system

Abstract

To provide a system that enables efficient news distribution with high user satisfaction. [Solution] A system including a news gathering means, a means for classifying the gathered news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for delivering the generated news titles and summaries to the user's terminal, a means for accepting a request for providing detailed information about a specific news item, a means for obtaining detailed information about the specified news item and regenerating it into a user-friendly format using generative artificial intelligence to provide it, and a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias towards a specific category.

Description

[Technical Field]

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

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

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

[0004] Today's businesspeople are busy and need to quickly collect and organize a large amount of information every day. However, there are currently few ways to efficiently access a wide range of news information and obtain only the information that best suits their interests. Furthermore, bias in information sources can lead to a risk of relying solely on specific categories, potentially resulting in decision-making based on a biased perspective. Another challenge is the difficulty of quickly obtaining detailed information about specific news items and digging deeper. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by providing a system including a news collection means, a means for classifying the collected news by topic and category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for delivering the generated news titles and summaries to a user terminal, a means for accepting a request for detailed information on a specific news item, a means for acquiring detailed information on the specified news item and regenerating it in a user-friendly format using generative artificial intelligence, and a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias toward a specific category. The system also includes a means for collecting user feedback regarding the content and personalization of the news distribution and a means for retraining the generative artificial intelligence based on the collected feedback, thereby improving the accuracy of the system and user satisfaction. Furthermore, the system includes a means for automatically calling the API of a news provider at a user-specified time interval to collect the latest news articles, enabling the system to quickly provide the latest information.

[0006] "News gathering means" refers to a means for obtaining the latest news data from news sources.

[0007] A "classification method by topic or category" is a method for classifying collected news articles into specific topics or categories.

[0008] "Generative AI" is AI that has the ability to automatically generate text and information based on input data.

[0009] The "means for generating titles and summaries" is a means for generating titles and summaries from collected news articles.

[0010] The "means for creating a personalized news distribution list based on the user's past browsing history" is a means for creating a personalized news distribution list based on the user's past news browsing history.

[0011] The "means for delivering the generated news titles and summaries to the user terminal" refers to a means for transmitting the generated news titles and summaries to the user's device.

[0012] The "means for accepting a request for providing detailed information" is a means for accepting a request for detailed information from a user.

[0013] "Means for acquiring detailed information and regenerating it into a user-friendly format using generative artificial intelligence for provision" refers to means for collecting detailed information about specified news, generating it into an easy-to-understand format using generative artificial intelligence for provision.

[0014] "Means for analyzing browsing history and generating and sending alerts if there is a bias in a category" refers to a means for periodically analyzing a user's news browsing history and generating and notifying a warning if there is a bias in a specific category.

[0015] "Means for collecting feedback" refers to means for collecting feedback and opinions from users.

[0016] "Means for retraining generative AI" refers to means for retraining AI to improve its accuracy and functionality based on collected feedback.

[0017] "Means for automatically calling the API of a news provider to collect the latest news articles" refers to a means for automatically calling the API of a news provider at specified time intervals to obtain the latest news data.

[0018] "User-friendly format" refers to providing information in an easy-to-understand format that users can easily understand. [Brief explanation of the drawings]

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

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

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

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

[0023] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

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

[0025] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0026] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0027] [First embodiment]

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

[0029] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

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

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

[0032] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0033] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0034] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

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

[0036] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

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

[0038] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

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

[0040] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[0041] First, the server periodically calls the news provider's API to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses the received news data and classifies it by topic or category (for example, "business," "technology," "sports," etc.).

[0042] Next, the server uses generative artificial intelligence (AI) to generate titles and summaries for each news article based on the collected news data. The AI ​​uses the collected news headlines and text as inputs to automatically generate titles and summaries that are easy for users to provide. These generated titles and summaries are then stored in a database by the server.

[0043] The server analyzes the user's past news browsing history and creates a personalized news delivery list based on this. It collects each user's browsing history data and identifies their interests in specific categories and topics. This analysis makes it possible to deliver news optimized for each user. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[0044] When a user wants to know more information about a particular news article, the device sends a request for more information to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[0045] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[0046] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[0047] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will immediately respond to the request and provide the details. Furthermore, if the user only views business news, the server will send an alert to encourage the user to also view news in the "technology" and "sports" categories, encouraging the user to gather information from multiple angles. In this way, the system of the present invention helps users gather information more efficiently and gain a more diverse perspective.

[0048] The processing flow will be explained below.

[0049] Step 1:

[0050] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[0051] Step 2:

[0052] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[0053] Step 3:

[0054] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[0055] Step 4:

[0056] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[0057] Step 5:

[0058] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[0059] Step 6:

[0060] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[0061] Step 7:

[0062] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[0063] Step 8:

[0064] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[0065] Step 9:

[0066] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[0067] Step 10:

[0068] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[0069] Step 11:

[0070] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[0071] Step 12:

[0072] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[0073] Step 13:

[0074] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[0075] Step 14:

[0076] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[0077] Step 15:

[0078] The server generates alerts and sends them to the user's device, delivering them as push notifications or in-app messages.

[0079] Step 16:

[0080] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[0081] Step 17:

[0082] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[0083] Step 18:

[0084] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[0085] Example 1

[0086] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0087] Conventional news distribution systems often require manual processes for news collection, classification, and personalized distribution, resulting in a time-consuming and labor-intensive process. Furthermore, users may only view news from specific categories, and the systems lack appropriate alert functions to address this. Furthermore, the systems lack a mechanism for retraining generative AI based on feedback, which hinders user satisfaction.

[0088] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0089] In this invention, the server includes a news collection means, a means for classifying the collected news by topic and category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a user terminal, a means for accepting a request for detailed information on a specific news item, a means for acquiring detailed information on the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias in a specific category, and a means for collecting user feedback and retraining. This automates the process from news collection to personalized distribution, provision of detailed information, generation of alerts for category bias, and retraining of AI based on feedback, enabling efficient news distribution with high user satisfaction.

[0090] A "news gathering method" is a method for periodically calling the API of a news provider to gather the latest news articles.

[0091] "Means for classifying collected news by topic or category" refers to a means that has the function of parsing collected news data and classifying each news article by topic or category (e.g., "business," "technology," "sports," etc.).

[0092] "Means for generating titles and summaries using generative artificial intelligence" refers to means for automatically generating titles and summaries from the headlines and body text of news articles using a generative AI model.

[0093] "Means for creating a personalized news distribution list based on a user's past browsing history" refers to means for analyzing a user's past news browsing history and creating a news list optimized for each user based on the results.

[0094] "Means for delivering the generated news titles and summaries to the user's device" refers to means for sending the generated news titles and summaries to the user's device via push notifications or in-app messages.

[0095] The "means for accepting a request for providing detailed information on a specific news item" refers to a means for receiving a request when a user wants to know detailed information on a specific news item.

[0096] "Means for obtaining detailed information about specified news items and regenerating and providing it in a user-friendly format using generative AI" refers to means for obtaining detailed information about news items in response to a request and regenerating and providing it in a user-friendly format using a generative AI model.

