System

Abstract

A system is provided.SOLUTION: A system comprising: means for receiving a question input by a user; means for generating an answer with a generative AI model based on the received question; means for evaluating an accuracy of the generated answer; and means for displaying the evaluated accuracy to the user.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] When using generative AI models, the answers may contain errors, making it difficult for users to intuitively judge their accuracy. This poses the risk of spreading misinformation and causing inefficiencies in work. Therefore, there is a need for a system that allows users to easily check the accuracy of generative AI models' answers. [Means for solving the problem]

[0005] The present invention relates to a system that includes a means for receiving a question entered by a user, a means for generating an answer using a generative AI model based on the received question, a means for evaluating the accuracy of the generated answer, and a means for displaying the evaluated accuracy to the user. The system uses an internal database and statistical models to evaluate the accuracy. The evaluated accuracy information is also integrated into the answer and displayed to the user. This allows the user to easily determine the accuracy of the answer, preventing the spread of misinformation and improving work efficiency.

[0006] "User" refers to an entity that uses the system to enter questions and receive answers.

[0007] A "question" refers to text data that linguistically expresses the information that a user wants to know.

[0008] "Means for receiving" refers to the function of acquiring questions sent by users and inputting them into the system.

[0009] A "generative AI model" refers to an artificial intelligence model that generates appropriate answers based on input questions.

[0010] "Answer" refers to the text data of the response generated by the generative AI model in response to a question.

[0011] "Means for assessing accuracy" refers to the function of quantitatively analyzing and assessing the accuracy and reliability of the generated answers.

[0012] "Internal database" refers to the accumulated collection of data used as reference by the generative AI model and accuracy assessment module.

[0013] A "statistical model" refers to a mathematical method or algorithm used to analyze or make predictions based on data.

[0014] "Means for displaying to the user" refers to an interface function for visually conveying the evaluated answers and their accuracy information to the user.

[0015] "Accuracy information" refers to quantitative evaluation results that indicate the reliability of the generated answers. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0024] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0037] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user.

[0038] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[0039] User Perspective

[0040] A user inputs a question using the interface of a device (e.g., a PC or smartphone), for example, "Tell me about the planets in the solar system."

[0041] Terminal perspective

[0042] The terminal receives the question entered by the user, sends it to the server, usually in the form of an HTTP request, and prepares to receive a response from the server.

[0043] Server Perspective

[0044] 1. Receiving Questions

[0045] The server receives the question sent from the device. This is called reception processing. After receiving the question, it analyzes it appropriately and prepares it in a format that can be passed to the generative AI model.

[0046] 2. Answer Generation

[0047] The server passes the question to the generative AI model, which generates an answer. The generative AI model uses its internal knowledge base and algorithms to create the best answer to the question. For example, it generates the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0048] 3. Evaluating the accuracy of answers

[0049] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[0050] 4. Submitting the evaluation results

[0051] The server combines the evaluated accuracy with the answer and sends it to the device. For example, it creates data such as "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0052] Terminal view (again)

[0053] The terminal receives the answer and accuracy information sent from the server and displays them to the user, who can then judge the reliability of the answer.

[0054] Specific examples

[0055] For example, if a user asks, "When was the first moon landing?", the server receives this question and passes it to the generative AI model. The generative AI model generates the answer, "The first moon landing took place in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then sends the data, "The first moon landing took place in 1969. (Answer accuracy: 90%)," to the device, which displays it to the user. The user sees this information and determines that the answer is relatively reliable.

[0056] This system allows users to easily determine the accuracy of the answers generated by the AI ​​model, preventing the spread of misinformation.

[0057] The processing flow will be explained below.

[0058] Step 1:

[0059] The user types a question into an input field on the device, for example, "Tell me about the planets in the solar system."

[0060] Step 2:

[0061] The terminal sends the user's input as text data to the server using an HTTP request.

[0062] Step 3:

[0063] The server receives the question sent from the device, analyzes it, and prepares it in a format that can be passed to the generative AI model.

[0064] Step 4:

[0065] The server passes the analyzed question to the generative AI model and requests it to generate an answer. For example, it passes a question like, "Tell me about the planets in the solar system."

[0066] Step 5:

[0067] A generative AI model generates an answer based on the question, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0068] Step 6:

[0069] The server receives the answers generated by the generative AI model.

[0070] Step 7:

[0071] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[0072] Step 8:

[0073] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[0074] Step 9:

[0075] The server receives the evaluation results from the accuracy evaluation module.

[0076] Step 10:

[0077] The server combines the generated answer and its accuracy assessment to create data to be sent to the device. For example, it generates data such as "Answer: There are eight planets in the solar system. (Answer accuracy: 95%)."

[0078] Step 11:

[0079] The server transmits the integrated data to the terminal.

[0080] Step 12:

[0081] The device displays the answer it received and its accuracy to the user, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0082] Step 13:

[0083] The user checks the information displayed on the device and determines the reliability of the answer.

[0084] Example 1

[0085] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0086] Previous generative AI models lacked mechanisms for evaluating the accuracy and reliability of their answers, making it difficult for users to determine how trustworthy the answers were. Furthermore, there was a lack of means to evaluate the accuracy of generated answers and display them in an easy-to-understand manner, making it impossible to prevent the spread of misinformation.

[0087] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0088] In this invention, the server includes means for receiving text input by a user through a terminal, means for formatting the received text so that it can be passed to a generative AI model, means for generating an answer using the generative AI model, means for passing the generated answer to an accuracy evaluation module to evaluate its accuracy, and means for integrating the evaluated accuracy into the answer and sending it to the terminal. This makes it possible to evaluate the accuracy of the answer of the generative AI model and display it in an easy-to-understand manner for the user.

[0089] "User" refers to the entity that operates the system and inputs questions.

[0090] A "terminal" is an input device used by a user, and includes a personal computer, a smartphone, etc.

[0091] The "means for receiving" refers to a mechanism for inputting a question entered by a user into the system.

[0092] "Server" refers to the back-end computer that parses questions, passes them to the generative AI model, and generates and evaluates answers.

[0093] "Means of analysis" refers to the mechanism that converts the question into a form that the generative AI model can understand.

[0094] A "generative AI model" refers to an artificial intelligence model that generates answers to input questions.

[0095] "Means for generating an answer" refers to a mechanism that uses a generative AI model to create an answer to a question.

[0096] "Accuracy Assessment Module" refers to a rating system for measuring the accuracy of generated answers.

[0097] The "means for evaluating accuracy" refers to a mechanism that outputs the reliability of the generated answer as an evaluation value.

[0098] "Means for sending" refers to a communication means for sending the evaluated answers to the terminal.

[0099] The "means for displaying" refers to a mechanism for visually showing the answer including the evaluation result to the user.

[0100] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system transmits a question entered by a user through a terminal to a server, generates an answer to the question using a generative AI model, evaluates the accuracy of the answer, and displays it to the user.

[0101] The overall system configuration consists of three main components: the user, the terminal, and the server. The specific operation of each component is explained below.

[0102] User Perspective

[0103] A user inputs a question using a device such as a personal computer or smartphone. For example, the user inputs a question such as, "Tell me about the planets in the solar system."

[0104] Terminal perspective

[0105] The terminal receives the question entered by the user and sends it to the server, usually using a protocol called an HTTP request, and prepares to receive a response from the server.

[0106] Server Perspective

[0107] The server proceeds with the process in the following steps.

[0108] 1. Receiving and analyzing questions

[0109] The server receives the query sent from the terminal.

[0110] The received question is analyzed and formatted into a form that can be passed to the generative AI model.

[0111] 2. Answer Generation

[0112] The server inputs the formulated question into a generative AI model to generate an answer.

[0113] Large-scale language models such as GPT-3 are used as generative AI models.

[0114] For example, generate the answer "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0115] 3. Evaluating the accuracy of answers

[0116] The generated answer is passed to an accuracy evaluation module to evaluate its accuracy.

[0117] Accuracy assessment uses internal databases and statistical models.

[0118] For example, the evaluation result may be "Answer accuracy: 95%."

[0119] 4. Submitting the evaluation results

[0120] The server generates the data "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)" and sends it to the terminal.

[0121] Specific examples

[0122] For example, if a user asks, "When was the first moon landing?" the system will act as follows:

[0123] 1. User types, "When was the first moon landing?"

[0124] 2. The device sends a question to the server.

[0125] 3. The server receives the question and formats it for passing to the generative AI model.

[0126] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[0127] 5. The accuracy evaluation module evaluates the accuracy of this answer and obtains the result "Answer accuracy: 90%".

[0128] 6. The server generates the data "The first moon landing took place in 1969. (Answer accuracy: 90%)" and sends it to the device.

[0129] 7. The terminal receives this and displays it to the user.

[0130] In this way, users can easily check the accuracy of the generative AI model's answers and obtain reliable information.

[0131] Prompt Sentence Examples

[0132] For example, if a user enters the question "Tell me about the planets in the solar system," the prompt to the generative AI model will be as follows:

[0133] User Question: "Tell me about the planets in the solar system."

[0134] Input to generative AI model: "How many planets are there in the solar system? What are their names?"

[0135] This system makes it possible to evaluate the accuracy of the generative AI model's answers and present them to the user in a clear manner.

[0136] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0137] Step 1:

[0138] The user inputs a question through the terminal.

[0139] Specific behavior:

[0140] Users enter their questions using a computer or smartphone interface.

[0141] For example, type "Tell me about the planets in the solar system" and press the send button.

[0142] input:

[0143] User question: "Tell me about the planets in the solar system."

[0144] output:

[0145] User-entered question data

[0146] Step 2:

[0147] The terminal receives the user's question and sends it to the server.

[0148] Specific behavior:

[0149] The terminal sends the user's question to the server in the form of an HTTP request.

[0150] It is necessary to specify the destination URL or API endpoint.

[0151] input:

[0152] User-entered question data

[0153] output:

[0154] HTTP request sent to the server

[0155] Step 3:

[0156] The server receives the question and formats it for passing to the generative AI model.

[0157] Specific behavior:

[0158] The server receives the HTTP request and analyzes the question.

[0159] The question content is converted into a format that is easy for the generative AI model to understand, such as JSON format.

[0160] For example, reframe the question as, "How many planets are there in the solar system? What are their names?"

[0161] input:

[0162] HTTP request sent to the server

[0163] output:

[0164] Formatted question data

[0165] Step 4:

[0166] The server generates an answer using a generative AI model.

[0167] Specific behavior:

[0168] The server passes the formatted question to the generative AI model.

[0169] Generative AI models (e.g., GPT-3) generate answers using internal knowledge bases and algorithms.

[0170] For example, you might get the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0171] input:

[0172] Formatted question data

[0173] output:

[0174] Answers generated by generative AI models

[0175] Step 5:

[0176] The server passes the generated answers to an accuracy assessment module.

[0177] Specific behavior:

[0178] The server sends the generated answers to an accuracy assessment module.

[0179] An accuracy assessment module examines the answers and compares them against various assessment metrics.

[0180] input:

[0181] Answers generated by generative AI models

[0182] output:

[0183] Data to be evaluated passed to the accuracy evaluation module

[0184] Step 6:

[0185] An accuracy assessment module assesses the accuracy of the answers.

[0186] Specific behavior:

[0187] We use internal databases and statistical models to quantify the accuracy of responses.

[0188] For example, the evaluation result may be "Answer accuracy: 95%."

[0189] input:

[0190] Evaluation data passed to the accuracy evaluation module

[0191] output:

[0192] Evaluation result data (e.g., "Answer accuracy: 95%")

[0193] Step 7:

[0194] The server integrates the evaluation results with the answers and sends them to the terminal.

[0195] Specific behavior:

[0196] The server integrates the evaluation results into an answer and generates the data, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0197] This data is sent to the terminal as an HTTP response.

[0198] input:

[0199] Assessment result data and generated answers

[0200] output:

[0201] Answer and evaluation data sent to the device

[0202] Step 8:

[0203] The device displays the answer and accuracy information to the user.

[0204] Specific behavior:

[0205] The terminal analyzes the data received from the server and displays it in an appropriate format on the user interface.

[0206] For example, the screen might display, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0207] input:

[0208] Answer and evaluation data sent to the device

[0209] output:

[0210] Answer and rating information displayed to users

[0211] (Application example 1)

[0212] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0213] There is a need for a system that makes it easy for users to understand the reliability of answers provided by generative AI models. In particular, in customer support applications on online shopping sites, it is necessary to accurately evaluate the accuracy of answers to user questions and display the results in an easy-to-understand manner to improve user trust and usability.

[0214] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0215] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer to the question using a generative AI model, means for evaluating the accuracy of the generated answer, and means for integrating the generated answer and its evaluated accuracy and displaying it to the user. This allows the user to easily determine the reliability of the information provided and receive high-quality support.

[0216] A "user" is an individual or entity that utilizes the system to enter questions and receive answers.

[0217] A "question" is a sentence or phrase that a user enters into the system to solicit an answer.

[0218] A "generative AI model" is an artificial intelligence model that generates the optimal answer to a given question.

[0219] An "answer" is information that a generative AI model generates based on a user's question.

[0220] "Accuracy" is an indicator of the reliability and accuracy of the generated answer.

[0221] "Evaluation" means measuring the accuracy of the generated answer and quantifying its reliability.

[0222] The "server" is the core computing system responsible for receiving user questions, generating answers using the generative AI model, evaluating their accuracy, and providing the results to the user.

[0223] "API communication" is a standardized means of communication for sending and receiving data between different software components.

[0224] "System" is a general term for a series of means and devices for realizing all of the above functions.

[0225] "Synthesis" is the process of combining generated answers and their accuracy information and presenting them to the user in a single format.

[0226] This invention shows an application example aimed at customer support on a forum website. This system uses a generative AI model to generate answers to questions entered by users, evaluates their accuracy, and displays them to the user. The following describes how this system is specifically implemented.

[0227] The entire system consists of a user, a device (e.g., a smartphone), and a server. The device basically receives questions from the user and provides a means for communication with the server. The server has a generative AI model and is capable of generating optimal answers based on the questions and evaluating their accuracy.

[0228] The device receives the user's question and sends it to the server. When the user enters a question such as "Is this product waterproof?" into the device, the question is sent to the server as an HTTP request. The server receives the question, parses it into an appropriate format, and passes it to the generative AI model. The generative AI model uses its internal knowledge base and algorithms to generate the optimal answer. In this case, OpenAI GPT-4 is used as the generative AI model, for example.

[0229] The generated answer will be in the form of "Yes, this product is waterproof." This answer is not passed on as is, but is further passed to the accuracy evaluation module. The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the generated answer. For example, it evaluates whether this answer can be trusted with 95% accuracy. In this case, it is evaluated as "Answer accuracy: 95%."

[0230] After the evaluation is complete, the server combines the answer and its accuracy to generate data. For example, it may be in the form of "Yes, this product is waterproof. (Answer accuracy: 95%)." This data is again sent to the device as an HTTP response. The device displays the received answer and its accuracy to the user, allowing the user to judge the reliability of the answer.

[0231] As a concrete example, consider a customer support application that works like this:

[0232] 1. The user types in a question. For example, "Is this product waterproof?"

[0233] 2. The application sends the question to the server as an HTTP request.

[0234] 3. The server generates an answer using the generative AI model and evaluates its accuracy.

[0235] 4. The combined answer and accuracy are sent back to the device.

[0236] 5. The application displays to the user, "Yes, this product is waterproof. (Answer accuracy: 95%)."

[0237] By implementing this system, users can easily judge the reliability of the information provided and receive high-quality support.

[0238] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0239] Step 1:

[0240] The user types a question into the terminal.

[0241] Input: A question typed by the user in natural language (e.g., "Is this product waterproof?").

[0242] Specific behavior: The user enters a question into the application's text box and presses the send button, which enters the question into the terminal.

[0243] Output: The question text is stored in the terminal's internal buffer.

[0244] Step 2:

[0245] The terminal sends the entered question to the server.

[0246] Input: The question entered by the user in step 1.

[0247] Specific operation: The terminal creates an HTTP request and sends the user's question to the server using the HTTP protocol.

[0248] Output: The question is sent to the server as an HTTP request.

[0249] Step 3:

[0250] The server receives the user's query.

[0251] Input: The HTTP request containing the question sent from the terminal in step 2.

[0252] What happens: The server receives an HTTP request and extracts the question text from the payload.

[0253] Output: The extracted question text is imported into the server's internal data.

[0254] Step 4:

[0255] The server passes the question to a generative AI model to generate an answer.

[0256] Input: The question text received by the server.

[0257] How it works: The server passes the question text to a generative AI model (e.g., OpenAI GPT-4), which then generates the optimal answer. The generative AI model uses its internal knowledge base to analyze the sentence and create an appropriate answer.

[0258] Output: The generated answer text (e.g., "Yes, this product is waterproof.").

[0259] Step 5:

[0260] The server evaluates the accuracy of the generated answer.

[0261] Input: The answer text generated in step 4.

[0262] What it does: Passes the generated answer to an accuracy assessment module, which uses an internal database and statistical models to quantify its confidence (e.g., "Answer accuracy: 95%").

[0263] Output: The accuracy value of the evaluated answer.

[0264] Step 6:

[0265] The server aggregates the answers and their accuracy to generate data.

[0266] Input: The answer text generated in step 4, and the accuracy value evaluated in step 5.

[0267] What it does: Combines the answer text and its accuracy and assembles the data into a format to present to the user (e.g., "Yes, this product is waterproof. (Answer accuracy: 95%)").

[0268] Output: Text data of the consolidated answers and accuracy.

[0269] Step 7:

[0270] The server sends the consolidated answer and accuracy data to the device.

[0271] Input: The consolidated data generated in step 6.

[0272] Specific operation: Generates integrated data as an HTTP response and sends it to the terminal.

[0273] Output: HTTP response containing the integration data.

[0274] Step 8:

[0275] The terminal receives the integrated data from the server and displays it to the user.

[0276] Input: The HTTP response containing the integration data sent by the server in step 7.

[0277] Specific operation: The device receives the HTTP response, extracts the answer and accuracy text from the payload, and displays them on the screen.

[0278] Output: The user will see the message "Yes, this product is waterproof. (Answer accuracy: 95%)."

[0279] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0280] The present invention relates to a system that combines a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user with an emotion engine that recognizes the user's emotions. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user, while also taking the user's emotions into consideration.

