system

Abstract

We provide the system. [Solution] A system for determining the accuracy of generated information, A means of receiving information entered by the user, Means for collecting additional information from relevant public data sources and trusted databases, A method for tokenizing the obtained data and extracting keywords and important information, A means of executing fact-checking algorithms to evaluate the accuracy of information, A means of visually displaying the results of fact-checking, A system that includes this.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In conventional information generation systems, the accuracy and reliability of the generated information are often not guaranteed, and there is a risk of misinformation spreading particularly in education and public institutions. As a result, there is a possibility that the recipient of the information may make an incorrect judgment, so there is a demand for a system that automatically determines the reliability of the generated information and provides it as accurate data.

Means for Solving the Problems

[0005] This invention provides a system for determining the accuracy of generated information. The system includes means for receiving information input from a user and collecting additional information from relevant public data sources and trusted databases. Furthermore, it includes means for tokenizing the obtained data, extracting keywords and important information, and executing fact-checking algorithms to evaluate the accuracy of the information. This allows for automatic evaluation of the reliability of the information and calculation of a reliability score. The system also provides means for visually displaying the fact-checking results, enabling the provision of accurate information by displaying the results in a user-friendly format.

[0006] A "user" is an individual or legal entity that uses the system and is the entity that inputs information and receives the results.

[0007] "Information" refers to text and data generated by generative AI, which users input into the system to determine its accuracy.

[0008] "Public data sources" refer to databases, APIs, and websites that are publicly available and whose reliability has been verified.

[0009] A "reliable database" is a database that aggregates reliable data provided by public institutions, academic institutions, or trustworthy companies.

[0010] "Tokenization" is a technique in natural language processing that divides text into units of words or phrases.

[0011] A "keyword" is a word or phrase that has particularly important meaning within text or data.

[0012] A "fact-checking algorithm" refers to a set of computational procedures or models designed to evaluate the accuracy of information.

[0013] A "reliability score" is a numerical value or evaluation criterion that indicates the reliability of information calculated by a fact-checking algorithm.

[0014] "Visual representation" refers to methods of presenting results in the form of graphs, charts, and other formats so that users can intuitively understand the accuracy of the information. [Brief explanation of the drawing]

[0015] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] This is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13]It is a sequence diagram showing the processing flow of the data processing system in Example 2 when the emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when the emotion engine is combined.

Embodiments for Carrying Out the Invention

[0016] Hereinafter, an example of an embodiment of the system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0017] First, the language used in the following description will be explained.

[0018] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0019] In the following embodiments, the numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor. <s

[0020] In the following embodiments, the numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0021] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0022] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0023] [First Embodiment]

[0024] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0025] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0026] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0027] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0028] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0029] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0030] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0031] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0032] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0033] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0034] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0035] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0036] This invention is a system for determining the accuracy of information generated by generative AI, and is primarily designed for educational and public institutions. The system consists of a series of processes: user input, data collection from public data sources and reliable databases, information analysis, fact-checking, and a visual display of the results.

[0037] Program Overview

[0038] The program for this system consists of the following modules:

[0039] 1. Data Acquisition Module

[0040] 2. Information Analysis Module

[0041] 3. Fact-checking module

[0042] 4. Result Display Module

[0043] Program Processing Description

[0044] Data Acquisition Module

[0045] Users input information into the system. Educators and government officials provide the system with lesson materials, draft press releases, and other materials.

[0046] The terminal retrieves information from the user input form and sends it to the server. The information is sent to the API endpoint in JSON format.

[0047] The server receives the input information and stores it in temporary storage. This ensures consistency for each piece of information.

[0048] The server makes API calls to collect additional information from relevant public data sources and trusted databases (e.g., Wikipedia or government data).

[0049] Information analysis module

[0050] The server tokenizes the collected data using a natural language processing library (e.g., NLTK, spaCy). This divides the text into words and phrases.

[0051] The server uses technologies such as TF-IDF and Word2Vec to extract keywords and important information. This clarifies the central points of the information.

[0052] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows the context of the information content to be understood.

[0053] Fact-checking module

[0054] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it might use models such as random forests, SVMs, or neural networks.

[0055] The server compares the acquired data with user input information and calculates a reliability score. The numerical value, based on the scoring function, indicates the reliability of the information.

[0056] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[0057] Result display module

[0058] The server compiles the fact-checking results and generates a comprehensive report. The report includes evaluation results, relevant sources, and recommended corrections.

[0059] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0060] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[0061] Specific example

[0062] Example 1: Use for educational institutions

[0063] The user (educator) inputs materials to be used in history lessons into the system.

[0064] The terminal sends the entered information to the server.

[0065] The server collects relevant information by making API calls to the relevant historical databases.

[0066] The server uses natural language processing techniques to tokenize information and extract keywords.

[0067] The server evaluates the accuracy of the information and calculates a reliability score based on fact-checking algorithms.

[0068] The server visualizes the results and displays them to the educator.

[0069] The device provides educators with accurate information sources and suggested revisions.

[0070] Example 2: Use by public institutions

[0071] The user (government official) enters a draft of a press release regarding the new policy.

[0072] The terminal sends information to the server.

[0073] The server collects additional information from trusted public databases.

[0074] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0075] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[0076] The server visualizes the results in graphs and charts and presents them to government officials.

[0077] The terminal displays detailed analysis results and suggested corrections to the staff.

[0078] The above describes the specific embodiments for implementing the system of the present invention.

[0079] The following describes the processing flow.

[0080] Step 1:

[0081] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[0082] Step 2:

[0083] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0084] Step 3:

[0085] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[0086] Step 4:

[0087] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[0088] Step 5:

[0089] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[0090] Step 6:

[0091] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[0092] Step 7:

[0093] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[0094] Step 8:

[0095] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[0096] Step 9:

[0097] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and suggested corrections.

[0098] Step 10:

[0099] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0100] Step 11:

[0101] The terminal displays the generated report in the user interface. Users can view detailed analysis results and recommended corrective actions.

[0102] (Example 1)

[0103] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0104] Determining the accuracy of information generated by generative AI is a critical issue, particularly in educational and public institutions. However, current systems lack the means to automatically and efficiently evaluate the accuracy of user-inputted information. As a result, reliable fact-checking is not performed, and there is a risk of misinformation spreading. Furthermore, the analysis and visual display of collected data are insufficient, making it difficult for users to intuitively understand the information. An effective system is needed to address these challenges.

[0105] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0106] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for performing dependency analysis and syntactic analysis to analyze the context of the information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, and means for visually displaying the results of the fact-checking. This makes it possible to evaluate the accuracy of user-inputted information with high precision and provide reliable data. Furthermore, the visual display of the results realizes a system that is easy for users to understand intuitively.

[0107] "Information" refers to data that users input into the system, as well as data collected from relevant public data sources and trusted databases.

[0108] A "user" refers to an employee of an educational institution or public organization, or an individual who uses the system and is the entity that inputs information into the system.

[0109] "Public data sources" refer to information sources that are publicly available on the internet or in other accessible locations, and include, for example, encyclopedias and government statistics.

[0110] A "reliable database" refers to a source of information managed by a highly credible institution or organization. This includes academic journal databases and official statistical databases.

[0111] "Tokenization" refers to the process of dividing words and phrases in a text into individual elements, and it is a fundamental process in natural language processing.

[0112] A "keyword" refers to a word or phrase that is considered particularly important within information, and is used to grasp the subject or content of the information.

[0113] "Dependency analysis" is a technique for understanding sentence structure and meaning by analyzing the dependencies between words in a text.

[0114] "Syntactic analysis" is a method of analyzing the grammatical structure of a text and evaluating its meaning and grammatical accuracy.

[0115] "Fact-checking" refers to the process of evaluating the accuracy and truthfulness of information, and is the work of verifying the accuracy of information based on collected data.

[0116] A "fact-checking algorithm" refers to an algorithm used to computationally evaluate the accuracy of information, often employing machine learning models or statistical methods.

[0117] A "reliability score" is a numerical representation of the accuracy and reliability of information, calculated using an algorithm.

[0118] "Visual display" refers to presenting information and data to users in the form of graphs and charts using visualization tools, making it easier for users to intuitively understand the data.

[0119] This invention is a system composed of multiple modules for determining the accuracy of information generated by a generative AI. The specific method for realizing this system is described below.

[0120] System Overview

[0121] This system consists of the following modules:

[0122] 1. User Input Module

[0123] 2. Data Acquisition Module

[0124] 3. Information Analysis Module

[0125] 4. Fact-checking module

[0126] 5. Result Display Module

[0127] User input module

[0128] Users enter information through the system's input forms. For example, this could involve educators entering lesson materials or government officials entering draft policy documents.

[0129] The terminal retrieves information from the input form and sends it to the server in JSON format.

[0130] Data Acquisition Module

[0131] The server stores the information received from the user in temporary storage.

[0132] The server collects additional information from relevant public data sources and trusted databases. These public data sources include category information sites and public databases. For example, it uses the Wikipedia API and government public data APIs.

[0133] Information analysis module

[0134] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). Specifically, it divides sentences into words and phrases and analyzes their structure.

[0135] The server uses TF-IDF and Word2Vec technologies to extract keywords and important information, thereby clarifying the central points of the information.

[0136] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows it to understand the contextual meaning of the text.

[0137] Fact-checking module

[0138] The server evaluates the accuracy of the information using fact-checking algorithms (such as random forests, SVMs, and neural networks).

[0139] The server compares the collected data with user input information and calculates a reliability score. The reliability score is a numerical indicator of the accuracy of the information; for example, a score of 0.9 or higher is considered reliable.

[0140] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[0141] Result display module

[0142] The server compiles the fact-checking results into a comprehensive report. The report includes evaluation results, relevant sources, and recommended revisions.

[0143] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). This allows users to intuitively understand the results.

[0144] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[0145] Specific example

[0146] Example 1: Use for educational institutions

[0147] The user (educator) inputs the material "Outline of Greek Mythology," which is used in history lessons, into the system.

[0148] The terminal sends the input information to the server.

[0149] The server collects and analyzes information by making API calls to relevant historical databases. For example, it might use "the content of Wikipedia articles on Greek mythology."

[0150] The server uses fact-checking algorithms to evaluate the accuracy of the information and calculates a reliability score.

[0151] The server visualizes the results and displays them to the educator.

[0152] The device provides educators with accurate information sources and suggested revisions.

[0153] Example 2: Use by public institutions

[0154] The user (government official) enters a draft of a press release regarding a new policy.

[0155] The terminal sends information to the server.

[0156] The server collects additional information from trusted public databases and performs analysis.

[0157] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[0158] The server visualizes the results in graphs and charts and presents them to government officials.

[0159] The terminal displays detailed analysis results and suggested corrections to the staff.

[0160] The above describes the embodiments for implementing the system of the present invention.

[0161] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0162] Step 1: Enter Information

[0163] The user enters information into the system's input form. For example, "An overview of Greek mythology."

[0164] The terminal retrieves the entered information and sends it to the server in JSON format. The input includes the title "Outline of Greek Mythology" and the body text.

[0165] The server stores the received information in temporary storage. This ensures consistency in the input information and facilitates subsequent processing.

[0166] Step 2: Data Collection

[0167] The server collects data from relevant public data sources and trusted databases based on information received from the user. For example, it retrieves data related to "Greek mythology" from the Wikipedia API and public databases.

[0168] The server temporarily stores the collected data in JSON format. Specifically, this includes information such as "details of Greek mythology" and "major characters."

[0169] The server synchronizes with publicly available data sources at regular intervals to reflect the latest information.

[0170] Step 3: Information Analysis

[0171] The server tokenizes the collected data using natural language processing libraries (NLTK, spaCy). Specifically, it divides sentences into words and phrases. The input "Zeus, the chief god of Greek mythology" is tokenized as ["Greek mythology", "chief god", "Zeus"].

[0172] The server uses TF-IDF and Word2Vec to extract keywords and important information. For example, it might extract keywords such as "Zeus" and "Olympian gods."

[0173] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This helps it understand the context of "Zeus is the king of the Olympian gods."

[0174] Step 4: Fact Check

[0175] The server uses fact-checking algorithms (such as random forests, SVMs, and neural networks) to evaluate the accuracy of the information. Specifically, it identifies discrepancies between the input information and the collected data.

[0176] The server calculates a reliability score based on the comparison results. The reliability score is displayed as a numerical value; for example, a score of 0.9 or higher is considered "high reliability."

[0177] The server determines whether the score meets the set criteria and assesses its accuracy. For example, it might determine that "Zeus is a god in Greek mythology" is correct, while "Zeus is a god in Roman mythology" is incorrect.

[0178] Step 5: Display Results

[0179] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and recommended corrections.

[0180] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). Confidence scores are visually displayed using bar graphs, etc.

[0181] The terminal displays the generated report in the user interface, including specific data points and links to suggested corrections.

[0182] The above outlines the specific processing steps of this system's program.

[0183] (Application Example 1)

[0184] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0185] Ensuring the reliability of information such as product details and user reviews is crucial for e-commerce websites. However, identifying misinformation and false reviews from a vast amount of data and providing only reliable information is not easy. Current systems lack the means to determine the accuracy of user-provided information in real time, which makes them prone to problems caused by misinformation.

[0186] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0187] In this invention, the server includes means for receiving information entered by a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for visually displaying the results of the fact-checking, means for providing a user interface for inputting product information and user reviews, means for collecting product information from an official database, means for analyzing the content of the information using natural language processing technology, and means for calculating a reliability score and visually indicating reliability. This makes it possible to determine the accuracy of information on an e-commerce site in real time and provide users with only highly reliable information.

[0188] A "user" is an individual or organization that uses the system to input product information and user reviews, and to verify the accuracy of that information.

[0189] "Public data sources" refer to information sources that can be obtained from publicly available databases and websites.

[0190] A "reliable database" is a database containing highly reliable information provided by a public institution or certified organization.

[0191] "Tokenization of information" is the process of dividing text into words or phrases using natural language processing techniques.

[0192] "Keyword extraction" is the process of selecting important words and phrases from a text.

[0193] A "fact-checking algorithm" is a computational method or model used to compare input information with acquired information and evaluate its accuracy and reliability.

[0194] A "user interface" is an interactive screen or system through which a user inputs information and receives results.

[0195] An "official database" is a reliable database used to provide official information about a product.

[0196] "Natural language processing technology" refers to the technology that enables computers to understand, analyze, and process human language.

[0197] A "reliability score" is an indicator that numerically represents the accuracy of information, and it concretely expresses the evaluation results.

[0198] This invention is a system for determining the accuracy of generated information. This system is primarily used to determine the reliability of product information and user reviews on e-commerce websites.

[0199] The server first uses a means to receive information entered by the user. Users can enter product information and user reviews through the user interface of their smartphone. The server receives this input information and temporarily stores it.

[0200] Next, the server employs means to collect additional information from relevant public data sources and trusted databases. Collecting product information from official databases ensures reliable data.

[0201] The server then tokenizes the obtained data and employs methods to extract keywords and important information. Natural language processing techniques (e.g., Spacy) are used to divide the text into words and phrases, and extract the important parts.

[0202] The server then runs fact-checking algorithms to assess the accuracy of the information. These algorithms include a function to calculate a reliability score, utilizing techniques such as random forests, SVMs, and neural networks. The information provided by the user is compared with official data, and a numerical value is calculated based on the scoring function.

[0203] Finally, the server employs a means to visually display the fact-checking results. Using visualization tools (e.g., Matplotlib), it displays reliability scores as charts and graphs, providing users with an intuitively understandable format.

[0204] For example, if a user enters a product description such as "This is a test product description for a smartwatch," the server will collect information about smartwatches from the official database and extract key keywords using natural language processing technology. Next, a fact-checking algorithm will calculate a reliability score, and finally, the results will be displayed as a graph.

[0205] This makes it possible to verify the accuracy of information on e-commerce sites in real time and provide users with reliable information.

[0206] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0207] Step 1:

[0208] Users input product information and user reviews through their smartphone's user interface. The entered information is sent to the server in JSON format. For example, the input data might be the text "This is a test product description for a smartwatch." The server receives this input information and stores it temporarily.

[0209] Step 2:

[0210] The server collects additional information from relevant public data sources and trusted databases. This collection process uses REST API calls to access databases and retrieve official product information (e.g., product specifications and official reviews). The retrieved data is returned to the server in JSON format and temporarily stored. This ensures reliable data for comparison with user input.

[0211] Step 3:

[0212] The server tokenizes the stored data using natural language processing techniques. Specifically, it uses the Spacy library to split text into words and phrases. This process converts the input text into a list of words, including attribute information for each word (e.g., part of speech).

[0213] Step 4:

[0214] The server extracts keywords and important information from the tokenized data. Algorithms such as TF-IDF and Word2Vec are used to identify particularly important words and phrases within the text. This process yields a list of extracted keywords. For example, "smartwatch" and "product description" might be extracted as keywords.