[0097] "Means for analyzing a user's news browsing history and generating and sending an alert if there is a bias towards a specific category" refers to a means for periodically analyzing a user's news browsing history, generating an alert if there is a bias towards a specific category, and sending the alert to the user's device.

[0098] "Means for collecting user feedback and retraining" refers to a means for collecting feedback from users regarding the content of news distribution and the degree of personalization, and retraining the generative artificial intelligence based on that feedback.

[0099] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[0100] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses this data and classifies each news article by topic or category (e.g., "business," "technology," "sports," etc.). This operation uses techniques such as perlject and field analysis.

[0101] Next, the server uses generative artificial intelligence (such as GPT-4 (registered trademark)) to generate titles and summaries for each news article based on the collected news data. The generative AI model uses the collected news headlines and text as inputs to automatically generate titles and summaries that are easy to provide to users. These generated titles and summaries are then stored in a database by the server.

[0102] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on this. Browsing history data is collected from databases such as MySQL (registered trademark) to identify interests in specific categories and topics. As a result, news distribution optimized for each user becomes possible. The personalized news list is sent from the server to the device and delivered to the user as a push notification or in-app message.

[0103] When a user wants to know more information about a particular news article, the device sends a detailed information request to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[0104] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[0105] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[0106] For example, if a businessperson receives the latest news at 7 a.m. and requests to know more about a particular business news item, the server will immediately respond to the request and provide more detailed information. Furthermore, if the person is only browsing business news, the server will send an alert to encourage them to also browse news in the "technology" and "sports" categories, encouraging them to gather information from a variety of angles.

[0107] The following are some examples of prompts that can be input to a generative AI model:

[0108] Write a summary of a news article that begins with "How new AI technologies will impact business." The news article should include the following:

[0109] [Enter news article text here]

[0110] Abstract length should be no more than 150 characters.

[0111] This system allows users to efficiently gather the latest news information and gain a multifaceted perspective.

[0112] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0113] Step 1:

[0114] The server periodically calls the news provider's API to collect the latest news articles. It uses the API endpoint and authentication token as input. The output is news data in JSON format. Specifically, it sends an HTTP request and receives a list of news articles as a response.

[0115] Step 2:

[0116] The server parses the collected news data and classifies it by topic and category. It uses the acquired JSON-formatted news data as input. The output is news data classified by category. Specifically, it parses the JSON data and sorts each news article into a category such as "business," "technology," or "sports."

[0117] Step 3:

[0118] The server uses generative artificial intelligence to generate a title and summary for each news article. It uses classified news data as input. The output is the generated title and summary. Specifically, it passes the headline and body of a news article to the AI ​​model and obtains the generated title and summary.

[0119] Step 4:

[0120] The server saves the title and summary in a database. It uses the generated title and summary as input. The output is news data saved in the database. The specific operation is to save the title and summary in the database together with the news data of the corresponding category.

[0121] Step 5:

[0122] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It uses the user's browsing history data as input, and the output is a personalized news list. Specifically, it extracts browsing history from a database and analyzes interest in specific categories and topics to create a news list.

[0123] Step 6:

[0124] The server delivers a personalized news list to the device. As input, it uses the personalized news list and the user's device token. The output is the news list sent to the user's device. Specific behavior is to send the news list as a push notification or in-app message.

[0125] Step 7:

[0126] When a device wants to know more information about a particular news article, it sends a request for more information to the server. The input is the ID of the requested news article. The output is receiving the requested information. In concrete terms, the user selects a news article and sends a request to the server.

[0127] Step 8:

[0128] The server receives the request and calls the news provider's API again to retrieve details about the specified news article. The input is the news article ID and API endpoint. The output is the retrieved details. The specific operation is to make an API call and receive the details.

[0129] Step 9:

[0130] The server uses generative artificial intelligence to regenerate the details into a user-friendly format and send it to the device. It uses the retrieved details as input. The output is the regenerated details in a user-friendly format. Specifically, it passes the details to the AI ​​model, retrieves the generated summary and clarification, and sends them.

[0131] Step 10:

[0132] The server periodically analyzes the user's news browsing history and generates and sends an alert if there is a bias in a particular category. The input is the user's browsing history data. The output is the generated alert message. Specifically, it extracts the browsing history from the database, analyzes the bias in categories, creates an alert message, and sends it to the device.

[0133] Step 11:

[0134] The server collects user feedback and retrains the generative AI based on that feedback. It uses user feedback as input and the output is a retrained AI model. Specifically, it collects feedback data and uses it as new training data to retrain the AI ​​model.

[0135] (Application example 1)

[0136] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0137] In modern society, users need to gather news from many sources, but in the process, they tend to be biased toward certain categories, making it difficult to gather information from multiple perspectives. Furthermore, systems that efficiently gather news and provide it to users in a personalized format are still insufficient. In particular, little consideration has been given to news distribution within virtual stores, leaving users with limited means to obtain important information in real time while shopping. To solve these issues, a personalized news distribution system is needed.

[0138] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0139] In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a virtual store user interface, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, and a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias toward a specific category, thereby enabling real-time news distribution and personalization in the virtual store.

[0140] "News gathering means" refers to a means of regularly gathering the latest news data using the API of a news provider.

[0141] "Means of classifying by topic or category" refers to a means of organizing collected news data into categories such as "business," "technology," and "sports."

[0142] "Generative AI" is an AI technology that automatically generates titles and summaries from collected news headlines and text.

[0143] The "means for creating a personalized news distribution list" is a means for analyzing a user's past news browsing history and generating an optimal news distribution list based on the user's interests.

[0144] A "virtual store user interface" is an interface device (such as smart glasses or a head-mounted display) that allows users to display and manipulate information within a virtual store.

[0145] The "means for delivering news titles and summaries" is a means for transmitting the generated news titles and summaries to a user terminal.

[0146] The "means for accepting a request for providing detailed information" is a means for accepting a request when a user wants to know detailed information about a particular news article.

[0147] "Means for regenerating using generative artificial intelligence" refers to artificial intelligence technology for regenerating the acquired detailed information in a format that is easy for the user to understand.

[0148] The "means for analyzing news browsing history" is a means for analyzing a user's past news browsing status and determining whether there is a bias toward a particular category.

[0149] "Means for generating and sending alerts" refers to the means for generating and sending alerts to the user interface when it is determined that there is a bias in a particular news category, encouraging the user to gather information from multiple angles.

[0150] "Means for collecting feedback from users" refers to means for collecting users' opinions and impressions regarding the content of news distribution and the degree of personalization.

[0151] The "means for retraining generative artificial intelligence" refers to a means for retraining generative artificial intelligence to improve its accuracy based on feedback data collected from users.

[0152] The "means for collecting the latest news articles" is a means for automatically calling the API of a news provider at a time interval specified by the user to collect the latest news articles.

[0153] This invention is a system that uses the API of news providers to collect the latest news articles, personalize them, and deliver them to a virtual store user interface. Below, we will explain the specific processing performed by the server and the hardware and software used.

[0154] Hardware and Software Overview

[0155] 1. Server: Communicates with news provider APIs to collect news data, parses and stores the news data, runs generative AI models, creates news distribution lists for each user, and collects user feedback.