[0281] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[0282] User Perspective

[0283] The user inputs a question using the interface of a device (e.g., a PC or smartphone). For example, they input a question such as, "Tell me about the planets in the solar system." While the user is inputting the question, the emotion engine recognizes the user's emotion in real time.

[0284] Terminal perspective

[0285] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server.

[0286] Server Perspective

[0287] 1. Receiving Questions

[0288] The server receives the question sent from the device, analyzes it appropriately, and prepares it in a format that can be passed to the generative AI model.

[0289] 2. Receiving Emotion Data

[0290] The server receives the user's emotion data sent from the device, which allows it to take emotion information into account when generating answers and how to handle them.

[0291] 3. Generate answers

[0292] The server passes the question and emotion data to the generative AI model and requests it to generate an answer. For example, in response to the question, "Tell me about the planets in the solar system," the model generates an answer that takes emotion data into account: "There are eight planets in the solar system. I'm feeling great today!: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0293] 4. Evaluating the accuracy of answers

[0294] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[0295] 5. Considering user emotions

[0296] The generated answer is evaluated for accuracy, and the user's sentiment data is taken into account. For example, if the user's sentiment is negative, appropriate follow-up information or support links can be added.

[0297] 6. Submitting the evaluation results

[0298] The server combines the evaluated accuracy with the answer to create data to be sent to the device, such as "Answer: There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[0299] Terminal view (again)

[0300] The device receives the answer and accuracy information sent from the server and displays it to the user. For example, it displays "There are eight planets in the solar system. You're feeling great today! (Answer accuracy: 95%)." By looking at this display, the user can confirm the reliability of the answer as well as the information that takes their feelings into consideration.

[0301] Specific examples

[0302] For example, if a user asks, "When was the first moon landing?", the server receives this question and the user's emotional data at the time. The generative AI model generates the answer, "The first moon landing occurred in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then takes the user's emotional data into consideration and sends the following data to the device: "The first moon landing occurred in 1969. (Answer accuracy: 90%) That's very interesting!" The device then displays this to the user. Seeing this information, the user determines that the answer is reliable and feels reassured that their emotional state has been taken into consideration.

[0303] This system not only allows users to easily judge the accuracy of the generated AI model's answers, but also provides information that takes emotions into consideration, resulting in a more user-friendly interface.

[0304] The processing flow will be explained below.

[0305] Step 1:

[0306] The user enters a question into the input field of the device. For example, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotions (e.g., joy, surprise, anxiety, etc.) in real time.

[0307] Step 2:

[0308] The device sends the question and emotion data entered by the user to the server as text data using an HTTP request.

[0309] Step 3:

[0310] The server receives the question sent from the device, then separates and analyzes the question and emotion data, and prepares them in a format that can be passed to the generative AI model.

[0311] Step 4:

[0312] The server passes the analyzed question and emotion data to the generation AI model and requests it to generate an answer. For example, it passes the question "Tell me about the planets in the solar system" and emotion data that "the user is excited."

[0313] Step 5:

[0314] A generative AI model generates an answer based on the question and sentiment data, such as "There are eight planets in the solar system. Amazing!"

[0315] Step 6:

[0316] The server receives the answers generated by the generative AI model.

[0317] Step 7:

[0318] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[0319] Step 8:

[0320] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[0321] Step 9:

[0322] The server receives the evaluation results from the accuracy evaluation module.

[0323] Step 10:

[0324] The server combines the generated answers and their accuracy ratings, and creates data to send to the device, taking into account emotional data. For example, it generates data such as "Answer: There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[0325] Step 11:

[0326] The server sends this integrated data to the terminal.

[0327] Step 12:

[0328] The device displays the received answer and a comment that takes into account the accuracy and sentiment of the user. For example, it displays "There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[0329] Step 13:

[0330] Users can check the information displayed on their device and receive information that takes their emotions into consideration along with the reliability of the answers, making it easier for them to accept the content.

[0331] As a result, users can obtain reliable answers and receive information that takes their emotions into consideration, resulting in a more natural and satisfying information experience.

[0332] Example 2

[0333] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0334] Conventional generative AI systems generate answers to user questions, but they have difficulty informing users of the accuracy and reliability of the answers. Furthermore, they provide answers without considering the user's emotions, resulting in a poor user experience. To solve these problems, it is necessary to evaluate the accuracy of the generated answers, clearly communicate this to the user, and provide information that takes the user's emotions into consideration.

[0335] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0336] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time using an emotion engine, means for transmitting the recognized emotion data to the server, and means for generating an answer taking the emotion data into consideration. This makes it possible to provide more appropriate information that takes the user's emotions into consideration while indicating the reliability of the generated answer to the user.

[0337] A "user" is an entity that uses the system to input questions and receives information via a terminal.

[0338] A "question" refers to the inquiry a user enters into the system, and is text data that is analyzed by the generative AI model to generate an answer.

[0339] A "generative AI model" is an algorithm or program that generates natural language answers based on questions entered by a user.

[0340] An "answer" refers to the response text that a generative AI model generates in response to a question, and is the information provided to the user.

[0341] "Accuracy" is an indicator that evaluates the accuracy and reliability of the generated answers, and is displayed numerically or qualitatively.

[0342] An "emotion engine" is hardware or software that recognizes and analyzes a user's emotions in real time.

[0343] "Emotion data" is information obtained as a result of the emotion engine recognizing and analyzing the user's emotions, and is data that indicates the user's psychological state.

[0344] The "server" is a computer system that processes questions and emotion data received from the device, generates answers using a generative AI model, evaluates them, and sends them to the device.

[0345] A "terminal" is a device used by a user to access the system, enter questions, and receive answers, and includes a personal computer, smartphone, etc.

[0346] An "HTTP request" is a communication protocol used to send questions and emotional data from a device to a server.

[0347] The "accuracy assessment module" is a software module for assessing the accuracy and reliability of generated answers, using an internal database and statistical models.

[0348] "Follow-up information" refers to additional information or assistance provided depending on the user's emotional state, and is information intended to improve the user experience.

[0349] "Natural language" refers to the language that people use on a daily basis, and is the subject of processing by systems in the form of text, speech, etc.

[0350] The present invention is a system that uses a generative AI model to generate answers to questions entered by users, evaluates the accuracy of the answers, and displays them after taking into account the user's emotions.

[0351] A means for receiving user-entered questions

[0352] A user inputs a question using a device such as a PC or smartphone. For example, a user might input a question such as, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotion in real time. The emotion engine includes software that analyzes facial expressions, tone of voice, and other factors.

[0353] Submitting questions and emotion data

[0354] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server at the same time.

[0355] Receiving and analyzing questions and emotion data

[0356] The server receives the question and emotion data sent from the device, parses the question into an appropriate format, and converts it into a format that can be passed to the generative AI model. This analysis includes text preprocessing and tokenization.

[0357] Generate answers

[0358] The server then passes the analyzed question and emotion data to the generative AI model to generate an answer. For example, in response to the question, "Tell me about the planets in our solar system," the model generates the answer, "There are eight planets in our solar system. I'm in a good mood today!"

[0359] Answer accuracy assessment

[0360] The generated answers are passed to an accuracy evaluation module in the server, which uses an internal database and statistical models to evaluate the accuracy and reliability of the answers, for example, rating them as "Answer accuracy: 95%."

[0361] Considering user emotions

[0362] The server integrates the user's sentiment data into the generated answer and its accuracy rating. If the user's sentiment is negative, appropriate follow-up information or support links can be added. For example, the data can be configured as "There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[0363] Server generates and transmits evaluation results

[0364] The server then sends the final answer, which integrates the evaluated accuracy and emotional data, to the device, allowing the user to confirm the reliability of the generated answer while receiving information that takes their emotions into consideration.

[0365] Displaying evaluation results on a device

[0366] The device receives the evaluation results sent from the server and displays them to the user. For example, it displays "There are eight planets in the solar system. You're in a good mood today! (Answer accuracy: 95%)."

[0367] Specific examples

[0368] For example, if a user asks, "When was the first moon landing?" the system operates as follows:

[0369] 1. A user types into a terminal, "When was the first moon landing?"

[0370] 2. The device sends the question and emotion data to the server.

[0371] 3. The server receives the question and emotion data and converts it into a format that can be passed to the generative AI model.

[0372] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[0373] 5. The server evaluates the accuracy of this answer as "Answer accuracy: 90%".

[0374] 6. The server considers the user's emotional data and integrates it into the data "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!"

[0375] 7. The server sends this integrated data to the terminal.

[0376] 8. The terminal displays the received data to the user.

[0377] This embodiment allows users to check the accuracy of answers provided by the generative AI model and receive information that takes emotions into account.

[0378] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0379] Step 1: User enters question

[0380] A user inputs a question using their own device (such as a PC or smartphone). For example, they might input a question like, "When was the first moon landing?" At this time, the emotion engine recognizes and analyzes the user's emotions in real time. The inputs are the user's question text and the recognized emotion data.

[0381] Specific behavior:

[0382] A user types into the question form, "When was the first moon landing?"

[0383] The emotion engine analyzes the user's facial expressions and tone of voice to generate emotion data (e.g., interesting).

[0384] Step 2: Device sends question and emotion data

[0385] The device receives the question entered by the user and the emotion data generated by the emotion engine, and sends them to the server using an HTTP request. The device receives the user's question text and emotion data as input, and sends them to the server as output.

[0386] Specific behavior:

[0387] The device builds question data and emotion data

[0388] Issue an HTTP request and send the question data and emotion data to the server

[0389] Step 3: Server receives question and emotion data

[0390] The server receives the question text and emotion data sent from the device, processes the received data, and obtains the question text and emotion data as input. As output, the server passes these data to the next processing step.

[0391] Specific behavior:

[0392] The server receives the HTTP request

[0393] Extract question text and sentiment data and pass them on to the next processing step

[0394] Step 4: Parsing the Question

[0395] The server properly parses the received question text and prepares it in a format that can be passed to the generative AI model. This parsing process includes text preprocessing (cleaning and tokenization). As input, it receives the user's question text and generates data in a format that can be passed to the generative AI model as output.

[0396] Specific behavior:

[0397] The server pre-processes the question text (removing unnecessary spaces, symbols, etc.)

[0398] Tokenization and conversion to a data format suitable for generative AI models

[0399] Step 5: Generate an answer

[0400] The server passes the analyzed question text and emotion data to the generative AI model to generate an answer. The generated answer takes the emotion data into account. The server receives the analyzed question text and emotion data as input and generates an answer text as output.

[0401] Specific behavior:

[0402] Input the user's question text and emotion data into the generative AI model

[0403] The model generates an answer (e.g., "The first moon landing took place in 1969")

[0404] Step 6: Evaluate the accuracy of your answers

[0405] The generated answer is passed to the accuracy evaluation module, where the accuracy is evaluated using an internal database and statistical models. The generated answer text is received as input, and the accuracy evaluation result (e.g., answer accuracy: 90%) is obtained as output.

[0406] Specific behavior:

[0407] The accuracy assessment module evaluates the accuracy and reliability of the answer text

[0408] Output evaluation results as numerical and qualitative information

[0409] Step 7: Consider user sentiment

[0410] The server integrates the user's emotional data with the generated answer and its accuracy assessment, and adds follow-up information as needed.The server receives the answer text, accuracy assessment results, and emotional data as input, and generates the integrated answer data as output.

[0411] Specific behavior:

[0412] Integrates answer text, accuracy evaluation results, and sentiment data

[0413] Add follow-up information or support links (e.g., "That's very interesting!")

[0414] Step 8: Server generates and sends evaluation results

[0415] The server converts the integrated response data into a final format and transmits it to the terminal. The server receives the integrated response data as input and generates data to be transmitted to the terminal as output.

[0416] Specific behavior:

[0417] Build aggregated response data (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[0418] Send data to the device

[0419] Step 9: Displaying the evaluation results on the device

[0420] The terminal receives the evaluation results sent from the server and displays them to the user. As input, it receives the data sent from the server and displays it to the user as output.

[0421] Specific behavior:

[0422] Display the received data on the screen (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[0423] The user checks the displayed information

[0424] Through the above processing steps, users can receive information that takes into account the accuracy and emotions of the generated AI model's answers.

[0425] (Application example 2)

[0426] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0427] Conventional systems using generative AI models focus on generating answers to user questions and evaluating their accuracy, but they lack information that takes into account the user's emotions. This results in a poor user experience, especially in situations where the user is feeling stressed or anxious, and a lack of appropriate follow-up. This makes it necessary to improve the user's sense of security and satisfaction.

[0428] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0429] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time, and means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration. This allows the user to receive not only the accuracy of the generated answer but also appropriate support information according to their emotions, thereby improving their sense of security and satisfaction.

[0430] The "means for receiving questions entered by a user" is a function for collecting questions entered by a user using a device as data and transmitting the data to a system such as a server.

[0431] "Means for generating answers using a generative AI model" is a function that uses artificial intelligence to generate appropriate answers based on received questions.

[0432] The "means for evaluating the accuracy of the generated answer" is a function for evaluating the accuracy and reliability of the generated answer and expressing it as a number, percentage, etc.

[0433] The "means for displaying the evaluated accuracy to the user" is a function for displaying the generated answer on the user's device so that the accuracy of the answer can be confirmed.

[0434] "Means for recognizing user emotions in real time" is a function for analyzing and recognizing the emotional state (for example, joy or anxiety) of the user when inputting a question in real time.

[0435] "Means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration" is a function for analyzing the user's emotional data and providing a response that adds an appropriate emotional response to the generated answer.

[0436] The present invention relates to a system that evaluates the accuracy of answers generated by a generative AI model and provides information taking into account the user's emotions. The system receives a user's question, generates an appropriate answer, evaluates the accuracy of the answer, and provides a response taking into account the user's emotions.

[0437] Hardware and software used

[0438] Hardware: Smartphones, servers

[0439] Software: Emotion recognition engine (e.g., Python library emotion_recognition), generative AI model, web server (e.g., Flask or Django)

[0440] Processing flow

[0441] The system operates in the following steps:

[0442] User Perspective

[0443] Users input questions using a smartphone interface. At the same time, an emotion recognition engine recognizes the user's emotions in real time. For example, a user might input, "I think my phone has been hacked."

[0444] Terminal perspective

[0445] The device receives the question and emotion data entered by the user and sends it to the server using an HTTP request.

[0446] Server Perspective

[0447] 1. Receiving Questions

[0448] The server receives questions sent from the device and generates answers using a generative AI model.

[0449] 2. Receiving Emotion Data

[0450] The server receives the emotion data sent from the terminal and takes the emotion information into consideration when generating a response based on this data and when handling the response.

[0451] 3. Answer generation and accuracy evaluation

[0452] The generative AI model responds to user questions by generating appropriate answers, such as, "Your phone may have been hacked. Please change your password immediately and check for suspicious activity."

[0453] Next, the accuracy analysis module evaluates the accuracy of the generated answers and calculates a value such as "Answer accuracy: 85%."

[0454] 4. Considering user emotions and generating responses

[0455] The generated answer is evaluated based on the user's sentiment data, along with the accuracy of the answer. For example, if the user's sentiment is negative, appropriate follow-up information or support links will be added. A response such as "Your phone may have been hacked. Please change your password immediately and check for suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." may be generated.

[0456] 5. Sending the Response

[0457] The server sends the generated response to the terminal.

[0458] Terminal view (again)

[0459] The device receives the response sent from the server and displays it to the user, who can then check the reliability of the generated AI model's answer as well as support information that takes their own emotions into consideration.

[0460] Specific examples

[0461] For example, if a user asks, "My phone may have been hacked," the server receives this question and the user's emotional data at the time. The generative AI model generates an answer: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity." The accuracy evaluation module evaluates the accuracy of this answer as "Answer accuracy: 85%." The server then takes the user's emotional data into consideration and generates an emotionally sensitive response: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." This is then sent to the device, where it is displayed to the user.

[0462] Prompt Sentence Examples

[0463] User Query: "I think my phone might have been hacked."

[0464] Generate a calm and reassuring response considering the potential anxiety of the user. Ensure the response is concise and provides actionable steps to secure against hacking.

[0465] In this way, users can receive appropriate answers from the generative AI model as well as support information that takes their own emotions into consideration, allowing them to use the system with peace of mind.

[0466] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0467] Step 1:

[0468] The user inputs a question using the smartphone interface. Based on the user's input, the emotion recognition engine recognizes the user's emotion in real time. In this process, the user's question text is used as input data and the recognized emotion data is generated as output.

[0469] Step 2:

[0470] The terminal receives the question and emotion data entered by the user. The received data is sent to the server as an HTTP request. In this process, an HTTP request is generated that includes the user question text and emotion data as input data and is sent to the server as output data.

[0471] Step 3:

[0472] The server receives the question and emotion data sent from the device. The received question is converted into a format that can be interpreted by the generative AI model, and then passed to the generative AI model. In this step, the input data includes the user question text and emotion data, and the output data is the question text formatted to be passed to the generative AI model.

[0473] Step 4:

[0474] The server uses a generative AI model to generate an answer to the received question. The server provides the user question text as input to the generative AI model and obtains the generated answer text as output.

[0475] Step 5:

[0476] The server evaluates the accuracy of the generated answer. Using an accuracy evaluation module, the reliability of the generated answer text is quantified. This process includes the generated answer text as input data, and an accuracy evaluation value indicating reliability is obtained as output data.

[0477] Step 6:

[0478] The server integrates the evaluated accuracy into the answer to the question and generates an emotion-sensitive response based on the user's emotion data. For example, if a negative emotion is recognized, additional supporting information is added. In this step, the input data includes the generated answer text, the evaluated accuracy, and the user's emotion data, and the output data is a final emotion-sensitive response sentence.

[0479] Step 7:

[0480] The server sends the final response to the terminal. The final response, which takes emotion into consideration, is included in the transmission data.

[0481] Step 8:

[0482] The terminal receives the response sent from the server and displays it to the user. The user confirms the displayed information. This process is completed when the terminal receives the data from the server and displays it on the screen.

[0483] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0484] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0485] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0486] [Second embodiment]

[0487] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0488] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0489] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0490] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0491] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0492] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0493] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0494] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0495] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0496] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0497] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0498] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0499] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user.

[0500] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[0501] User Perspective

[0502] A user inputs a question using the interface of a device (e.g., a PC or smartphone), for example, "Tell me about the planets in the solar system."