[0215] Step 5:

[0216] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it compares user input information with official data using random forests or neural network models. This algorithm evaluates the reliability of each keyword in the input information and calculates an overall reliability score. This score is output as a numerical value ranging from 0 to 1.

[0217] Step 6:

[0218] The server visually displays the fact-checking results based on the calculated reliability score. Visualization tools such as Matplotlib are used to convert the reliability score into charts and graphs. This display is then sent back to the smartphone user interface, presenting it in an intuitively understandable format. For example, a pie chart might show a reliability of 70%.

[0219] Step 7:

[0220] The user reviews the displayed fact-check results and corrects the input information as needed. The corrected information is sent back to the server for re-evaluation. This loop ensures that accurate and reliable information is ultimately provided to the user.

[0221] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0222] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[0223] Program Overview

[0224] The program for this system consists of the following modules:

[0225] 1. Data Acquisition Module

[0226] 2. Information Analysis Module

[0227] 3. Fact-checking module

[0228] 4. Emotion Recognition Module

[0229] 5. Result Display Module

[0230] Program Processing Description

[0231] Data Acquisition Module

[0232] The user logs into the system and enters the information they want to analyze. They enter text into the input form and click the submit button.

[0233] The terminal packages the received information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0234] The server temporarily stores the received information in storage and makes API calls to retrieve additional information from relevant public data sources or trusted databases.

[0235] Information analysis module

[0236] The server tokenizes the collected data using a natural language processing library. This divides the text into words and phrases, facilitating data analysis.

[0237] The server extracts keywords and important information using TF-IDF and Word2Vec. This clarifies the central points of the information.

[0238] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. The context of the text is understood in detail.

[0239] Fact-checking module

[0240] The server runs fact-checking algorithms to assess the accuracy of the information. It calculates a reliability score and compares the retrieved data with the input information.

[0241] The server evaluates the accuracy of the information based on a reliability score.

[0242] Emotion recognition module

[0243] Users express emotions by providing voice input or facial expressions to the system. The system measures how users in educational and public institutions react emotionally while evaluating information.

[0244] The device acquires audio or video data and sends it to the server.

[0245] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[0246] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is feeling anxious or suspicious, the server adjusts how the results are displayed and the content of the feedback accordingly.

[0247] Result display module

[0248] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[0249] The server converts the results into graphs and charts using visualization tools, allowing users to intuitively understand the results.

[0250] The device displays the generated report in the user interface, allowing the user to view detailed analysis results and sentiment-based feedback.

[0251] Specific example

[0252] Example 1: Use for educational institutions

[0253] The user (educator) inputs materials to be used in history lessons into the system.

[0254] The terminal sends the entered information to the server.

[0255] The server collects relevant information by making API calls to the relevant historical databases.

[0256] The server uses natural language processing techniques to tokenize information and extract keywords.

[0257] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[0258] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[0259] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0260] The device displays the generated report and recommended feedback to the educator.

[0261] Example 2: Use by public institutions

[0262] The user (government official) enters a draft of a press release regarding the new policy.

[0263] The terminal sends information to the server.

[0264] The server collects additional information from trusted public databases.

[0265] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0266] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[0267] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[0268] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0269] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[0270] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[0271] The following describes the processing flow.

[0272] Step 1:

[0273] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[0274] Step 2:

[0275] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0276] Step 3:

[0277] The server receives the information received and stores it in temporary storage. This temporary storage is managed by a database, and the consistency of the information is maintained in subsequent processing.

[0278] Step 4:

[0279] The server calls an external API to obtain additional information from related public data sources or reliable databases. For example, it accesses Wikipedia or government public databases to obtain relevant data.

[0280] Step 5:

[0281] The server tokenizes the collected data using a natural language processing library (e.g., NLTK, spaCy). This is to split the text into words or phrases to improve the ease of analyzing the data.

[0282] Step 6:

[0283] The server uses techniques such as TF-IDF or Word2Vec to identify keywords and important information from the extracted tokens. This technique is used to identify important words and phrases in the information.

[0284] Step 7:

[0285] The server performs dependency parsing and syntactic analysis to analyze the relationship between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[0286] Step 8:

[0287] The server runs a fact-checking algorithm to evaluate the accuracy of the information. The algorithm calculates a reliability score for each data point and evaluates the accuracy of the information based on this.

[0288] Step 9:

[0289] The server receives voice input or facial recognition data to recognize the user's emotions. The user can communicate their emotions to the system through voice or facial expressions.

[0290] Step 10:

[0291] The device sends the collected audio or video data to the server, which then provides data for sentiment analysis.

[0292] Step 11:

[0293] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[0294] Step 12:

[0295] The server adjusts the output of the fact-checking algorithm based on the analyzed emotional state. For example, if the user is feeling anxious or suspicious, it flexibly changes how the results are displayed and the content of the feedback.

[0296] Step 13:

[0297] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[0298] Step 14:

[0299] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0300] Step 15:

[0301] The device displays the generated report in the user interface. Users can view detailed analysis results and sentiment-based feedback.

[0302] (Example 2)

[0303] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0304] Traditionally, systems that judged the accuracy of information generated by generative AI lacked the ability to consider user emotions. As a result, evaluation results sometimes did not align with the user's emotional state, leading to problems in understanding and accepting the results. To address this issue, a system is needed that provides comprehensive feedback that considers user emotions in addition to accurate information judgment.

[0305] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotions, means for adjusting the output results of the fact-checking algorithm based on emotion recognition, and means for compiling the fact-checking results and emotion recognition results into a comprehensive report and displaying it visually. This makes it possible to evaluate the accuracy of the information and provide feedback that takes the user's emotions into consideration.

[0306] A "user" refers to the entity that uses this system to input information and receive the results.

[0307] "Open data source" refers to a reliable information source publicly available on the Internet. For example, encyclopedias and academic papers are applicable.

[0308] "Reliable database" refers to a collection of information with guaranteed accuracy and reliability. For example, government statistical databases and databases by official research institutions are applicable.

[0309] "Tokenization" is one of the preprocessing steps in natural language processing and refers to the process of splitting a text into words or phrases.

[0310] "Keyword" refers to a word or phrase that is particularly important among the input information or collected data.

[0311] "Fact-checking algorithm" refers to an algorithm for evaluating the accuracy and reliability of information.

[0312] "Emotion recognition engine" refers to a technology that analyzes voice or video data to identify the user's emotional state (e.g., joy, sadness, anger, etc.).

[0313] "Visually display" refers to the process of converting text information into a visual form such as graphs, charts, or diagrams for display.

[0314] "Reliability score" refers to an evaluation indicator that numerically represents the reliability of information.

[0315] The present invention combines an emotion engine that recognizes the user's emotions with a system for determining the accuracy of information generated by generative AI. This system aims to evaluate the reliability of the information input by the user and provide emotion - considerate feedback for educational institutions and public institutions.

[0316] Program summary

[0317] This system's program consists of the following modules:

[0318] 1. Data Acquisition Module

[0319] 2. Information Analysis Module

[0320] 3. Fact-checking module

[0321] 4. Emotion Recognition Module

[0322] 5. Result Display Module

[0323] Hardware and software to use

[0324] The following hardware and software are required to implement this system:

[0325] Server: A server that performs data processing and analysis.

[0326] Device: A device used by a user to input information (e.g., a PC or tablet).

[0327] Database: Storage for storing received data and additional information.

[0328] Natural language processing libraries: For example, NLTK and spaCy

[0329] Fact-checking tools: For example, FactCheck Tools API

[0330] Emotion recognition engines: For example, Google® Cloud Speech-to-Text and Microsoft® Azure® Face API.

[0331] Visualization tools: For example, D3.js and Chart.js

[0332] Specific example

[0333] Specific example 1: Use in educational institutions

[0334] The user (educator) inputs materials to be used in history lessons into the system. For example, they might input information such as, "The American Civil War ended in 1865."

[0335] The terminal sends the entered information to the server.

[0336] The server collects relevant information by making API calls to the relevant historical databases.

[0337] The server uses natural language processing techniques to tokenize information and extract keywords.

[0338] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[0339] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[0340] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0341] The device displays the generated report and recommended feedback to the educator.

[0342] Specific example 2: Use by public institutions

[0343] The user (government official) enters a draft of a press release regarding a new policy. For example, they might enter information such as, "A new environmental policy will reduce carbon dioxide emissions by 30% by 2030."

[0344] The terminal sends information to the server.

[0345] The server collects additional information from trusted public databases.

[0346] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0347] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[0348] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[0349] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0350] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[0351] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[0352] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0353] Step 1:

[0354] The user logs into the system. On the login screen, they enter their username and password and submit their authentication information. Input: Username, Password. Output: Authentication token.

[0355] Step 2:

[0356] The user enters the information to be evaluated using a text input form. For example, they might enter the contents of materials used in a history class into the text field. Input: Text data of the information to be evaluated. Output: Click event of the submit button.

[0357] Step 3:

[0358] The terminal retrieves the information entered by the user when they click the submit button. It then packages the retrieved information into JSON format and sends it to the server. Input: Text data. Output: Sending JSON data.

[0359] Step 4:

[0360] The server temporarily stores the received JSON data in storage. This storage uses a database or object storage. Input: JSON data. Output: Temporarily stored data.

[0361] Step 5:

[0362] The server makes API calls to public data sources and trusted databases to collect relevant information from the database. For example, it uses Wikidata and other reliable sources. Input: Temporarily stored data. Output: Additional data collected.

[0363] Step 6:

[0364] The server tokenizes the collected data using a natural language processing library (e.g., NLTK or spaCy). This divides the text into words and phrases. Input: Collected data. Output: Tokenized data.

[0365] Step 7:

[0366] The server extracts keywords from tokenized data using TF-IDF or Word2Vec. This identifies important information. Input: Tokenized data. Output: Keywords.

[0367] Step 8:

[0368] The server performs dependency analysis and syntactic analysis to analyze the relationships between the obtained keywords. For example, it uses the Stanford NLP dependency analysis model. Input: Keywords. Output: Relationships between keywords.

[0369] Step 9:

[0370] The server executes a fact-checking algorithm, comparing the acquired data with the input information to calculate a reliability score. Input: Relationships between keywords. Output: Reliability score.

[0371] Step 10:

[0372] The user provides audio or video input to the system, for example, using a webcam or microphone. Input: audio data, video data. Output: transmission of emotion recognition data.

[0373] Step 11:

[0374] The terminal acquires audio or video data and sends it to the server. Input: Audio data, video data. Output: Data transmission to the server.

[0375] Step 12:

[0376] The server analyzes the user's emotions using an emotion recognition engine (e.g., Google Cloud Speech-to-Text or Microsoft Azure Face API). Input: Audio data, video data. Output: Emotion analysis results.

[0377] Step 13:

[0378] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is expressing anxiety, it selects a way to present the results in a clearer manner. Input: Sentiment analysis results, reliability score. Output: Adjusted feedback content.

[0379] Step 14:

[0380] The server compiles the fact-checking and sentiment assessment results into a comprehensive report and converts them into graphs and charts using visualization tools (e.g., D3.js or Chart.js). Input: Adjusted feedback. Output: Visualized report.

[0381] Step 15:

[0382] The terminal displays the generated report in the user interface. On this screen, the user can view detailed analysis results and sentiment-based feedback. Input: Visualized report. Output: Displayed in the user interface.

[0383] The above outlines the specific processing steps of the system program.

[0384] (Application Example 2)

[0385] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0386] When determining the accuracy of information generated by generative AI, conventional systems have a problem in that they cannot provide feedback that takes user emotions into account. The feedback users receive is uniform, and for example, if a user feels anxious or doubtful, appropriate support is not provided. This raises concerns that it increases the psychological burden on users and hinders their acceptance and understanding of the information.

[0387] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0388] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotional state, means for visually displaying the fact-checking results, and means for adjusting feedback based on the user's emotional state. This enables the evaluation of the reliability of the information input by the user and the provision of appropriate feedback based on the user's emotional state. In this way, the psychological burden on the user can be reduced and the acceptance and understanding of information can be promoted.

[0389] A "user" is someone who inputs information into a system and receives the results.

[0390] A "public data source" refers to a collection of data that is widely made public to provide reliable information.

[0391] A "reliable database" is a database that stores information whose reliability has been verified.

[0392] "Tokenization" is the process of dividing an input text into its smallest units, such as words or phrases.

[0393] A "keyword" refers to a word or phrase that has particularly important meaning within a text or data.

[0394] A "fact-checking algorithm" refers to a set of computational procedures designed to evaluate the accuracy of input information.

[0395] An "emotion recognition engine" is a technology that analyzes a user's voice and facial expression data to identify their emotional state.

[0396] A "reliability score" is an indicator that quantifies the accuracy of information.

[0397] "Visually displaying" refers to providing users with data and information in a visual format, such as graphs and charts.

[0398] "Adjusting feedback" means changing the content and format of the feedback output based on the user's emotional state.

[0399] To implement this invention, a system is required in which a server, terminal, and user work together. This system includes modules for data collection, information analysis, fact-checking, sentiment recognition, and result display.

[0400] The server receives information entered by the user. For example, if a user enters a URL for a news article, the server packages that information in JSON format and saves it to storage. It also makes API calls to collect additional information from relevant public data sources and trusted databases. This allows for centralized management of necessary data.

[0401] Next, the server tokenizes the collected data using a natural language processing library. Specifically, it uses the Python nltk library to make it easier to divide sentences into words and phrases. Furthermore, it uses the scikit-learn library to extract keywords and important information using methods such as TF-IDF. This clarifies the central points of the information and improves the accuracy of fact-checking.

[0402] The fact-checking algorithm is executed on the server side. The server performs comparisons to calculate a reliability score based on the collected and analyzed data. It cross-references the acquired data with the input information to evaluate the accuracy of the information.

[0403] Furthermore, an emotion recognition engine is used to analyze the user's emotional state. The user provides voice input and facial expressions, and the device sends this data to the server. The server analyzes the emotional state using voice analysis and facial recognition technology, and adjusts the fact-check output based on the results. Here, the emotion_recognition library is used as the emotion recognition engine.

[0404] Finally, the server generates a report integrating fact-checking results and sentiment recognition results, and uses visualization tools to convert the results into graphs and charts. This allows users to intuitively understand the results and reduces the psychological burden of receiving feedback. Users can view detailed analysis results and sentiment-based feedback through their devices.

[0405] As a concrete example, let's explain how the news application "Emofact News" works. The user enters the URL of a news article and provides audio and facial expression data. On the server side, the input information is analyzed and a reliability score and emotional state are evaluated. Finally, the results are displayed visually, allowing the user to intuitively understand the reliability of the information and receive emotionally sensitive feedback.

[0406] An example of the generated prompt statement is as follows:

[0407] Please evaluate the reliability of the following article and the user's sentiment-based feedback. Visit "https: / / example.com / news / some-article", check the accuracy of the information, and display the results while taking into consideration any concerns or doubts the user may have.

[0408] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0409] Step 1:

[0410] User enters information

[0411] The user accesses the system interface and enters information such as the URL of a news article. The entered information is packaged in JSON format on the terminal and sent to the server.

[0412] Input: URL of the news article

[0413] Output: Information in JSON format

[0414] ---

[0415] Step 2:

[0416] Data collection

[0417] The server temporarily stores the JSON-formatted information received from the terminal in storage. Furthermore, it executes API calls to collect additional information from relevant public data sources and trusted databases.

[0418] Input: Information in JSON format

[0419] Output: Stored information and additionally collected data

[0420] ---

[0421] Step 3:

[0422] Information analysis

[0423] The server tokenizes the collected data using a natural language processing library (e.g., the nltk library). From the tokenized data, keywords and important information are extracted using TF-IDF.

[0424] Input: Collected data

[0425] Output: Tokenized data and extracted keywords

[0426] ---

[0427] Step 4:

[0428] Fact Check

[0429] The server runs fact-checking algorithms and calculates a reliability score by cross-referencing extracted keywords with data collected from publicly available data sources and trusted databases.

[0430] Input: Extracted keywords and collected additional information

[0431] Output: Reliability score

[0432] ---

[0433] Step 5:

[0434] emotion recognition

[0435] When entering information, the user provides voice input or facial expression data. The terminal sends this data to the server, which uses an emotion recognition engine to analyze the user's emotional state. For example, the emotion_recognition library can be used.

[0436] Input: Voice data or facial expression data

[0437] Output: Analyzed emotional state

[0438] ---

[0439] Step 6:

[0440] Adjustment and integration of results

[0441] The server integrates the reliability score and emotional state, and adjusts the feedback content according to the user's emotions. This generates appropriate feedback that reduces the user's psychological burden.

[0442] Input: Reliability score and analyzed emotional state

[0443] Output: Adjusted feedback content

[0444] ---

[0445] Step 7:

[0446] Visualization and display of results

[0447] The server converts the integrated information into graphs and charts using visualization tools. This allows the terminal to display the results in a format that is intuitively easy for the user to understand.