[0156] 2. User interface device: Smart glasses or head-mounted display used in the virtual store. Users receive news feeds through this device.

[0157] 3. Generative AI model: Automatically generates titles and summaries for news data. The AI ​​model receives prompts to generate titles and summaries based on the latest news articles and generates responses.

[0158] 4. Database: Stores collected news data, generated titles and summaries, user browsing history, and feedback data.

[0159] Specific examples of processing

[0160] News gathering

[0161] The server periodically calls the news provider's API to collect the latest news data, allowing the server to always maintain the latest information.

[0162] Parsing and categorizing news

[0163] The acquired news data is generally provided in structured data format such as JSON. The server parses this data and organizes it into categories such as "business," "technology," and "sports."

[0164] Title and Abstract Generation

[0165] A generative AI model generates user-friendly titles and summaries from collected news data. Here are some example prompts:

[0166] "Create a summary of the latest business news."

[0167] This produces the following summary:

[0168] News: (Business Trends)

[0169] Summary: A Japanese company has announced plans to expand into a new market, which is expected to have a lot of potential.

[0170] Personalized news delivery

[0171] The server analyzes the user's news browsing history and creates an optimal news distribution list for each user based on that information. This list is then distributed to the user interface within the virtual store.

[0172] Providing more information

[0173] When a user wants to know more information about a particular news article, the user interface device sends a request to the server, which then calls the news provider's API again to obtain more information about the specified news article, regenerates it in a user-friendly format using the generative AI model, and provides it to the user.

[0174] Alerts for bias in news categories

[0175] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, sending it to the user interface, making it easier for users to gather information from multiple angles.

[0176] In this way, the system of the present invention provides news gathering, personalization, real-time distribution, detailed information provision, and alert functions for multifaceted information gathering in an integrated manner, allowing users to efficiently gather information even within a virtual store.

[0177] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0178] Step 1:

[0179] The server calls the news provider's API to collect the latest news data. The input is the API URL and any necessary authentication information, and the output is the news data in JSON format. Specifically, a regularly scheduled task sends requests to the news API and saves the retrieved JSON data.

[0180] Step 2:

[0181] The server parses the collected news data and classifies it by category. The input is news data in JSON format, and the output is a news list for each category. Specifically, it parses the JSON data and sorts it into lists based on the category information of each news article.

[0182] Step 3:

[0183] The server uses a generative AI model to generate a title and summary for each news article. The input is the organized news list, and the output is the generated title and summary. Specifically, for each news article, the server inputs the following prompt sentence into the generative AI model:

[0184] "Create a summary of the latest business news."

[0185] The generative AI model generates a summary as a response.

[0186] Step 4:

[0187] The server stores the generated title and summary in a database. The input is the generated title and summary, and the output is the data stored in the database. Specifically, the server associates the title and summary with the news article ID and category information and stores them in the database.

[0188] Step 5:

[0189] The server analyzes the user's past news browsing history and creates a personalized news distribution list. The input is the user's browsing history data, and the output is a personalized news distribution list. Specifically, it uses a machine learning algorithm to predict the user's interests and select the most suitable news articles.

[0190] Step 6:

[0191] The server delivers the generated news titles and summaries to the virtual store user interface. The input is a personalized news distribution list, and the output is the news displayed on the user interface. Specifically, the server sends the list contents to smart glasses or a head-mounted display, which notifies the user.

[0192] Step 7:

[0193] When a user wants to know more information about a particular news article, they send a request for more information from their device to the server. The input is the news article ID, and the output is detailed data from the news source. Specifically, the user operates the interface and clicks the "More Information" button for a particular article.

[0194] Step 8:

[0195] After receiving the request, the server calls the news provider's API again to obtain detailed information about the specified news article. The input is the news article ID, API URL, and authentication information, and the output is the detailed data of the news article. Specifically, the server makes the appropriate API call to obtain the detailed information.

[0196] Step 9:

[0197] The server uses the generative AI model to regenerate the acquired detailed information into a user-friendly format. The input is the acquired detailed information, and the output is the regenerated detailed information. Specifically, the acquired detailed information is input back into the generative AI model and reconstructed into an easy-to-understand format.

[0198] Step 10:

[0199] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, and sends it to the user interface. The input is news browsing history data, and the output is an alert message. Specifically, it uses a machine learning algorithm to analyze the history, and generates and sends an alert to users who only browse a particular category, such as "Please try browsing articles in other categories."

[0200] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0201] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by combining it with an emotion engine, it recognizes the user's emotions, and based on that information, it enhances the degree of personalization of the news delivery list and adjusts the content and format of alerts to address biases in news categories.

[0202] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles. This allows information such as the title, body, and category of each news article to be extracted.

[0203] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[0204] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[0205] One of the features of this invention is the incorporation of an emotion engine. The emotion engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[0206] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[0207] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[0208] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing the emotion engine makes it possible to deliver news that responds appropriately to the user's emotional state.

[0209] The system of the present invention makes users' information gathering more efficient, supports them in gaining a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[0210] The processing flow will be explained below.

[0211] Step 1:

[0212] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[0213] Step 2:

[0214] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[0215] Step 3:

[0216] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[0217] Step 4:

[0218] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[0219] Step 5:

[0220] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[0221] Step 6:

[0222] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[0223] Step 7:

[0224] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[0225] Step 8:

[0226] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[0227] Step 9:

[0228] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[0229] Step 10:

[0230] The emotion engine analyzes users' reactions to news articles, and the reaction data is provided by the device as facial expression recognition and interaction data.

[0231] Step 11:

[0232] The emotion engine recognizes the user's emotional state based on the reaction data it acquires, which can be categorized into categories such as stress, joy, and excitement.

[0233] Step 12:

[0234] The server updates the user profile with the recognized emotion data and uses that information to further personalize the news distribution list, selecting news articles according to the user's emotional state.

[0235] Step 13:

[0236] The server then transmits a personalized news distribution list based on the emotion engine back to the user's terminal, the distribution including news optimized based on the emotional state.

[0237] Step 14:

[0238] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[0239] Step 15:

[0240] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[0241] Step 16:

[0242] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[0243] Step 17:

[0244] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[0245] Step 18:

[0246] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[0247] Step 19:

[0248] The server generates alerts and sends them to the user's device. Alerts are delivered as push notifications or in-app messages. The content and format of the alerts are adjusted according to the user's emotional state, as recognized by the emotion engine.

[0249] Step 20:

[0250] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[0251] Step 21:

[0252] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[0253] Step 22:

[0254] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[0255] Example 2

[0256] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0257] Currently, many users gather news online, but personalized news delivery is not yet fully implemented. Furthermore, news recommendations do not take into account the emotional state of individual users, which can lead to information overload or bias. As a result, users feel stressed or dissatisfied with the news content. Furthermore, when users want to know more about a news item, they must manually search for additional information, which is time-consuming. There is a need to solve these issues and realize a more efficient and user-friendly news delivery system.