[0503] Terminal perspective

[0504] The terminal receives the question entered by the user, sends it to the server, usually in the form of an HTTP request, and prepares to receive a response from the server.

[0505] Server Perspective

[0506] 1. Receiving Questions

[0507] The server receives the question sent from the device. This is called reception processing. After receiving the question, it analyzes it appropriately and prepares it in a format that can be passed to the generative AI model.

[0508] 2. Answer Generation

[0509] The server passes the question to the generative AI model, which generates an answer. The generative AI model uses its internal knowledge base and algorithms to create the best answer to the question. For example, it generates the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0510] 3. Evaluating the accuracy of answers

[0511] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[0512] 4. Submitting the evaluation results

[0513] The server combines the evaluated accuracy with the answer and sends it to the device. For example, it creates data such as "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0514] Terminal view (again)

[0515] The terminal receives the answer and accuracy information sent from the server and displays them to the user, who can then judge the reliability of the answer.

[0516] Specific examples

[0517] For example, if a user asks, "When was the first moon landing?", the server receives this question and passes it to the generative AI model. The generative AI model generates the answer, "The first moon landing took place in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then sends the data, "The first moon landing took place in 1969. (Answer accuracy: 90%)," to the device, which displays it to the user. The user sees this information and determines that the answer is relatively reliable.

[0518] This system allows users to easily determine the accuracy of the answers generated by the AI ​​model, preventing the spread of misinformation.

[0519] The processing flow will be explained below.

[0520] Step 1:

[0521] The user types a question into an input field on the device, for example, "Tell me about the planets in the solar system."

[0522] Step 2:

[0523] The terminal sends the user's input as text data to the server using an HTTP request.

[0524] Step 3:

[0525] The server receives the question sent from the device, analyzes it, and prepares it in a format that can be passed to the generative AI model.

[0526] Step 4:

[0527] The server passes the analyzed question to the generative AI model and requests it to generate an answer. For example, it passes a question like, "Tell me about the planets in the solar system."

[0528] Step 5:

[0529] A generative AI model generates an answer based on the question, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0530] Step 6:

[0531] The server receives the answers generated by the generative AI model.

[0532] Step 7:

[0533] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[0534] Step 8:

[0535] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[0536] Step 9:

[0537] The server receives the evaluation results from the accuracy evaluation module.

[0538] Step 10:

[0539] The server combines the generated answer and its accuracy assessment to create data to be sent to the device. For example, it generates data such as "Answer: There are eight planets in the solar system. (Answer accuracy: 95%)."

[0540] Step 11:

[0541] The server transmits the integrated data to the terminal.

[0542] Step 12:

[0543] The device displays the answer it received and its accuracy to the user, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0544] Step 13:

[0545] The user checks the information displayed on the device and determines the reliability of the answer.

[0546] Example 1

[0547] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0548] Previous generative AI models lacked mechanisms for evaluating the accuracy and reliability of their answers, making it difficult for users to determine how trustworthy the answers were. Furthermore, there was a lack of means to evaluate the accuracy of generated answers and display them in an easy-to-understand manner, making it impossible to prevent the spread of misinformation.

[0549] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0550] In this invention, the server includes means for receiving text input by a user through a terminal, means for formatting the received text so that it can be passed to a generative AI model, means for generating an answer using the generative AI model, means for passing the generated answer to an accuracy evaluation module to evaluate its accuracy, and means for integrating the evaluated accuracy into the answer and sending it to the terminal. This makes it possible to evaluate the accuracy of the answer of the generative AI model and display it in an easy-to-understand manner for the user.

[0551] "User" refers to the entity that operates the system and inputs questions.

[0552] A "terminal" is an input device used by a user, and includes a personal computer, a smartphone, etc.

[0553] The "means for receiving" refers to a mechanism for inputting a question entered by a user into the system.

[0554] "Server" refers to the back-end computer that parses questions, passes them to the generative AI model, and generates and evaluates answers.

[0555] "Means of analysis" refers to the mechanism that converts the question into a form that the generative AI model can understand.

[0556] A "generative AI model" refers to an artificial intelligence model that generates answers to input questions.

[0557] "Means for generating an answer" refers to a mechanism that uses a generative AI model to create an answer to a question.

[0558] "Accuracy Assessment Module" refers to a rating system for measuring the accuracy of generated answers.

[0559] The "means for evaluating accuracy" refers to a mechanism that outputs the reliability of the generated answer as an evaluation value.

[0560] "Means for sending" refers to a communication means for sending the evaluated answers to the terminal.

[0561] The "means for displaying" refers to a mechanism for visually showing the answer including the evaluation result to the user.

[0562] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system transmits a question entered by a user through a terminal to a server, generates an answer to the question using a generative AI model, evaluates the accuracy of the answer, and displays it to the user.

[0563] The overall system configuration consists of three main components: the user, the terminal, and the server. The specific operation of each component is explained below.

[0564] User Perspective

[0565] A user inputs a question using a device such as a personal computer or smartphone. For example, the user inputs a question such as, "Tell me about the planets in the solar system."

[0566] Terminal perspective

[0567] The terminal receives the question entered by the user and sends it to the server, usually using a protocol called an HTTP request, and prepares to receive a response from the server.

[0568] Server Perspective

[0569] The server proceeds with the process in the following steps.

[0570] 1. Receiving and analyzing questions

[0571] The server receives the query sent from the terminal.

[0572] The received question is analyzed and formatted into a form that can be passed to the generative AI model.

[0573] 2. Answer Generation

[0574] The server inputs the formulated question into a generative AI model to generate an answer.

[0575] Large-scale language models such as GPT-3 are used as generative AI models.

[0576] For example, generate the answer "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0577] 3. Evaluating the accuracy of answers

[0578] The generated answer is passed to an accuracy evaluation module to evaluate its accuracy.

[0579] Accuracy assessment uses internal databases and statistical models.

[0580] For example, the evaluation result may be "Answer accuracy: 95%."

[0581] 4. Submitting the evaluation results

[0582] The server generates the data "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)" and sends it to the terminal.

[0583] Specific examples

[0584] For example, if a user asks, "When was the first moon landing?" the system will act as follows:

[0585] 1. User types, "When was the first moon landing?"

[0586] 2. The device sends a question to the server.

[0587] 3. The server receives the question and formats it for passing to the generative AI model.

[0588] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[0589] 5. The accuracy evaluation module evaluates the accuracy of this answer and obtains the result "Answer accuracy: 90%".

[0590] 6. The server generates the data "The first moon landing took place in 1969. (Answer accuracy: 90%)" and sends it to the device.

[0591] 7. The terminal receives this and displays it to the user.

[0592] In this way, users can easily check the accuracy of the generative AI model's answers and obtain reliable information.

[0593] Prompt Sentence Examples

[0594] For example, if a user enters the question "Tell me about the planets in the solar system," the prompt to the generative AI model will be as follows:

[0595] User Question: "Tell me about the planets in the solar system."

[0596] Input to generative AI model: "How many planets are there in the solar system? What are their names?"

[0597] This system makes it possible to evaluate the accuracy of the generative AI model's answers and present them to the user in a clear manner.

[0598] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0599] Step 1:

[0600] The user inputs a question through the terminal.

[0601] Specific behavior:

[0602] Users enter their questions using a computer or smartphone interface.

[0603] For example, type "Tell me about the planets in the solar system" and press the send button.

[0604] input:

[0605] User question: "Tell me about the planets in the solar system."

[0606] output:

[0607] User-entered question data

[0608] Step 2:

[0609] The terminal receives the user's question and sends it to the server.

[0610] Specific behavior:

[0611] The terminal sends the user's question to the server in the form of an HTTP request.

[0612] It is necessary to specify the destination URL or API endpoint.

[0613] input:

[0614] User-entered question data

[0615] output:

[0616] HTTP request sent to the server

[0617] Step 3:

[0618] The server receives the question and formats it for passing to the generative AI model.

[0619] Specific behavior:

[0620] The server receives the HTTP request and analyzes the question.

[0621] The question content is converted into a format that is easy for the generative AI model to understand, such as JSON format.

[0622] For example, reframe the question as, "How many planets are there in the solar system? What are their names?"

[0623] input:

[0624] HTTP request sent to the server

[0625] output:

[0626] Formatted question data

[0627] Step 4:

[0628] The server generates an answer using a generative AI model.

[0629] Specific behavior:

[0630] The server passes the formatted question to the generative AI model.

[0631] Generative AI models (e.g., GPT-3) generate answers using internal knowledge bases and algorithms.

[0632] For example, you might get the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0633] input:

[0634] Formatted question data

[0635] output:

[0636] Answers generated by generative AI models

[0637] Step 5:

[0638] The server passes the generated answers to an accuracy assessment module.

[0639] Specific behavior:

[0640] The server sends the generated answers to an accuracy assessment module.

[0641] An accuracy assessment module examines the answers and compares them against various assessment metrics.

[0642] input:

[0643] Answers generated by generative AI models

[0644] output:

[0645] Data to be evaluated passed to the accuracy evaluation module

[0646] Step 6:

[0647] An accuracy assessment module assesses the accuracy of the answers.

[0648] Specific behavior:

[0649] We use internal databases and statistical models to quantify the accuracy of responses.

[0650] For example, the evaluation result may be "Answer accuracy: 95%."

[0651] input:

[0652] Evaluation data passed to the accuracy evaluation module

[0653] output:

[0654] Evaluation result data (e.g., "Answer accuracy: 95%")

[0655] Step 7:

[0656] The server integrates the evaluation results with the answers and sends them to the terminal.

[0657] Specific behavior:

[0658] The server integrates the evaluation results into an answer and generates the data, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0659] This data is sent to the terminal as an HTTP response.

[0660] input:

[0661] Assessment result data and generated answers

[0662] output:

[0663] Answer and evaluation data sent to the device

[0664] Step 8:

[0665] The device displays the answer and accuracy information to the user.

[0666] Specific behavior:

[0667] The terminal analyzes the data received from the server and displays it in an appropriate format on the user interface.

[0668] For example, the screen might display, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0669] input:

[0670] Answer and evaluation data sent to the device

[0671] output:

[0672] Answer and rating information displayed to users

[0673] (Application example 1)

[0674] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0675] There is a need for a system that makes it easy for users to understand the reliability of answers provided by generative AI models. In particular, in customer support applications on online shopping sites, it is necessary to accurately evaluate the accuracy of answers to user questions and display the results in an easy-to-understand manner to improve user trust and usability.

[0676] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0677] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer to the question using a generative AI model, means for evaluating the accuracy of the generated answer, and means for integrating the generated answer and its evaluated accuracy and displaying it to the user. This allows the user to easily determine the reliability of the information provided and receive high-quality support.

[0678] A "user" is an individual or entity that utilizes the system to enter questions and receive answers.

[0679] A "question" is a sentence or phrase that a user enters into the system to solicit an answer.

[0680] A "generative AI model" is an artificial intelligence model that generates the optimal answer to a given question.

[0681] An "answer" is information that a generative AI model generates based on a user's question.

[0682] "Accuracy" is an indicator of the reliability and accuracy of the generated answer.

[0683] "Evaluation" means measuring the accuracy of the generated answer and quantifying its reliability.

[0684] The "server" is the core computing system responsible for receiving user questions, generating answers using the generative AI model, evaluating their accuracy, and providing the results to the user.

[0685] "API communication" is a standardized means of communication for sending and receiving data between different software components.

[0686] "System" is a general term for a series of means and devices for realizing all of the above functions.

[0687] "Synthesis" is the process of combining generated answers and their accuracy information and presenting them to the user in a single format.

[0688] This invention shows an application example aimed at customer support on a forum website. This system uses a generative AI model to generate answers to questions entered by users, evaluates their accuracy, and displays them to the user. The following describes how this system is specifically implemented.

[0689] The entire system consists of a user, a device (e.g., a smartphone), and a server. The device basically receives questions from the user and provides a means for communication with the server. The server has a generative AI model and is capable of generating optimal answers based on the questions and evaluating their accuracy.

[0690] The device receives the user's question and sends it to the server. When the user enters a question such as "Is this product waterproof?" into the device, the question is sent to the server as an HTTP request. The server receives the question, parses it into an appropriate format, and passes it to the generative AI model. The generative AI model uses its internal knowledge base and algorithms to generate the optimal answer. In this case, OpenAI GPT-4 is used as the generative AI model, for example.

[0691] The generated answer will be in the form of "Yes, this product is waterproof." This answer is not passed on as is, but is further passed to the accuracy evaluation module. The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the generated answer. For example, it evaluates whether this answer can be trusted with 95% accuracy. In this case, it is evaluated as "Answer accuracy: 95%."

[0692] After the evaluation is complete, the server combines the answer and its accuracy to generate data. For example, it may be in the form of "Yes, this product is waterproof. (Answer accuracy: 95%)." This data is again sent to the device as an HTTP response. The device displays the received answer and its accuracy to the user, allowing the user to judge the reliability of the answer.

[0693] As a concrete example, consider a customer support application that works like this:

[0694] 1. The user types in a question. For example, "Is this product waterproof?"

[0695] 2. The application sends the question to the server as an HTTP request.

[0696] 3. The server generates an answer using the generative AI model and evaluates its accuracy.

[0697] 4. The combined answer and accuracy are sent back to the device.

[0698] 5. The application displays to the user, "Yes, this product is waterproof. (Answer accuracy: 95%)."

[0699] By implementing this system, users can easily judge the reliability of the information provided and receive high-quality support.

[0700] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0701] Step 1:

[0702] The user types a question into the terminal.

[0703] Input: A question typed by the user in natural language (e.g., "Is this product waterproof?").

[0704] Specific behavior: The user enters a question into the application's text box and presses the send button, which enters the question into the terminal.

[0705] Output: The question text is stored in the terminal's internal buffer.

[0706] Step 2:

[0707] The terminal sends the entered question to the server.

[0708] Input: The question entered by the user in step 1.

[0709] Specific operation: The terminal creates an HTTP request and sends the user's question to the server using the HTTP protocol.

[0710] Output: The question is sent to the server as an HTTP request.

[0711] Step 3:

[0712] The server receives the user's query.

[0713] Input: The HTTP request containing the question sent from the terminal in step 2.

[0714] What happens: The server receives an HTTP request and extracts the question text from the payload.

[0715] Output: The extracted question text is imported into the server's internal data.

[0716] Step 4:

[0717] The server passes the question to a generative AI model to generate an answer.

[0718] Input: The question text received by the server.

[0719] How it works: The server passes the question text to a generative AI model (e.g., OpenAI GPT-4), which then generates the optimal answer. The generative AI model uses its internal knowledge base to analyze the sentence and create an appropriate answer.

[0720] Output: The generated answer text (e.g., "Yes, this product is waterproof.").

[0721] Step 5:

[0722] The server evaluates the accuracy of the generated answer.

[0723] Input: The answer text generated in step 4.

[0724] What it does: Passes the generated answer to an accuracy assessment module, which uses an internal database and statistical models to quantify its confidence (e.g., "Answer accuracy: 95%").

[0725] Output: The accuracy value of the evaluated answer.

[0726] Step 6:

[0727] The server aggregates the answers and their accuracy to generate data.

[0728] Input: The answer text generated in step 4, and the accuracy value evaluated in step 5.

[0729] What it does: Combines the answer text and its accuracy and assembles the data into a format to present to the user (e.g., "Yes, this product is waterproof. (Answer accuracy: 95%)").

[0730] Output: Text data of the consolidated answers and accuracy.

[0731] Step 7:

[0732] The server sends the consolidated answer and accuracy data to the device.

[0733] Input: The consolidated data generated in step 6.

[0734] Specific operation: Generates integrated data as an HTTP response and sends it to the terminal.

[0735] Output: HTTP response containing the integration data.

[0736] Step 8:

[0737] The terminal receives the integrated data from the server and displays it to the user.

[0738] Input: The HTTP response containing the integration data sent by the server in step 7.

[0739] Specific operation: The device receives the HTTP response, extracts the answer and accuracy text from the payload, and displays them on the screen.

[0740] Output: The user will see the message "Yes, this product is waterproof. (Answer accuracy: 95%)."

[0741] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0742] The present invention relates to a system that combines a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user with an emotion engine that recognizes the user's emotions. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user, while also taking the user's emotions into consideration.

[0743] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[0744] User Perspective

[0745] The user inputs a question using the interface of a device (e.g., a PC or smartphone). For example, they input a question such as, "Tell me about the planets in the solar system." While the user is inputting the question, the emotion engine recognizes the user's emotion in real time.

[0746] Terminal perspective

[0747] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server.

[0748] Server Perspective

[0749] 1. Receiving Questions

[0750] The server receives the question sent from the device, analyzes it appropriately, and prepares it in a format that can be passed to the generative AI model.

[0751] 2. Receiving Emotion Data

[0752] The server receives the user's emotion data sent from the device, which allows it to take emotion information into account when generating answers and how to handle them.

[0753] 3. Generate answers

[0754] The server passes the question and emotion data to the generative AI model and requests it to generate an answer. For example, in response to the question, "Tell me about the planets in the solar system," the model generates an answer that takes emotion data into account: "There are eight planets in the solar system. I'm feeling great today!: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0755] 4. Evaluating the accuracy of answers

[0756] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[0757] 5. Considering user emotions

[0758] The generated answer is evaluated for accuracy, and the user's sentiment data is taken into account. For example, if the user's sentiment is negative, appropriate follow-up information or support links can be added.

[0759] 6. Submitting the evaluation results

[0760] The server combines the evaluated accuracy with the answer to create data to be sent to the device, such as "Answer: There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[0761] Terminal view (again)

[0762] The device receives the answer and accuracy information sent from the server and displays it to the user. For example, it displays "There are eight planets in the solar system. You're feeling great today! (Answer accuracy: 95%)." By looking at this display, the user can confirm the reliability of the answer as well as the information that takes their feelings into consideration.

[0763] Specific examples

[0764] For example, if a user asks, "When was the first moon landing?", the server receives this question and the user's emotional data at the time. The generative AI model generates the answer, "The first moon landing occurred in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then takes the user's emotional data into consideration and sends the following data to the device: "The first moon landing occurred in 1969. (Answer accuracy: 90%) That's very interesting!" The device then displays this to the user. Seeing this information, the user determines that the answer is reliable and feels reassured that their emotional state has been taken into consideration.