[0448] Input: Integrated result

[0449] Output: Visualized results and feedback

[0450] ---

[0451] Step 8:

[0452] Provide feedback

[0453] The device displays visualized results and refined feedback sent from the server to the user. The user reviews the results and receives feedback that is reliable and considerate of their feelings.

[0454] Input: Visualized results and feedback content

[0455] Output: Display to the user

[0456] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0457] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0458] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0459] [Second Embodiment]

[0460] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0461] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0462] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0463] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0464] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0465] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0466] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0467] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0468] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0469] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0470] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0471] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0472] This invention is a system for determining the accuracy of information generated by generative AI, and is primarily designed for educational and public institutions. The system consists of a series of processes: user input, data collection from public data sources and reliable databases, information analysis, fact-checking, and a visual display of the results.

[0473] Program Overview

[0474] The program for this system consists of the following modules:

[0475] 1. Data Acquisition Module

[0476] 2. Information Analysis Module

[0477] 3. Fact-checking module

[0478] 4. Result Display Module

[0479] Program Processing Description

[0480] Data Acquisition Module

[0481] Users input information into the system. Educators and government officials provide the system with lesson materials, draft press releases, and other materials.

[0482] The terminal retrieves information from the user input form and sends it to the server. The information is sent to the API endpoint in JSON format.

[0483] The server receives the input information and stores it in temporary storage. This ensures consistency for each piece of information.

[0484] The server makes API calls to collect additional information from relevant public data sources and trusted databases (e.g., Wikipedia or government data).

[0485] Information analysis module

[0486] The server tokenizes the collected data using a natural language processing library (e.g., NLTK, spaCy). This divides the text into words and phrases.

[0487] The server uses technologies such as TF-IDF and Word2Vec to extract keywords and important information. This clarifies the central points of the information.

[0488] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows the context of the information content to be understood.

[0489] Fact-checking module

[0490] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it might use models such as random forests, SVMs, or neural networks.

[0491] The server compares the acquired data with user input information and calculates a reliability score. The numerical value, based on the scoring function, indicates the reliability of the information.

[0492] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[0493] Result display module

[0494] The server compiles the fact-checking results and generates a comprehensive report. The report includes evaluation results, relevant sources, and recommended corrections.

[0495] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0496] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[0497] Specific example

[0498] Example 1: Use for educational institutions

[0499] The user (educator) inputs materials to be used in history lessons into the system.

[0500] The terminal sends the entered information to the server.

[0501] The server collects relevant information by making API calls to the relevant historical databases.

[0502] The server uses natural language processing techniques to tokenize information and extract keywords.

[0503] The server evaluates the accuracy of the information and calculates a reliability score based on fact-checking algorithms.

[0504] The server visualizes the results and displays them to the educator.

[0505] The device provides educators with accurate information sources and suggested revisions.

[0506] Example 2: Use by public institutions

[0507] The user (government official) enters a draft of a press release regarding the new policy.

[0508] The terminal sends information to the server.

[0509] The server collects additional information from trusted public databases.

[0510] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0511] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[0512] The server visualizes the results in graphs and charts and presents them to government officials.

[0513] The terminal displays detailed analysis results and suggested corrections to the staff.

[0514] The above describes the specific embodiments for implementing the system of the present invention.

[0515] The following describes the processing flow.

[0516] Step 1:

[0517] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[0518] Step 2:

[0519] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0520] Step 3:

[0521] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[0522] Step 4:

[0523] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[0524] Step 5:

[0525] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[0526] Step 6:

[0527] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[0528] Step 7:

[0529] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[0530] Step 8:

[0531] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[0532] Step 9:

[0533] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and suggested corrections.

[0534] Step 10:

[0535] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0536] Step 11:

[0537] The terminal displays the generated report in the user interface. Users can view detailed analysis results and recommended corrective actions.

[0538] (Example 1)

[0539] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0540] Determining the accuracy of information generated by generative AI is a critical issue, particularly in educational and public institutions. However, current systems lack the means to automatically and efficiently evaluate the accuracy of user-inputted information. As a result, reliable fact-checking is not performed, and there is a risk of misinformation spreading. Furthermore, the analysis and visual display of collected data are insufficient, making it difficult for users to intuitively understand the information. An effective system is needed to address these challenges.

[0541] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0542] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for performing dependency analysis and syntactic analysis to analyze the context of the information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, and means for visually displaying the results of the fact-checking. This makes it possible to evaluate the accuracy of user-inputted information with high precision and provide reliable data. Furthermore, the visual display of the results realizes a system that is easy for users to understand intuitively.

[0543] "Information" refers to data that users input into the system, as well as data collected from relevant public data sources and trusted databases.

[0544] A "user" refers to an employee of an educational institution or public organization, or an individual who uses the system and is the entity that inputs information into the system.

[0545] "Public data sources" refer to information sources that are publicly available on the internet or in other accessible locations, and include, for example, encyclopedias and government statistics.

[0546] A "reliable database" refers to a source of information managed by a highly credible institution or organization. This includes academic journal databases and official statistical databases.

[0547] "Tokenization" refers to the process of dividing words and phrases in a text into individual elements, and it is a fundamental process in natural language processing.

[0548] A "keyword" refers to a word or phrase that is considered particularly important within information, and is used to grasp the subject or content of the information.

[0549] "Dependency analysis" is a technique for understanding sentence structure and meaning by analyzing the dependencies between words in a text.

[0550] "Syntactic analysis" is a method of analyzing the grammatical structure of a text and evaluating its meaning and grammatical accuracy.

[0551] "Fact-checking" refers to the process of evaluating the accuracy and truthfulness of information, and is the work of verifying the accuracy of information based on collected data.

[0552] A "fact-checking algorithm" refers to an algorithm used to computationally evaluate the accuracy of information, often employing machine learning models or statistical methods.

[0553] A "reliability score" is a numerical representation of the accuracy and reliability of information, calculated using an algorithm.

[0554] "Visual display" refers to presenting information and data to users in the form of graphs and charts using visualization tools, making it easier for users to intuitively understand the data.

[0555] This invention is a system composed of multiple modules for determining the accuracy of information generated by a generative AI. The specific method for realizing this system is described below.

[0556] System Overview

[0557] This system consists of the following modules:

[0558] 1. User Input Module

[0559] 2. Data Acquisition Module

[0560] 3. Information Analysis Module

[0561] 4. Fact-checking module

[0562] 5. Result Display Module

[0563] User input module

[0564] Users enter information through the system's input forms. For example, this could involve educators entering lesson materials or government officials entering draft policy documents.

[0565] The terminal retrieves information from the input form and sends it to the server in JSON format.

[0566] Data Acquisition Module

[0567] The server stores the information received from the user in temporary storage.

[0568] The server collects additional information from relevant public data sources and trusted databases. These public data sources include category information sites and public databases. For example, it uses the Wikipedia API and government public data APIs.

[0569] Information analysis module

[0570] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). Specifically, it divides sentences into words and phrases and analyzes their structure.

[0571] The server uses TF-IDF and Word2Vec technologies to extract keywords and important information, thereby clarifying the central points of the information.

[0572] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows it to understand the contextual meaning of the text.

[0573] Fact-checking module

[0574] The server evaluates the accuracy of the information using fact-checking algorithms (such as random forests, SVMs, and neural networks).

[0575] The server compares the collected data with user input information and calculates a reliability score. The reliability score is a numerical indicator of the accuracy of the information; for example, a score of 0.9 or higher is considered reliable.

[0576] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[0577] Result display module

[0578] The server compiles the fact-checking results into a comprehensive report. The report includes evaluation results, relevant sources, and recommended revisions.

[0579] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). This allows users to intuitively understand the results.

[0580] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[0581] Specific example

[0582] Example 1: Use for educational institutions

[0583] The user (educator) inputs the material "Outline of Greek Mythology," which is used in history lessons, into the system.

[0584] The terminal sends the input information to the server.

[0585] The server collects and analyzes information by making API calls to relevant historical databases. For example, it might use "the content of Wikipedia articles on Greek mythology."

[0586] The server uses fact-checking algorithms to evaluate the accuracy of the information and calculates a reliability score.

[0587] The server visualizes the results and displays them to the educator.

[0588] The device provides educators with accurate information sources and suggested revisions.

[0589] Example 2: Use by public institutions

[0590] The user (government official) enters a draft of a press release regarding a new policy.

[0591] The terminal sends information to the server.

[0592] The server collects additional information from trusted public databases and performs analysis.

[0593] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[0594] The server visualizes the results in graphs and charts and presents them to government officials.

[0595] The terminal displays detailed analysis results and suggested corrections to the staff.

[0596] The above describes the embodiments for implementing the system of the present invention.

[0597] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0598] Step 1: Enter Information

[0599] The user enters information into the system's input form. For example, "An overview of Greek mythology."

[0600] The terminal retrieves the entered information and sends it to the server in JSON format. The input includes the title "Outline of Greek Mythology" and the body text.

[0601] The server stores the received information in temporary storage. This ensures consistency in the input information and facilitates subsequent processing.

[0602] Step 2: Data Collection

[0603] The server collects data from relevant public data sources and trusted databases based on information received from the user. For example, it retrieves data related to "Greek mythology" from the Wikipedia API and public databases.

[0604] The server temporarily stores the collected data in JSON format. Specifically, this includes information such as "details of Greek mythology" and "major characters."

[0605] The server synchronizes with publicly available data sources at regular intervals to reflect the latest information.

[0606] Step 3: Information Analysis

[0607] The server tokenizes the collected data using natural language processing libraries (NLTK, spaCy). Specifically, it divides sentences into words and phrases. The input "Zeus, the chief god of Greek mythology" is tokenized as ["Greek mythology", "chief god", "Zeus"].

[0608] The server uses TF-IDF and Word2Vec to extract keywords and important information. For example, it might extract keywords such as "Zeus" and "Olympian gods."

[0609] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This helps it understand the context of "Zeus is the king of the Olympian gods."

[0610] Step 4: Fact Check

[0611] The server uses fact-checking algorithms (such as random forests, SVMs, and neural networks) to evaluate the accuracy of the information. Specifically, it identifies discrepancies between the input information and the collected data.

[0612] The server calculates a reliability score based on the comparison results. The reliability score is displayed as a numerical value; for example, a score of 0.9 or higher is considered "high reliability."

[0613] The server determines whether the score meets the set criteria and assesses its accuracy. For example, it might determine that "Zeus is a god in Greek mythology" is correct, while "Zeus is a god in Roman mythology" is incorrect.

[0614] Step 5: Display Results

[0615] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and recommended corrections.

[0616] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). Confidence scores are visually displayed using bar graphs, etc.

[0617] The terminal displays the generated report in the user interface, including specific data points and links to suggested corrections.

[0618] The above outlines the specific processing steps of this system's program.

[0619] (Application Example 1)

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

[0621] Ensuring the reliability of information such as product details and user reviews is crucial for e-commerce websites. However, identifying misinformation and false reviews from a vast amount of data and providing only reliable information is not easy. Current systems lack the means to determine the accuracy of user-provided information in real time, which makes them prone to problems caused by misinformation.

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

[0623] In this invention, the server includes means for receiving information entered by a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for visually displaying the results of the fact-checking, means for providing a user interface for inputting product information and user reviews, means for collecting product information from an official database, means for analyzing the content of the information using natural language processing technology, and means for calculating a reliability score and visually indicating reliability. This makes it possible to determine the accuracy of information on an e-commerce site in real time and provide users with only highly reliable information.

[0624] A "user" is an individual or organization that uses the system to input product information and user reviews, and to verify the accuracy of that information.

[0625] "Public data sources" refer to information sources that can be obtained from publicly available databases and websites.

[0626] A "reliable database" is a database containing highly reliable information provided by a public institution or certified organization.

[0627] "Tokenization of information" is the process of dividing text into words or phrases using natural language processing techniques.

[0628] "Keyword extraction" is the process of selecting important words and phrases from a text.

[0629] A "fact-checking algorithm" is a computational method or model used to compare input information with acquired information and evaluate its accuracy and reliability.

[0630] A "user interface" is an interactive screen or system through which a user inputs information and receives results.

[0631] An "official database" is a reliable database used to provide official information about a product.

[0632] "Natural language processing technology" refers to the technology that enables computers to understand, analyze, and process human language.

[0633] A "reliability score" is an indicator that numerically represents the accuracy of information, and it concretely expresses the evaluation results.

[0634] This invention is a system for determining the accuracy of generated information. This system is primarily used to determine the reliability of product information and user reviews on e-commerce websites.

[0635] The server first uses a means to receive information entered by the user. Users can enter product information and user reviews through the user interface of their smartphone. The server receives this input information and temporarily stores it.

[0636] Next, the server employs means to collect additional information from relevant public data sources and trusted databases. Collecting product information from official databases ensures reliable data.

[0637] The server then tokenizes the obtained data and employs methods to extract keywords and important information. Natural language processing techniques (e.g., Spacy) are used to divide the text into words and phrases, and extract the important parts.

[0638] The server then runs fact-checking algorithms to assess the accuracy of the information. These algorithms include a function to calculate a reliability score, utilizing techniques such as random forests, SVMs, and neural networks. The information provided by the user is compared with official data, and a numerical value is calculated based on the scoring function.

[0639] Finally, the server employs a means to visually display the fact-checking results. Using visualization tools (e.g., Matplotlib), it displays reliability scores as charts and graphs, providing users with an intuitively understandable format.

[0640] For example, if a user enters a product description such as "This is a test product description for a smartwatch," the server will collect information about smartwatches from the official database and extract key keywords using natural language processing technology. Next, a fact-checking algorithm will calculate a reliability score, and finally, the results will be displayed as a graph.

[0641] This makes it possible to verify the accuracy of information on e-commerce sites in real time and provide users with reliable information.

[0642] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0643] Step 1:

[0644] Users input product information and user reviews through their smartphone's user interface. The entered information is sent to the server in JSON format. For example, the input data might be the text "This is a test product description for a smartwatch." The server receives this input information and stores it temporarily.

[0645] Step 2:

[0646] The server collects additional information from relevant public data sources and trusted databases. This collection process uses REST API calls to access databases and retrieve official product information (e.g., product specifications and official reviews). The retrieved data is returned to the server in JSON format and temporarily stored. This ensures reliable data for comparison with user input.

[0647] Step 3:

[0648] The server tokenizes the stored data using natural language processing techniques. Specifically, it uses the Spacy library to split text into words and phrases. This process converts the input text into a list of words, including attribute information for each word (e.g., part of speech).

[0649] Step 4:

[0650] The server extracts keywords and important information from the tokenized data. Algorithms such as TF-IDF and Word2Vec are used to identify particularly important words and phrases within the text. This process yields a list of extracted keywords. For example, "smartwatch" and "product description" might be extracted as keywords.

[0651] Step 5:

[0652] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it compares user input information with official data using random forests or neural network models. This algorithm evaluates the reliability of each keyword in the input information and calculates an overall reliability score. This score is output as a numerical value ranging from 0 to 1.

[0653] Step 6:

[0654] The server visually displays the fact-checking results based on the calculated reliability score. Visualization tools such as Matplotlib are used to convert the reliability score into charts and graphs. This display is then sent back to the smartphone user interface, presenting it in an intuitively understandable format. For example, a pie chart might show a reliability of 70%.

[0655] Step 7:

[0656] The user reviews the displayed fact-check results and corrects the input information as needed. The corrected information is sent back to the server for re-evaluation. This loop ensures that accurate and reliable information is ultimately provided to the user.

[0657] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0658] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[0659] Program Overview

[0660] The program for this system consists of the following modules:

[0661] 1. Data Acquisition Module

[0662] 2. Information Analysis Module

[0663] 3. Fact-checking module

[0664] 4. Emotion Recognition Module

[0665] 5. Result Display Module

[0666] Program Processing Description

[0667] Data Acquisition Module

[0668] The user logs into the system and enters the information they want to analyze. They enter text into the input form and click the submit button.

[0669] The terminal packages the received information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0670] The server temporarily stores the received information in storage and makes API calls to retrieve additional information from relevant public data sources or trusted databases.

[0671] Information analysis module

[0672] The server tokenizes the collected data using a natural language processing library. This divides the text into words and phrases, facilitating data analysis.

[0673] The server extracts keywords and important information using TF-IDF and Word2Vec. This clarifies the central points of the information.

[0674] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. The context of the text is understood in detail.

[0675] Fact-checking module

[0676] The server runs fact-checking algorithms to assess the accuracy of the information. It calculates a reliability score and compares the retrieved data with the input information.

[0677] The server evaluates the accuracy of the information based on a reliability score.

[0678] Emotion recognition module

[0679] Users express emotions by providing voice input or facial expressions to the system. The system measures how users in educational and public institutions react emotionally while evaluating information.

[0680] The device acquires audio or video data and sends it to the server.