[0258] The identification process 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 a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to the user terminal, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for processing the user's reactions while browsing the news using an emotion analysis engine and adjusting the news distribution list based on the user's emotional state, and a means for analyzing the user's news browsing history and generating and sending an alert based on the emotion analysis results if there is a bias toward a specific category. This not only personalizes the user's news browsing experience, but also enables flexible news distribution that takes the user's emotional state into consideration.

[0259] A "news gathering means" is a mechanism for periodically calling the interface of a news provider to obtain the latest news articles.

[0260] "Classification method" refers to the algorithm or machine learning model used to sort collected news articles into specific topics or categories.

[0261] "Generative AI" refers to AI that uses natural language processing technology to automatically generate titles and summaries of news articles.

[0262] A "personalized news delivery list" refers to a list of news articles organized in an optimized manner based on a user's past browsing history and interests.

[0263] "User Terminal" means a device (e.g., a smartphone or tablet) used by a User to receive news articles and notifications.

[0264] A "means for accepting requests for more information" is a mechanism for receiving and processing user requests for additional information about a particular news article.

[0265] "Means of regenerating into a user-friendly format" refers to a mechanism that uses generative artificial intelligence to convert news articles containing academic or specialized information into a form that is easy for the general user to understand.

[0266] "Emotion analysis engine" refers to technology that analyzes reaction data (e.g., facial expressions and operational behavior) when a user browses the news and identifies their emotional state.

[0267] The "means for generating and sending alerts" refers to a mechanism for generating alerts that notify users of bias in specific categories based on their news viewing history and sentiment analysis results, and sending the alerts to the users' devices.

[0268] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by incorporating a sentiment analysis engine, it recognizes user emotions and uses that information to enhance the degree of personalization of news distribution lists and adjust the content and format of alerts to address biases in news categories. A specific embodiment for implementing this system is described below.

[0269] News gathering

[0270] First, the server periodically calls the interface of the news provider to collect the latest news articles. This collection process is performed using, for example, a news API. The collected news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles.

[0271] Category Classification and Generative AI

[0272] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[0273] Personalized News Generation

[0274] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[0275] Utilizing a sentiment analysis engine

[0276] One of the features of this invention is the incorporation of an emotion analysis engine. The emotion analysis engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[0277] Providing more information

[0278] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[0279] Alert function

[0280] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion analysis engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[0281] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion analysis engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing an emotion analysis engine makes it possible to deliver news that appropriately responds to the user's emotional state.

[0282] Example prompt sentence:

[0283] "Analyze the news browsing history of user A and identify areas of interest. Also, generate a news distribution list for when user A has recently been feeling stressed."

[0284] In this way, the system of the present invention makes the user's information gathering more efficient, supports the user in having a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[0285] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0286] Step 1:

[0287] The server periodically calls the interface of the news provider. For example, it can be configured to issue a request to the news API every hour. It specifies the URL endpoint to use for the call and sends the request with any necessary parameters (e.g., API key, country code, etc.).

[0288] Input: API endpoint URL, API key, request parameters

[0289] Output: News data in JSON format

[0290] Specifically, it sends an HTTP GET request to a URL such as https: / / newsapi.org / v2 / top-headlines?country=jp&apiKey=YOUR_API_KEY.

[0291] Step 2:

[0292] The server parses the received JSON-formatted news data to extract individual news articles. It then uses a JSON library to analyze the received data and obtain the title, body, and category of each news article.

[0293] Input: News data in JSON format

[0294] Output: List of news articles (title, body, category)

[0295] Specifically, it uses code such as json.loads(response.text) to convert the JSON data into a dictionary format and extract the required data fields.

[0296] Step 3:

[0297] The server categorizes the collected news articles by topic or category, and based on a predefined list of categories (e.g., business, technology, entertainment), it assigns the news articles to the corresponding category.

[0298] Input: List of news articles (title, body, category)

[0299] Output: A list of news articles sorted by category

[0300] Specifically, it classifies news articles into categories using conditional statements such as if article["category"] == "business":

[0301] Step 4:

[0302] The server uses a generative device to generate titles and summaries for news articles. It uses a generative AI model (e.g., GPT-3 (registered trademark)) to automatically generate optimal titles and summaries from the text of news articles.

[0303] Input: News article body

[0304] Output: Generated title and summary

[0305] Specifically, it calls functions such as gpt3.generate_summary(article["content"]) to obtain the generated results from the AI ​​model.

[0306] Step 5:

[0307] The server stores the generated title and abstract in a database. It executes an INSERT query to the database to store data such as title, abstract, body, and category.

[0308] Input: Generated title and summary, news article body, category

[0309] Output: News article information stored in a database

[0310] Specifically, it executes SQL statements such as INSERT INTO news_articles (title, summary, content, category) VALUES (...).

[0311] Step 6:

[0312] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It retrieves the user's browsing data from a database and uses analytical algorithms to extract personal preferences.

[0313] Input: User news browsing history data

[0314] Output: A personalized news distribution list

[0315] Specifically, the system aggregates data such as the number of category views per user and time periods, and calculates a priority score.

[0316] Step 7:

[0317] The server sends the created news distribution list to the device and delivers it to the user as a push notification or in-app message. The distribution list is packaged in JSON format and delivered using a messaging service (e.g., Firebase Cloud Messaging).

[0318] Input: Personalized news distribution list

[0319] Output: Notification to user terminal

[0320] Specifically, message sending code such as fcm.send_message(user_device_token, news_list_json) is used.

[0321] Step 8:

[0322] The device sends the user's reactions to the news to the emotion analysis engine, including facial expression recognition results and device touch data.

[0323] Input: User reaction data when browsing news

[0324] Output: Emotion analysis results

[0325] Specifically, the camera captures facial expressions and sends the images to an emotion analysis API.

[0326] Step 9:

[0327] The server processes the transmitted emotional data using an emotion analysis engine. The emotion analysis engine is used to identify the user's emotional state. The analysis results (e.g., stress level) are used to influence the next news delivery.

[0328] Input: User reaction data, results of sentiment analysis engine

[0329] Output: News distribution list based on emotional state

[0330] Specifically, it calls the sentiment analysis API and obtains the results.

[0331] Step 10:

[0332] Based on the analysis results, the server selects appropriate news articles and recommends them to the user. It also regenerates a news distribution list based on the emotional state and sends it to the device.

[0333] Input: Sentiment analysis results, news article list

[0334] Output: News distribution list based on emotional state

[0335] Specifically, the system generates and transmits a personalized news list again.

[0336] Step 11:

[0337] A user requests more information about a particular news item. Through the device interface, the user sends a request for more information about the particular article.

[0338] Input: News article ID

[0339] Output: Request status

[0340] Specifically, the request is sent when the "View details" button is clicked.

[0341] Step 12:

[0342] The server calls the news API again to get more information, which is then regenerated into a user-friendly format using generative artificial intelligence.

[0343] Input: News article ID

[0344] Output: Detailed information generated

[0345] Specifically, the API is called by adding the news article ID as a query parameter, and functions such as gpt3.generate_readable_format(article_detail_content) are used.

