they might enter, "When are the next Olympics?"

[1714] Input: The user enters a question into the input form.

[1715] Output: The terminal retrieves the user's question data.

[1716] Step 2:

[1717] The device analyzes the question, voice tone, and facial expressions.

[1718] Operation: The device uses an emotion engine to analyze the user's voice tone and facial expressions in real time. Specifically, it captures voice with a microphone and analyzes facial expressions with a camera. Software used includes speech recognition tools (e.g., Google Speech-to-Text) and face recognition APIs (e.g., Microsoft Azure Face API).

[1719] Input: User's questions and data on voice tone and facial expressions.

[1720] Output: Emotional status analyzed by the emotion engine (e.g., excitement, anticipation, etc.).

[1721] Step 3:

[1722] The device sends the question and sentiment status to the server.

[1723] Operation: The terminal sends the question and the obtained sentiment status to the server as data packets. Possible transmission formats include JSON.

[1724] Input: User's question and analyzed sentiment status.

[1725] Output: Question and sentiment data sent to the server in JSON format, etc.

[1726] Step 4:

[1727] The server sends questions to multiple generative AIs.

[1728] Operation: The server sends received question and sentiment data as asynchronous requests to multiple generative artificial intelligence systems. Recipients include OpenAI's GPT models, Google's BERT, and Microsoft's Turing.

[1729] Input: User question and sentiment data received by the server.

[1730] Output: Question and sentiment data sent to a generative AI service.

[1731] Step 5:

[1732] Receive and evaluate responses from each generative AI.

[1733] Operation: The server receives responses from each generative AI and stores each response in a list. It sets evaluation criteria considering the emotional status and evaluates the best response.

[1734] Input: Responses and emotional status from each generative AI.

[1735] Output: Store the most appropriate answer in a list and calculate the evaluation score.

[1736] Step 6:

[1737] Select the most appropriate answer and perform grammatical checks and fact-checks.

[1738] Operation: The server selects the most appropriate answer based on the evaluation score and checks it using a grammar checking tool (e.g., Grammarly API) and fact-checking sources (e.g., the official Olympic website).

[1739] Input: The answer with the highest rating score.

[1740] Output: Corrected answer after grammar check and fact-checking (e.g., "from July 26, 2024").

[1741] Step 7:

[1742] Get relevant ads based on your emotional status

[1743] Operation: The server considers the user's emotional state and retrieves relevant ads from the ad delivery system (e.g., Google AdSense API).

[1744] Input: Emotional status and optimal response.

[1745] Output: Relevant ads selected based on emotional status.

[1746] Step 8:

[1747] Send the final answer and related advertisements to your device and display them.

[1748] Operation: The corrected final answer and related advertisements are sent to the device, which then displays them to the user. For example, an advertisement for sports equipment might be displayed along with the answer, "The next Olympics are from July 26, 2024."

[1749] Input: Corrected final answer and related ads.

[1750] Output: The final answer and related advertisements displayed on the user's device.

[1751] (Application Example 2)

[1752] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1753] Conventional generative artificial intelligence systems often lacked sufficient accuracy and quality in their responses to user questions. A particular challenge was the decrease in user satisfaction due to responses generated without considering the user's emotions. Furthermore, the process of appropriately evaluating responses from multiple generative AI systems and selecting the optimal answer was complex, and efficiency improvements were needed. Additionally, there was the difficulty in effectively presenting relevant products that responded to the user's emotions when suggesting products in virtual stores.

[1754] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1755] In this invention, the server includes means for receiving questions from a user, means for transmitting questions to multiple generative artificial intelligence systems, means for receiving answers from each generative artificial intelligence system, means for evaluating the received multiple answers and selecting the most appropriate answer, means for automatically correcting the selected answer, means for displaying the corrected final answer and related advertisements, means for recognizing the user's emotions, means for transmitting questions containing emotion data to the generative artificial intelligence systems, and means for selecting the optimal answer based on the emotion data. This enables the generation of optimal answers that take the user's emotions into consideration and the effective suggestion of related products based on those answers.

[1756] "Means for receiving user questions" refers to a device or program that has the function of receiving question information entered by a user.

[1757] "Means for sending questions to multiple generative artificial intelligence systems" refers to a device or program that has the function of sending a user's questions to multiple generative AI services.

[1758] "Means for receiving responses from each generative artificial intelligence" refers to a device or program that has the function of receiving response information returned from a generative AI service.