[0765] This system not only allows users to easily judge the accuracy of the generated AI model's answers, but also provides information that takes emotions into consideration, resulting in a more user-friendly interface.

[0766] The processing flow will be explained below.

[0767] Step 1:

[0768] The user enters a question into the input field of the device. For example, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotions (e.g., joy, surprise, anxiety, etc.) in real time.

[0769] Step 2:

[0770] The device sends the question and emotion data entered by the user to the server as text data using an HTTP request.

[0771] Step 3:

[0772] The server receives the question sent from the device, then separates and analyzes the question and emotion data, and prepares them in a format that can be passed to the generative AI model.

[0773] Step 4:

[0774] The server passes the analyzed question and emotion data to the generation AI model and requests it to generate an answer. For example, it passes the question "Tell me about the planets in the solar system" and emotion data that "the user is excited."

[0775] Step 5:

[0776] A generative AI model generates an answer based on the question and sentiment data, such as "There are eight planets in the solar system. Amazing!"

[0777] Step 6:

[0778] The server receives the answers generated by the generative AI model.

[0779] Step 7:

[0780] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[0781] Step 8:

[0782] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[0783] Step 9:

[0784] The server receives the evaluation results from the accuracy evaluation module.

[0785] Step 10:

[0786] The server combines the generated answers and their accuracy ratings, and creates data to send to the device, taking into account emotional data. For example, it generates data such as "Answer: There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[0787] Step 11:

[0788] The server sends this integrated data to the terminal.

[0789] Step 12:

[0790] The device displays the received answer and a comment that takes into account the accuracy and sentiment of the user. For example, it displays "There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[0791] Step 13:

[0792] Users can check the information displayed on their device and receive information that takes their emotions into consideration along with the reliability of the answers, making it easier for them to accept the content.

[0793] As a result, users can obtain reliable answers and receive information that takes their emotions into consideration, resulting in a more natural and satisfying information experience.

[0794] Example 2

[0795] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0796] Conventional generative AI systems generate answers to user questions, but they have difficulty informing users of the accuracy and reliability of the answers. Furthermore, they provide answers without considering the user's emotions, resulting in a poor user experience. To solve these problems, it is necessary to evaluate the accuracy of the generated answers, clearly communicate this to the user, and provide information that takes the user's emotions into consideration.

[0797] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0798] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time using an emotion engine, means for transmitting the recognized emotion data to the server, and means for generating an answer taking the emotion data into consideration. This makes it possible to provide more appropriate information that takes the user's emotions into consideration while indicating the reliability of the generated answer to the user.

[0799] A "user" is an entity that uses the system to input questions and receives information via a terminal.

[0800] A "question" refers to the inquiry a user enters into the system, and is text data that is analyzed by the generative AI model to generate an answer.

[0801] A "generative AI model" is an algorithm or program that generates natural language answers based on questions entered by a user.

[0802] An "answer" refers to the response text that a generative AI model generates in response to a question, and is the information provided to the user.

[0803] "Accuracy" is an indicator that evaluates the accuracy and reliability of the generated answers, and is displayed numerically or qualitatively.

[0804] An "emotion engine" is hardware or software that recognizes and analyzes a user's emotions in real time.

[0805] "Emotion data" is information obtained as a result of the emotion engine recognizing and analyzing the user's emotions, and is data that indicates the user's psychological state.

[0806] The "server" is a computer system that processes questions and emotion data received from the device, generates answers using a generative AI model, evaluates them, and sends them to the device.

[0807] A "terminal" is a device used by a user to access the system, enter questions, and receive answers, and includes a personal computer, smartphone, etc.

[0808] An "HTTP request" is a communication protocol used to send questions and emotional data from a device to a server.

[0809] The "accuracy assessment module" is a software module for assessing the accuracy and reliability of generated answers, using an internal database and statistical models.

[0810] "Follow-up information" refers to additional information or assistance provided depending on the user's emotional state, and is information intended to improve the user experience.

[0811] "Natural language" refers to the language that people use on a daily basis, and is the subject of processing by systems in the form of text, speech, etc.

[0812] The present invention is a system that uses a generative AI model to generate answers to questions entered by users, evaluates the accuracy of the answers, and displays them after taking into account the user's emotions.

[0813] A means for receiving user-entered questions

[0814] A user inputs a question using a device such as a PC or smartphone. For example, a user might input a question such as, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotion in real time. The emotion engine includes software that analyzes facial expressions, tone of voice, and other factors.

[0815] Submitting questions and emotion data

[0816] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server at the same time.

[0817] Receiving and analyzing questions and emotion data

[0818] The server receives the question and emotion data sent from the device, parses the question into an appropriate format, and converts it into a format that can be passed to the generative AI model. This analysis includes text preprocessing and tokenization.

[0819] Generate answers

[0820] The server then passes the analyzed question and emotion data to the generative AI model to generate an answer. For example, in response to the question, "Tell me about the planets in our solar system," the model generates the answer, "There are eight planets in our solar system. I'm in a good mood today!"

[0821] Answer accuracy assessment

[0822] The generated answers are passed to an accuracy evaluation module in the server, which uses an internal database and statistical models to evaluate the accuracy and reliability of the answers, for example, rating them as "Answer accuracy: 95%."

[0823] Considering user emotions

[0824] The server integrates the user's sentiment data into the generated answer and its accuracy rating. If the user's sentiment is negative, appropriate follow-up information or support links can be added. For example, the data can be configured as "There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[0825] Server generates and transmits evaluation results

[0826] The server then sends the final answer, which integrates the evaluated accuracy and emotional data, to the device, allowing the user to confirm the reliability of the generated answer while receiving information that takes their emotions into consideration.

[0827] Displaying evaluation results on a device

[0828] The device receives the evaluation results sent from the server and displays them to the user. For example, it displays "There are eight planets in the solar system. You're in a good mood today! (Answer accuracy: 95%)."

[0829] Specific examples

[0830] For example, if a user asks, "When was the first moon landing?" the system operates as follows:

[0831] 1. A user types into a terminal, "When was the first moon landing?"

[0832] 2. The device sends the question and emotion data to the server.

[0833] 3. The server receives the question and emotion data and converts it into a format that can be passed to the generative AI model.

[0834] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[0835] 5. The server evaluates the accuracy of this answer as "Answer accuracy: 90%".

[0836] 6. The server considers the user's emotional data and integrates it into the data "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!"

[0837] 7. The server sends this integrated data to the terminal.

[0838] 8. The terminal displays the received data to the user.

[0839] This embodiment allows users to check the accuracy of answers provided by the generative AI model and receive information that takes emotions into account.

[0840] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0841] Step 1: User enters question

[0842] A user inputs a question using their own device (such as a PC or smartphone). For example, they might input a question like, "When was the first moon landing?" At this time, the emotion engine recognizes and analyzes the user's emotions in real time. The inputs are the user's question text and the recognized emotion data.

[0843] Specific behavior:

[0844] A user types into the question form, "When was the first moon landing?"

[0845] The emotion engine analyzes the user's facial expressions and tone of voice to generate emotion data (e.g., interesting).

[0846] Step 2: Device sends question and emotion data

[0847] The device receives the question entered by the user and the emotion data generated by the emotion engine, and sends them to the server using an HTTP request. The device receives the user's question text and emotion data as input, and sends them to the server as output.

[0848] Specific behavior:

[0849] The device builds question data and emotion data

[0850] Issue an HTTP request and send the question data and emotion data to the server

[0851] Step 3: Server receives question and emotion data

[0852] The server receives the question text and emotion data sent from the device, processes the received data, and obtains the question text and emotion data as input. As output, the server passes these data to the next processing step.

[0853] Specific behavior:

[0854] The server receives the HTTP request

[0855] Extract question text and sentiment data and pass them on to the next processing step

[0856] Step 4: Parsing the Question

[0857] The server properly parses the received question text and prepares it in a format that can be passed to the generative AI model. This parsing process includes text preprocessing (cleaning and tokenization). As input, it receives the user's question text and generates data in a format that can be passed to the generative AI model as output.

[0858] Specific behavior:

[0859] The server pre-processes the question text (removing unnecessary spaces, symbols, etc.)

[0860] Tokenization and conversion to a data format suitable for generative AI models

[0861] Step 5: Generate an answer

[0862] The server passes the analyzed question text and emotion data to the generative AI model to generate an answer. The generated answer takes the emotion data into account. The server receives the analyzed question text and emotion data as input and generates an answer text as output.

[0863] Specific behavior:

[0864] Input the user's question text and emotion data into the generative AI model

[0865] The model generates an answer (e.g., "The first moon landing took place in 1969")

[0866] Step 6: Evaluate the accuracy of your answers

[0867] The generated answer is passed to the accuracy evaluation module, where the accuracy is evaluated using an internal database and statistical models. The generated answer text is received as input, and the accuracy evaluation result (e.g., answer accuracy: 90%) is obtained as output.

[0868] Specific behavior:

[0869] The accuracy assessment module evaluates the accuracy and reliability of the answer text

[0870] Output evaluation results as numerical and qualitative information

[0871] Step 7: Consider user sentiment

[0872] The server integrates the user's emotional data with the generated answer and its accuracy assessment, and adds follow-up information as needed.The server receives the answer text, accuracy assessment results, and emotional data as input, and generates the integrated answer data as output.

[0873] Specific behavior:

[0874] Integrates answer text, accuracy evaluation results, and sentiment data

[0875] Add follow-up information or support links (e.g., "That's very interesting!")

[0876] Step 8: Server generates and sends evaluation results

[0877] The server converts the integrated response data into a final format and transmits it to the terminal. The server receives the integrated response data as input and generates data to be transmitted to the terminal as output.

[0878] Specific behavior:

[0879] Build aggregated response data (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[0880] Send data to the device

[0881] Step 9: Displaying the evaluation results on the device

[0882] The terminal receives the evaluation results sent from the server and displays them to the user. As input, it receives the data sent from the server and displays it to the user as output.

[0883] Specific behavior:

[0884] Display the received data on the screen (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[0885] The user checks the displayed information

[0886] Through the above processing steps, users can receive information that takes into account the accuracy and emotions of the generated AI model's answers.

[0887] (Application example 2)

[0888] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0889] Conventional systems using generative AI models focus on generating answers to user questions and evaluating their accuracy, but they lack information that takes into account the user's emotions. This results in a poor user experience, especially in situations where the user is feeling stressed or anxious, and a lack of appropriate follow-up. This makes it necessary to improve the user's sense of security and satisfaction.

[0890] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0891] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time, and means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration. This allows the user to receive not only the accuracy of the generated answer but also appropriate support information according to their emotions, thereby improving their sense of security and satisfaction.

[0892] The "means for receiving questions entered by a user" is a function for collecting questions entered by a user using a device as data and transmitting the data to a system such as a server.

[0893] "Means for generating answers using a generative AI model" is a function that uses artificial intelligence to generate appropriate answers based on received questions.

[0894] The "means for evaluating the accuracy of the generated answer" is a function for evaluating the accuracy and reliability of the generated answer and expressing it as a number, percentage, etc.

[0895] The "means for displaying the evaluated accuracy to the user" is a function for displaying the generated answer on the user's device so that the accuracy of the answer can be confirmed.

[0896] "Means for recognizing user emotions in real time" is a function for analyzing and recognizing the emotional state (for example, joy or anxiety) of the user when inputting a question in real time.

[0897] "Means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration" is a function for analyzing the user's emotional data and providing a response that adds an appropriate emotional response to the generated answer.

[0898] The present invention relates to a system that evaluates the accuracy of answers generated by a generative AI model and provides information taking into account the user's emotions. The system receives a user's question, generates an appropriate answer, evaluates the accuracy of the answer, and provides a response taking into account the user's emotions.

[0899] Hardware and software used

[0900] Hardware: Smartphones, servers

[0901] Software: Emotion recognition engine (e.g., Python library emotion_recognition), generative AI model, web server (e.g., Flask or Django)

[0902] Processing flow

[0903] The system operates in the following steps:

[0904] User Perspective

[0905] Users input questions using a smartphone interface. At the same time, an emotion recognition engine recognizes the user's emotions in real time. For example, a user might input, "I think my phone has been hacked."

[0906] Terminal perspective

[0907] The device receives the question and emotion data entered by the user and sends it to the server using an HTTP request.

[0908] Server Perspective

[0909] 1. Receiving Questions

[0910] The server receives questions sent from the device and generates answers using a generative AI model.

[0911] 2. Receiving Emotion Data

[0912] The server receives the emotion data sent from the terminal and takes the emotion information into consideration when generating a response based on this data and when handling the response.

[0913] 3. Answer generation and accuracy evaluation

[0914] The generative AI model responds to user questions by generating appropriate answers, such as, "Your phone may have been hacked. Please change your password immediately and check for suspicious activity."

[0915] Next, the accuracy analysis module evaluates the accuracy of the generated answers and calculates a value such as "Answer accuracy: 85%."

[0916] 4. Considering user emotions and generating responses

[0917] The generated answer is evaluated based on the user's sentiment data, along with the accuracy of the answer. For example, if the user's sentiment is negative, appropriate follow-up information or support links will be added. A response such as "Your phone may have been hacked. Please change your password immediately and check for suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." may be generated.

[0918] 5. Sending the Response

[0919] The server sends the generated response to the terminal.

[0920] Terminal view (again)

[0921] The device receives the response sent from the server and displays it to the user, who can then check the reliability of the generated AI model's answer as well as support information that takes their own emotions into consideration.

[0922] Specific examples

[0923] For example, if a user asks, "My phone may have been hacked," the server receives this question and the user's emotional data at the time. The generative AI model generates an answer: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity." The accuracy evaluation module evaluates the accuracy of this answer as "Answer accuracy: 85%." The server then takes the user's emotional data into consideration and generates an emotionally sensitive response: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." This is then sent to the device, where it is displayed to the user.

[0924] Prompt Sentence Examples

[0925] User Query: "I think my phone might have been hacked."

[0926] Generate a calm and reassuring response considering the potential anxiety of the user. Ensure the response is concise and provides actionable steps to secure against hacking.

[0927] In this way, users can receive appropriate answers from the generative AI model as well as support information that takes their own emotions into consideration, allowing them to use the system with peace of mind.

[0928] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0929] Step 1:

[0930] The user inputs a question using the smartphone interface. Based on the user's input, the emotion recognition engine recognizes the user's emotion in real time. In this process, the user's question text is used as input data and the recognized emotion data is generated as output.

[0931] Step 2:

[0932] The terminal receives the question and emotion data entered by the user. The received data is sent to the server as an HTTP request. In this process, an HTTP request is generated that includes the user question text and emotion data as input data and is sent to the server as output data.

[0933] Step 3:

[0934] The server receives the question and emotion data sent from the device. The received question is converted into a format that can be interpreted by the generative AI model, and then passed to the generative AI model. In this step, the input data includes the user question text and emotion data, and the output data is the question text formatted to be passed to the generative AI model.

[0935] Step 4:

[0936] The server uses a generative AI model to generate an answer to the received question. The server provides the user question text as input to the generative AI model and obtains the generated answer text as output.

[0937] Step 5:

[0938] The server evaluates the accuracy of the generated answer. Using an accuracy evaluation module, the reliability of the generated answer text is quantified. This process includes the generated answer text as input data, and an accuracy evaluation value indicating reliability is obtained as output data.

[0939] Step 6:

[0940] The server integrates the evaluated accuracy into the answer to the question and generates an emotion-sensitive response based on the user's emotion data. For example, if a negative emotion is recognized, additional supporting information is added. In this step, the input data includes the generated answer text, the evaluated accuracy, and the user's emotion data, and the output data is a final emotion-sensitive response sentence.

[0941] Step 7:

[0942] The server sends the final response to the terminal. The final response, which takes emotion into consideration, is included in the transmission data.

[0943] Step 8:

[0944] The terminal receives the response sent from the server and displays it to the user. The user confirms the displayed information. This process is completed when the terminal receives the data from the server and displays it on the screen.

[0945] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0946] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0947] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0948] [Third embodiment]

[0949] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0950] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0951] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0952] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0953] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0954] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0955] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0956] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0957] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0958] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0959] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0960] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0961] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user.

[0962] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[0963] User Perspective

[0964] A user inputs a question using the interface of a device (e.g., a PC or smartphone), for example, "Tell me about the planets in the solar system."

[0965] Terminal perspective

[0966] The terminal receives the question entered by the user, sends it to the server, usually in the form of an HTTP request, and prepares to receive a response from the server.

[0967] Server Perspective

[0968] 1. Receiving Questions

[0969] The server receives the question sent from the device. This is called reception processing. After receiving the question, it analyzes it appropriately and prepares it in a format that can be passed to the generative AI model.

[0970] 2. Answer Generation

[0971] The server passes the question to the generative AI model, which generates an answer. The generative AI model uses its internal knowledge base and algorithms to create the best answer to the question. For example, it generates the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0972] 3. Evaluating the accuracy of answers

[0973] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[0974] 4. Submitting the evaluation results

[0975] The server combines the evaluated accuracy with the answer and sends it to the device. For example, it creates data such as "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[0976] Terminal view (again)

[0977] The terminal receives the answer and accuracy information sent from the server and displays them to the user, who can then judge the reliability of the answer.

[0978] Specific examples

[0979] For example, if a user asks, "When was the first moon landing?", the server receives this question and passes it to the generative AI model. The generative AI model generates the answer, "The first moon landing took place in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then sends the data, "The first moon landing took place in 1969. (Answer accuracy: 90%)," to the device, which displays it to the user. The user sees this information and determines that the answer is relatively reliable.

[0980] This system allows users to easily determine the accuracy of the answers generated by the AI ​​model, preventing the spread of misinformation.

[0981] The processing flow will be explained below.

[0982] Step 1:

[0983] The user types a question into an input field on the device, for example, "Tell me about the planets in the solar system."

[0984] Step 2:

[0985] The terminal sends the user's input as text data to the server using an HTTP request.

[0986] Step 3:

[0987] The server receives the question sent from the device, analyzes it, and prepares it in a format that can be passed to the generative AI model.

[0988] Step 4:

[0989] The server passes the analyzed question to the generative AI model and requests it to generate an answer. For example, it passes a question like, "Tell me about the planets in the solar system."

[0990] Step 5:

[0991] A generative AI model generates an answer based on the question, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[0992] Step 6:

[0993] The server receives the answers generated by the generative AI model.