[0346] Step 13:

[0347] The server sends the regenerated details to the device for the user to view, and packages the generated details in JSON format and sends it.

[0348] Input: Generated details

[0349] Output: Notification to terminal

[0350] Specific operations include code such as send_to_device(user_device_token, detailed_article_json).

[0351] Step 14:

[0352] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category. It combines a sentiment analysis engine to adjust the content and format of the alert according to the user's emotional state.

[0353] Input: User news browsing history, sentiment analysis results

[0354] Output: Bias alert

[0355] Specifically, it compares the viewing frequency of each category on a weekly basis and generates an alert if there is a bias.

[0356] Step 15:

[0357] The server sends the generated alert to the device to alert the user, and packages the alert content in JSON format and sends it to the device.

[0358] Input: Bias Alert

[0359] Output: Alert notification to terminal

[0360] Specific actions include using codes such as send_alert(user_device_token, alert_message_json).

[0361] (Application example 2)

[0362] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0363] In today's world, many users get their information from online news, but this presents several challenges. First, it is often impossible to efficiently gather the information that is truly important to users from the vast amount of news articles. Second, it is difficult to personalize news based on the user's interests and emotions. In particular, if a user is feeling stressed or has a bias toward certain categories, appropriate news delivery may not be provided. Furthermore, there is a lack of a mechanism for utilizing user feedback to improve the system.

[0364] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, and a means for generating titles and summaries using generative artificial intelligence. This enables efficient processing of collected news articles and personalized news delivery according to the user's needs. The server also includes an emotion engine that recognizes the user's emotional state and a means for dynamically adjusting the news delivery list based on the recognized emotion, a means for regenerating and providing news in a user-friendly format using generative artificial intelligence, and a means for accepting requests for detailed information on specific news. This configuration enables flexible and effective news delivery based on the user's emotions and interests, improving the user's information gathering experience.

[0365] A "news gathering means" is a means for gathering the latest news articles from news sources.

[0366] "Topic or category classification means" refers to a means for automatically classifying collected news articles into specific topics or categories, such as business, technology, entertainment, etc.

[0367] "Generative AI" is an AI technology used to automatically generate titles and summaries from collected news articles.

[0368] The "means for creating a personalized news distribution list" is a means for creating a news distribution list optimized for an individual user based on the user's past browsing history.

[0369] The "means for delivering to a user terminal" refers to a means for delivering the generated news title and summary to a device such as a user's smartphone or computer.

[0370] The "means for accepting a request for providing detailed information" is a means for accepting a request from a user for detailed information about a specific news article.

[0371] "Means for regenerating and providing in a user-friendly format" refers to means for obtaining detailed information about specified news and regenerating it in a user-friendly format using generative artificial intelligence.

[0372] "Means for generating and sending alerts" refers to means for analyzing a user's news browsing history, generating an alert if there is a bias in a specific category, and notifying the user.

[0373] The "emotion engine" is an engine that recognizes the user's emotional state (e.g., stress or frustration) and adjusts the content of news distribution lists and alerts based on that data.

[0374] A system embodying the invention includes the following major elements:

[0375] 1. News collection and classification

[0376] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The retrieved news articles are then classified by topic or category. For example, categories such as business, technology, and entertainment are set.

[0377] 2. Title and Abstract Generation Using Generative AI

[0378] The server uses generative artificial intelligence (AI) to automatically generate titles and summaries from collected news articles, which are then stored in a database.

[0379] 3. Create a personalized news distribution list

[0380] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on that information. The generated news distribution list is sent to the user's device as a push notification or in-app message.

[0381] 4. Emotional Engine Adjustment

[0382] The server is equipped with an emotion engine that analyzes users' reactions while browsing the news (e.g., facial expression recognition and device interaction data). This allows the server to recognize the user's emotional state and recommend the most appropriate news articles for each individual user. For example, if a user is feeling stressed, it will prioritize the delivery of relaxing entertainment articles.

[0383] 5. Requesting and Providing More Information

[0384] When a user wants to check the details of a specific article, they send a request for more information from their device. Upon receiving this request, the server calls the news API again to retrieve the details, and then uses generative artificial intelligence to regenerate them into a user-friendly format. The regenerated details are then sent to the device, where they can be viewed by the user.

[0385] 6. Generate and send alerts

[0386] The server periodically analyzes a user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function also incorporates an emotion engine, which adjusts the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotion into consideration.

[0387] Hardware and software used

[0388] Emotion recognition: Keras (TENSORFLOW(R) backend), OpenCV

[0389] News gathering: requests library, news API

[0390] Data storage and management: Any database (e.g., PostgreSQL)

[0391] Emotion engine: Generative AI models (e.g. GPT-3)

[0392] Specific examples

[0393] A user launches a news app on their smartphone and browses recent news. If the emotion engine analyzes the user's facial expressions and recognizes that they are feeling stressed, the server automatically prioritizes relaxing entertainment news. If the user requests more information about a particular business news item, the server instantly generates and provides the detailed information.

[0394] Prompt Sentence Examples

[0395] When users are feeling stressed, introduce them to today's relaxing entertainment news.

[0396] This will realize a news distribution system that recommends appropriate news and products based on the user's emotions and history, and provides personalized feedback.

[0397] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0398] Step 1:

[0399] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The input is the response data from the news API, and the output is a list of extracted news articles.

[0400] Step 2:

[0401] The server classifies the collected news articles by topic and category. The input is the list of news articles obtained in the previous step, and the output is the news articles classified by category. Keyword analysis and natural language processing techniques are used for classification.

[0402] Step 3:

[0403] The server uses generative artificial intelligence to generate titles and summaries from news articles. The generated titles and summaries are stored in a database. The input is the news articles classified in step 2, and the output is the generated title and summary for each article. A generative artificial intelligence model (e.g., GPT-3) performs the text generation.

[0404] Step 4:

[0405] The server analyzes the user's past browsing history and creates a personalized news distribution list. The input is the user's browsing history data and the news articles obtained in step 3, and the output is a personalized news distribution list. A machine learning algorithm is used for the analysis.

[0406] Step 5:

[0407] The server uses an emotion engine to recognize the user's emotional state in real time. The input is the user's facial expression recognition data and interaction data, and the output is the user's emotional state information. Keras models and OpenCV are used for emotion recognition.

[0408] Step 6:

[0409] The server dynamically adjusts the news distribution list based on the user's emotional state. The input is the emotional state information obtained in step 5 and the news distribution list created in step 4, and the output is the adjusted news distribution list. The list priority is changed depending on the emotional state.

[0410] Step 7:

[0411] The device delivers the personalized news feed list to the user as push notifications or in-app messages. The input is the news feed list adjusted in step 6, and the output is the news articles displayed on the user's device.

[0412] Step 8:

[0413] When a user requests detailed information about a specific article, the terminal sends a detailed information request to the server. The input is the user's request for detailed information, and the output is the request being sent to the server.

[0414] Step 9:

[0415] When the server receives the request for more information, it calls the news API again to obtain the details and regenerates them in a user-friendly format using generative artificial intelligence. The input is the request for more information and the response data from the news API, and the output is the regenerated details.