[0681] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[0682] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is feeling anxious or suspicious, the server adjusts how the results are displayed and the content of the feedback accordingly.

[0683] Result display module

[0684] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[0685] The server converts the results into graphs and charts using visualization tools, allowing users to intuitively understand the results.

[0686] The device displays the generated report in the user interface, allowing the user to view detailed analysis results and sentiment-based feedback.

[0687] Specific example

[0688] Example 1: Use for educational institutions

[0689] The user (educator) inputs materials to be used in history lessons into the system.

[0690] The terminal sends the entered information to the server.

[0691] The server collects relevant information by making API calls to the relevant historical databases.

[0692] The server uses natural language processing techniques to tokenize information and extract keywords.

[0693] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[0694] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[0695] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0696] The device displays the generated report and recommended feedback to the educator.

[0697] Example 2: Use by public institutions

[0698] The user (government official) enters a draft of a press release regarding the new policy.

[0699] The terminal sends information to the server.

[0700] The server collects additional information from trusted public databases.

[0701] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0702] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[0703] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[0704] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0705] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[0706] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[0707] The following describes the processing flow.

[0708] Step 1:

[0709] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[0710] Step 2:

[0711] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0712] Step 3:

[0713] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[0714] Step 4:

[0715] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[0716] Step 5:

[0717] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[0718] Step 6:

[0719] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[0720] Step 7:

[0721] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[0722] Step 8:

[0723] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[0724] Step 9:

[0725] The server receives voice input or facial recognition data to recognize the user's emotions. The user can communicate their emotions to the system through voice or facial expressions.

[0726] Step 10:

[0727] The device sends the collected audio or video data to the server, which then provides data for sentiment analysis.

[0728] Step 11:

[0729] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[0730] Step 12:

[0731] The server adjusts the output of the fact-checking algorithm based on the analyzed emotional state. For example, if the user is feeling anxious or suspicious, it flexibly changes how the results are displayed and the content of the feedback.

[0732] Step 13:

[0733] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[0734] Step 14:

[0735] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0736] Step 15:

[0737] The device displays the generated report in the user interface. Users can view detailed analysis results and sentiment-based feedback.

[0738] (Example 2)

[0739] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0740] Traditionally, systems that judged the accuracy of information generated by generative AI lacked the ability to consider user emotions. As a result, evaluation results sometimes did not align with the user's emotional state, leading to problems in understanding and accepting the results. To address this issue, a system is needed that provides comprehensive feedback that considers user emotions in addition to accurate information judgment.

[0741] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotions, means for adjusting the output results of the fact-checking algorithm based on emotion recognition, and means for compiling the fact-checking results and emotion recognition results into a comprehensive report and displaying it visually. This makes it possible to evaluate the accuracy of the information and provide feedback that takes the user's emotions into consideration.

[0742] A "user" refers to the entity that uses this system to input information and receive the results.

[0743] "Public data sources" refer to reliable sources of information that are publicly available on the internet. Examples include encyclopedias and academic papers.

[0744] A "reliable database" refers to a collection of information whose accuracy and reliability are guaranteed. Examples include government statistical databases and databases from official research institutions.

[0745] "Tokenization" is a preprocessing step in natural language processing, referring to the process of dividing a text into words or phrases.

[0746] A "keyword" refers to a word or phrase that is considered particularly important from the input information or collected data.

[0747] A "fact-checking algorithm" refers to an algorithm used to evaluate the accuracy and reliability of information.

[0748] An "emotion recognition engine" refers to a technology that analyzes audio and video data to identify the user's emotional state (for example, joy, sadness, anger, etc.).

[0749] "Visual display" refers to the process of converting text information into visual formats such as graphs, charts, and diagrams for display.

[0750] A "reliability score" refers to an evaluation index that quantifies and shows the reliability of information.

[0751] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[0752] Program Overview

[0753] This system's program consists of the following modules:

[0754] 1. Data Acquisition Module

[0755] 2. Information Analysis Module

[0756] 3. Fact-checking module

[0757] 4. Emotion Recognition Module

[0758] 5. Result Display Module

[0759] Hardware and software to use

[0760] The following hardware and software are required to implement this system:

[0761] Server: A server that performs data processing and analysis.

[0762] Device: A device used by a user to input information (e.g., a PC or tablet).

[0763] Database: Storage for storing received data and additional information.

[0764] Natural language processing libraries: For example, NLTK and spaCy

[0765] Fact-checking tools: For example, FactCheck Tools API

[0766] Emotion recognition engines: For example, Google Cloud Speech-to-Text and Microsoft Azure Face API

[0767] Visualization tools: For example, D3.js and Chart.js

[0768] Specific example

[0769] Specific example 1: Use in educational institutions

[0770] The user (educator) inputs materials to be used in history lessons into the system. For example, they might input information such as, "The American Civil War ended in 1865."

[0771] The terminal sends the entered information to the server.

[0772] The server collects relevant information by making API calls to the relevant historical databases.

[0773] The server uses natural language processing techniques to tokenize information and extract keywords.

[0774] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[0775] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[0776] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0777] The device displays the generated report and recommended feedback to the educator.

[0778] Specific example 2: Use by public institutions

[0779] The user (government official) enters a draft of a press release regarding a new policy. For example, they might enter information such as, "A new environmental policy will reduce carbon dioxide emissions by 30% by 2030."

[0780] The terminal sends information to the server.

[0781] The server collects additional information from trusted public databases.

[0782] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0783] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[0784] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[0785] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[0786] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[0787] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[0788] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0789] Step 1:

[0790] The user logs into the system. On the login screen, they enter their username and password and submit their authentication information. Input: Username, Password. Output: Authentication token.

[0791] Step 2:

[0792] The user enters the information to be evaluated using a text input form. For example, they might enter the contents of materials used in a history class into the text field. Input: Text data of the information to be evaluated. Output: Click event of the submit button.

[0793] Step 3:

[0794] The terminal retrieves the information entered by the user when they click the submit button. It then packages the retrieved information into JSON format and sends it to the server. Input: Text data. Output: Sending JSON data.

[0795] Step 4:

[0796] The server temporarily stores the received JSON data in storage. This storage uses a database or object storage. Input: JSON data. Output: Temporarily stored data.

[0797] Step 5:

[0798] The server makes API calls to public data sources and trusted databases to collect relevant information from the database. For example, it uses Wikidata and other reliable sources. Input: Temporarily stored data. Output: Additional data collected.

[0799] Step 6:

[0800] The server tokenizes the collected data using a natural language processing library (e.g., NLTK or spaCy). This divides the text into words and phrases. Input: Collected data. Output: Tokenized data.

[0801] Step 7:

[0802] The server extracts keywords from tokenized data using TF-IDF or Word2Vec. This identifies important information. Input: Tokenized data. Output: Keywords.

[0803] Step 8:

[0804] The server performs dependency analysis and syntactic analysis to analyze the relationships between the obtained keywords. For example, it uses the Stanford NLP dependency analysis model. Input: Keywords. Output: Relationships between keywords.

[0805] Step 9:

[0806] The server executes a fact-checking algorithm, comparing the acquired data with the input information to calculate a reliability score. Input: Relationships between keywords. Output: Reliability score.

[0807] Step 10:

[0808] The user provides audio or video input to the system, for example, using a webcam or microphone. Input: audio data, video data. Output: transmission of emotion recognition data.

[0809] Step 11:

[0810] The terminal acquires audio or video data and sends it to the server. Input: Audio data, video data. Output: Data transmission to the server.

[0811] Step 12:

[0812] The server analyzes the user's emotions using an emotion recognition engine (e.g., Google Cloud Speech-to-Text or Microsoft Azure Face API). Input: Audio data, video data. Output: Emotion analysis results.

[0813] Step 13:

[0814] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is expressing anxiety, it selects a way to present the results in a clearer manner. Input: Sentiment analysis results, reliability score. Output: Adjusted feedback content.

[0815] Step 14:

[0816] The server compiles the fact-checking and sentiment assessment results into a comprehensive report and converts them into graphs and charts using visualization tools (e.g., D3.js or Chart.js). Input: Adjusted feedback. Output: Visualized report.

[0817] Step 15:

[0818] The terminal displays the generated report in the user interface. On this screen, the user can view detailed analysis results and sentiment-based feedback. Input: Visualized report. Output: Displayed in the user interface.

[0819] The above outlines the specific processing steps of the system program.

[0820] (Application Example 2)

[0821] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0822] When determining the accuracy of information generated by generative AI, conventional systems have a problem in that they cannot provide feedback that takes user emotions into account. The feedback users receive is uniform, and for example, if a user feels anxious or doubtful, appropriate support is not provided. This raises concerns that it increases the psychological burden on users and hinders their acceptance and understanding of the information.

[0823] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0824] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotional state, means for visually displaying the fact-checking results, and means for adjusting feedback based on the user's emotional state. This enables the evaluation of the reliability of the information input by the user and the provision of appropriate feedback based on the user's emotional state. In this way, the psychological burden on the user can be reduced and the acceptance and understanding of information can be promoted.

[0825] A "user" is someone who inputs information into a system and receives the results.

[0826] A "public data source" refers to a collection of data that is widely made public to provide reliable information.

[0827] A "reliable database" is a database that stores information whose reliability has been verified.

[0828] "Tokenization" is the process of dividing an input text into its smallest units, such as words or phrases.

[0829] A "keyword" refers to a word or phrase that has particularly important meaning within a text or data.

[0830] A "fact-checking algorithm" refers to a set of computational procedures designed to evaluate the accuracy of input information.

[0831] An "emotion recognition engine" is a technology that analyzes a user's voice and facial expression data to identify their emotional state.

[0832] A "reliability score" is an indicator that quantifies the accuracy of information.

[0833] "Visually displaying" refers to providing users with data and information in a visual format, such as graphs and charts.

[0834] "Adjusting feedback" means changing the content and format of the feedback output based on the user's emotional state.

[0835] To implement this invention, a system is required in which a server, terminal, and user work together. This system includes modules for data collection, information analysis, fact-checking, sentiment recognition, and result display.

[0836] The server receives information entered by the user. For example, if a user enters a URL for a news article, the server packages that information in JSON format and saves it to storage. It also makes API calls to collect additional information from relevant public data sources and trusted databases. This allows for centralized management of necessary data.

[0837] Next, the server tokenizes the collected data using a natural language processing library. Specifically, it uses the Python nltk library to make it easier to divide sentences into words and phrases. Furthermore, it uses the scikit-learn library to extract keywords and important information using methods such as TF-IDF. This clarifies the central points of the information and improves the accuracy of fact-checking.

[0838] The fact-checking algorithm is executed on the server side. The server performs comparisons to calculate a reliability score based on the collected and analyzed data. It cross-references the acquired data with the input information to evaluate the accuracy of the information.

[0839] Furthermore, an emotion recognition engine is used to analyze the user's emotional state. The user provides voice input and facial expressions, and the device sends this data to the server. The server analyzes the emotional state using voice analysis and facial recognition technology, and adjusts the fact-check output based on the results. Here, the emotion_recognition library is used as the emotion recognition engine.

[0840] Finally, the server generates a report integrating fact-checking results and sentiment recognition results, and uses visualization tools to convert the results into graphs and charts. This allows users to intuitively understand the results and reduces the psychological burden of receiving feedback. Users can view detailed analysis results and sentiment-based feedback through their devices.

[0841] As a concrete example, let's explain how the news application "Emofact News" works. The user enters the URL of a news article and provides audio and facial expression data. On the server side, the input information is analyzed and a reliability score and emotional state are evaluated. Finally, the results are displayed visually, allowing the user to intuitively understand the reliability of the information and receive emotionally sensitive feedback.

[0842] An example of the generated prompt statement is as follows:

[0843] Please evaluate the reliability of the following article and the user's sentiment-based feedback. Visit "https: / / example.com / news / some-article", check the accuracy of the information, and display the results while taking into consideration any concerns or doubts the user may have.

[0844] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0845] Step 1:

[0846] User enters information

[0847] The user accesses the system interface and enters information such as the URL of a news article. The entered information is packaged in JSON format on the terminal and sent to the server.

[0848] Input: URL of the news article

[0849] Output: Information in JSON format

[0850] ---

[0851] Step 2:

[0852] Data collection

[0853] The server temporarily stores the JSON-formatted information received from the terminal in storage. Furthermore, it executes API calls to collect additional information from relevant public data sources and trusted databases.

[0854] Input: Information in JSON format

[0855] Output: Stored information and additionally collected data

[0856] ---

[0857] Step 3:

[0858] Information analysis

[0859] The server tokenizes the collected data using a natural language processing library (e.g., the nltk library). From the tokenized data, keywords and important information are extracted using TF-IDF.

[0860] Input: Collected data

[0861] Output: Tokenized data and extracted keywords

[0862] ---

[0863] Step 4:

[0864] Fact Check

[0865] The server runs fact-checking algorithms and calculates a reliability score by cross-referencing extracted keywords with data collected from publicly available data sources and trusted databases.

[0866] Input: Extracted keywords and collected additional information

[0867] Output: Reliability score

[0868] ---

[0869] Step 5:

[0870] emotion recognition

[0871] When entering information, the user provides voice input or facial expression data. The terminal sends this data to the server, which uses an emotion recognition engine to analyze the user's emotional state. For example, the emotion_recognition library can be used.

[0872] Input: Voice data or facial expression data

[0873] Output: Analyzed emotional state

[0874] ---

[0875] Step 6:

[0876] Adjustment and integration of results

[0877] The server integrates the reliability score and emotional state, and adjusts the feedback content according to the user's emotions. This generates appropriate feedback that reduces the user's psychological burden.

[0878] Input: Reliability score and analyzed emotional state

[0879] Output: Adjusted feedback content

[0880] ---

[0881] Step 7:

[0882] Visualization and display of results

[0883] The server converts the integrated information into graphs and charts using visualization tools. This allows the terminal to display the results in a format that is intuitively easy for the user to understand.

[0884] Input: Integrated result

[0885] Output: Visualized results and feedback

[0886] ---

[0887] Step 8:

[0888] Provide feedback

[0889] The device displays visualized results and refined feedback sent from the server to the user. The user reviews the results and receives feedback that is reliable and considerate of their feelings.

[0890] Input: Visualized results and feedback content

[0891] Output: Display to the user

[0892] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0893] The data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One 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">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0894] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0895] [Third Embodiment]

[0896] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0897] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0898] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0899] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0900] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0901] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0902] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0903] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0904] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0905] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0906] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0907] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0908] This invention is a system for determining the accuracy of information generated by generative AI, and is primarily designed for educational and public institutions. The system consists of a series of processes: user input, data collection from public data sources and reliable databases, information analysis, fact-checking, and a visual display of the results.

[0909] Program Overview

[0910] The program for this system consists of the following modules:

[0911] 1. Data Acquisition Module

[0912] 2. Information Analysis Module

[0913] 3. Fact-checking module

[0914] 4. Result Display Module

[0915] Program Processing Description

[0916] Data Acquisition Module

[0917] Users input information into the system. Educators and government officials provide the system with lesson materials, draft press releases, and other materials.

[0918] The terminal retrieves information from the user input form and sends it to the server. The information is sent to the API endpoint in JSON format.

[0919] The server receives the input information and stores it in temporary storage. This ensures consistency for each piece of information.

[0920] The server makes API calls to collect additional information from relevant public data sources and trusted databases (e.g., Wikipedia or government data).

[0921] Information analysis module

[0922] The server tokenizes the collected data using a natural language processing library (e.g., NLTK, spaCy). This divides the text into words and phrases.

[0923] The server uses technologies such as TF-IDF and Word2Vec to extract keywords and important information. This clarifies the central points of the information.

[0924] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows the context of the information content to be understood.

[0925] Fact-checking module

[0926] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it might use models such as random forests, SVMs, or neural networks.

[0927] The server compares the acquired data with user input information and calculates a reliability score. The numerical value, based on the scoring function, indicates the reliability of the information.

[0928] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[0929] Result display module

[0930] The server compiles the fact-checking results and generates a comprehensive report. The report includes evaluation results, relevant sources, and recommended corrections.

[0931] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0932] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[0933] Specific example

[0934] Example 1: Use for educational institutions

[0935] The user (educator) inputs materials to be used in history lessons into the system.

[0936] The terminal sends the entered information to the server.

[0937] The server collects relevant information by making API calls to the relevant historical databases.

[0938] The server uses natural language processing techniques to tokenize information and extract keywords.

[0939] The server evaluates the accuracy of the information and calculates a reliability score based on fact-checking algorithms.

[0940] The server visualizes the results and displays them to the educator.

[0941] The device provides educators with accurate information sources and suggested revisions.

[0942] Example 2: Use by public institutions

[0943] The user (government official) enters a draft of a press release regarding the new policy.

[0944] The terminal sends information to the server.

[0945] The server collects additional information from trusted public databases.