[1759] "A means of evaluating multiple received responses and selecting the most appropriate response" refers to a device or program that has the function of analyzing and evaluating responses from multiple generative AI systems and selecting the most appropriate one.

[1760] "Means for automatically correcting selected answers" refers to a device or program that has the function of automatically performing grammatical checks and factual verification on selected answers and correcting them as necessary.

[1761] "Means for displaying corrected final answers and related advertisements" refers to a device or program that has the function of presenting corrected answers and related advertisements to a user.

[1762] "Means for recognizing user emotions" refers to a device or program that analyzes a user's voice and facial expressions to determine their emotions.

[1763] "Means for sending questions containing emotional data to a generative artificial intelligence system" refers to a device or program that has the function of adding emotional data to a user's question and sending it to a generative AI service.

[1764] "Means for selecting the optimal response based on emotional data" refers to a device or program that has the function of evaluating responses from multiple generative AI systems, taking emotional data into consideration, and selecting the best one.

[1765] "Means for performing grammatical checks" refers to a device or program that has the function of verifying and correcting the grammatical consistency of a text.

[1766] "Means of fact-checking" refers to a device or program that has the function of verifying the content of a received response based on reliable sources.

[1767] "Means for setting evaluation criteria for selecting the optimal answer" refers to a device or program that has the function of setting criteria (e.g., accuracy, relevance, etc.) for selecting the best answer from among multiple answers.

[1768] "Means for calculating the evaluation score for each response" refers to a device or program that has the function of assigning a score to each response based on established evaluation criteria.

[1769] "Means for selecting the highest-rated response based on scores" refers to a device or program that has the function of selecting the highest-rated response based on evaluation scores.

[1770] "Means for setting evaluation criteria that take emotional data into consideration" refers to a device or program that has the function of setting criteria for evaluating responses while taking into consideration the user's emotional data.

[1771] This invention is a system that improves the quality of responses by using multiple generative artificial intelligence systems to provide the most appropriate answers to user questions, and by combining this with an emotion engine that recognizes the user's emotions. This system mainly consists of a server and terminals.

[1772] Hardware and software configuration

[1773] Hardware to use

[1774] Smart glasses, head-mounted displays, or smartphones

[1775] Software to use

[1776] Emotion recognition libraries (e.g., Microsoft Azure Emotion Recognition API)

[1777] Generative AI libraries (e.g., OpenAI, Google AI)

[1778] Programming language: Python

[1779] System operation

[1780] User question received

[1781] Users ask questions about products within the virtual store using a device (smart glasses, head-mounted display, or smartphone). The device receives the question via an input form, and an emotion engine analyzes the user's voice tone and facial expressions to generate an emotion status. This emotion status is then sent to the server.

[1782] Sending questions and sentiment data

[1783] The server receives the question and sentiment status, converts them into the appropriate format, and sends them to multiple generative artificial intelligences. Asynchronous requests are made to the API endpoint of each generative AI, and each AI generates an answer to the question using its own algorithm.

[1784] Receiving and evaluating responses

[1785] The server receives responses from each generative AI system and stores them in a list. The server evaluates each response, taking into account the user's emotional status, and selects the most appropriate response. The evaluation also takes emotional status into account; if the user is excited, a more detailed and future-oriented response is more likely to be selected.

[1786] Correcting responses and generating ads

[1787] The server performs grammar checks and fact-checks on the selected answers. It then references reliable sources to add further details. Additionally, it utilizes ad delivery service APIs to retrieve relevant ads based on the sentiment status.

[1788] Final answer and ad display

[1789] The corrected final answers and relevant advertisements are sent to the device and displayed to the user. This provides the user with accurate, reliable answers and relevant advertisements that are tailored to their emotions.

[1790] Specific example

[1791] If a user asks "Is this jacket waterproof?" in a virtual store, the following process will occur.

[1792] 1. Use the microphone and camera on the smart glasses to capture the user's questions and facial expressions.

[1793] 2. The emotion engine analyzes the user's voice tone and facial expressions and generates an emotion status of "excited".

[1794] 3. The question and emotion status are sent to the server, and requests are sent to multiple generative AI systems.

[1795] 4. Each generation AI generates a response such as "Yes, it's waterproof," "Yes, it's waterproof tested," or "Yes, it's waterproof and windproof."

[1796] 5. The server, considering its "excited" emotional status, selects the detailed answer "Yes, it is waterproof and windproof."

[1797] 6. As related advertisements, suggest items such as waterproof jackets and waterproof sprays.