[0416] Step 10:

[0417] The terminal displays the details received from the server to the user. The input is the details regenerated in step 9, and the output is the details displayed on the terminal.

[0418] Step 11:

[0419] The server periodically analyzes the user's news browsing history and generates and sends alerts if there is a bias towards a specific category. The input is the user's news browsing history data and current browsing data, and the output is the generated alert. An algorithm is used for the analysis, and the content and format of the alert reflects an emotional engine.

[0420] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0421] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0422] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0423] [Second embodiment]

[0424] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0425] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0426] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0427] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0428] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0429] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0430] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0431] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0432] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0433] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0434] In the smart glasses 214, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0435] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0436] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[0437] First, the server periodically calls the news provider's API to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses the received news data and classifies it by topic or category (for example, "business," "technology," "sports," etc.).

[0438] Next, the server uses generative artificial intelligence (AI) to generate titles and summaries for each news article based on the collected news data. The AI ​​uses the collected news headlines and text as inputs to automatically generate titles and summaries that are easy for users to provide. These generated titles and summaries are then stored in a database by the server.

[0439] The server analyzes the user's past news browsing history and creates a personalized news delivery list based on this. It collects each user's browsing history data and identifies their interests in specific categories and topics. This analysis makes it possible to deliver news optimized for each user. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[0440] When a user wants to know more information about a particular news article, the device sends a request for more information to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[0441] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[0442] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[0443] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will immediately respond to the request and provide the details. Furthermore, if the user only views business news, the server will send an alert to encourage the user to also view news in the "technology" and "sports" categories, encouraging the user to gather information from multiple angles. In this way, the system of the present invention helps users gather information more efficiently and gain a more diverse perspective.

[0444] The processing flow will be explained below.

[0445] Step 1:

[0446] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[0447] Step 2:

[0448] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[0449] Step 3:

[0450] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[0451] Step 4:

[0452] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[0453] Step 5:

[0454] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[0455] Step 6:

[0456] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[0457] Step 7:

[0458] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[0459] Step 8:

[0460] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[0461] Step 9:

[0462] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[0463] Step 10:

[0464] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[0465] Step 11:

[0466] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[0467] Step 12:

[0468] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[0469] Step 13:

[0470] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[0471] Step 14:

[0472] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[0473] Step 15:

[0474] The server generates alerts and sends them to the user's device, delivering them as push notifications or in-app messages.

[0475] Step 16:

[0476] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[0477] Step 17:

[0478] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[0479] Step 18:

[0480] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[0481] Example 1

[0482] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0483] Conventional news distribution systems often require manual processes for news collection, classification, and personalized distribution, resulting in a time-consuming and labor-intensive process. Furthermore, users may only view news from specific categories, and the systems lack appropriate alert functions to address this. Furthermore, the systems lack a mechanism for retraining generative AI based on feedback, which hinders user satisfaction.

[0484] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0485] In this invention, the server includes a news collection means, a means for classifying the collected news by topic and category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a user terminal, a means for accepting a request for detailed information on a specific news item, a means for acquiring detailed information on the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias in a specific category, and a means for collecting user feedback and retraining. This automates the process from news collection to personalized distribution, provision of detailed information, generation of alerts for category bias, and retraining of AI based on feedback, enabling efficient news distribution with high user satisfaction.

[0486] A "news gathering method" is a method for periodically calling the API of a news provider to gather the latest news articles.

[0487] "Means for classifying collected news by topic or category" refers to a means that has the function of parsing collected news data and classifying each news article by topic or category (e.g., "business," "technology," "sports," etc.).

[0488] "Means for generating titles and summaries using generative artificial intelligence" refers to means for automatically generating titles and summaries from the headlines and body text of news articles using a generative AI model.

[0489] "Means for creating a personalized news distribution list based on a user's past browsing history" refers to means for analyzing a user's past news browsing history and creating a news list optimized for each user based on the results.

[0490] "Means for delivering the generated news titles and summaries to the user's device" refers to means for sending the generated news titles and summaries to the user's device via push notifications or in-app messages.

[0491] The "means for accepting a request for providing detailed information on a specific news item" refers to a means for receiving a request when a user wants to know detailed information on a specific news item.

[0492] "Means for obtaining detailed information about specified news items and regenerating and providing it in a user-friendly format using generative AI" refers to means for obtaining detailed information about news items in response to a request and regenerating and providing it in a user-friendly format using a generative AI model.

[0493] "Means for analyzing a user's news browsing history and generating and sending an alert if there is a bias towards a specific category" refers to a means for periodically analyzing a user's news browsing history, generating an alert if there is a bias towards a specific category, and sending the alert to the user's device.

[0494] "Means for collecting user feedback and retraining" refers to a means for collecting feedback from users regarding the content of news distribution and the degree of personalization, and retraining the generative artificial intelligence based on that feedback.

[0495] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[0496] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses this data and classifies each news article by topic or category (e.g., "business," "technology," "sports," etc.). This operation uses techniques such as perlject and field analysis.

[0497] Next, the server uses a generative AI model (such as GPT-4) to generate a title and summary for each news article based on the collected news data. The generative AI model uses the collected news headlines and text as inputs to automatically generate a title and summary that is easy to provide to users. The generated title and summary are then stored in a database by the server.

[0498] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on this. Browsing history data is collected from databases such as MySQL to identify interests in specific categories and topics. As a result, news distribution optimized for each user becomes possible. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[0499] When a user wants to know more information about a particular news article, the device sends a detailed information request to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[0500] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[0501] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[0502] For example, if a businessperson receives the latest news at 7 a.m. and requests to know more about a particular business news item, the server will immediately respond to the request and provide more detailed information. Furthermore, if the person is only browsing business news, the server will send an alert to encourage them to also browse news in the "technology" and "sports" categories, encouraging them to gather information from a variety of angles.

[0503] The following are some examples of prompts that can be input to a generative AI model:

[0504] Write a summary of a news article that begins with "How new AI technologies will impact business." The news article should include the following:

[0505] [Enter news article text here]

[0506] Abstract length should be no more than 150 characters.

[0507] This system allows users to efficiently gather the latest news information and gain a multifaceted perspective.

[0508] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0509] Step 1:

[0510] The server periodically calls the news provider's API to collect the latest news articles. It uses the API endpoint and authentication token as input. The output is news data in JSON format. Specifically, it sends an HTTP request and receives a list of news articles as a response.

[0511] Step 2:

[0512] The server parses the collected news data and classifies it by topic and category. It uses the acquired JSON-formatted news data as input. The output is news data classified by category. Specifically, it parses the JSON data and sorts each news article into a category such as "business," "technology," or "sports."

[0513] Step 3:

[0514] The server uses generative artificial intelligence to generate a title and summary for each news article. It uses classified news data as input. The output is the generated title and summary. Specifically, it passes the headline and body of a news article to the AI ​​model and obtains the generated title and summary.

[0515] Step 4:

[0516] The server saves the title and summary in a database. It uses the generated title and summary as input. The output is news data saved in the database. The specific operation is to save the title and summary in the database together with the news data of the corresponding category.