[0946] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[0947] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[0948] The server visualizes the results in graphs and charts and presents them to government officials.

[0949] The terminal displays detailed analysis results and suggested corrections to the staff.

[0950] The above describes the specific embodiments for implementing the system of the present invention.

[0951] The following describes the processing flow.

[0952] Step 1:

[0953] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[0954] Step 2:

[0955] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[0956] Step 3:

[0957] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[0958] Step 4:

[0959] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[0960] Step 5:

[0961] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[0962] Step 6:

[0963] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[0964] Step 7:

[0965] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[0966] Step 8:

[0967] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[0968] Step 9:

[0969] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and suggested corrections.

[0970] Step 10:

[0971] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[0972] Step 11:

[0973] The terminal displays the generated report in the user interface. Users can view detailed analysis results and recommended corrective actions.

[0974] (Example 1)

[0975] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0976] Determining the accuracy of information generated by generative AI is a critical issue, particularly in educational and public institutions. However, current systems lack the means to automatically and efficiently evaluate the accuracy of user-inputted information. As a result, reliable fact-checking is not performed, and there is a risk of misinformation spreading. Furthermore, the analysis and visual display of collected data are insufficient, making it difficult for users to intuitively understand the information. An effective system is needed to address these challenges.

[0977] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0978] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for performing dependency analysis and syntactic analysis to analyze the context of the information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, and means for visually displaying the results of the fact-checking. This makes it possible to evaluate the accuracy of user-inputted information with high precision and provide reliable data. Furthermore, the visual display of the results realizes a system that is easy for users to understand intuitively.

[0979] "Information" refers to data that users input into the system, as well as data collected from relevant public data sources and trusted databases.

[0980] A "user" refers to an employee of an educational institution or public organization, or an individual who uses the system and is the entity that inputs information into the system.

[0981] "Public data sources" refer to information sources that are publicly available on the internet or in other accessible locations, and include, for example, encyclopedias and government statistics.

[0982] A "reliable database" refers to a source of information managed by a highly credible institution or organization. This includes academic journal databases and official statistical databases.

[0983] "Tokenization" refers to the process of dividing words and phrases in a text into individual elements, and it is a fundamental process in natural language processing.

[0984] A "keyword" refers to a word or phrase that is considered particularly important within information, and is used to grasp the subject or content of the information.

[0985] "Dependency analysis" is a technique for understanding sentence structure and meaning by analyzing the dependencies between words in a text.

[0986] "Syntactic analysis" is a method of analyzing the grammatical structure of a text and evaluating its meaning and grammatical accuracy.

[0987] "Fact-checking" refers to the process of evaluating the accuracy and truthfulness of information, and is the work of verifying the accuracy of information based on collected data.

[0988] A "fact-checking algorithm" refers to an algorithm used to computationally evaluate the accuracy of information, often employing machine learning models or statistical methods.

[0989] A "reliability score" is a numerical representation of the accuracy and reliability of information, calculated using an algorithm.

[0990] "Visual display" refers to presenting information and data to users in the form of graphs and charts using visualization tools, making it easier for users to intuitively understand the data.

[0991] This invention is a system composed of multiple modules for determining the accuracy of information generated by a generative AI. The specific method for realizing this system is described below.

[0992] System Overview

[0993] This system consists of the following modules:

[0994] 1. User Input Module

[0995] 2. Data Acquisition Module

[0996] 3. Information Analysis Module

[0997] 4. Fact-checking module

[0998] 5. Result Display Module

[0999] User input module

[1000] Users enter information through the system's input forms. For example, this could involve educators entering lesson materials or government officials entering draft policy documents.

[1001] The terminal retrieves information from the input form and sends it to the server in JSON format.

[1002] Data Acquisition Module

[1003] The server stores the information received from the user in temporary storage.

[1004] The server collects additional information from relevant public data sources and trusted databases. These public data sources include category information sites and public databases. For example, it uses the Wikipedia API and government public data APIs.

[1005] Information analysis module

[1006] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). Specifically, it divides sentences into words and phrases and analyzes their structure.

[1007] The server uses TF-IDF and Word2Vec technologies to extract keywords and important information, thereby clarifying the central points of the information.

[1008] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows it to understand the contextual meaning of the text.

[1009] Fact-checking module

[1010] The server evaluates the accuracy of the information using fact-checking algorithms (such as random forests, SVMs, and neural networks).

[1011] The server compares the collected data with user input information and calculates a reliability score. The reliability score is a numerical indicator of the accuracy of the information; for example, a score of 0.9 or higher is considered reliable.

[1012] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[1013] Result display module

[1014] The server compiles the fact-checking results into a comprehensive report. The report includes evaluation results, relevant sources, and recommended revisions.

[1015] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). This allows users to intuitively understand the results.

[1016] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[1017] Specific example

[1018] Example 1: Use for educational institutions

[1019] The user (educator) inputs the material "Outline of Greek Mythology," which is used in history lessons, into the system.

[1020] The terminal sends the input information to the server.

[1021] The server collects and analyzes information by making API calls to relevant historical databases. For example, it might use "the content of Wikipedia articles on Greek mythology."

[1022] The server uses fact-checking algorithms to evaluate the accuracy of the information and calculates a reliability score.

[1023] The server visualizes the results and displays them to the educator.

[1024] The device provides educators with accurate information sources and suggested revisions.

[1025] Example 2: Use by public institutions

[1026] The user (government official) enters a draft of a press release regarding a new policy.

[1027] The terminal sends information to the server.

[1028] The server collects additional information from trusted public databases and performs analysis.

[1029] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[1030] The server visualizes the results in graphs and charts and presents them to government officials.

[1031] The terminal displays detailed analysis results and suggested corrections to the staff.

[1032] The above describes the embodiments for implementing the system of the present invention.

[1033] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1034] Step 1: Enter Information

[1035] The user enters information into the system's input form. For example, "An overview of Greek mythology."

[1036] The terminal retrieves the entered information and sends it to the server in JSON format. The input includes the title "Outline of Greek Mythology" and the body text.

[1037] The server stores the received information in temporary storage. This ensures consistency in the input information and facilitates subsequent processing.

[1038] Step 2: Data Collection

[1039] The server collects data from relevant public data sources and trusted databases based on information received from the user. For example, it retrieves data related to "Greek mythology" from the Wikipedia API and public databases.

[1040] The server temporarily stores the collected data in JSON format. Specifically, this includes information such as "details of Greek mythology" and "major characters."

[1041] The server synchronizes with publicly available data sources at regular intervals to reflect the latest information.

[1042] Step 3: Information Analysis

[1043] The server tokenizes the collected data using natural language processing libraries (NLTK, spaCy). Specifically, it divides sentences into words and phrases. The input "Zeus, the chief god of Greek mythology" is tokenized as ["Greek mythology", "chief god", "Zeus"].

[1044] The server uses TF-IDF and Word2Vec to extract keywords and important information. For example, it might extract keywords such as "Zeus" and "Olympian gods."

[1045] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This helps it understand the context of "Zeus is the king of the Olympian gods."

[1046] Step 4: Fact Check

[1047] The server uses fact-checking algorithms (such as random forests, SVMs, and neural networks) to evaluate the accuracy of the information. Specifically, it identifies discrepancies between the input information and the collected data.

[1048] The server calculates a reliability score based on the comparison results. The reliability score is displayed as a numerical value; for example, a score of 0.9 or higher is considered "high reliability."

[1049] The server determines whether the score meets the set criteria and assesses its accuracy. For example, it might determine that "Zeus is a god in Greek mythology" is correct, while "Zeus is a god in Roman mythology" is incorrect.

[1050] Step 5: Display Results

[1051] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and recommended corrections.

[1052] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). Confidence scores are visually displayed using bar graphs, etc.

[1053] The terminal displays the generated report in the user interface, including specific data points and links to suggested corrections.

[1054] The above outlines the specific processing steps of this system's program.

[1055] (Application Example 1)

[1056] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1057] Ensuring the reliability of information such as product details and user reviews is crucial for e-commerce websites. However, identifying misinformation and false reviews from a vast amount of data and providing only reliable information is not easy. Current systems lack the means to determine the accuracy of user-provided information in real time, which makes them prone to problems caused by misinformation.

[1058] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1059] In this invention, the server includes means for receiving information entered by a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for visually displaying the results of the fact-checking, means for providing a user interface for inputting product information and user reviews, means for collecting product information from an official database, means for analyzing the content of the information using natural language processing technology, and means for calculating a reliability score and visually indicating reliability. This makes it possible to determine the accuracy of information on an e-commerce site in real time and provide users with only highly reliable information.

[1060] A "user" is an individual or organization that uses the system to input product information and user reviews, and to verify the accuracy of that information.

[1061] "Public data sources" refer to information sources that can be obtained from publicly available databases and websites.

[1062] A "reliable database" is a database containing highly reliable information provided by a public institution or certified organization.

[1063] "Tokenization of information" is the process of dividing text into words or phrases using natural language processing techniques.

[1064] "Keyword extraction" is the process of selecting important words and phrases from a text.

[1065] A "fact-checking algorithm" is a computational method or model used to compare input information with acquired information and evaluate its accuracy and reliability.

[1066] A "user interface" is an interactive screen or system through which a user inputs information and receives results.

[1067] An "official database" is a reliable database used to provide official information about a product.

[1068] "Natural language processing technology" refers to the technology that enables computers to understand, analyze, and process human language.

[1069] A "reliability score" is an indicator that numerically represents the accuracy of information, and it concretely expresses the evaluation results.

[1070] This invention is a system for determining the accuracy of generated information. This system is primarily used to determine the reliability of product information and user reviews on e-commerce websites.

[1071] The server first uses a means to receive information entered by the user. Users can enter product information and user reviews through the user interface of their smartphone. The server receives this input information and temporarily stores it.

[1072] Next, the server employs means to collect additional information from relevant public data sources and trusted databases. Collecting product information from official databases ensures reliable data.

[1073] The server then tokenizes the obtained data and employs methods to extract keywords and important information. Natural language processing techniques (e.g., Spacy) are used to divide the text into words and phrases, and extract the important parts.

[1074] The server then runs fact-checking algorithms to assess the accuracy of the information. These algorithms include a function to calculate a reliability score, utilizing techniques such as random forests, SVMs, and neural networks. The information provided by the user is compared with official data, and a numerical value is calculated based on the scoring function.

[1075] Finally, the server employs a means to visually display the fact-checking results. Using visualization tools (e.g., Matplotlib), it displays reliability scores as charts and graphs, providing users with an intuitively understandable format.

[1076] For example, if a user enters a product description such as "This is a test product description for a smartwatch," the server will collect information about smartwatches from the official database and extract key keywords using natural language processing technology. Next, a fact-checking algorithm will calculate a reliability score, and finally, the results will be displayed as a graph.

[1077] This makes it possible to verify the accuracy of information on e-commerce sites in real time and provide users with reliable information.

[1078] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1079] Step 1:

[1080] Users input product information and user reviews through their smartphone's user interface. The entered information is sent to the server in JSON format. For example, the input data might be the text "This is a test product description for a smartwatch." The server receives this input information and stores it temporarily.

[1081] Step 2:

[1082] The server collects additional information from relevant public data sources and trusted databases. This collection process uses REST API calls to access databases and retrieve official product information (e.g., product specifications and official reviews). The retrieved data is returned to the server in JSON format and temporarily stored. This ensures reliable data for comparison with user input.

[1083] Step 3:

[1084] The server tokenizes the stored data using natural language processing techniques. Specifically, it uses the Spacy library to split text into words and phrases. This process converts the input text into a list of words, including attribute information for each word (e.g., part of speech).

[1085] Step 4:

[1086] The server extracts keywords and important information from the tokenized data. Algorithms such as TF-IDF and Word2Vec are used to identify particularly important words and phrases within the text. This process yields a list of extracted keywords. For example, "smartwatch" and "product description" might be extracted as keywords.

[1087] Step 5:

[1088] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it compares user input information with official data using random forests or neural network models. This algorithm evaluates the reliability of each keyword in the input information and calculates an overall reliability score. This score is output as a numerical value ranging from 0 to 1.

[1089] Step 6:

[1090] The server visually displays the fact-checking results based on the calculated reliability score. Visualization tools such as Matplotlib are used to convert the reliability score into charts and graphs. This display is then sent back to the smartphone user interface, presenting it in an intuitively understandable format. For example, a pie chart might show a reliability of 70%.

[1091] Step 7:

[1092] The user reviews the displayed fact-check results and corrects the input information as needed. The corrected information is sent back to the server for re-evaluation. This loop ensures that accurate and reliable information is ultimately provided to the user.

[1093] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1094] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[1095] Program Overview

[1096] The program for this system consists of the following modules:

[1097] 1. Data Acquisition Module

[1098] 2. Information Analysis Module

[1099] 3. Fact-checking module

[1100] 4. Emotion Recognition Module

[1101] 5. Result Display Module

[1102] Program Processing Description

[1103] Data Acquisition Module

[1104] The user logs into the system and enters the information they want to analyze. They enter text into the input form and click the submit button.

[1105] The terminal packages the received information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[1106] The server temporarily stores the received information in storage and makes API calls to retrieve additional information from relevant public data sources or trusted databases.

[1107] Information analysis module

[1108] The server tokenizes the collected data using a natural language processing library. This divides the text into words and phrases, facilitating data analysis.

[1109] The server extracts keywords and important information using TF-IDF and Word2Vec. This clarifies the central points of the information.

[1110] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. The context of the text is understood in detail.

[1111] Fact-checking module

[1112] The server runs fact-checking algorithms to assess the accuracy of the information. It calculates a reliability score and compares the retrieved data with the input information.

[1113] The server evaluates the accuracy of the information based on a reliability score.

[1114] Emotion recognition module

[1115] Users express emotions by providing voice input or facial expressions to the system. The system measures how users in educational and public institutions react emotionally while evaluating information.

[1116] The device acquires audio or video data and sends it to the server.

[1117] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[1118] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is feeling anxious or suspicious, the server adjusts how the results are displayed and the content of the feedback accordingly.

[1119] Result display module

[1120] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[1121] The server converts the results into graphs and charts using visualization tools, allowing users to intuitively understand the results.

[1122] The device displays the generated report in the user interface, allowing the user to view detailed analysis results and sentiment-based feedback.

[1123] Specific example

[1124] Example 1: Use for educational institutions

[1125] The user (educator) inputs materials to be used in history lessons into the system.

[1126] The terminal sends the entered information to the server.

[1127] The server collects relevant information by making API calls to the relevant historical databases.

[1128] The server uses natural language processing techniques to tokenize information and extract keywords.

[1129] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[1130] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[1131] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1132] The device displays the generated report and recommended feedback to the educator.

[1133] Example 2: Use by public institutions

[1134] The user (government official) enters a draft of a press release regarding the new policy.

[1135] The terminal sends information to the server.

[1136] The server collects additional information from trusted public databases.

[1137] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[1138] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[1139] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[1140] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1141] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[1142] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[1143] The following describes the processing flow.

[1144] Step 1:

[1145] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[1146] Step 2:

[1147] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[1148] Step 3:

[1149] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[1150] Step 4:

[1151] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[1152] Step 5:

[1153] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[1154] Step 6:

[1155] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[1156] Step 7:

[1157] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[1158] Step 8:

[1159] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[1160] Step 9:

[1161] The server receives voice input or facial recognition data to recognize the user's emotions. The user can communicate their emotions to the system through voice or facial expressions.

[1162] Step 10:

[1163] The device sends the collected audio or video data to the server, which then provides data for sentiment analysis.

[1164] Step 11:

[1165] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[1166] Step 12:

[1167] The server adjusts the output of the fact-checking algorithm based on the analyzed emotional state. For example, if the user is feeling anxious or suspicious, it flexibly changes how the results are displayed and the content of the feedback.

[1168] Step 13:

[1169] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[1170] Step 14:

[1171] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[1172] Step 15:

[1173] The device displays the generated report in the user interface. Users can view detailed analysis results and sentiment-based feedback.

[1174] (Example 2)

[1175] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1176] Traditionally, systems that judged the accuracy of information generated by generative AI lacked the ability to consider user emotions. As a result, evaluation results sometimes did not align with the user's emotional state, leading to problems in understanding and accepting the results. To address this issue, a system is needed that provides comprehensive feedback that considers user emotions in addition to accurate information judgment.

[1177] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotions, means for adjusting the output results of the fact-checking algorithm based on emotion recognition, and means for compiling the fact-checking results and emotion recognition results into a comprehensive report and displaying it visually. This makes it possible to evaluate the accuracy of the information and provide feedback that takes the user's emotions into consideration.

[1178] A "user" refers to the entity that uses this system to input information and receive the results.

[1179] "Public data sources" refer to reliable sources of information that are publicly available on the internet. Examples include encyclopedias and academic papers.

[1180] A "reliable database" refers to a collection of information whose accuracy and reliability are guaranteed. Examples include government statistical databases and databases from official research institutions.