[0994] Step 7:

[0995] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[0996] Step 8:

[0997] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[0998] Step 9:

[0999] The server receives the evaluation results from the accuracy evaluation module.

[1000] Step 10:

[1001] The server combines the generated answer and its accuracy assessment to create data to be sent to the device. For example, it generates data such as "Answer: There are eight planets in the solar system. (Answer accuracy: 95%)."

[1002] Step 11:

[1003] The server transmits the integrated data to the terminal.

[1004] Step 12:

[1005] The device displays the answer it received and its accuracy to the user, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1006] Step 13:

[1007] The user checks the information displayed on the device and determines the reliability of the answer.

[1008] Example 1

[1009] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1010] Previous generative AI models lacked mechanisms for evaluating the accuracy and reliability of their answers, making it difficult for users to determine how trustworthy the answers were. Furthermore, there was a lack of means to evaluate the accuracy of generated answers and display them in an easy-to-understand manner, making it impossible to prevent the spread of misinformation.

[1011] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1012] In this invention, the server includes means for receiving text input by a user through a terminal, means for formatting the received text so that it can be passed to a generative AI model, means for generating an answer using the generative AI model, means for passing the generated answer to an accuracy evaluation module to evaluate its accuracy, and means for integrating the evaluated accuracy into the answer and sending it to the terminal. This makes it possible to evaluate the accuracy of the answer of the generative AI model and display it in an easy-to-understand manner for the user.

[1013] "User" refers to the entity that operates the system and inputs questions.

[1014] A "terminal" is an input device used by a user, and includes a personal computer, a smartphone, etc.

[1015] The "means for receiving" refers to a mechanism for inputting a question entered by a user into the system.

[1016] "Server" refers to the back-end computer that parses questions, passes them to the generative AI model, and generates and evaluates answers.

[1017] "Means of analysis" refers to the mechanism that converts the question into a form that the generative AI model can understand.

[1018] A "generative AI model" refers to an artificial intelligence model that generates answers to input questions.

[1019] "Means for generating an answer" refers to a mechanism that uses a generative AI model to create an answer to a question.

[1020] "Accuracy Assessment Module" refers to a rating system for measuring the accuracy of generated answers.

[1021] The "means for evaluating accuracy" refers to a mechanism that outputs the reliability of the generated answer as an evaluation value.

[1022] "Means for sending" refers to a communication means for sending the evaluated answers to the terminal.

[1023] The "means for displaying" refers to a mechanism for visually showing the answer including the evaluation result to the user.

[1024] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system transmits a question entered by a user through a terminal to a server, generates an answer to the question using a generative AI model, evaluates the accuracy of the answer, and displays it to the user.

[1025] The overall system configuration consists of three main components: the user, the terminal, and the server. The specific operation of each component is explained below.

[1026] User Perspective

[1027] A user inputs a question using a device such as a personal computer or smartphone. For example, the user inputs a question such as, "Tell me about the planets in the solar system."

[1028] Terminal perspective

[1029] The terminal receives the question entered by the user and sends it to the server, usually using a protocol called an HTTP request, and prepares to receive a response from the server.

[1030] Server Perspective

[1031] The server proceeds with the process in the following steps.

[1032] 1. Receiving and analyzing questions

[1033] The server receives the query sent from the terminal.

[1034] The received question is analyzed and formatted into a form that can be passed to the generative AI model.

[1035] 2. Answer Generation

[1036] The server inputs the formulated question into a generative AI model to generate an answer.

[1037] Large-scale language models such as GPT-3 are used as generative AI models.

[1038] For example, generate the answer "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1039] 3. Evaluating the accuracy of answers

[1040] The generated answer is passed to an accuracy evaluation module to evaluate its accuracy.

[1041] Accuracy assessment uses internal databases and statistical models.

[1042] For example, the evaluation result may be "Answer accuracy: 95%."

[1043] 4. Submitting the evaluation results

[1044] The server generates the data "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)" and sends it to the terminal.

[1045] Specific examples

[1046] For example, if a user asks, "When was the first moon landing?" the system will act as follows:

[1047] 1. User types, "When was the first moon landing?"

[1048] 2. The device sends a question to the server.

[1049] 3. The server receives the question and formats it for passing to the generative AI model.

[1050] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[1051] 5. The accuracy evaluation module evaluates the accuracy of this answer and obtains the result "Answer accuracy: 90%".

[1052] 6. The server generates the data "The first moon landing took place in 1969. (Answer accuracy: 90%)" and sends it to the device.

[1053] 7. The terminal receives this and displays it to the user.

[1054] In this way, users can easily check the accuracy of the generative AI model's answers and obtain reliable information.

[1055] Prompt Sentence Examples

[1056] For example, if a user enters the question "Tell me about the planets in the solar system," the prompt to the generative AI model will be as follows:

[1057] User Question: "Tell me about the planets in the solar system."

[1058] Input to generative AI model: "How many planets are there in the solar system? What are their names?"

[1059] This system makes it possible to evaluate the accuracy of the generative AI model's answers and present them to the user in a clear manner.

[1060] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1061] Step 1:

[1062] The user inputs a question through the terminal.

[1063] Specific behavior:

[1064] Users enter their questions using a computer or smartphone interface.

[1065] For example, type "Tell me about the planets in the solar system" and press the send button.

[1066] input:

[1067] User question: "Tell me about the planets in the solar system."

[1068] output:

[1069] User-entered question data

[1070] Step 2:

[1071] The terminal receives the user's question and sends it to the server.

[1072] Specific behavior:

[1073] The terminal sends the user's question to the server in the form of an HTTP request.

[1074] It is necessary to specify the destination URL or API endpoint.

[1075] input:

[1076] User-entered question data

[1077] output:

[1078] HTTP request sent to the server

[1079] Step 3:

[1080] The server receives the question and formats it for passing to the generative AI model.

[1081] Specific behavior:

[1082] The server receives the HTTP request and analyzes the question.

[1083] The question content is converted into a format that is easy for the generative AI model to understand, such as JSON format.

[1084] For example, reframe the question as, "How many planets are there in the solar system? What are their names?"

[1085] input:

[1086] HTTP request sent to the server

[1087] output:

[1088] Formatted question data

[1089] Step 4:

[1090] The server generates an answer using a generative AI model.

[1091] Specific behavior:

[1092] The server passes the formatted question to the generative AI model.

[1093] Generative AI models (e.g., GPT-3) generate answers using internal knowledge bases and algorithms.

[1094] For example, you might get the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1095] input:

[1096] Formatted question data

[1097] output:

[1098] Answers generated by generative AI models

[1099] Step 5:

[1100] The server passes the generated answers to an accuracy assessment module.

[1101] Specific behavior:

[1102] The server sends the generated answers to an accuracy assessment module.

[1103] An accuracy assessment module examines the answers and compares them against various assessment metrics.

[1104] input:

[1105] Answers generated by generative AI models

[1106] output:

[1107] Data to be evaluated passed to the accuracy evaluation module

[1108] Step 6:

[1109] An accuracy assessment module assesses the accuracy of the answers.

[1110] Specific behavior:

[1111] We use internal databases and statistical models to quantify the accuracy of responses.

[1112] For example, the evaluation result may be "Answer accuracy: 95%."

[1113] input:

[1114] Evaluation data passed to the accuracy evaluation module

[1115] output:

[1116] Evaluation result data (e.g., "Answer accuracy: 95%")

[1117] Step 7:

[1118] The server integrates the evaluation results with the answers and sends them to the terminal.

[1119] Specific behavior:

[1120] The server integrates the evaluation results into an answer and generates the data, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1121] This data is sent to the terminal as an HTTP response.

[1122] input:

[1123] Assessment result data and generated answers

[1124] output:

[1125] Answer and evaluation data sent to the device

[1126] Step 8:

[1127] The device displays the answer and accuracy information to the user.

[1128] Specific behavior:

[1129] The terminal analyzes the data received from the server and displays it in an appropriate format on the user interface.

[1130] For example, the screen might display, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1131] input:

[1132] Answer and evaluation data sent to the device

[1133] output:

[1134] Answer and rating information displayed to users

[1135] (Application example 1)

[1136] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1137] There is a need for a system that makes it easy for users to understand the reliability of answers provided by generative AI models. In particular, in customer support applications on online shopping sites, it is necessary to accurately evaluate the accuracy of answers to user questions and display the results in an easy-to-understand manner to improve user trust and usability.

[1138] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1139] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer to the question using a generative AI model, means for evaluating the accuracy of the generated answer, and means for integrating the generated answer and its evaluated accuracy and displaying it to the user. This allows the user to easily determine the reliability of the information provided and receive high-quality support.

[1140] A "user" is an individual or entity that utilizes the system to enter questions and receive answers.

[1141] A "question" is a sentence or phrase that a user enters into the system to solicit an answer.

[1142] A "generative AI model" is an artificial intelligence model that generates the optimal answer to a given question.

[1143] An "answer" is information that a generative AI model generates based on a user's question.

[1144] "Accuracy" is an indicator of the reliability and accuracy of the generated answer.

[1145] "Evaluation" means measuring the accuracy of the generated answer and quantifying its reliability.

[1146] The "server" is the core computing system responsible for receiving user questions, generating answers using the generative AI model, evaluating their accuracy, and providing the results to the user.

[1147] "API communication" is a standardized means of communication for sending and receiving data between different software components.

[1148] "System" is a general term for a series of means and devices for realizing all of the above functions.

[1149] "Synthesis" is the process of combining generated answers and their accuracy information and presenting them to the user in a single format.

[1150] This invention shows an application example aimed at customer support on a forum website. This system uses a generative AI model to generate answers to questions entered by users, evaluates their accuracy, and displays them to the user. The following describes how this system is specifically implemented.

[1151] The entire system consists of a user, a device (e.g., a smartphone), and a server. The device basically receives questions from the user and provides a means for communication with the server. The server has a generative AI model and is capable of generating optimal answers based on the questions and evaluating their accuracy.

[1152] The device receives the user's question and sends it to the server. When the user enters a question such as "Is this product waterproof?" into the device, the question is sent to the server as an HTTP request. The server receives the question, parses it into an appropriate format, and passes it to the generative AI model. The generative AI model uses its internal knowledge base and algorithms to generate the optimal answer. In this case, OpenAI GPT-4 is used as the generative AI model, for example.

[1153] The generated answer will be in the form of "Yes, this product is waterproof." This answer is not passed on as is, but is further passed to the accuracy evaluation module. The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the generated answer. For example, it evaluates whether this answer can be trusted with 95% accuracy. In this case, it is evaluated as "Answer accuracy: 95%."

[1154] After the evaluation is complete, the server combines the answer and its accuracy to generate data. For example, it may be in the form of "Yes, this product is waterproof. (Answer accuracy: 95%)." This data is again sent to the device as an HTTP response. The device displays the received answer and its accuracy to the user, allowing the user to judge the reliability of the answer.

[1155] As a concrete example, consider a customer support application that works like this:

[1156] 1. The user types in a question. For example, "Is this product waterproof?"

[1157] 2. The application sends the question to the server as an HTTP request.

[1158] 3. The server generates an answer using the generative AI model and evaluates its accuracy.

[1159] 4. The combined answer and accuracy are sent back to the device.

[1160] 5. The application displays to the user, "Yes, this product is waterproof. (Answer accuracy: 95%)."

[1161] By implementing this system, users can easily judge the reliability of the information provided and receive high-quality support.

[1162] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1163] Step 1:

[1164] The user types a question into the terminal.

[1165] Input: A question typed by the user in natural language (e.g., "Is this product waterproof?").

[1166] Specific behavior: The user enters a question into the application's text box and presses the send button, which enters the question into the terminal.

[1167] Output: The question text is stored in the terminal's internal buffer.

[1168] Step 2:

[1169] The terminal sends the entered question to the server.

[1170] Input: The question entered by the user in step 1.

[1171] Specific operation: The terminal creates an HTTP request and sends the user's question to the server using the HTTP protocol.

[1172] Output: The question is sent to the server as an HTTP request.

[1173] Step 3:

[1174] The server receives the user's query.

[1175] Input: The HTTP request containing the question sent from the terminal in step 2.

[1176] What happens: The server receives an HTTP request and extracts the question text from the payload.

[1177] Output: The extracted question text is imported into the server's internal data.

[1178] Step 4:

[1179] The server passes the question to a generative AI model to generate an answer.

[1180] Input: The question text received by the server.

[1181] How it works: The server passes the question text to a generative AI model (e.g., OpenAI GPT-4), which then generates the optimal answer. The generative AI model uses its internal knowledge base to analyze the sentence and create an appropriate answer.

[1182] Output: The generated answer text (e.g., "Yes, this product is waterproof.").

[1183] Step 5:

[1184] The server evaluates the accuracy of the generated answer.

[1185] Input: The answer text generated in step 4.

[1186] What it does: Passes the generated answer to an accuracy assessment module, which uses an internal database and statistical models to quantify its confidence (e.g., "Answer accuracy: 95%").

[1187] Output: The accuracy value of the evaluated answer.

[1188] Step 6:

[1189] The server aggregates the answers and their accuracy to generate data.

[1190] Input: The answer text generated in step 4, and the accuracy value evaluated in step 5.

[1191] What it does: Combines the answer text and its accuracy and assembles the data into a format to present to the user (e.g., "Yes, this product is waterproof. (Answer accuracy: 95%)").

[1192] Output: Text data of the consolidated answers and accuracy.

[1193] Step 7:

[1194] The server sends the consolidated answer and accuracy data to the device.

[1195] Input: The consolidated data generated in step 6.

[1196] Specific operation: Generates integrated data as an HTTP response and sends it to the terminal.

[1197] Output: HTTP response containing the integration data.

[1198] Step 8:

[1199] The terminal receives the integrated data from the server and displays it to the user.

[1200] Input: The HTTP response containing the integration data sent by the server in step 7.

[1201] Specific operation: The device receives the HTTP response, extracts the answer and accuracy text from the payload, and displays them on the screen.

[1202] Output: The user will see the message "Yes, this product is waterproof. (Answer accuracy: 95%)."

[1203] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1204] The present invention relates to a system that combines a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user with an emotion engine that recognizes the user's emotions. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user, while also taking the user's emotions into consideration.

[1205] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[1206] User Perspective

[1207] The user inputs a question using the interface of a device (e.g., a PC or smartphone). For example, they input a question such as, "Tell me about the planets in the solar system." While the user is inputting the question, the emotion engine recognizes the user's emotion in real time.

[1208] Terminal perspective

[1209] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server.

[1210] Server Perspective

[1211] 1. Receiving Questions

[1212] The server receives the question sent from the device, analyzes it appropriately, and prepares it in a format that can be passed to the generative AI model.

[1213] 2. Receiving Emotion Data

[1214] The server receives the user's emotion data sent from the device, which allows it to take emotion information into account when generating answers and how to handle them.

[1215] 3. Generate answers

[1216] The server passes the question and emotion data to the generative AI model and requests it to generate an answer. For example, in response to the question, "Tell me about the planets in the solar system," the model generates an answer that takes emotion data into account: "There are eight planets in the solar system. I'm feeling great today!: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1217] 4. Evaluating the accuracy of answers

[1218] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[1219] 5. Considering user emotions

[1220] The generated answer is evaluated for accuracy, and the user's sentiment data is taken into account. For example, if the user's sentiment is negative, appropriate follow-up information or support links can be added.

[1221] 6. Submitting the evaluation results

[1222] The server combines the evaluated accuracy with the answer to create data to be sent to the device, such as "Answer: There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[1223] Terminal view (again)

[1224] The device receives the answer and accuracy information sent from the server and displays it to the user. For example, it displays "There are eight planets in the solar system. You're feeling great today! (Answer accuracy: 95%)." By looking at this display, the user can confirm the reliability of the answer as well as the information that takes their feelings into consideration.

[1225] Specific examples

[1226] For example, if a user asks, "When was the first moon landing?", the server receives this question and the user's emotional data at the time. The generative AI model generates the answer, "The first moon landing occurred in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then takes the user's emotional data into consideration and sends the following data to the device: "The first moon landing occurred in 1969. (Answer accuracy: 90%) That's very interesting!" The device then displays this to the user. Seeing this information, the user determines that the answer is reliable and feels reassured that their emotional state has been taken into consideration.

[1227] This system not only allows users to easily judge the accuracy of the generated AI model's answers, but also provides information that takes emotions into consideration, resulting in a more user-friendly interface.

[1228] The processing flow will be explained below.

[1229] Step 1:

[1230] The user enters a question into the input field of the device. For example, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotions (e.g., joy, surprise, anxiety, etc.) in real time.

[1231] Step 2:

[1232] The device sends the question and emotion data entered by the user to the server as text data using an HTTP request.

[1233] Step 3:

[1234] The server receives the question sent from the device, then separates and analyzes the question and emotion data, and prepares them in a format that can be passed to the generative AI model.

[1235] Step 4:

[1236] The server passes the analyzed question and emotion data to the generation AI model and requests it to generate an answer. For example, it passes the question "Tell me about the planets in the solar system" and emotion data that "the user is excited."

[1237] Step 5:

[1238] A generative AI model generates an answer based on the question and sentiment data, such as "There are eight planets in the solar system. Amazing!"

[1239] Step 6:

[1240] The server receives the answers generated by the generative AI model.

[1241] Step 7:

[1242] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[1243] Step 8:

[1244] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[1245] Step 9:

[1246] The server receives the evaluation results from the accuracy evaluation module.

[1247] Step 10:

[1248] The server combines the generated answers and their accuracy ratings, and creates data to send to the device, taking into account emotional data. For example, it generates data such as "Answer: There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[1249] Step 11:

[1250] The server sends this integrated data to the terminal.

[1251] Step 12:

[1252] The device displays the received answer and a comment that takes into account the accuracy and sentiment of the user. For example, it displays "There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[1253] Step 13:

[1254] Users can check the information displayed on their device and receive information that takes their emotions into consideration along with the reliability of the answers, making it easier for them to accept the content.

[1255] As a result, users can obtain reliable answers and receive information that takes their emotions into consideration, resulting in a more natural and satisfying information experience.

[1256] Example 2

[1257] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1258] Conventional generative AI systems generate answers to user questions, but they have difficulty informing users of the accuracy and reliability of the answers. Furthermore, they provide answers without considering the user's emotions, resulting in a poor user experience. To solve these problems, it is necessary to evaluate the accuracy of the generated answers, clearly communicate this to the user, and provide information that takes the user's emotions into consideration.