[0517] Step 5:

[0518] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It uses the user's browsing history data as input, and the output is a personalized news list. Specifically, it extracts browsing history from a database and analyzes interest in specific categories and topics to create a news list.

[0519] Step 6:

[0520] The server delivers a personalized news list to the device. As input, it uses the personalized news list and the user's device token. The output is the news list sent to the user's device. Specific behavior is to send the news list as a push notification or in-app message.

[0521] Step 7:

[0522] When a device wants to know more information about a particular news article, it sends a request for more information to the server. The input is the ID of the requested news article. The output is receiving the requested information. In concrete terms, the user selects a news article and sends a request to the server.

[0523] Step 8:

[0524] The server receives the request and calls the news provider's API again to retrieve details about the specified news article. The input is the news article ID and API endpoint. The output is the retrieved details. The specific operation is to make an API call and receive the details.

[0525] Step 9:

[0526] The server uses generative artificial intelligence to regenerate the details into a user-friendly format and send it to the device. It uses the retrieved details as input. The output is the regenerated details in a user-friendly format. Specifically, it passes the details to the AI ​​model, retrieves the generated summary and clarification, and sends them.

[0527] Step 10:

[0528] The server periodically analyzes the user's news browsing history and generates and sends an alert if there is a bias in a particular category. The input is the user's browsing history data. The output is the generated alert message. Specifically, it extracts the browsing history from the database, analyzes the bias in categories, creates an alert message, and sends it to the device.

[0529] Step 11:

[0530] The server collects user feedback and retrains the generative AI based on that feedback. It uses user feedback as input and the output is a retrained AI model. Specifically, it collects feedback data and uses it as new training data to retrain the AI ​​model.

[0531] (Application example 1)

[0532] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0533] In modern society, users need to gather news from many sources, but in the process, they tend to be biased toward certain categories, making it difficult to gather information from multiple perspectives. Furthermore, systems that efficiently gather news and provide it to users in a personalized format are still insufficient. In particular, little consideration has been given to news distribution within virtual stores, leaving users with limited means to obtain important information in real time while shopping. To solve these issues, a personalized news distribution system is needed.

[0534] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0535] In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a virtual store user interface, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, and a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias toward a specific category, thereby enabling real-time news distribution and personalization in the virtual store.

[0536] "News gathering means" refers to a means of regularly gathering the latest news data using the API of a news provider.

[0537] "Means of classifying by topic or category" refers to a means of organizing collected news data into categories such as "business," "technology," and "sports."

[0538] "Generative AI" is an AI technology that automatically generates titles and summaries from collected news headlines and text.

[0539] The "means for creating a personalized news distribution list" is a means for analyzing a user's past news browsing history and generating an optimal news distribution list based on the user's interests.

[0540] A "virtual store user interface" is an interface device (such as smart glasses or a head-mounted display) that allows users to display and manipulate information within a virtual store.

[0541] The "means for delivering news titles and summaries" is a means for transmitting the generated news titles and summaries to a user terminal.

[0542] The "means for accepting a request for providing detailed information" is a means for accepting a request when a user wants to know detailed information about a particular news article.

[0543] "Means for regenerating using generative artificial intelligence" refers to artificial intelligence technology for regenerating the acquired detailed information in a format that is easy for the user to understand.

[0544] The "means for analyzing news browsing history" is a means for analyzing a user's past news browsing status and determining whether there is a bias toward a particular category.

[0545] "Means for generating and sending alerts" refers to the means for generating and sending alerts to the user interface when it is determined that there is a bias in a particular news category, encouraging the user to gather information from multiple angles.

[0546] "Means for collecting feedback from users" refers to means for collecting users' opinions and impressions regarding the content of news distribution and the degree of personalization.

[0547] The "means for retraining generative artificial intelligence" refers to a means for retraining generative artificial intelligence to improve its accuracy based on feedback data collected from users.

[0548] The "means for collecting the latest news articles" is a means for automatically calling the API of a news provider at a time interval specified by the user to collect the latest news articles.

[0549] This invention is a system that uses the API of news providers to collect the latest news articles, personalize them, and deliver them to a virtual store user interface. Below, we will explain the specific processing performed by the server and the hardware and software used.

[0550] Hardware and Software Overview

[0551] 1. Server: Communicates with news provider APIs to collect news data, parses and stores the news data, runs generative AI models, creates news distribution lists for each user, and collects user feedback.

[0552] 2. User interface device: Smart glasses or head-mounted display used in the virtual store. Users receive news feeds through this device.

[0553] 3. Generative AI model: Automatically generates titles and summaries for news data. The AI ​​model receives prompts to generate titles and summaries based on the latest news articles and generates responses.

[0554] 4. Database: Stores collected news data, generated titles and summaries, user browsing history, and feedback data.

[0555] Specific examples of processing

[0556] News gathering

[0557] The server periodically calls the news provider's API to collect the latest news data, allowing the server to always maintain the latest information.

[0558] Parsing and categorizing news

[0559] The acquired news data is generally provided in structured data format such as JSON. The server parses this data and organizes it into categories such as "business," "technology," and "sports."

[0560] Title and Abstract Generation

[0561] A generative AI model generates user-friendly titles and summaries from collected news data. Here are some example prompts:

[0562] "Create a summary of the latest business news."

[0563] This produces the following summary:

[0564] News: (Business Trends)

[0565] Summary: A Japanese company has announced plans to expand into a new market, which is expected to have a lot of potential.

[0566] Personalized news delivery

[0567] The server analyzes the user's news browsing history and creates an optimal news distribution list for each user based on that information. This list is then distributed to the user interface within the virtual store.

[0568] Providing more information

[0569] When a user wants to know more information about a particular news article, the user interface device sends a request to the server, which then calls the news provider's API again to obtain more information about the specified news article, regenerates it in a user-friendly format using the generative AI model, and provides it to the user.

[0570] Alerts for bias in news categories

[0571] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, sending it to the user interface, making it easier for users to gather information from multiple angles.

[0572] In this way, the system of the present invention provides news gathering, personalization, real-time distribution, detailed information provision, and alert functions for multifaceted information gathering in an integrated manner, allowing users to efficiently gather information even within a virtual store.

[0573] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0574] Step 1:

[0575] The server calls the news provider's API to collect the latest news data. The input is the API URL and any necessary authentication information, and the output is the news data in JSON format. Specifically, a regularly scheduled task sends requests to the news API and saves the retrieved JSON data.

[0576] Step 2:

[0577] The server parses the collected news data and classifies it by category. The input is news data in JSON format, and the output is a news list for each category. Specifically, it parses the JSON data and sorts it into lists based on the category information of each news article.

[0578] Step 3:

[0579] The server uses a generative AI model to generate a title and summary for each news article. The input is the organized news list, and the output is the generated title and summary. Specifically, for each news article, the server inputs the following prompt sentence into the generative AI model:

[0580] "Create a summary of the latest business news."

[0581] The generative AI model generates a summary as a response.

[0582] Step 4:

[0583] The server stores the generated title and summary in a database. The input is the generated title and summary, and the output is the data stored in the database. Specifically, the server associates the title and summary with the news article ID and category information and stores them in the database.