[1181] "Tokenization" is a preprocessing step in natural language processing, referring to the process of dividing a text into words or phrases.

[1182] A "keyword" refers to a word or phrase that is considered particularly important from the input information or collected data.

[1183] A "fact-checking algorithm" refers to an algorithm used to evaluate the accuracy and reliability of information.

[1184] An "emotion recognition engine" refers to a technology that analyzes audio and video data to identify the user's emotional state (for example, joy, sadness, anger, etc.).

[1185] "Visual display" refers to the process of converting text information into visual formats such as graphs, charts, and diagrams for display.

[1186] A "reliability score" refers to an evaluation index that quantifies and shows the reliability of information.

[1187] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[1188] Program Overview

[1189] This system's program consists of the following modules:

[1190] 1. Data Acquisition Module

[1191] 2. Information Analysis Module

[1192] 3. Fact-checking module

[1193] 4. Emotion Recognition Module

[1194] 5. Result Display Module

[1195] Hardware and software to use

[1196] The following hardware and software are required to implement this system:

[1197] Server: A server that performs data processing and analysis.

[1198] Device: A device used by a user to input information (e.g., a PC or tablet).

[1199] Database: Storage for storing received data and additional information.

[1200] Natural language processing libraries: For example, NLTK and spaCy

[1201] Fact-checking tools: For example, FactCheck Tools API

[1202] Emotion recognition engines: For example, Google Cloud Speech-to-Text and Microsoft Azure Face API

[1203] Visualization tools: For example, D3.js and Chart.js

[1204] Specific example

[1205] Specific example 1: Use in educational institutions

[1206] The user (educator) inputs materials to be used in history lessons into the system. For example, they might input information such as, "The American Civil War ended in 1865."

[1207] The terminal sends the entered information to the server.

[1208] The server collects relevant information by making API calls to the relevant historical databases.

[1209] The server uses natural language processing techniques to tokenize information and extract keywords.

[1210] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[1211] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[1212] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1213] The device displays the generated report and recommended feedback to the educator.

[1214] Specific example 2: Use by public institutions

[1215] The user (government official) enters a draft of a press release regarding a new policy. For example, they might enter information such as, "A new environmental policy will reduce carbon dioxide emissions by 30% by 2030."

[1216] The terminal sends information to the server.

[1217] The server collects additional information from trusted public databases.

[1218] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[1219] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[1220] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[1221] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1222] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[1223] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[1224] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1225] Step 1:

[1226] The user logs into the system. On the login screen, they enter their username and password and submit their authentication information. Input: Username, Password. Output: Authentication token.

[1227] Step 2:

[1228] The user enters the information to be evaluated using a text input form. For example, they might enter the contents of materials used in a history class into the text field. Input: Text data of the information to be evaluated. Output: Click event of the submit button.

[1229] Step 3:

[1230] The terminal retrieves the information entered by the user when they click the submit button. It then packages the retrieved information into JSON format and sends it to the server. Input: Text data. Output: Sending JSON data.

[1231] Step 4:

[1232] The server temporarily stores the received JSON data in storage. This storage uses a database or object storage. Input: JSON data. Output: Temporarily stored data.

[1233] Step 5:

[1234] The server makes API calls to public data sources and trusted databases to collect relevant information from the database. For example, it uses Wikidata and other reliable sources. Input: Temporarily stored data. Output: Additional data collected.

[1235] Step 6:

[1236] The server tokenizes the collected data using a natural language processing library (e.g., NLTK or spaCy). This divides the text into words and phrases. Input: Collected data. Output: Tokenized data.

[1237] Step 7:

[1238] The server extracts keywords from tokenized data using TF-IDF or Word2Vec. This identifies important information. Input: Tokenized data. Output: Keywords.

[1239] Step 8:

[1240] The server performs dependency analysis and syntactic analysis to analyze the relationships between the obtained keywords. For example, it uses the Stanford NLP dependency analysis model. Input: Keywords. Output: Relationships between keywords.

[1241] Step 9:

[1242] The server executes a fact-checking algorithm, comparing the acquired data with the input information to calculate a reliability score. Input: Relationships between keywords. Output: Reliability score.

[1243] Step 10:

[1244] The user provides audio or video input to the system, for example, using a webcam or microphone. Input: audio data, video data. Output: transmission of emotion recognition data.

[1245] Step 11:

[1246] The terminal acquires audio or video data and sends it to the server. Input: Audio data, video data. Output: Data transmission to the server.

[1247] Step 12:

[1248] The server analyzes the user's emotions using an emotion recognition engine (e.g., Google Cloud Speech-to-Text or Microsoft Azure Face API). Input: Audio data, video data. Output: Emotion analysis results.

[1249] Step 13:

[1250] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is expressing anxiety, it selects a way to present the results in a clearer manner. Input: Sentiment analysis results, reliability score. Output: Adjusted feedback content.

[1251] Step 14:

[1252] The server compiles the fact-checking and sentiment assessment results into a comprehensive report and converts them into graphs and charts using visualization tools (e.g., D3.js or Chart.js). Input: Adjusted feedback. Output: Visualized report.

[1253] Step 15:

[1254] The terminal displays the generated report in the user interface. On this screen, the user can view detailed analysis results and sentiment-based feedback. Input: Visualized report. Output: Displayed in the user interface.

[1255] The above outlines the specific processing steps of the system program.

[1256] (Application Example 2)

[1257] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1258] When determining the accuracy of information generated by generative AI, conventional systems have a problem in that they cannot provide feedback that takes user emotions into account. The feedback users receive is uniform, and for example, if a user feels anxious or doubtful, appropriate support is not provided. This raises concerns that it increases the psychological burden on users and hinders their acceptance and understanding of the information.

[1259] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1260] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotional state, means for visually displaying the fact-checking results, and means for adjusting feedback based on the user's emotional state. This enables the evaluation of the reliability of the information input by the user and the provision of appropriate feedback based on the user's emotional state. In this way, the psychological burden on the user can be reduced and the acceptance and understanding of information can be promoted.

[1261] A "user" is someone who inputs information into a system and receives the results.

[1262] A "public data source" refers to a collection of data that is widely made public to provide reliable information.

[1263] A "reliable database" is a database that stores information whose reliability has been verified.

[1264] "Tokenization" is the process of dividing an input text into its smallest units, such as words or phrases.

[1265] A "keyword" refers to a word or phrase that has particularly important meaning within a text or data.

[1266] A "fact-checking algorithm" refers to a set of computational procedures designed to evaluate the accuracy of input information.

[1267] An "emotion recognition engine" is a technology that analyzes a user's voice and facial expression data to identify their emotional state.

[1268] A "reliability score" is an indicator that quantifies the accuracy of information.

[1269] "Visually displaying" refers to providing users with data and information in a visual format, such as graphs and charts.

[1270] "Adjusting feedback" means changing the content and format of the feedback output based on the user's emotional state.

[1271] To implement this invention, a system is required in which a server, terminal, and user work together. This system includes modules for data collection, information analysis, fact-checking, sentiment recognition, and result display.

[1272] The server receives information entered by the user. For example, if a user enters a URL for a news article, the server packages that information in JSON format and saves it to storage. It also makes API calls to collect additional information from relevant public data sources and trusted databases. This allows for centralized management of necessary data.

[1273] Next, the server tokenizes the collected data using a natural language processing library. Specifically, it uses the Python nltk library to make it easier to divide sentences into words and phrases. Furthermore, it uses the scikit-learn library to extract keywords and important information using methods such as TF-IDF. This clarifies the central points of the information and improves the accuracy of fact-checking.

[1274] The fact-checking algorithm is executed on the server side. The server performs comparisons to calculate a reliability score based on the collected and analyzed data. It cross-references the acquired data with the input information to evaluate the accuracy of the information.

[1275] Furthermore, an emotion recognition engine is used to analyze the user's emotional state. The user provides voice input and facial expressions, and the device sends this data to the server. The server analyzes the emotional state using voice analysis and facial recognition technology, and adjusts the fact-check output based on the results. Here, the emotion_recognition library is used as the emotion recognition engine.

[1276] Finally, the server generates a report integrating fact-checking results and sentiment recognition results, and uses visualization tools to convert the results into graphs and charts. This allows users to intuitively understand the results and reduces the psychological burden of receiving feedback. Users can view detailed analysis results and sentiment-based feedback through their devices.

[1277] As a concrete example, let's explain how the news application "Emofact News" works. The user enters the URL of a news article and provides audio and facial expression data. On the server side, the input information is analyzed and a reliability score and emotional state are evaluated. Finally, the results are displayed visually, allowing the user to intuitively understand the reliability of the information and receive emotionally sensitive feedback.

[1278] An example of the generated prompt statement is as follows:

[1279] Please evaluate the reliability of the following article and the user's sentiment-based feedback. Visit "https: / / example.com / news / some-article", check the accuracy of the information, and display the results while taking into consideration any concerns or doubts the user may have.

[1280] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1281] Step 1:

[1282] User enters information

[1283] The user accesses the system interface and enters information such as the URL of a news article. The entered information is packaged in JSON format on the terminal and sent to the server.

[1284] Input: URL of the news article

[1285] Output: Information in JSON format

[1286] ---

[1287] Step 2:

[1288] Data collection

[1289] The server temporarily stores the JSON-formatted information received from the terminal in storage. Furthermore, it executes API calls to collect additional information from relevant public data sources and trusted databases.

[1290] Input: Information in JSON format

[1291] Output: Stored information and additionally collected data

[1292] ---

[1293] Step 3:

[1294] Information analysis

[1295] The server tokenizes the collected data using a natural language processing library (e.g., the nltk library). From the tokenized data, keywords and important information are extracted using TF-IDF.

[1296] Input: Collected data

[1297] Output: Tokenized data and extracted keywords

[1298] ---

[1299] Step 4:

[1300] Fact Check

[1301] The server runs fact-checking algorithms and calculates a reliability score by cross-referencing extracted keywords with data collected from publicly available data sources and trusted databases.

[1302] Input: Extracted keywords and collected additional information

[1303] Output: Reliability score

[1304] ---

[1305] Step 5:

[1306] emotion recognition

[1307] When entering information, the user provides voice input or facial expression data. The terminal sends this data to the server, which uses an emotion recognition engine to analyze the user's emotional state. For example, the emotion_recognition library can be used.

[1308] Input: Voice data or facial expression data

[1309] Output: Analyzed emotional state

[1310] ---

[1311] Step 6:

[1312] Adjustment and integration of results

[1313] The server integrates the reliability score and emotional state, and adjusts the feedback content according to the user's emotions. This generates appropriate feedback that reduces the user's psychological burden.

[1314] Input: Reliability score and analyzed emotional state

[1315] Output: Adjusted feedback content

[1316] ---

[1317] Step 7:

[1318] Visualization and display of results

[1319] The server converts the integrated information into graphs and charts using visualization tools. This allows the terminal to display the results in a format that is intuitively easy for the user to understand.

[1320] Input: Integrated result

[1321] Output: Visualized results and feedback

[1322] ---

[1323] Step 8:

[1324] Provide feedback

[1325] The device displays visualized results and refined feedback sent from the server to the user. The user reviews the results and receives feedback that is reliable and considerate of their feelings.

[1326] Input: Visualized results and feedback content

[1327] Output: Display to the user

[1328] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1329] The data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One 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">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1330] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1331] [Fourth Embodiment]

[1332] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1333] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1334] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1335] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1336] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1337] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1338] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1339] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1340] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1341] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1342] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1343] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1344] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1345] This invention is a system for determining the accuracy of information generated by generative AI, and is primarily designed for educational and public institutions. The system consists of a series of processes: user input, data collection from public data sources and reliable databases, information analysis, fact-checking, and a visual display of the results.

[1346] Program Overview

[1347] The program for this system consists of the following modules:

[1348] 1. Data Acquisition Module

[1349] 2. Information Analysis Module

[1350] 3. Fact-checking module

[1351] 4. Result Display Module

[1352] Program Processing Description

[1353] Data Acquisition Module

[1354] Users input information into the system. Educators and government officials provide the system with lesson materials, draft press releases, and other materials.

[1355] The terminal retrieves information from the user input form and sends it to the server. The information is sent to the API endpoint in JSON format.

[1356] The server receives the input information and stores it in temporary storage. This ensures consistency for each piece of information.

[1357] The server makes API calls to collect additional information from relevant public data sources and trusted databases (e.g., Wikipedia or government data).

[1358] Information analysis module

[1359] The server tokenizes the collected data using a natural language processing library (e.g., NLTK, spaCy). This divides the text into words and phrases.

[1360] The server uses technologies such as TF-IDF and Word2Vec to extract keywords and important information. This clarifies the central points of the information.

[1361] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows the context of the information content to be understood.

[1362] Fact-checking module

[1363] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it might use models such as random forests, SVMs, or neural networks.

[1364] The server compares the acquired data with user input information and calculates a reliability score. The numerical value, based on the scoring function, indicates the reliability of the information.

[1365] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[1366] Result display module

[1367] The server compiles the fact-checking results and generates a comprehensive report. The report includes evaluation results, relevant sources, and recommended corrections.

[1368] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[1369] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[1370] Specific example

[1371] Example 1: Use for educational institutions

[1372] The user (educator) inputs materials to be used in history lessons into the system.

[1373] The terminal sends the entered information to the server.

[1374] The server collects relevant information by making API calls to the relevant historical databases.

[1375] The server uses natural language processing techniques to tokenize information and extract keywords.

[1376] The server evaluates the accuracy of the information and calculates a reliability score based on fact-checking algorithms.

[1377] The server visualizes the results and displays them to the educator.

[1378] The device provides educators with accurate information sources and suggested revisions.

[1379] Example 2: Use by public institutions

[1380] The user (government official) enters a draft of a press release regarding the new policy.

[1381] The terminal sends information to the server.

[1382] The server collects additional information from trusted public databases.

[1383] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[1384] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[1385] The server visualizes the results in graphs and charts and presents them to government officials.

[1386] The terminal displays detailed analysis results and suggested corrections to the staff.

[1387] The above describes the specific embodiments for implementing the system of the present invention.

[1388] The following describes the processing flow.

[1389] Step 1:

[1390] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[1391] Step 2:

[1392] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[1393] Step 3:

[1394] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[1395] Step 4:

[1396] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[1397] Step 5:

[1398] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[1399] Step 6:

[1400] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[1401] Step 7:

[1402] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[1403] Step 8:

[1404] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[1405] Step 9:

[1406] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and suggested corrections.

[1407] Step 10:

[1408] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[1409] Step 11:

[1410] The terminal displays the generated report in the user interface. Users can view detailed analysis results and recommended corrective actions.

[1411] (Example 1)

[1412] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1413] Determining the accuracy of information generated by generative AI is a critical issue, particularly in educational and public institutions. However, current systems lack the means to automatically and efficiently evaluate the accuracy of user-inputted information. As a result, reliable fact-checking is not performed, and there is a risk of misinformation spreading. Furthermore, the analysis and visual display of collected data are insufficient, making it difficult for users to intuitively understand the information. An effective system is needed to address these challenges.

[1414] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1415] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for performing dependency analysis and syntactic analysis to analyze the context of the information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, and means for visually displaying the results of the fact-checking. This makes it possible to evaluate the accuracy of user-inputted information with high precision and provide reliable data. Furthermore, the visual display of the results realizes a system that is easy for users to understand intuitively.

[1416] "Information" refers to data that users input into the system, as well as data collected from relevant public data sources and trusted databases.

[1417] A "user" refers to an employee of an educational institution or public organization, or an individual who uses the system and is the entity that inputs information into the system.

[1418] "Public data sources" refer to information sources that are publicly available on the internet or in other accessible locations, and include, for example, encyclopedias and government statistics.

[1419] A "reliable database" refers to a source of information managed by a highly credible institution or organization. This includes academic journal databases and official statistical databases.

[1420] "Tokenization" refers to the process of dividing words and phrases in a text into individual elements, and it is a fundamental process in natural language processing.

[1421] A "keyword" refers to a word or phrase that is considered particularly important within information, and is used to grasp the subject or content of the information.

[1422] "Dependency analysis" is a technique for understanding sentence structure and meaning by analyzing the dependencies between words in a text.

[1423] "Syntactic analysis" is a method of analyzing the grammatical structure of a text and evaluating its meaning and grammatical accuracy.

[1424] "Fact-checking" refers to the process of evaluating the accuracy and truthfulness of information, and is the work of verifying the accuracy of information based on collected data.

[1425] A "fact-checking algorithm" refers to an algorithm used to computationally evaluate the accuracy of information, often employing machine learning models or statistical methods.

[1426] A "reliability score" is a numerical representation of the accuracy and reliability of information, calculated using an algorithm.

[1427] "Visual display" refers to presenting information and data to users in the form of graphs and charts using visualization tools, making it easier for users to intuitively understand the data.