[1798] 7. Display the final answer and related advertisements on the smart glasses' display.

[1799] Example input prompts for a generative AI model

[1800] User question: "Is this jacket waterproof?"

[1801] User's emotion: "Excited"

[1802] Please generate variations of the answer.

[1803] This invention makes it possible to generate optimal responses that take user emotions into consideration, and to effectively suggest related products based on those responses.

[1804] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1805] Step 1:

[1806] The user enters a question about the product. A device (smart glasses, head-mounted display, or smartphone) receives this question, and simultaneously, an emotion engine analyzes the user's voice tone and facial expressions. The device generates analyzed emotion data and sends it to the server.

[1807] Input: User's question, user's voice tone, facial expression

[1808] Data processing: The emotion engine analyzes voice tone and facial expressions to generate emotion data.

[1809] Output: Questionnaire data, sentiment data

[1810] Step 2:

[1811] The server receives user questions and sentiment data. It converts the received data into an appropriate format (e.g., JSON) and sends asynchronous requests to the API endpoints of each generative artificial intelligence system.

[1812] Input: Question data, sentiment data

[1813] Data processing: Convert question data and sentiment data to JSON format.

[1814] Output: Request to generative AI service

[1815] Step 3:

[1816] Each generative artificial intelligence generates an answer to a question and returns the answer data to the server. The server receives this answer data.

[1817] Input: Response from a generative AI service

[1818] Data processing: Each generative AI generates answers using its own unique algorithm.

[1819] Output: List of response data

[1820] Step 4:

[1821] The server evaluates multiple response data received. Considering user sentiment data, it calculates a score for each response based on evaluation criteria to select the best answer. The response with the highest score is then selected.

[1822] Input: List of response data, sentiment data

[1823] Data processing: Assign a score to each response based on evaluation criteria.

[1824] Output: Optimal response data

[1825] Step 5:

[1826] The server performs grammar checks and fact-checks on the selected optimal response data. Corrections are made as needed, and the final response is generated.

[1827] Input: Optimal response data

[1828] Data processing: grammar check, fact-checking, correction.

[1829] Output: Corrected final response data

[1830] Step 6:

[1831] The server retrieves relevant advertisements based on the final response data. Using the ad delivery service's API, it searches for advertisements that match the user's sentiment data and selects the most suitable advertisement.

[1832] Input: Final response data, sentiment data

[1833] Data processing: Obtain advertising data from advertising distribution services.

[1834] Output: Related ad data

[1835] Step 7:

[1836] The corrected final response data and related advertising data are sent from the server to the device. The device then displays them to the user, specifically on the display of smart glasses or a head-mounted display.

[1837] Input: Final response data, related ad data

[1838] Data processing: Conversion to display format

[1839] Output: User-viewable responses and advertisements

[1840] This process allows users to quickly receive detailed and reliable answers that address their emotions, and relevant advertisements are presented effectively.

[1841] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1842] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1843] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1844] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1845] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1846] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1847] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1848] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1849] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1850] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1851] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1852] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1853] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1854] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1855] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1856] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1857] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1858] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1859] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1860] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1861] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[1862] The following is further disclosed regarding the embodiments described above.

[1863] (Claim 1)

[1864] A means of receiving questions from users,

[1865] A means of sending questions to multiple generative artificial intelligence systems,

[1866] A means of receiving responses from each generative artificial intelligence system,

[1867] A means of evaluating multiple received responses and selecting the most appropriate response,

[1868] A means to automatically correct the selected answer,

[1869] A means of displaying the corrected final answer and related advertisements,

[1870] A system that includes this.

[1871] (Claim 2)

[1872] A means of performing grammatical checks on the selected answers,

[1873] Means for verifying the facts,

[1874] The system according to claim 1, including the following:

[1875] (Claim 3)

[1876] A means of setting evaluation criteria for selecting the optimal answer,

[1877] A means for calculating the evaluation score for each response,

[1878] A means of selecting the answer with the highest rating based on the score,

[1879] The system according to claim 1, including the following:

[1880] "Example 1"

[1881] (Claim 1)

[1882] A means of receiving questions from users,

[1883] A means of forwarding received questions to the server,

[1884] A means of sending questions to multiple generative artificial intelligence systems,

[1885] A means of receiving responses from each generative artificial intelligence system,

[1886] A means of evaluating multiple received responses and selecting the most appropriate response,

[1887] A means to automatically correct the selected answer,

[1888] A means of displaying the corrected final answer and related advertisements,

[1889] A system that includes this.