[1259] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1260] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time using an emotion engine, means for transmitting the recognized emotion data to the server, and means for generating an answer taking the emotion data into consideration. This makes it possible to provide more appropriate information that takes the user's emotions into consideration while indicating the reliability of the generated answer to the user.

[1261] A "user" is an entity that uses the system to input questions and receives information via a terminal.

[1262] A "question" refers to the inquiry a user enters into the system, and is text data that is analyzed by the generative AI model to generate an answer.

[1263] A "generative AI model" is an algorithm or program that generates natural language answers based on questions entered by a user.

[1264] An "answer" refers to the response text that a generative AI model generates in response to a question, and is the information provided to the user.

[1265] "Accuracy" is an indicator that evaluates the accuracy and reliability of the generated answers, and is displayed numerically or qualitatively.

[1266] An "emotion engine" is hardware or software that recognizes and analyzes a user's emotions in real time.

[1267] "Emotion data" is information obtained as a result of the emotion engine recognizing and analyzing the user's emotions, and is data that indicates the user's psychological state.

[1268] The "server" is a computer system that processes questions and emotion data received from the device, generates answers using a generative AI model, evaluates them, and sends them to the device.

[1269] A "terminal" is a device used by a user to access the system, enter questions, and receive answers, and includes a personal computer, smartphone, etc.

[1270] An "HTTP request" is a communication protocol used to send questions and emotional data from a device to a server.

[1271] The "accuracy assessment module" is a software module for assessing the accuracy and reliability of generated answers, using an internal database and statistical models.

[1272] "Follow-up information" refers to additional information or assistance provided depending on the user's emotional state, and is information intended to improve the user experience.

[1273] "Natural language" refers to the language that people use on a daily basis, and is the subject of processing by systems in the form of text, speech, etc.

[1274] The present invention is a system that uses a generative AI model to generate answers to questions entered by users, evaluates the accuracy of the answers, and displays them after taking into account the user's emotions.

[1275] A means for receiving user-entered questions

[1276] A user inputs a question using a device such as a PC or smartphone. For example, a user might input a question such as, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotion in real time. The emotion engine includes software that analyzes facial expressions, tone of voice, and other factors.

[1277] Submitting questions and emotion data

[1278] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server at the same time.

[1279] Receiving and analyzing questions and emotion data

[1280] The server receives the question and emotion data sent from the device, parses the question into an appropriate format, and converts it into a format that can be passed to the generative AI model. This analysis includes text preprocessing and tokenization.

[1281] Generate answers

[1282] The server then passes the analyzed question and emotion data to the generative AI model to generate an answer. For example, in response to the question, "Tell me about the planets in our solar system," the model generates the answer, "There are eight planets in our solar system. I'm in a good mood today!"

[1283] Answer accuracy assessment

[1284] The generated answers are passed to an accuracy evaluation module in the server, which uses an internal database and statistical models to evaluate the accuracy and reliability of the answers, for example, rating them as "Answer accuracy: 95%."

[1285] Considering user emotions

[1286] The server integrates the user's sentiment data into the generated answer and its accuracy rating. If the user's sentiment is negative, appropriate follow-up information or support links can be added. For example, the data can be configured as "There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[1287] Server generates and transmits evaluation results

[1288] The server then sends the final answer, which integrates the evaluated accuracy and emotional data, to the device, allowing the user to confirm the reliability of the generated answer while receiving information that takes their emotions into consideration.

[1289] Displaying evaluation results on a device

[1290] The device receives the evaluation results sent from the server and displays them to the user. For example, it displays "There are eight planets in the solar system. You're in a good mood today! (Answer accuracy: 95%)."

[1291] Specific examples

[1292] For example, if a user asks, "When was the first moon landing?" the system operates as follows:

[1293] 1. A user types into a terminal, "When was the first moon landing?"

[1294] 2. The device sends the question and emotion data to the server.

[1295] 3. The server receives the question and emotion data and converts it into a format that can be passed to the generative AI model.

[1296] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[1297] 5. The server evaluates the accuracy of this answer as "Answer accuracy: 90%".

[1298] 6. The server considers the user's emotional data and integrates it into the data "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!"

[1299] 7. The server sends this integrated data to the terminal.

[1300] 8. The terminal displays the received data to the user.

[1301] This embodiment allows users to check the accuracy of answers provided by the generative AI model and receive information that takes emotions into account.

[1302] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1303] Step 1: User enters question

[1304] A user inputs a question using their own device (such as a PC or smartphone). For example, they might input a question like, "When was the first moon landing?" At this time, the emotion engine recognizes and analyzes the user's emotions in real time. The inputs are the user's question text and the recognized emotion data.

[1305] Specific behavior:

[1306] A user types into the question form, "When was the first moon landing?"

[1307] The emotion engine analyzes the user's facial expressions and tone of voice to generate emotion data (e.g., interesting).

[1308] Step 2: Device sends question and emotion data

[1309] The device receives the question entered by the user and the emotion data generated by the emotion engine, and sends them to the server using an HTTP request. The device receives the user's question text and emotion data as input, and sends them to the server as output.

[1310] Specific behavior:

[1311] The device builds question data and emotion data

[1312] Issue an HTTP request and send the question data and emotion data to the server

[1313] Step 3: Server receives question and emotion data

[1314] The server receives the question text and emotion data sent from the device, processes the received data, and obtains the question text and emotion data as input. As output, the server passes these data to the next processing step.

[1315] Specific behavior:

[1316] The server receives the HTTP request

[1317] Extract question text and sentiment data and pass them on to the next processing step

[1318] Step 4: Parsing the Question

[1319] The server properly parses the received question text and prepares it in a format that can be passed to the generative AI model. This parsing process includes text preprocessing (cleaning and tokenization). As input, it receives the user's question text and generates data in a format that can be passed to the generative AI model as output.

[1320] Specific behavior:

[1321] The server pre-processes the question text (removing unnecessary spaces, symbols, etc.)

[1322] Tokenization and conversion to a data format suitable for generative AI models

[1323] Step 5: Generate an answer

[1324] The server passes the analyzed question text and emotion data to the generative AI model to generate an answer. The generated answer takes the emotion data into account. The server receives the analyzed question text and emotion data as input and generates an answer text as output.

[1325] Specific behavior:

[1326] Input the user's question text and emotion data into the generative AI model

[1327] The model generates an answer (e.g., "The first moon landing took place in 1969")

[1328] Step 6: Evaluate the accuracy of your answers

[1329] The generated answer is passed to the accuracy evaluation module, where the accuracy is evaluated using an internal database and statistical models. The generated answer text is received as input, and the accuracy evaluation result (e.g., answer accuracy: 90%) is obtained as output.

[1330] Specific behavior:

[1331] The accuracy assessment module evaluates the accuracy and reliability of the answer text

[1332] Output evaluation results as numerical and qualitative information

[1333] Step 7: Consider user sentiment

[1334] The server integrates the user's emotional data with the generated answer and its accuracy assessment, and adds follow-up information as needed.The server receives the answer text, accuracy assessment results, and emotional data as input, and generates the integrated answer data as output.

[1335] Specific behavior:

[1336] Integrates answer text, accuracy evaluation results, and sentiment data

[1337] Add follow-up information or support links (e.g., "That's very interesting!")

[1338] Step 8: Server generates and sends evaluation results

[1339] The server converts the integrated response data into a final format and transmits it to the terminal. The server receives the integrated response data as input and generates data to be transmitted to the terminal as output.

[1340] Specific behavior:

[1341] Build aggregated response data (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[1342] Send data to the device

[1343] Step 9: Displaying the evaluation results on the device

[1344] The terminal receives the evaluation results sent from the server and displays them to the user. As input, it receives the data sent from the server and displays it to the user as output.

[1345] Specific behavior:

[1346] Display the received data on the screen (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[1347] The user checks the displayed information

[1348] Through the above processing steps, users can receive information that takes into account the accuracy and emotions of the generated AI model's answers.

[1349] (Application example 2)

[1350] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1351] Conventional systems using generative AI models focus on generating answers to user questions and evaluating their accuracy, but they lack information that takes into account the user's emotions. This results in a poor user experience, especially in situations where the user is feeling stressed or anxious, and a lack of appropriate follow-up. This makes it necessary to improve the user's sense of security and satisfaction.

[1352] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1353] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time, and means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration. This allows the user to receive not only the accuracy of the generated answer but also appropriate support information according to their emotions, thereby improving their sense of security and satisfaction.

[1354] The "means for receiving questions entered by a user" is a function for collecting questions entered by a user using a device as data and transmitting the data to a system such as a server.

[1355] "Means for generating answers using a generative AI model" is a function that uses artificial intelligence to generate appropriate answers based on received questions.

[1356] The "means for evaluating the accuracy of the generated answer" is a function for evaluating the accuracy and reliability of the generated answer and expressing it as a number, percentage, etc.

[1357] The "means for displaying the evaluated accuracy to the user" is a function for displaying the generated answer on the user's device so that the accuracy of the answer can be confirmed.

[1358] "Means for recognizing user emotions in real time" is a function for analyzing and recognizing the emotional state (for example, joy or anxiety) of the user when inputting a question in real time.

[1359] "Means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration" is a function for analyzing the user's emotional data and providing a response that adds an appropriate emotional response to the generated answer.

[1360] The present invention relates to a system that evaluates the accuracy of answers generated by a generative AI model and provides information taking into account the user's emotions. The system receives a user's question, generates an appropriate answer, evaluates the accuracy of the answer, and provides a response taking into account the user's emotions.

[1361] Hardware and software used

[1362] Hardware: Smartphones, servers

[1363] Software: Emotion recognition engine (e.g., Python library emotion_recognition), generative AI model, web server (e.g., Flask or Django)

[1364] Processing flow

[1365] The system operates in the following steps:

[1366] User Perspective

[1367] Users input questions using a smartphone interface. At the same time, an emotion recognition engine recognizes the user's emotions in real time. For example, a user might input, "I think my phone has been hacked."

[1368] Terminal perspective

[1369] The device receives the question and emotion data entered by the user and sends it to the server using an HTTP request.

[1370] Server Perspective

[1371] 1. Receiving Questions

[1372] The server receives questions sent from the device and generates answers using a generative AI model.

[1373] 2. Receiving Emotion Data

[1374] The server receives the emotion data sent from the terminal and takes the emotion information into consideration when generating a response based on this data and when handling the response.

[1375] 3. Answer generation and accuracy evaluation

[1376] The generative AI model responds to user questions by generating appropriate answers, such as, "Your phone may have been hacked. Please change your password immediately and check for suspicious activity."

[1377] Next, the accuracy analysis module evaluates the accuracy of the generated answers and calculates a value such as "Answer accuracy: 85%."

[1378] 4. Considering user emotions and generating responses

[1379] The generated answer is evaluated based on the user's sentiment data, along with the accuracy of the answer. For example, if the user's sentiment is negative, appropriate follow-up information or support links will be added. A response such as "Your phone may have been hacked. Please change your password immediately and check for suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." may be generated.

[1380] 5. Sending the Response

[1381] The server sends the generated response to the terminal.

[1382] Terminal view (again)

[1383] The device receives the response sent from the server and displays it to the user, who can then check the reliability of the generated AI model's answer as well as support information that takes their own emotions into consideration.

[1384] Specific examples

[1385] For example, if a user asks, "My phone may have been hacked," the server receives this question and the user's emotional data at the time. The generative AI model generates an answer: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity." The accuracy evaluation module evaluates the accuracy of this answer as "Answer accuracy: 85%." The server then takes the user's emotional data into consideration and generates an emotionally sensitive response: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." This is then sent to the device, where it is displayed to the user.

[1386] Prompt Sentence Examples

[1387] User Query: "I think my phone might have been hacked."

[1388] Generate a calm and reassuring response considering the potential anxiety of the user. Ensure the response is concise and provides actionable steps to secure against hacking.

[1389] In this way, users can receive appropriate answers from the generative AI model as well as support information that takes their own emotions into consideration, allowing them to use the system with peace of mind.

[1390] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1391] Step 1:

[1392] The user inputs a question using the smartphone interface. Based on the user's input, the emotion recognition engine recognizes the user's emotion in real time. In this process, the user's question text is used as input data and the recognized emotion data is generated as output.

[1393] Step 2:

[1394] The terminal receives the question and emotion data entered by the user. The received data is sent to the server as an HTTP request. In this process, an HTTP request is generated that includes the user question text and emotion data as input data and is sent to the server as output data.

[1395] Step 3:

[1396] The server receives the question and emotion data sent from the device. The received question is converted into a format that can be interpreted by the generative AI model, and then passed to the generative AI model. In this step, the input data includes the user question text and emotion data, and the output data is the question text formatted to be passed to the generative AI model.

[1397] Step 4:

[1398] The server uses a generative AI model to generate an answer to the received question. The server provides the user question text as input to the generative AI model and obtains the generated answer text as output.

[1399] Step 5:

[1400] The server evaluates the accuracy of the generated answer. Using an accuracy evaluation module, the reliability of the generated answer text is quantified. This process includes the generated answer text as input data, and an accuracy evaluation value indicating reliability is obtained as output data.

[1401] Step 6:

[1402] The server integrates the evaluated accuracy into the answer to the question and generates an emotion-sensitive response based on the user's emotion data. For example, if a negative emotion is recognized, additional supporting information is added. In this step, the input data includes the generated answer text, the evaluated accuracy, and the user's emotion data, and the output data is a final emotion-sensitive response sentence.

[1403] Step 7:

[1404] The server sends the final response to the terminal. The final response, which takes emotion into consideration, is included in the transmission data.

[1405] Step 8:

[1406] The terminal receives the response sent from the server and displays it to the user. The user confirms the displayed information. This process is completed when the terminal receives the data from the server and displays it on the screen.

[1407] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1408] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1409] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1410] [Fourth embodiment]

[1411] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1412] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1413] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1414] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1415] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1416] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1417] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1418] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1419] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1420] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1421] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1422] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1423] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1424] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user.

[1425] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[1426] User Perspective

[1427] A user inputs a question using the interface of a device (e.g., a PC or smartphone), for example, "Tell me about the planets in the solar system."

[1428] Terminal perspective

[1429] The terminal receives the question entered by the user, sends it to the server, usually in the form of an HTTP request, and prepares to receive a response from the server.

[1430] Server Perspective

[1431] 1. Receiving Questions

[1432] The server receives the question sent from the device. This is called reception processing. After receiving the question, it analyzes it appropriately and prepares it in a format that can be passed to the generative AI model.

[1433] 2. Answer Generation

[1434] The server passes the question to the generative AI model, which generates an answer. The generative AI model uses its internal knowledge base and algorithms to create the best answer to the question. For example, it generates the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1435] 3. Evaluating the accuracy of answers

[1436] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[1437] 4. Submitting the evaluation results

[1438] The server combines the evaluated accuracy with the answer and sends it to the device. For example, it creates data such as "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1439] Terminal view (again)

[1440] The terminal receives the answer and accuracy information sent from the server and displays them to the user, who can then judge the reliability of the answer.

[1441] Specific examples

[1442] For example, if a user asks, "When was the first moon landing?", the server receives this question and passes it to the generative AI model. The generative AI model generates the answer, "The first moon landing took place in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then sends the data, "The first moon landing took place in 1969. (Answer accuracy: 90%)," to the device, which displays it to the user. The user sees this information and determines that the answer is relatively reliable.

[1443] This system allows users to easily determine the accuracy of the answers generated by the AI ​​model, preventing the spread of misinformation.

[1444] The processing flow will be explained below.

[1445] Step 1:

[1446] The user types a question into an input field on the device, for example, "Tell me about the planets in the solar system."

[1447] Step 2:

[1448] The terminal sends the user's input as text data to the server using an HTTP request.

[1449] Step 3:

[1450] The server receives the question sent from the device, analyzes it, and prepares it in a format that can be passed to the generative AI model.

[1451] Step 4:

[1452] The server passes the analyzed question to the generative AI model and requests it to generate an answer. For example, it passes a question like, "Tell me about the planets in the solar system."

[1453] Step 5:

[1454] A generative AI model generates an answer based on the question, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1455] Step 6:

[1456] The server receives the answers generated by the generative AI model.

[1457] Step 7:

[1458] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[1459] Step 8:

[1460] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[1461] Step 9:

[1462] The server receives the evaluation results from the accuracy evaluation module.

[1463] Step 10:

[1464] The server combines the generated answer and its accuracy assessment to create data to be sent to the device. For example, it generates data such as "Answer: There are eight planets in the solar system. (Answer accuracy: 95%)."

[1465] Step 11:

[1466] The server transmits the integrated data to the terminal.

[1467] Step 12:

[1468] The device displays the answer it received and its accuracy to the user, for example, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1469] Step 13:

[1470] The user checks the information displayed on the device and determines the reliability of the answer.

[1471] Example 1

[1472] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1473] Previous generative AI models lacked mechanisms for evaluating the accuracy and reliability of their answers, making it difficult for users to determine how trustworthy the answers were. Furthermore, there was a lack of means to evaluate the accuracy of generated answers and display them in an easy-to-understand manner, making it impossible to prevent the spread of misinformation.

[1474] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1475] In this invention, the server includes means for receiving text input by a user through a terminal, means for formatting the received text so that it can be passed to a generative AI model, means for generating an answer using the generative AI model, means for passing the generated answer to an accuracy evaluation module to evaluate its accuracy, and means for integrating the evaluated accuracy into the answer and sending it to the terminal. This makes it possible to evaluate the accuracy of the answer of the generative AI model and display it in an easy-to-understand manner for the user.

[1476] "User" refers to the entity that operates the system and inputs questions.

[1477] A "terminal" is an input device used by a user, and includes a personal computer, a smartphone, etc.

[1478] The "means for receiving" refers to a mechanism for inputting a question entered by a user into the system.

[1479] "Server" refers to the back-end computer that parses questions, passes them to the generative AI model, and generates and evaluates answers.

[1480] "Means of analysis" refers to the mechanism that converts the question into a form that the generative AI model can understand.

[1481] A "generative AI model" refers to an artificial intelligence model that generates answers to input questions.

[1482] "Means for generating an answer" refers to a mechanism that uses a generative AI model to create an answer to a question.