[1428] This invention is a system composed of multiple modules for determining the accuracy of information generated by a generative AI. The specific method for realizing this system is described below.

[1429] System Overview

[1430] This system consists of the following modules:

[1431] 1. User Input Module

[1432] 2. Data Acquisition Module

[1433] 3. Information Analysis Module

[1434] 4. Fact-checking module

[1435] 5. Result Display Module

[1436] User input module

[1437] Users enter information through the system's input forms. For example, this could involve educators entering lesson materials or government officials entering draft policy documents.

[1438] The terminal retrieves information from the input form and sends it to the server in JSON format.

[1439] Data Acquisition Module

[1440] The server stores the information received from the user in temporary storage.

[1441] The server collects additional information from relevant public data sources and trusted databases. These public data sources include category information sites and public databases. For example, it uses the Wikipedia API and government public data APIs.

[1442] Information analysis module

[1443] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). Specifically, it divides sentences into words and phrases and analyzes their structure.

[1444] The server uses TF-IDF and Word2Vec technologies to extract keywords and important information, thereby clarifying the central points of the information.

[1445] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This allows it to understand the contextual meaning of the text.

[1446] Fact-checking module

[1447] The server evaluates the accuracy of the information using fact-checking algorithms (such as random forests, SVMs, and neural networks).

[1448] The server compares the collected data with user input information and calculates a reliability score. The reliability score is a numerical indicator of the accuracy of the information; for example, a score of 0.9 or higher is considered reliable.

[1449] The server determines whether the information is accurate or inaccurate based on the configured baseline score.

[1450] Result display module

[1451] The server compiles the fact-checking results into a comprehensive report. The report includes evaluation results, relevant sources, and recommended revisions.

[1452] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). This allows users to intuitively understand the results.

[1453] The terminal displays the generated report in the user interface. Specific data points and links to suggested corrections are also provided.

[1454] Specific example

[1455] Example 1: Use for educational institutions

[1456] The user (educator) inputs the material "Outline of Greek Mythology," which is used in history lessons, into the system.

[1457] The terminal sends the input information to the server.

[1458] The server collects and analyzes information by making API calls to relevant historical databases. For example, it might use "the content of Wikipedia articles on Greek mythology."

[1459] The server uses fact-checking algorithms to evaluate the accuracy of the information and calculates a reliability score.

[1460] The server visualizes the results and displays them to the educator.

[1461] The device provides educators with accurate information sources and suggested revisions.

[1462] Example 2: Use by public institutions

[1463] The user (government official) enters a draft of a press release regarding a new policy.

[1464] The terminal sends information to the server.

[1465] The server collects additional information from trusted public databases and performs analysis.

[1466] The server uses an algorithm to evaluate the accuracy of the information and calculates a reliability score.

[1467] The server visualizes the results in graphs and charts and presents them to government officials.

[1468] The terminal displays detailed analysis results and suggested corrections to the staff.

[1469] The above describes the embodiments for implementing the system of the present invention.

[1470] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1471] Step 1: Enter Information

[1472] The user enters information into the system's input form. For example, "An overview of Greek mythology."

[1473] The terminal retrieves the entered information and sends it to the server in JSON format. The input includes the title "Outline of Greek Mythology" and the body text.

[1474] The server stores the received information in temporary storage. This ensures consistency in the input information and facilitates subsequent processing.

[1475] Step 2: Data Collection

[1476] The server collects data from relevant public data sources and trusted databases based on information received from the user. For example, it retrieves data related to "Greek mythology" from the Wikipedia API and public databases.

[1477] The server temporarily stores the collected data in JSON format. Specifically, this includes information such as "details of Greek mythology" and "major characters."

[1478] The server synchronizes with publicly available data sources at regular intervals to reflect the latest information.

[1479] Step 3: Information Analysis

[1480] The server tokenizes the collected data using natural language processing libraries (NLTK, spaCy). Specifically, it divides sentences into words and phrases. The input "Zeus, the chief god of Greek mythology" is tokenized as ["Greek mythology", "chief god", "Zeus"].

[1481] The server uses TF-IDF and Word2Vec to extract keywords and important information. For example, it might extract keywords such as "Zeus" and "Olympian gods."

[1482] The server performs dependency analysis and syntactic analysis to analyze the relationships between keywords. This helps it understand the context of "Zeus is the king of the Olympian gods."

[1483] Step 4: Fact Check

[1484] The server uses fact-checking algorithms (such as random forests, SVMs, and neural networks) to evaluate the accuracy of the information. Specifically, it identifies discrepancies between the input information and the collected data.

[1485] The server calculates a reliability score based on the comparison results. The reliability score is displayed as a numerical value; for example, a score of 0.9 or higher is considered "high reliability."

[1486] The server determines whether the score meets the set criteria and assesses its accuracy. For example, it might determine that "Zeus is a god in Greek mythology" is correct, while "Zeus is a god in Roman mythology" is incorrect.

[1487] Step 5: Display Results

[1488] The server compiles the fact-checking results into a comprehensive report. The report includes the evaluation results, relevant sources, and recommended corrections.

[1489] The server converts the results into graphs and charts using visualization tools (D3.js, Chart.js). Confidence scores are visually displayed using bar graphs, etc.

[1490] The terminal displays the generated report in the user interface, including specific data points and links to suggested corrections.

[1491] The above outlines the specific processing steps of this system's program.

[1492] (Application Example 1)

[1493] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1494] Ensuring the reliability of information such as product details and user reviews is crucial for e-commerce websites. However, identifying misinformation and false reviews from a vast amount of data and providing only reliable information is not easy. Current systems lack the means to determine the accuracy of user-provided information in real time, which makes them prone to problems caused by misinformation.

[1495] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1496] In this invention, the server includes means for receiving information entered by a user, means for collecting additional information from relevant public data sources and trusted databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for visually displaying the results of the fact-checking, means for providing a user interface for inputting product information and user reviews, means for collecting product information from an official database, means for analyzing the content of the information using natural language processing technology, and means for calculating a reliability score and visually indicating reliability. This makes it possible to determine the accuracy of information on an e-commerce site in real time and provide users with only highly reliable information.

[1497] A "user" is an individual or organization that uses the system to input product information and user reviews, and to verify the accuracy of that information.

[1498] "Public data sources" refer to information sources that can be obtained from publicly available databases and websites.

[1499] A "reliable database" is a database containing highly reliable information provided by a public institution or certified organization.

[1500] "Tokenization of information" is the process of dividing text into words or phrases using natural language processing techniques.

[1501] "Keyword extraction" is the process of selecting important words and phrases from a text.

[1502] A "fact-checking algorithm" is a computational method or model used to compare input information with acquired information and evaluate its accuracy and reliability.

[1503] A "user interface" is an interactive screen or system through which a user inputs information and receives results.

[1504] An "official database" is a reliable database used to provide official information about a product.

[1505] "Natural language processing technology" refers to the technology that enables computers to understand, analyze, and process human language.

[1506] A "reliability score" is an indicator that numerically represents the accuracy of information, and it concretely expresses the evaluation results.

[1507] This invention is a system for determining the accuracy of generated information. This system is primarily used to determine the reliability of product information and user reviews on e-commerce websites.

[1508] The server first uses a means to receive information entered by the user. Users can enter product information and user reviews through the user interface of their smartphone. The server receives this input information and temporarily stores it.

[1509] Next, the server employs means to collect additional information from relevant public data sources and trusted databases. Collecting product information from official databases ensures reliable data.

[1510] The server then tokenizes the obtained data and employs methods to extract keywords and important information. Natural language processing techniques (e.g., Spacy) are used to divide the text into words and phrases, and extract the important parts.

[1511] The server then runs fact-checking algorithms to assess the accuracy of the information. These algorithms include a function to calculate a reliability score, utilizing techniques such as random forests, SVMs, and neural networks. The information provided by the user is compared with official data, and a numerical value is calculated based on the scoring function.

[1512] Finally, the server employs a means to visually display the fact-checking results. Using visualization tools (e.g., Matplotlib), it displays reliability scores as charts and graphs, providing users with an intuitively understandable format.

[1513] For example, if a user enters a product description such as "This is a test product description for a smartwatch," the server will collect information about smartwatches from the official database and extract key keywords using natural language processing technology. Next, a fact-checking algorithm will calculate a reliability score, and finally, the results will be displayed as a graph.

[1514] This makes it possible to verify the accuracy of information on e-commerce sites in real time and provide users with reliable information.

[1515] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1516] Step 1:

[1517] Users input product information and user reviews through their smartphone's user interface. The entered information is sent to the server in JSON format. For example, the input data might be the text "This is a test product description for a smartwatch." The server receives this input information and stores it temporarily.

[1518] Step 2:

[1519] The server collects additional information from relevant public data sources and trusted databases. This collection process uses REST API calls to access databases and retrieve official product information (e.g., product specifications and official reviews). The retrieved data is returned to the server in JSON format and temporarily stored. This ensures reliable data for comparison with user input.

[1520] Step 3:

[1521] The server tokenizes the stored data using natural language processing techniques. Specifically, it uses the Spacy library to split text into words and phrases. This process converts the input text into a list of words, including attribute information for each word (e.g., part of speech).

[1522] Step 4:

[1523] The server extracts keywords and important information from the tokenized data. Algorithms such as TF-IDF and Word2Vec are used to identify particularly important words and phrases within the text. This process yields a list of extracted keywords. For example, "smartwatch" and "product description" might be extracted as keywords.

[1524] Step 5:

[1525] The server uses fact-checking algorithms to evaluate the accuracy of the information. For example, it compares user input information with official data using random forests or neural network models. This algorithm evaluates the reliability of each keyword in the input information and calculates an overall reliability score. This score is output as a numerical value ranging from 0 to 1.

[1526] Step 6:

[1527] The server visually displays the fact-checking results based on the calculated reliability score. Visualization tools such as Matplotlib are used to convert the reliability score into charts and graphs. This display is then sent back to the smartphone user interface, presenting it in an intuitively understandable format. For example, a pie chart might show a reliability of 70%.

[1528] Step 7:

[1529] The user reviews the displayed fact-check results and corrects the input information as needed. The corrected information is sent back to the server for re-evaluation. This loop ensures that accurate and reliable information is ultimately provided to the user.

[1530] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1531] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[1532] Program Overview

[1533] The program for this system consists of the following modules:

[1534] 1. Data Acquisition Module

[1535] 2. Information Analysis Module

[1536] 3. Fact-checking module

[1537] 4. Emotion Recognition Module

[1538] 5. Result Display Module

[1539] Program Processing Description

[1540] Data Acquisition Module

[1541] The user logs into the system and enters the information they want to analyze. They enter text into the input form and click the submit button.

[1542] The terminal packages the received information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[1543] The server temporarily stores the received information in storage and makes API calls to retrieve additional information from relevant public data sources or trusted databases.

[1544] Information analysis module

[1545] The server tokenizes the collected data using a natural language processing library. This divides the text into words and phrases, facilitating data analysis.

[1546] The server extracts keywords and important information using TF-IDF and Word2Vec. This clarifies the central points of the information.

[1547] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. The context of the text is understood in detail.

[1548] Fact-checking module

[1549] The server runs fact-checking algorithms to assess the accuracy of the information. It calculates a reliability score and compares the retrieved data with the input information.

[1550] The server evaluates the accuracy of the information based on a reliability score.

[1551] Emotion recognition module

[1552] Users express emotions by providing voice input or facial expressions to the system. The system measures how users in educational and public institutions react emotionally while evaluating information.

[1553] The device acquires audio or video data and sends it to the server.

[1554] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[1555] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is feeling anxious or suspicious, the server adjusts how the results are displayed and the content of the feedback accordingly.

[1556] Result display module

[1557] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[1558] The server converts the results into graphs and charts using visualization tools, allowing users to intuitively understand the results.

[1559] The device displays the generated report in the user interface, allowing the user to view detailed analysis results and sentiment-based feedback.

[1560] Specific example

[1561] Example 1: Use for educational institutions

[1562] The user (educator) inputs materials to be used in history lessons into the system.

[1563] The terminal sends the entered information to the server.

[1564] The server collects relevant information by making API calls to the relevant historical databases.

[1565] The server uses natural language processing techniques to tokenize information and extract keywords.

[1566] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[1567] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[1568] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1569] The device displays the generated report and recommended feedback to the educator.

[1570] Example 2: Use by public institutions

[1571] The user (government official) enters a draft of a press release regarding the new policy.

[1572] The terminal sends information to the server.

[1573] The server collects additional information from trusted public databases.

[1574] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[1575] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[1576] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[1577] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1578] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[1579] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[1580] The following describes the processing flow.

[1581] Step 1:

[1582] The user logs into the system and enters the information they want to check. The user enters text into the input form and clicks the submit button.

[1583] Step 2:

[1584] The terminal packages the entered information in JSON format and sends it to the server. It then executes a POST request to the API endpoint.

[1585] Step 3:

[1586] The server receives the information it receives and stores it in temporary storage. This temporary storage is managed by a database, ensuring the consistency of the information in subsequent processing.

[1587] Step 4:

[1588] The server calls external APIs to retrieve additional information from relevant public data sources and trusted databases. For example, it might access Wikipedia or government databases to retrieve relevant data.

[1589] Step 5:

[1590] The server tokenizes the collected data using natural language processing libraries (e.g., NLTK, spaCy). This is done to break down sentences into words and phrases, improving the ease of data analysis.

[1591] Step 6:

[1592] The server uses techniques such as TF-IDF and Word2Vec to identify keywords and important information from the extracted tokens. This method identifies important words and phrases within the information.

[1593] Step 7:

[1594] The server performs dependency analysis and syntactic analysis, analyzing the relationships between keywords. This forms the basis for accurately understanding the context of the text and evaluating the accuracy of the information.

[1595] Step 8:

[1596] The server runs a fact-checking algorithm to assess the accuracy of the information. The algorithm calculates a confidence score for each data point and uses this to evaluate the accuracy of the information.

[1597] Step 9:

[1598] The server receives voice input or facial recognition data to recognize the user's emotions. The user can communicate their emotions to the system through voice or facial expressions.

[1599] Step 10:

[1600] The device sends the collected audio or video data to the server, which then provides data for sentiment analysis.

[1601] Step 11:

[1602] The server analyzes the user's emotions using an emotion recognition engine. It identifies the user's emotional state using voice analysis and facial recognition technology.

[1603] Step 12:

[1604] The server adjusts the output of the fact-checking algorithm based on the analyzed emotional state. For example, if the user is feeling anxious or suspicious, it flexibly changes how the results are displayed and the content of the feedback.

[1605] Step 13:

[1606] The server compiles the fact-checking results and sentiment recognition results into a comprehensive report. The report includes evaluation results, relevant sources, suggested revisions, and recommendation feedback based on sentiment analysis.

[1607] Step 14:

[1608] The server converts the results into graphs and charts using visualization tools (e.g., D3.js, Chart.js). This allows users to intuitively understand the results.

[1609] Step 15:

[1610] The device displays the generated report in the user interface. Users can view detailed analysis results and sentiment-based feedback.

[1611] (Example 2)

[1612] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1613] Traditionally, systems that judged the accuracy of information generated by generative AI lacked the ability to consider user emotions. As a result, evaluation results sometimes did not align with the user's emotional state, leading to problems in understanding and accepting the results. To address this issue, a system is needed that provides comprehensive feedback that considers user emotions in addition to accurate information judgment.

[1614] The identification processing performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotions, means for adjusting the output results of the fact-checking algorithm based on emotion recognition, and means for compiling the fact-checking results and emotion recognition results into a comprehensive report and displaying it visually. This makes it possible to evaluate the accuracy of the information and provide feedback that takes the user's emotions into consideration.

[1615] A "user" refers to the entity that uses this system to input information and receive the results.

[1616] "Public data sources" refer to reliable sources of information that are publicly available on the internet. Examples include encyclopedias and academic papers.

[1617] A "reliable database" refers to a collection of information whose accuracy and reliability are guaranteed. Examples include government statistical databases and databases from official research institutions.

[1618] "Tokenization" is a preprocessing step in natural language processing, referring to the process of dividing a text into words or phrases.

[1619] A "keyword" refers to a word or phrase that is considered particularly important from the input information or collected data.

[1620] A "fact-checking algorithm" refers to an algorithm used to evaluate the accuracy and reliability of information.

[1621] An "emotion recognition engine" refers to a technology that analyzes audio and video data to identify the user's emotional state (for example, joy, sadness, anger, etc.).

[1622] "Visual display" refers to the process of converting text information into visual formats such as graphs, charts, and diagrams for display.

[1623] A "reliability score" refers to an evaluation index that quantifies and shows the reliability of information.

[1624] This invention combines a system for determining the accuracy of information generated by generative AI with an emotion engine that recognizes user emotions. The system is intended for educational institutions and public organizations, aiming to evaluate the reliability of user-inputted information and provide emotion-sensitive feedback.