[1890] (Claim 2)

[1891] A means of performing grammatical checks on the selected answers,

[1892] Means for verifying the facts,

[1893] The system according to claim 1, including the following:

[1894] (Claim 3)

[1895] A means of setting evaluation criteria for selecting the optimal answer,

[1896] A means for calculating the evaluation score for each response,

[1897] A means of selecting the answer with the highest rating based on the score,

[1898] The system according to claim 1, including the following:

[1899] "Application Example 1"

[1900] (Claim 1)

[1901] A means of receiving questions from users,

[1902] A means of sending questions to multiple generative artificial intelligence systems,

[1903] A means of receiving responses from each generative artificial intelligence system,

[1904] A means of evaluating multiple received responses and selecting the most appropriate response,

[1905] A means to automatically correct the selected answer,

[1906] A means of displaying the corrected final answer and related information,

[1907] Means for providing related content based on the displayed information,

[1908] A system that includes this.

[1909] (Claim 2)

[1910] A means of performing grammatical checks on the selected answers,

[1911] Means for verifying the facts,

[1912] The system according to claim 1, including the following:

[1913] (Claim 3)

[1914] A means of setting evaluation criteria for selecting the optimal answer,

[1915] A means for calculating the evaluation score for each response,

[1916] A means of selecting the answer with the highest rating based on the score,

[1917] Means for obtaining and displaying related content,

[1918] The system according to claim 1, including the following:

[1919] "Example 2 of combining an emotion engine"

[1920] (Claim 1)

[1921] A means of receiving questions from users,

[1922] Means of recognizing user emotions,

[1923] A means of attaching emotional data to a question and sending it,

[1924] A means of sending questions to multiple generative artificial intelligence systems,

[1925] A means of receiving responses from each generative artificial intelligence system,

[1926] A means of evaluating multiple received responses and selecting the most appropriate response,

[1927] A means of performing grammatical checks and fact-checking on the selected answers,

[1928] A means to automatically correct the selected answer,

[1929] A means of displaying the corrected final answer and related advertisements,

[1930] A system that includes this.

[1931] (Claim 2)

[1932] A means of considering user sentiment data in the evaluation criteria,

[1933] A means of obtaining relevant ads based on sentiment data,

[1934] The system according to claim 1, including the following:

[1935] (Claim 3)

[1936] A means of setting evaluation criteria for selecting the optimal answer,

[1937] A means for calculating the evaluation score for each response,

[1938] A means of selecting the answer with the highest rating based on the score,

[1939] The system according to claim 1, including the following:

[1940] "Application example 2 of combining emotional engines"

[1941] (Claim 1)

[1942] A means of receiving questions from users,

[1943] A means of sending questions to multiple generative artificial intelligence systems,

[1944] A means of receiving responses from each generative artificial intelligence system,

[1945] A means of evaluating multiple received responses and selecting the most appropriate response,

[1946] A means to automatically correct the selected answer,

[1947] A means of displaying the corrected final answer and related advertisements,

[1948] Means of recognizing user emotions,

[1949] A means for sending questions containing emotional data to a generative artificial intelligence,

[1950] A means of selecting the optimal answer based on emotional data,

[1951] A system that includes this.

[1952] (Claim 2)

[1953] A means of performing grammatical checks on the selected answers,

[1954] Means for verifying the facts,

[1955] The system according to claim 1, including the following:

[1956] (Claim 3)

[1957] A means of setting evaluation criteria for selecting the optimal answer,

[1958] A means for calculating the evaluation score for each response,

[1959] A means of selecting the answer with the highest rating based on the score,

[1960] A means of setting evaluation criteria that take emotional data into consideration,

[1961] The system according to claim 1, including the following: [Explanation of Symbols]

[1962] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of receiving questions from users, A means of sending questions to multiple generative artificial intelligence systems, A means of receiving responses from each generative artificial intelligence system, A means of evaluating multiple received responses and selecting the most appropriate response, A means to automatically correct the selected answer, A means of displaying the corrected final answer and related advertisements, A system that includes this.

2. A means of performing grammatical checks on the selected answers, Means for verifying the facts, The system according to claim 1, including the following:

3. A means of setting evaluation criteria for selecting the optimal answer, A means for calculating the evaluation score for each response, A means of selecting the answer with the highest rating based on the score, The system according to claim 1, including the following:

Images