[1483] "Accuracy Assessment Module" refers to a rating system for measuring the accuracy of generated answers.

[1484] The "means for evaluating accuracy" refers to a mechanism that outputs the reliability of the generated answer as an evaluation value.

[1485] "Means for sending" refers to a communication means for sending the evaluated answers to the terminal.

[1486] The "means for displaying" refers to a mechanism for visually showing the answer including the evaluation result to the user.

[1487] The present invention relates to a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user. This system transmits a question entered by a user through a terminal to a server, generates an answer to the question using a generative AI model, evaluates the accuracy of the answer, and displays it to the user.

[1488] The overall system configuration consists of three main components: the user, the terminal, and the server. The specific operation of each component is explained below.

[1489] User Perspective

[1490] A user inputs a question using a device such as a personal computer or smartphone. For example, the user inputs a question such as, "Tell me about the planets in the solar system."

[1491] Terminal perspective

[1492] The terminal receives the question entered by the user and sends it to the server, usually using a protocol called an HTTP request, and prepares to receive a response from the server.

[1493] Server Perspective

[1494] The server proceeds with the process in the following steps.

[1495] 1. Receiving and analyzing questions

[1496] The server receives the query sent from the terminal.

[1497] The received question is analyzed and formatted into a form that can be passed to the generative AI model.

[1498] 2. Answer Generation

[1499] The server inputs the formulated question into a generative AI model to generate an answer.

[1500] Large-scale language models such as GPT-3 are used as generative AI models.

[1501] For example, generate the answer "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1502] 3. Evaluating the accuracy of answers

[1503] The generated answer is passed to an accuracy evaluation module to evaluate its accuracy.

[1504] Accuracy assessment uses internal databases and statistical models.

[1505] For example, the evaluation result may be "Answer accuracy: 95%."

[1506] 4. Submitting the evaluation results

[1507] The server generates the data "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)" and sends it to the terminal.

[1508] Specific examples

[1509] For example, if a user asks, "When was the first moon landing?" the system will act as follows:

[1510] 1. User types, "When was the first moon landing?"

[1511] 2. The device sends a question to the server.

[1512] 3. The server receives the question and formats it for passing to the generative AI model.

[1513] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[1514] 5. The accuracy evaluation module evaluates the accuracy of this answer and obtains the result "Answer accuracy: 90%".

[1515] 6. The server generates the data "The first moon landing took place in 1969. (Answer accuracy: 90%)" and sends it to the device.

[1516] 7. The terminal receives this and displays it to the user.

[1517] In this way, users can easily check the accuracy of the generative AI model's answers and obtain reliable information.

[1518] Prompt Sentence Examples

[1519] For example, if a user enters the question "Tell me about the planets in the solar system," the prompt to the generative AI model will be as follows:

[1520] User Question: "Tell me about the planets in the solar system."

[1521] Input to generative AI model: "How many planets are there in the solar system? What are their names?"

[1522] This system makes it possible to evaluate the accuracy of the generative AI model's answers and present them to the user in a clear manner.

[1523] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1524] Step 1:

[1525] The user inputs a question through the terminal.

[1526] Specific behavior:

[1527] Users enter their questions using a computer or smartphone interface.

[1528] For example, type "Tell me about the planets in the solar system" and press the send button.

[1529] input:

[1530] User question: "Tell me about the planets in the solar system."

[1531] output:

[1532] User-entered question data

[1533] Step 2:

[1534] The terminal receives the user's question and sends it to the server.

[1535] Specific behavior:

[1536] The terminal sends the user's question to the server in the form of an HTTP request.

[1537] It is necessary to specify the destination URL or API endpoint.

[1538] input:

[1539] User-entered question data

[1540] output:

[1541] HTTP request sent to the server

[1542] Step 3:

[1543] The server receives the question and formats it for passing to the generative AI model.

[1544] Specific behavior:

[1545] The server receives the HTTP request and analyzes the question.

[1546] The question content is converted into a format that is easy for the generative AI model to understand, such as JSON format.

[1547] For example, reframe the question as, "How many planets are there in the solar system? What are their names?"

[1548] input:

[1549] HTTP request sent to the server

[1550] output:

[1551] Formatted question data

[1552] Step 4:

[1553] The server generates an answer using a generative AI model.

[1554] Specific behavior:

[1555] The server passes the formatted question to the generative AI model.

[1556] Generative AI models (e.g., GPT-3) generate answers using internal knowledge bases and algorithms.

[1557] For example, you might get the answer, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1558] input:

[1559] Formatted question data

[1560] output:

[1561] Answers generated by generative AI models

[1562] Step 5:

[1563] The server passes the generated answers to an accuracy assessment module.

[1564] Specific behavior:

[1565] The server sends the generated answers to an accuracy assessment module.

[1566] An accuracy assessment module examines the answers and compares them against various assessment metrics.

[1567] input:

[1568] Answers generated by generative AI models

[1569] output:

[1570] Data to be evaluated passed to the accuracy evaluation module

[1571] Step 6:

[1572] An accuracy assessment module assesses the accuracy of the answers.

[1573] Specific behavior:

[1574] We use internal databases and statistical models to quantify the accuracy of responses.

[1575] For example, the evaluation result may be "Answer accuracy: 95%."

[1576] input:

[1577] Evaluation data passed to the accuracy evaluation module

[1578] output:

[1579] Evaluation result data (e.g., "Answer accuracy: 95%")

[1580] Step 7:

[1581] The server integrates the evaluation results with the answers and sends them to the terminal.

[1582] Specific behavior:

[1583] The server integrates the evaluation results into an answer and generates the data, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1584] This data is sent to the terminal as an HTTP response.

[1585] input:

[1586] Assessment result data and generated answers

[1587] output:

[1588] Answer and evaluation data sent to the device

[1589] Step 8:

[1590] The device displays the answer and accuracy information to the user.

[1591] Specific behavior:

[1592] The terminal analyzes the data received from the server and displays it in an appropriate format on the user interface.

[1593] For example, the screen might display, "There are eight planets in the solar system: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune. (Answer accuracy: 95%)."

[1594] input:

[1595] Answer and evaluation data sent to the device

[1596] output:

[1597] Answer and rating information displayed to users

[1598] (Application example 1)

[1599] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1600] There is a need for a system that makes it easy for users to understand the reliability of answers provided by generative AI models. In particular, in customer support applications on online shopping sites, it is necessary to accurately evaluate the accuracy of answers to user questions and display the results in an easy-to-understand manner to improve user trust and usability.

[1601] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1602] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer to the question using a generative AI model, means for evaluating the accuracy of the generated answer, and means for integrating the generated answer and its evaluated accuracy and displaying it to the user. This allows the user to easily determine the reliability of the information provided and receive high-quality support.

[1603] A "user" is an individual or entity that utilizes the system to enter questions and receive answers.

[1604] A "question" is a sentence or phrase that a user enters into the system to solicit an answer.

[1605] A "generative AI model" is an artificial intelligence model that generates the optimal answer to a given question.

[1606] An "answer" is information that a generative AI model generates based on a user's question.

[1607] "Accuracy" is an indicator of the reliability and accuracy of the generated answer.

[1608] "Evaluation" means measuring the accuracy of the generated answer and quantifying its reliability.

[1609] The "server" is the core computing system responsible for receiving user questions, generating answers using the generative AI model, evaluating their accuracy, and providing the results to the user.

[1610] "API communication" is a standardized means of communication for sending and receiving data between different software components.

[1611] "System" is a general term for a series of means and devices for realizing all of the above functions.

[1612] "Synthesis" is the process of combining generated answers and their accuracy information and presenting them to the user in a single format.

[1613] This invention shows an application example aimed at customer support on a forum website. This system uses a generative AI model to generate answers to questions entered by users, evaluates their accuracy, and displays them to the user. The following describes how this system is specifically implemented.

[1614] The entire system consists of a user, a device (e.g., a smartphone), and a server. The device basically receives questions from the user and provides a means for communication with the server. The server has a generative AI model and is capable of generating optimal answers based on the questions and evaluating their accuracy.

[1615] The device receives the user's question and sends it to the server. When the user enters a question such as "Is this product waterproof?" into the device, the question is sent to the server as an HTTP request. The server receives the question, parses it into an appropriate format, and passes it to the generative AI model. The generative AI model uses its internal knowledge base and algorithms to generate the optimal answer. In this case, OpenAI GPT-4 is used as the generative AI model, for example.

[1616] The generated answer will be in the form of "Yes, this product is waterproof." This answer is not passed on as is, but is further passed to the accuracy evaluation module. The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the generated answer. For example, it evaluates whether this answer can be trusted with 95% accuracy. In this case, it is evaluated as "Answer accuracy: 95%."

[1617] After the evaluation is complete, the server combines the answer and its accuracy to generate data. For example, it may be in the form of "Yes, this product is waterproof. (Answer accuracy: 95%)." This data is again sent to the device as an HTTP response. The device displays the received answer and its accuracy to the user, allowing the user to judge the reliability of the answer.

[1618] As a concrete example, consider a customer support application that works like this:

[1619] 1. The user types in a question. For example, "Is this product waterproof?"

[1620] 2. The application sends the question to the server as an HTTP request.

[1621] 3. The server generates an answer using the generative AI model and evaluates its accuracy.

[1622] 4. The combined answer and accuracy are sent back to the device.

[1623] 5. The application displays to the user, "Yes, this product is waterproof. (Answer accuracy: 95%)."

[1624] By implementing this system, users can easily judge the reliability of the information provided and receive high-quality support.

[1625] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1626] Step 1:

[1627] The user types a question into the terminal.

[1628] Input: A question typed by the user in natural language (e.g., "Is this product waterproof?").

[1629] Specific behavior: The user enters a question into the application's text box and presses the send button, which enters the question into the terminal.

[1630] Output: The question text is stored in the terminal's internal buffer.

[1631] Step 2:

[1632] The terminal sends the entered question to the server.

[1633] Input: The question entered by the user in step 1.

[1634] Specific operation: The terminal creates an HTTP request and sends the user's question to the server using the HTTP protocol.

[1635] Output: The question is sent to the server as an HTTP request.

[1636] Step 3:

[1637] The server receives the user's query.

[1638] Input: The HTTP request containing the question sent from the terminal in step 2.

[1639] What happens: The server receives an HTTP request and extracts the question text from the payload.

[1640] Output: The extracted question text is imported into the server's internal data.

[1641] Step 4:

[1642] The server passes the question to a generative AI model to generate an answer.

[1643] Input: The question text received by the server.

[1644] How it works: The server passes the question text to a generative AI model (e.g., OpenAI GPT-4), which then generates the optimal answer. The generative AI model uses its internal knowledge base to analyze the sentence and create an appropriate answer.

[1645] Output: The generated answer text (e.g., "Yes, this product is waterproof.").

[1646] Step 5:

[1647] The server evaluates the accuracy of the generated answer.

[1648] Input: The answer text generated in step 4.

[1649] What it does: Passes the generated answer to an accuracy assessment module, which uses an internal database and statistical models to quantify its confidence (e.g., "Answer accuracy: 95%").

[1650] Output: The accuracy value of the evaluated answer.

[1651] Step 6:

[1652] The server aggregates the answers and their accuracy to generate data.

[1653] Input: The answer text generated in step 4, and the accuracy value evaluated in step 5.

[1654] What it does: Combines the answer text and its accuracy and assembles the data into a format to present to the user (e.g., "Yes, this product is waterproof. (Answer accuracy: 95%)").

[1655] Output: Text data of the consolidated answers and accuracy.

[1656] Step 7:

[1657] The server sends the consolidated answer and accuracy data to the device.

[1658] Input: The consolidated data generated in step 6.

[1659] Specific operation: Generates integrated data as an HTTP response and sends it to the terminal.

[1660] Output: HTTP response containing the integration data.

[1661] Step 8:

[1662] The terminal receives the integrated data from the server and displays it to the user.

[1663] Input: The HTTP response containing the integration data sent by the server in step 7.

[1664] Specific operation: The device receives the HTTP response, extracts the answer and accuracy text from the payload, and displays them on the screen.

[1665] Output: The user will see the message "Yes, this product is waterproof. (Answer accuracy: 95%)."

[1666] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1667] The present invention relates to a system that combines a system that displays the accuracy of a generative AI model's answer in an easy-to-understand manner to a user with an emotion engine that recognizes the user's emotions. This system receives a question entered by a user, generates an answer using a generative AI model, evaluates the accuracy of the answer, and displays the evaluation result to the user, while also taking the user's emotions into consideration.

[1668] The specific operation of this system will be explained below from the viewpoints of the user, the terminal, and the server.

[1669] User Perspective

[1670] The user inputs a question using the interface of a device (e.g., a PC or smartphone). For example, they input a question such as, "Tell me about the planets in the solar system." While the user is inputting the question, the emotion engine recognizes the user's emotion in real time.

[1671] Terminal perspective

[1672] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server.

[1673] Server Perspective

[1674] 1. Receiving Questions

[1675] The server receives the question sent from the device, analyzes it appropriately, and prepares it in a format that can be passed to the generative AI model.

[1676] 2. Receiving Emotion Data

[1677] The server receives the user's emotion data sent from the device, which allows it to take emotion information into account when generating answers and how to handle them.

[1678] 3. Generate answers

[1679] The server passes the question and emotion data to the generative AI model and requests it to generate an answer. For example, in response to the question, "Tell me about the planets in the solar system," the model generates an answer that takes emotion data into account: "There are eight planets in the solar system. I'm feeling great today!: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune."

[1680] 4. Evaluating the accuracy of answers

[1681] The generated answer is passed to the accuracy evaluation module, where the accuracy of the answer is evaluated. An internal database and statistical models are used for accuracy evaluation. For example, the reliability of the answer is quantified and evaluated as "Answer accuracy: 95%."

[1682] 5. Considering user emotions

[1683] The generated answer is evaluated for accuracy, and the user's sentiment data is taken into account. For example, if the user's sentiment is negative, appropriate follow-up information or support links can be added.

[1684] 6. Submitting the evaluation results

[1685] The server combines the evaluated accuracy with the answer to create data to be sent to the device, such as "Answer: There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[1686] Terminal view (again)

[1687] The device receives the answer and accuracy information sent from the server and displays it to the user. For example, it displays "There are eight planets in the solar system. You're feeling great today! (Answer accuracy: 95%)." By looking at this display, the user can confirm the reliability of the answer as well as the information that takes their feelings into consideration.

[1688] Specific examples

[1689] For example, if a user asks, "When was the first moon landing?", the server receives this question and the user's emotional data at the time. The generative AI model generates the answer, "The first moon landing occurred in 1969," and the accuracy evaluation module evaluates the accuracy of this answer, setting it at "Answer accuracy: 90%." The server then takes the user's emotional data into consideration and sends the following data to the device: "The first moon landing occurred in 1969. (Answer accuracy: 90%) That's very interesting!" The device then displays this to the user. Seeing this information, the user determines that the answer is reliable and feels reassured that their emotional state has been taken into consideration.

[1690] This system not only allows users to easily judge the accuracy of the generated AI model's answers, but also provides information that takes emotions into consideration, resulting in a more user-friendly interface.

[1691] The processing flow will be explained below.

[1692] Step 1:

[1693] The user enters a question into the input field of the device. For example, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotions (e.g., joy, surprise, anxiety, etc.) in real time.

[1694] Step 2:

[1695] The device sends the question and emotion data entered by the user to the server as text data using an HTTP request.

[1696] Step 3:

[1697] The server receives the question sent from the device, then separates and analyzes the question and emotion data, and prepares them in a format that can be passed to the generative AI model.

[1698] Step 4:

[1699] The server passes the analyzed question and emotion data to the generation AI model and requests it to generate an answer. For example, it passes the question "Tell me about the planets in the solar system" and emotion data that "the user is excited."

[1700] Step 5:

[1701] A generative AI model generates an answer based on the question and sentiment data, such as "There are eight planets in the solar system. Amazing!"

[1702] Step 6:

[1703] The server receives the answers generated by the generative AI model.

[1704] Step 7:

[1705] The server passes the generated answer to the answer accuracy evaluation module and requests an accuracy evaluation. For example, it passes the answer "There are eight planets in the solar system."

[1706] Step 8:

[1707] The accuracy evaluation module uses an internal database and statistical models to evaluate the accuracy of the answer, for example, "Answer accuracy: 95%."

[1708] Step 9:

[1709] The server receives the evaluation results from the accuracy evaluation module.

[1710] Step 10:

[1711] The server combines the generated answers and their accuracy ratings, and creates data to send to the device, taking into account emotional data. For example, it generates data such as "Answer: There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[1712] Step 11:

[1713] The server sends this integrated data to the terminal.

[1714] Step 12:

[1715] The device displays the received answer and a comment that takes into account the accuracy and sentiment of the user. For example, it displays "There are eight planets in the solar system. Amazing! (Answer accuracy: 95%)."

[1716] Step 13:

[1717] Users can check the information displayed on their device and receive information that takes their emotions into consideration along with the reliability of the answers, making it easier for them to accept the content.

[1718] As a result, users can obtain reliable answers and receive information that takes their emotions into consideration, resulting in a more natural and satisfying information experience.

[1719] Example 2

[1720] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1721] Conventional generative AI systems generate answers to user questions, but they have difficulty informing users of the accuracy and reliability of the answers. Furthermore, they provide answers without considering the user's emotions, resulting in a poor user experience. To solve these problems, it is necessary to evaluate the accuracy of the generated answers, clearly communicate this to the user, and provide information that takes the user's emotions into consideration.

[1722] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1723] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time using an emotion engine, means for transmitting the recognized emotion data to the server, and means for generating an answer taking the emotion data into consideration. This makes it possible to provide more appropriate information that takes the user's emotions into consideration while indicating the reliability of the generated answer to the user.

[1724] A "user" is an entity that uses the system to input questions and receives information via a terminal.

[1725] A "question" refers to the inquiry a user enters into the system, and is text data that is analyzed by the generative AI model to generate an answer.

[1726] A "generative AI model" is an algorithm or program that generates natural language answers based on questions entered by a user.

[1727] An "answer" refers to the response text that a generative AI model generates in response to a question, and is the information provided to the user.

[1728] "Accuracy" is an indicator that evaluates the accuracy and reliability of the generated answers, and is displayed numerically or qualitatively.

[1729] An "emotion engine" is hardware or software that recognizes and analyzes a user's emotions in real time.