[1625] Program Overview

[1626] This system's program consists of the following modules:

[1627] 1. Data Acquisition Module

[1628] 2. Information Analysis Module

[1629] 3. Fact-checking module

[1630] 4. Emotion Recognition Module

[1631] 5. Result Display Module

[1632] Hardware and software to use

[1633] The following hardware and software are required to implement this system:

[1634] Server: A server that performs data processing and analysis.

[1635] Device: A device used by a user to input information (e.g., a PC or tablet).

[1636] Database: Storage for storing received data and additional information.

[1637] Natural language processing libraries: For example, NLTK and spaCy

[1638] Fact-checking tools: For example, FactCheck Tools API

[1639] Emotion recognition engines: For example, Google Cloud Speech-to-Text and Microsoft Azure Face API

[1640] Visualization tools: For example, D3.js and Chart.js

[1641] Specific example

[1642] Specific example 1: Use in educational institutions

[1643] The user (educator) inputs materials to be used in history lessons into the system. For example, they might input information such as, "The American Civil War ended in 1865."

[1644] The terminal sends the entered information to the server.

[1645] The server collects relevant information by making API calls to the relevant historical databases.

[1646] The server uses natural language processing techniques to tokenize information and extract keywords.

[1647] The server evaluates the accuracy of the information using fact-checking algorithms and calculates a reliability score.

[1648] The server analyzes the educator's voice input and facial expression data to check their emotional state.

[1649] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1650] The device displays the generated report and recommended feedback to the educator.

[1651] Specific example 2: Use by public institutions

[1652] The user (government official) enters a draft of a press release regarding a new policy. For example, they might enter information such as, "A new environmental policy will reduce carbon dioxide emissions by 30% by 2030."

[1653] The terminal sends information to the server.

[1654] The server collects additional information from trusted public databases.

[1655] The server analyzes the collected data, performing keyword extraction and contextual analysis.

[1656] The server evaluates the accuracy of the information using a fact-checking algorithm and calculates a reliability score.

[1657] The server analyzes the facial expressions and voice data of government officials to determine their emotional state.

[1658] The server adjusts how results are displayed and the content of feedback based on the emotional state.

[1659] The terminal displays detailed analysis results and emotionally sensitive feedback to the staff.

[1660] The above describes a specific embodiment of the system of the present invention that combines an emotion engine.

[1661] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1662] Step 1:

[1663] The user logs into the system. On the login screen, they enter their username and password and submit their authentication information. Input: Username, Password. Output: Authentication token.

[1664] Step 2:

[1665] The user enters the information to be evaluated using a text input form. For example, they might enter the contents of materials used in a history class into the text field. Input: Text data of the information to be evaluated. Output: Click event of the submit button.

[1666] Step 3:

[1667] The terminal retrieves the information entered by the user when they click the submit button. It then packages the retrieved information into JSON format and sends it to the server. Input: Text data. Output: Sending JSON data.

[1668] Step 4:

[1669] The server temporarily stores the received JSON data in storage. This storage uses a database or object storage. Input: JSON data. Output: Temporarily stored data.

[1670] Step 5:

[1671] The server makes API calls to public data sources and trusted databases to collect relevant information from the database. For example, it uses Wikidata and other reliable sources. Input: Temporarily stored data. Output: Additional data collected.

[1672] Step 6:

[1673] The server tokenizes the collected data using a natural language processing library (e.g., NLTK or spaCy). This divides the text into words and phrases. Input: Collected data. Output: Tokenized data.

[1674] Step 7:

[1675] The server extracts keywords from tokenized data using TF-IDF or Word2Vec. This identifies important information. Input: Tokenized data. Output: Keywords.

[1676] Step 8:

[1677] The server performs dependency analysis and syntactic analysis to analyze the relationships between the obtained keywords. For example, it uses the Stanford NLP dependency analysis model. Input: Keywords. Output: Relationships between keywords.

[1678] Step 9:

[1679] The server executes a fact-checking algorithm, comparing the acquired data with the input information to calculate a reliability score. Input: Relationships between keywords. Output: Reliability score.

[1680] Step 10:

[1681] The user provides audio or video input to the system, for example, using a webcam or microphone. Input: audio data, video data. Output: transmission of emotion recognition data.

[1682] Step 11:

[1683] The terminal acquires audio or video data and sends it to the server. Input: Audio data, video data. Output: Data transmission to the server.

[1684] Step 12:

[1685] The server analyzes the user's emotions using an emotion recognition engine (e.g., Google Cloud Speech-to-Text or Microsoft Azure Face API). Input: Audio data, video data. Output: Emotion analysis results.

[1686] Step 13:

[1687] The server adjusts the output of the fact-checking algorithm based on the user's emotional state. For example, if the user is expressing anxiety, it selects a way to present the results in a clearer manner. Input: Sentiment analysis results, reliability score. Output: Adjusted feedback content.

[1688] Step 14:

[1689] The server compiles the fact-checking and sentiment assessment results into a comprehensive report and converts them into graphs and charts using visualization tools (e.g., D3.js or Chart.js). Input: Adjusted feedback. Output: Visualized report.

[1690] Step 15:

[1691] The terminal displays the generated report in the user interface. On this screen, the user can view detailed analysis results and sentiment-based feedback. Input: Visualized report. Output: Displayed in the user interface.

[1692] The above outlines the specific processing steps of the system program.

[1693] (Application Example 2)

[1694] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1695] When determining the accuracy of information generated by generative AI, conventional systems have a problem in that they cannot provide feedback that takes user emotions into account. The feedback users receive is uniform, and for example, if a user feels anxious or doubtful, appropriate support is not provided. This raises concerns that it increases the psychological burden on users and hinders their acceptance and understanding of the information.

[1696] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1697] In this invention, the server includes means for receiving information input from a user, means for collecting additional information from relevant public data sources and reliable databases, means for tokenizing the obtained data and extracting keywords and important information, means for executing a fact-checking algorithm to evaluate the accuracy of the information, means for using an emotion recognition engine to analyze the user's emotional state, means for visually displaying the fact-checking results, and means for adjusting feedback based on the user's emotional state. This enables the evaluation of the reliability of the information input by the user and the provision of appropriate feedback based on the user's emotional state. In this way, the psychological burden on the user can be reduced and the acceptance and understanding of information can be promoted.

[1698] A "user" is someone who inputs information into a system and receives the results.

[1699] A "public data source" refers to a collection of data that is widely made public to provide reliable information.

[1700] A "reliable database" is a database that stores information whose reliability has been verified.

[1701] "Tokenization" is the process of dividing an input text into its smallest units, such as words or phrases.

[1702] A "keyword" refers to a word or phrase that has particularly important meaning within a text or data.

[1703] A "fact-checking algorithm" refers to a set of computational procedures designed to evaluate the accuracy of input information.

[1704] An "emotion recognition engine" is a technology that analyzes a user's voice and facial expression data to identify their emotional state.

[1705] A "reliability score" is an indicator that quantifies the accuracy of information.

[1706] "Visually displaying" refers to providing users with data and information in a visual format, such as graphs and charts.

[1707] "Adjusting feedback" means changing the content and format of the feedback output based on the user's emotional state.

[1708] To implement this invention, a system is required in which a server, terminal, and user work together. This system includes modules for data collection, information analysis, fact-checking, sentiment recognition, and result display.

[1709] The server receives information entered by the user. For example, if a user enters a URL for a news article, the server packages that information in JSON format and saves it to storage. It also makes API calls to collect additional information from relevant public data sources and trusted databases. This allows for centralized management of necessary data.

[1710] Next, the server tokenizes the collected data using a natural language processing library. Specifically, it uses the Python nltk library to make it easier to divide sentences into words and phrases. Furthermore, it uses the scikit-learn library to extract keywords and important information using methods such as TF-IDF. This clarifies the central points of the information and improves the accuracy of fact-checking.

[1711] The fact-checking algorithm is executed on the server side. The server performs comparisons to calculate a reliability score based on the collected and analyzed data. It cross-references the acquired data with the input information to evaluate the accuracy of the information.

[1712] Furthermore, an emotion recognition engine is used to analyze the user's emotional state. The user provides voice input and facial expressions, and the device sends this data to the server. The server analyzes the emotional state using voice analysis and facial recognition technology, and adjusts the fact-check output based on the results. Here, the emotion_recognition library is used as the emotion recognition engine.

[1713] Finally, the server generates a report integrating fact-checking results and sentiment recognition results, and uses visualization tools to convert the results into graphs and charts. This allows users to intuitively understand the results and reduces the psychological burden of receiving feedback. Users can view detailed analysis results and sentiment-based feedback through their devices.

[1714] As a concrete example, let's explain how the news application "Emofact News" works. The user enters the URL of a news article and provides audio and facial expression data. On the server side, the input information is analyzed and a reliability score and emotional state are evaluated. Finally, the results are displayed visually, allowing the user to intuitively understand the reliability of the information and receive emotionally sensitive feedback.

[1715] An example of the generated prompt statement is as follows:

[1716] Please evaluate the reliability of the following article and the user's sentiment-based feedback. Visit "https: / / example.com / news / some-article", check the accuracy of the information, and display the results while taking into consideration any concerns or doubts the user may have.

[1717] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1718] Step 1:

[1719] User enters information

[1720] The user accesses the system interface and enters information such as the URL of a news article. The entered information is packaged in JSON format on the terminal and sent to the server.

[1721] Input: URL of the news article

[1722] Output: Information in JSON format

[1723] ---

[1724] Step 2:

[1725] Data collection

[1726] The server temporarily stores the JSON-formatted information received from the terminal in storage. Furthermore, it executes API calls to collect additional information from relevant public data sources and trusted databases.

[1727] Input: Information in JSON format

[1728] Output: Stored information and additionally collected data

[1729] ---

[1730] Step 3:

[1731] Information analysis

[1732] The server tokenizes the collected data using a natural language processing library (e.g., the nltk library). From the tokenized data, keywords and important information are extracted using TF-IDF.

[1733] Input: Collected data

[1734] Output: Tokenized data and extracted keywords

[1735] ---

[1736] Step 4:

[1737] Fact Check

[1738] The server runs fact-checking algorithms and calculates a reliability score by cross-referencing extracted keywords with data collected from publicly available data sources and trusted databases.

[1739] Input: Extracted keywords and collected additional information

[1740] Output: Reliability score

[1741] ---

[1742] Step 5:

[1743] emotion recognition

[1744] When entering information, the user provides voice input or facial expression data. The terminal sends this data to the server, which uses an emotion recognition engine to analyze the user's emotional state. For example, the emotion_recognition library can be used.

[1745] Input: Voice data or facial expression data

[1746] Output: Analyzed emotional state

[1747] ---

[1748] Step 6:

[1749] Adjustment and integration of results

[1750] The server integrates the reliability score and emotional state, and adjusts the feedback content according to the user's emotions. This generates appropriate feedback that reduces the user's psychological burden.

[1751] Input: Reliability score and analyzed emotional state

[1752] Output: Adjusted feedback content

[1753] ---

[1754] Step 7:

[1755] Visualization and display of results

[1756] The server converts the integrated information into graphs and charts using visualization tools. This allows the terminal to display the results in a format that is intuitively easy for the user to understand.

[1757] Input: Integrated result

[1758] Output: Visualized results and feedback

[1759] ---

[1760] Step 8:

[1761] Provide feedback

[1762] The device displays visualized results and refined feedback sent from the server to the user. The user reviews the results and receives feedback that is reliable and considerate of their feelings.

[1763] Input: Visualized results and feedback content

[1764] Output: Display to the user

[1765] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1766] The data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One 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">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1767] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1768] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1769] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1770] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1771] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1772] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1773] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1774] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1775] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1776] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1777] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1778] 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.

[1779] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1780] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1781] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1782] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1783] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1784] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1785] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[1786] The following is further disclosed regarding the embodiments described above.

[1787] (Claim 1)

[1788] A system for determining the accuracy of generated information,

[1789] A means of receiving information entered by the user,

[1790] Means for collecting additional information from relevant public data sources and trusted databases,

[1791] A method for tokenizing the obtained data and extracting keywords and important information,

[1792] A means of executing fact-checking algorithms to evaluate the accuracy of information,

[1793] A means of visually displaying the results of fact-checking,

[1794] A system that includes this.

[1795] (Claim 2)

[1796] The system according to claim 1, wherein the fact-checking algorithm includes means for calculating a reliability score.

[1797] (Claim 3)

[1798] The system according to claim 1, wherein the tokenization of the aforementioned information and keyword extraction are performed using natural language processing techniques.

[1799] (Claim 4)

[1800] The system according to claim 1, wherein the visual display includes means for displaying the results in graph or chart format.

[1801] "Example 1"

[1802] (Claim 1)

[1803] A system for determining the accuracy of generated information,

[1804] A means of receiving information entered by the user,

[1805] Means for collecting additional information from relevant public data sources and trusted databases,

[1806] A method for tokenizing the obtained data and extracting keywords and important information,

[1807] Means for performing dependency analysis and syntactic analysis in order to analyze the context of information,

[1808] A means of executing fact-checking algorithms to evaluate the accuracy of information,

[1809] A means of visually displaying the results of fact-checking,

[1810] A system that includes this.

[1811] (Claim 2)

[1812] The system according to claim 1, wherein the fact-checking algorithm includes means for calculating a reliability score.

[1813] (Claim 3)

[1814] The system according to claim 1, wherein the tokenization of the aforementioned information and keyword extraction are performed using natural language processing techniques.

[1815] "Application Example 1"

[1816] (Claim 1)

[1817] A system for determining the accuracy of generated information,

[1818] A means of receiving information entered by the user,

[1819] Means for collecting additional information from relevant public data sources and trusted databases,

[1820] A method for tokenizing the obtained data and extracting keywords and important information,

[1821] A means of executing fact-checking algorithms to evaluate the accuracy of information,

[1822] A means of visually displaying the results of fact-checking,

[1823] A means of providing a user interface for entering product information and user reviews,

[1824] Methods for collecting product information from the official database,

[1825] A means for analyzing the content of information using natural language processing technology,

[1826] A means of calculating a reliability score and visually demonstrating reliability,

[1827] A system that includes this.

[1828] (Claim 2)

[1829] The system according to claim 1, wherein the fact-checking algorithm includes means for calculating a reliability score.

[1830] (Claim 3)

[1831] The system according to claim 1, wherein the tokenization of the aforementioned information and keyword extraction are performed using natural language processing techniques.

[1832] "Example 2 of combining an emotion engine"

[1833] (Claim 1)

[1834] A means of receiving information entered by the user,

[1835] Means for collecting additional information from relevant public data sources and trusted databases,

[1836] A method for tokenizing the obtained data and extracting keywords and important information,

[1837] A means of executing fact-checking algorithms to evaluate the accuracy of information,

[1838] A means of using an emotion recognition engine to analyze the user's emotions,

[1839] A means of adjusting the output of a fact-checking algorithm based on sentiment recognition,

[1840] A means of compiling fact-checking results and sentiment recognition results into a comprehensive report and displaying it visually,

[1841] A system that includes this.

[1842] (Claim 2)

[1843] The system according to claim 1, wherein the fact-checking algorithm includes means for calculating a reliability score.

[1844] (Claim 3)

[1845] The system according to claim 1, wherein the tokenization of the aforementioned information and keyword extraction are performed using natural language processing techniques.

[1846] "Application example 2 when combining with an emotional engine"

[1847] (Claim 1)

[1848] A system for determining the accuracy of generated information,

[1849] A means of receiving information entered by the user,

[1850] Means for collecting additional information from relevant public data sources and trusted databases,

[1851] A method for tokenizing the obtained data and extracting keywords and important information,

[1852] A means of executing fact-checking algorithms to evaluate the accuracy of information,

[1853] A method using an emotion recognition engine to analyze the user's emotional state,

[1854] A means of visually displaying the results of fact-checking,

[1855] A means of adjusting feedback based on the user's emotional state,

[1856] A system that includes this.

[1857] (Claim 2)

[1858] The system according to claim 1, wherein the fact-checking algorithm includes means for calculating a reliability score.

[1859] (Claim 3)

[1860] The system according to claim 1, wherein the tokenization of the aforementioned information and keyword extraction are performed using natural language processing techniques. [Explanation of symbols]

[1861] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A system for determining the accuracy of generated information, A means of receiving information entered by the user, Means for collecting additional information from relevant public data sources and trusted databases, A method for tokenizing the obtained data and extracting keywords and important information, A means of executing fact-checking algorithms to evaluate the accuracy of information, A means of visually displaying the results of fact-checking, A system that includes this.

2. The system according to claim 1, wherein the fact-checking algorithm includes means for calculating a reliability score.

3. The system according to claim 1, wherein the tokenization of the aforementioned information and keyword extraction are performed using natural language processing techniques.

4. The system according to claim 1, wherein the visual display includes means for displaying the results in graph or chart format.

Images