[1730] "Emotion data" is information obtained as a result of the emotion engine recognizing and analyzing the user's emotions, and is data that indicates the user's psychological state.

[1731] The "server" is a computer system that processes questions and emotion data received from the device, generates answers using a generative AI model, evaluates them, and sends them to the device.

[1732] A "terminal" is a device used by a user to access the system, enter questions, and receive answers, and includes a personal computer, smartphone, etc.

[1733] An "HTTP request" is a communication protocol used to send questions and emotional data from a device to a server.

[1734] The "accuracy assessment module" is a software module for assessing the accuracy and reliability of generated answers, using an internal database and statistical models.

[1735] "Follow-up information" refers to additional information or assistance provided depending on the user's emotional state, and is information intended to improve the user experience.

[1736] "Natural language" refers to the language that people use on a daily basis, and is the subject of processing by systems in the form of text, speech, etc.

[1737] The present invention is a system that uses a generative AI model to generate answers to questions entered by users, evaluates the accuracy of the answers, and displays them after taking into account the user's emotions.

[1738] A means for receiving user-entered questions

[1739] A user inputs a question using a device such as a PC or smartphone. For example, a user might input a question such as, "Tell me about the planets in the solar system." At this time, the emotion engine recognizes the user's emotion in real time. The emotion engine includes software that analyzes facial expressions, tone of voice, and other factors.

[1740] Submitting questions and emotion data

[1741] The device receives the question entered by the user as text data and sends it to the server using an HTTP request. The device also sends the recognition results of the emotion engine to the server at the same time.

[1742] Receiving and analyzing questions and emotion data

[1743] The server receives the question and emotion data sent from the device, parses the question into an appropriate format, and converts it into a format that can be passed to the generative AI model. This analysis includes text preprocessing and tokenization.

[1744] Generate answers

[1745] The server then passes the analyzed question and emotion data to the generative AI model to generate an answer. For example, in response to the question, "Tell me about the planets in our solar system," the model generates the answer, "There are eight planets in our solar system. I'm in a good mood today!"

[1746] Answer accuracy assessment

[1747] The generated answers are passed to an accuracy evaluation module in the server, which uses an internal database and statistical models to evaluate the accuracy and reliability of the answers, for example, rating them as "Answer accuracy: 95%."

[1748] Considering user emotions

[1749] The server integrates the user's sentiment data into the generated answer and its accuracy rating. If the user's sentiment is negative, appropriate follow-up information or support links can be added. For example, the data can be configured as "There are eight planets in the solar system. I'm in a good mood today! (Answer accuracy: 95%)."

[1750] Server generates and transmits evaluation results

[1751] The server then sends the final answer, which integrates the evaluated accuracy and emotional data, to the device, allowing the user to confirm the reliability of the generated answer while receiving information that takes their emotions into consideration.

[1752] Displaying evaluation results on a device

[1753] The device receives the evaluation results sent from the server and displays them to the user. For example, it displays "There are eight planets in the solar system. You're in a good mood today! (Answer accuracy: 95%)."

[1754] Specific examples

[1755] For example, if a user asks, "When was the first moon landing?" the system operates as follows:

[1756] 1. A user types into a terminal, "When was the first moon landing?"

[1757] 2. The device sends the question and emotion data to the server.

[1758] 3. The server receives the question and emotion data and converts it into a format that can be passed to the generative AI model.

[1759] 4. The generative AI model generates the answer, "The first moon landing took place in 1969."

[1760] 5. The server evaluates the accuracy of this answer as "Answer accuracy: 90%".

[1761] 6. The server considers the user's emotional data and integrates it into the data "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!"

[1762] 7. The server sends this integrated data to the terminal.

[1763] 8. The terminal displays the received data to the user.

[1764] This embodiment allows users to check the accuracy of answers provided by the generative AI model and receive information that takes emotions into account.

[1765] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1766] Step 1: User enters question

[1767] A user inputs a question using their own device (such as a PC or smartphone). For example, they might input a question like, "When was the first moon landing?" At this time, the emotion engine recognizes and analyzes the user's emotions in real time. The inputs are the user's question text and the recognized emotion data.

[1768] Specific behavior:

[1769] A user types into the question form, "When was the first moon landing?"

[1770] The emotion engine analyzes the user's facial expressions and tone of voice to generate emotion data (e.g., interesting).

[1771] Step 2: Device sends question and emotion data

[1772] The device receives the question entered by the user and the emotion data generated by the emotion engine, and sends them to the server using an HTTP request. The device receives the user's question text and emotion data as input, and sends them to the server as output.

[1773] Specific behavior:

[1774] The device builds question data and emotion data

[1775] Issue an HTTP request and send the question data and emotion data to the server

[1776] Step 3: Server receives question and emotion data

[1777] The server receives the question text and emotion data sent from the device, processes the received data, and obtains the question text and emotion data as input. As output, the server passes these data to the next processing step.

[1778] Specific behavior:

[1779] The server receives the HTTP request

[1780] Extract question text and sentiment data and pass them on to the next processing step

[1781] Step 4: Parsing the Question

[1782] The server properly parses the received question text and prepares it in a format that can be passed to the generative AI model. This parsing process includes text preprocessing (cleaning and tokenization). As input, it receives the user's question text and generates data in a format that can be passed to the generative AI model as output.

[1783] Specific behavior:

[1784] The server pre-processes the question text (removing unnecessary spaces, symbols, etc.)

[1785] Tokenization and conversion to a data format suitable for generative AI models

[1786] Step 5: Generate an answer

[1787] The server passes the analyzed question text and emotion data to the generative AI model to generate an answer. The generated answer takes the emotion data into account. The server receives the analyzed question text and emotion data as input and generates an answer text as output.

[1788] Specific behavior:

[1789] Input the user's question text and emotion data into the generative AI model

[1790] The model generates an answer (e.g., "The first moon landing took place in 1969")

[1791] Step 6: Evaluate the accuracy of your answers

[1792] The generated answer is passed to the accuracy evaluation module, where the accuracy is evaluated using an internal database and statistical models. The generated answer text is received as input, and the accuracy evaluation result (e.g., answer accuracy: 90%) is obtained as output.

[1793] Specific behavior:

[1794] The accuracy assessment module evaluates the accuracy and reliability of the answer text

[1795] Output evaluation results as numerical and qualitative information

[1796] Step 7: Consider user sentiment

[1797] The server integrates the user's emotional data with the generated answer and its accuracy assessment, and adds follow-up information as needed.The server receives the answer text, accuracy assessment results, and emotional data as input, and generates the integrated answer data as output.

[1798] Specific behavior:

[1799] Integrates answer text, accuracy evaluation results, and sentiment data

[1800] Add follow-up information or support links (e.g., "That's very interesting!")

[1801] Step 8: Server generates and sends evaluation results

[1802] The server converts the integrated response data into a final format and transmits it to the terminal. The server receives the integrated response data as input and generates data to be transmitted to the terminal as output.

[1803] Specific behavior:

[1804] Build aggregated response data (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[1805] Send data to the device

[1806] Step 9: Displaying the evaluation results on the device

[1807] The terminal receives the evaluation results sent from the server and displays them to the user. As input, it receives the data sent from the server and displays it to the user as output.

[1808] Specific behavior:

[1809] Display the received data on the screen (e.g., "The first moon landing took place in 1969. (Answer accuracy: 90%) Very interesting!")

[1810] The user checks the displayed information

[1811] Through the above processing steps, users can receive information that takes into account the accuracy and emotions of the generated AI model's answers.

[1812] (Application example 2)

[1813] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1814] Conventional systems using generative AI models focus on generating answers to user questions and evaluating their accuracy, but they lack information that takes into account the user's emotions. This results in a poor user experience, especially in situations where the user is feeling stressed or anxious, and a lack of appropriate follow-up. This makes it necessary to improve the user's sense of security and satisfaction.

[1815] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1816] In this invention, the server includes means for receiving a question entered by a user, means for generating an answer using a generative AI model based on the received question, means for evaluating the accuracy of the generated answer, means for displaying the evaluated accuracy to the user, means for recognizing the user's emotions in real time, and means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration. This allows the user to receive not only the accuracy of the generated answer but also appropriate support information according to their emotions, thereby improving their sense of security and satisfaction.

[1817] The "means for receiving questions entered by a user" is a function for collecting questions entered by a user using a device as data and transmitting the data to a system such as a server.

[1818] "Means for generating answers using a generative AI model" is a function that uses artificial intelligence to generate appropriate answers based on received questions.

[1819] The "means for evaluating the accuracy of the generated answer" is a function for evaluating the accuracy and reliability of the generated answer and expressing it as a number, percentage, etc.

[1820] The "means for displaying the evaluated accuracy to the user" is a function for displaying the generated answer on the user's device so that the accuracy of the answer can be confirmed.

[1821] "Means for recognizing user emotions in real time" is a function for analyzing and recognizing the emotional state (for example, joy or anxiety) of the user when inputting a question in real time.

[1822] "Means for processing the user's emotional data based on the generated answer and providing a response that takes the user's emotions into consideration" is a function for analyzing the user's emotional data and providing a response that adds an appropriate emotional response to the generated answer.

[1823] The present invention relates to a system that evaluates the accuracy of answers generated by a generative AI model and provides information taking into account the user's emotions. The system receives a user's question, generates an appropriate answer, evaluates the accuracy of the answer, and provides a response taking into account the user's emotions.

[1824] Hardware and software used

[1825] Hardware: Smartphones, servers

[1826] Software: Emotion recognition engine (e.g., Python library emotion_recognition), generative AI model, web server (e.g., Flask or Django)

[1827] Processing flow

[1828] The system operates in the following steps:

[1829] User Perspective

[1830] Users input questions using a smartphone interface. At the same time, an emotion recognition engine recognizes the user's emotions in real time. For example, a user might input, "I think my phone has been hacked."

[1831] Terminal perspective

[1832] The device receives the question and emotion data entered by the user and sends it to the server using an HTTP request.

[1833] Server Perspective

[1834] 1. Receiving Questions

[1835] The server receives questions sent from the device and generates answers using a generative AI model.

[1836] 2. Receiving Emotion Data

[1837] The server receives the emotion data sent from the terminal and takes the emotion information into consideration when generating a response based on this data and when handling the response.

[1838] 3. Answer generation and accuracy evaluation

[1839] The generative AI model responds to user questions by generating appropriate answers, such as, "Your phone may have been hacked. Please change your password immediately and check for suspicious activity."

[1840] Next, the accuracy analysis module evaluates the accuracy of the generated answers and calculates a value such as "Answer accuracy: 85%."

[1841] 4. Considering user emotions and generating responses

[1842] The generated answer is evaluated based on the user's sentiment data, along with the accuracy of the answer. For example, if the user's sentiment is negative, appropriate follow-up information or support links will be added. A response such as "Your phone may have been hacked. Please change your password immediately and check for suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." may be generated.

[1843] 5. Sending the Response

[1844] The server sends the generated response to the terminal.

[1845] Terminal view (again)

[1846] The device receives the response sent from the server and displays it to the user, who can then check the reliability of the generated AI model's answer as well as support information that takes their own emotions into consideration.

[1847] Specific examples

[1848] For example, if a user asks, "My phone may have been hacked," the server receives this question and the user's emotional data at the time. The generative AI model generates an answer: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity." The accuracy evaluation module evaluates the accuracy of this answer as "Answer accuracy: 85%." The server then takes the user's emotional data into consideration and generates an emotionally sensitive response: "Your phone may have been hacked. Please change your password immediately and check for any suspicious activity. (Answer accuracy: 85%) Please stay calm. Our team is here to help you." This is then sent to the device, where it is displayed to the user.

[1849] Prompt Sentence Examples

[1850] User Query: "I think my phone might have been hacked."

[1851] Generate a calm and reassuring response considering the potential anxiety of the user. Ensure the response is concise and provides actionable steps to secure against hacking.

[1852] In this way, users can receive appropriate answers from the generative AI model as well as support information that takes their own emotions into consideration, allowing them to use the system with peace of mind.

[1853] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1854] Step 1:

[1855] The user inputs a question using the smartphone interface. Based on the user's input, the emotion recognition engine recognizes the user's emotion in real time. In this process, the user's question text is used as input data and the recognized emotion data is generated as output.

[1856] Step 2:

[1857] The terminal receives the question and emotion data entered by the user. The received data is sent to the server as an HTTP request. In this process, an HTTP request is generated that includes the user question text and emotion data as input data and is sent to the server as output data.

[1858] Step 3:

[1859] The server receives the question and emotion data sent from the device. The received question is converted into a format that can be interpreted by the generative AI model, and then passed to the generative AI model. In this step, the input data includes the user question text and emotion data, and the output data is the question text formatted to be passed to the generative AI model.

[1860] Step 4:

[1861] The server uses a generative AI model to generate an answer to the received question. The server provides the user question text as input to the generative AI model and obtains the generated answer text as output.

[1862] Step 5:

[1863] The server evaluates the accuracy of the generated answer. Using an accuracy evaluation module, the reliability of the generated answer text is quantified. This process includes the generated answer text as input data, and an accuracy evaluation value indicating reliability is obtained as output data.

[1864] Step 6:

[1865] The server integrates the evaluated accuracy into the answer to the question and generates an emotion-sensitive response based on the user's emotion data. For example, if a negative emotion is recognized, additional supporting information is added. In this step, the input data includes the generated answer text, the evaluated accuracy, and the user's emotion data, and the output data is a final emotion-sensitive response sentence.

[1866] Step 7:

[1867] The server sends the final response to the terminal. The final response, which takes emotion into consideration, is included in the transmission data.

[1868] Step 8:

[1869] The terminal receives the response sent from the server and displays it to the user. The user confirms the displayed information. This process is completed when the terminal receives the data from the server and displays it on the screen.

[1870] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1871] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1872] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1873] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1874] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1875] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1876] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1877] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1878] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1879] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1880] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1881] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1882] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1883] 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.

[1884] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1885] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1886] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1887] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1888] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1889] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1890] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1891] The following is further disclosed regarding the above embodiment.

[1892] (Claim 1)

[1893] means for receiving a user-entered question;

[1894] a means for generating an answer using a generative AI model based on the received question;

[1895] a means for assessing the accuracy of the generated answers;

[1896] means for displaying the assessed accuracy to a user;

[1897] A system including:

[1898] (Claim 2)

[1899] 2. The system of claim 1, wherein the means for evaluating the accuracy of the generated answers is a means for evaluating the accuracy using an internal database or a statistical model.

[1900] (Claim 3)

[1901] 10. The system of claim 1, further comprising means for integrating assessed accuracy information into the answer and displaying it to the user.

[1902]

[1903] "Example 1"

[1904] (Claim 1)

[1905] means for receiving a text input by a user through a terminal;

[1906] means for transmitting the received text to a server;

[1907] A means for the server to analyze the received text and prepare it in a format that can be passed to the generative AI model.

[1908] A means for generating answers using a generative AI model; and

[1909] a means for passing the generated answer to an accuracy evaluation module to evaluate the accuracy;

[1910] a means for integrating the assessed accuracy into an answer and transmitting the answer to the terminal;

[1911] means for displaying the assessed accuracy to a user;

[1912] A system including:

[1913] (Claim 2)

[1914] 2. The system of claim 1, wherein the means for evaluating the accuracy of the generated answers is a means for evaluating the accuracy using an internal database or a statistical model.

[1915] (Claim 3)

[1916] 10. The system of claim 1, further comprising means for integrating assessed accuracy information into the answer and displaying it to the user.

[1917] "Application Example 1"

[1918] (Claim 1)

[1919] means for receiving a user-entered question;

[1920] a means for generating an answer using a generative AI model based on the received question;

[1921] a means for assessing the accuracy of the generated answers;

[1922] means for displaying the assessed accuracy to a user;

[1923] a means for integrating the generated answers and their evaluation accuracy and displaying them to the user;

[1924] A system including:

[1925] (Claim 2)

[1926] 2. The system of claim 1, wherein the means for evaluating the accuracy of the generated answers is a means for evaluating the accuracy using an internal database or a statistical model.

[1927] (Claim 3)

[1928] 10. The system of claim 1, further comprising means for receiving information on the generated answers and their accuracy via API communication and displaying the information to the user.

[1929] "Example 2: Combining Emotion Engines"

[1930] (Claim 1)

[1931] means for receiving a user-entered question;

[1932] a means for generating an answer using a generative AI model based on the received question;

[1933] a means for assessing the accuracy of the generated answers;

[1934] means for displaying the assessed accuracy to a user;

[1935] A means for recognizing a user's emotions in real time using an emotion engine;

[1936] means for transmitting the recognized emotion data to a server;

[1937] a means for generating an answer taking into account the emotion data;

[1938] A system including:

[1939] (Claim 2)

[1940] 2. The system of claim 1, wherein the means for evaluating the accuracy of the generated answers is a means for evaluating the accuracy using an internal database or a statistical model.

[1941] (Claim 3)

[1942] 10. The system of claim 1, further comprising means for integrating assessed accuracy information into answers and displaying them to the user taking into account user sentiment data.

[1943] "Application example 2 when combining emotion engines"

[1944] (Claim 1)

[1945] means for receiving a user-entered question;

[1946] a means for generating an answer using a generative AI model based on the received question;

[1947] a means for assessing the accuracy of the generated answers;

[1948] means for displaying the assessed accuracy to a user;

[1949] a means for recognizing a user's emotions in real time;

[1950] means for processing the user's emotion data based on the generated answers to provide an emotion-sensitive response;

[1951] A system including:

[1952] (Claim 2)

[1953] 2. The system of claim 1, wherein the means for evaluating the accuracy of the generated answers is a means for evaluating the accuracy using an internal database or a statistical model.

[1954] (Claim 3)

[1955] 10. The system of claim 1, further comprising means for integrating assessed accuracy information into the answer and displaying it to the user. [Explanation of symbols]

[1956] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. means for receiving a user-entered question; a means for generating an answer using a generative AI model based on the received question; a means for assessing the accuracy of the generated answers; means for displaying the assessed accuracy to a user; A system including:

2. 2. The system according to claim 1, wherein the means for evaluating the accuracy of the generated answers is a means for evaluating the accuracy using an internal database or a statistical model.

3. The system of claim 1 , further comprising means for integrating the assessed accuracy information into the answer and displaying it to the user.

Images