[0584] Step 5:

[0585] The server analyzes the user's past news browsing history and creates a personalized news distribution list. The input is the user's browsing history data, and the output is a personalized news distribution list. Specifically, it uses a machine learning algorithm to predict the user's interests and select the most suitable news articles.

[0586] Step 6:

[0587] The server delivers the generated news titles and summaries to the virtual store user interface. The input is a personalized news distribution list, and the output is the news displayed on the user interface. Specifically, the server sends the list contents to smart glasses or a head-mounted display, which notifies the user.

[0588] Step 7:

[0589] When a user wants to know more information about a particular news article, they send a request for more information from their device to the server. The input is the news article ID, and the output is detailed data from the news source. Specifically, the user operates the interface and clicks the "More Information" button for a particular article.

[0590] Step 8:

[0591] After receiving the request, the server calls the news provider's API again to obtain detailed information about the specified news article. The input is the news article ID, API URL, and authentication information, and the output is the detailed data of the news article. Specifically, the server makes the appropriate API call to obtain the detailed information.

[0592] Step 9:

[0593] The server uses the generative AI model to regenerate the acquired detailed information into a user-friendly format. The input is the acquired detailed information, and the output is the regenerated detailed information. Specifically, the acquired detailed information is input back into the generative AI model and reconstructed into an easy-to-understand format.

[0594] Step 10:

[0595] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, and sends it to the user interface. The input is news browsing history data, and the output is an alert message. Specifically, it uses a machine learning algorithm to analyze the history, and generates and sends an alert to users who only browse a particular category, such as "Please try browsing articles in other categories."

[0596] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0597] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by combining it with an emotion engine, it recognizes the user's emotions, and based on that information, it enhances the degree of personalization of the news delivery list and adjusts the content and format of alerts to address biases in news categories.

[0598] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles. This allows information such as the title, body, and category of each news article to be extracted.

[0599] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[0600] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[0601] One of the features of this invention is the incorporation of an emotion engine. The emotion engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[0602] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[0603] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[0604] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing the emotion engine makes it possible to deliver news that responds appropriately to the user's emotional state.

[0605] The system of the present invention makes users' information gathering more efficient, supports them in gaining a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[0606] The processing flow will be explained below.

[0607] Step 1:

[0608] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[0609] Step 2:

[0610] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[0611] Step 3:

[0612] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[0613] Step 4:

[0614] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[0615] Step 5:

[0616] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[0617] Step 6:

[0618] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[0619] Step 7:

[0620] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[0621] Step 8:

[0622] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[0623] Step 9:

[0624] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[0625] Step 10:

[0626] The emotion engine analyzes users' reactions to news articles, and the reaction data is provided by the device as facial expression recognition and interaction data.

[0627] Step 11:

[0628] The emotion engine recognizes the user's emotional state based on the reaction data it acquires, which can be categorized into categories such as stress, joy, and excitement.

[0629] Step 12:

[0630] The server updates the user profile with the recognized emotion data and uses that information to further personalize the news distribution list, selecting news articles according to the user's emotional state.

[0631] Step 13:

[0632] The server then transmits a personalized news distribution list based on the emotion engine back to the user's terminal, the distribution including news optimized based on the emotional state.

[0633] Step 14:

[0634] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[0635] Step 15:

[0636] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[0637] Step 16:

[0638] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[0639] Step 17:

[0640] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[0641] Step 18:

[0642] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[0643] Step 19:

[0644] The server generates alerts and sends them to the user's device. Alerts are delivered as push notifications or in-app messages. The content and format of the alerts are adjusted according to the user's emotional state, as recognized by the emotion engine.

[0645] Step 20:

[0646] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[0647] Step 21:

[0648] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[0649] Step 22:

[0650] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[0651] Example 2

[0652] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0653] Currently, many users gather news online, but personalized news delivery is not yet fully implemented. Furthermore, news recommendations do not take into account the emotional state of individual users, which can lead to information overload or bias. As a result, users feel stressed or dissatisfied with the news content. Furthermore, when users want to know more about a news item, they must manually search for additional information, which is time-consuming. There is a need to solve these issues and realize a more efficient and user-friendly news delivery system.

[0654] The identification process 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 a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to the user terminal, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for processing the user's reactions while browsing the news using an emotion analysis engine and adjusting the news distribution list based on the user's emotional state, and a means for analyzing the user's news browsing history and generating and sending an alert based on the emotion analysis results if there is a bias toward a specific category. This not only personalizes the user's news browsing experience, but also enables flexible news distribution that takes the user's emotional state into consideration.

[0655] A "news gathering means" is a mechanism for periodically calling the interface of a news provider to obtain the latest news articles.

[0656] "Classification method" refers to the algorithm or machine learning model used to sort collected news articles into specific topics or categories.

[0657] "Generative AI" refers to AI that uses natural language processing technology to automatically generate titles and summaries of news articles.

[0658] A "personalized news delivery list" refers to a list of news articles organized in an optimized manner based on a user's past browsing history and interests.

[0659] "User Terminal" means a device (e.g., a smartphone or tablet) used by a User to receive news articles and notifications.

[0660] A "means for accepting requests for more information" is a mechanism for receiving and processing user requests for additional information about a particular news article.

[0661] "Means of regenerating into a user-friendly format" refers to a mechanism that uses generative artificial intelligence to convert news articles containing academic or specialized information into a form that is easy for the general user to understand.

[0662] "Emotion analysis engine" refers to technology that analyzes reaction data (e.g., facial expressions and operational behavior) when a user browses the news and identifies their emotional state.

[0663] The "means for generating and sending alerts" refers to a mechanism for generating alerts that notify users of bias in specific categories based on their news viewing history and sentiment analysis results, and sending the alerts to the users' devices.

[0664] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by incorporating a sentiment analysis engine, it recognizes user emotions and uses that information to enhance the degree of personalization of news distribution lists and adjust the content and format of alerts to address biases in news categories. A specific embodiment for implementing this system is described below.

[0665] News gathering

[0666] First, the server periodically calls the interface of the news provider to collect the latest news articles. This collection process is performed using, for example, a news API. The collected news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles.

[0667] Category Classification and Generative AI

[0668] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[0669] Personalized News Generation

[0670] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[0671] Utilizing a sentiment analysis engine

[0672] One of the features of this invention is the incorporation of an emotion analysis engine. The emotion analysis engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[0673] Providing more information

[0674] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[0675] Alert function

[0676] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion analysis engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[0677] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion analysis engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing an emotion analysis engine makes it possible to deliver news that appropriately responds to the user's emotional state.

[0678] Example prompt sentence:

[0679] "Analyze the news browsing history of user A and identify areas of interest. Also, generate a news distribution list for when user A has recently been feeling stressed."

[0680] In this way, the system of the present invention makes the user's information gathering more efficient, supports the user in having a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[0681] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0682] Step 1:

[0683] The server periodically calls the interface of the news provider. For example, it can be configured to issue a request to the news API every hour. It specifies the URL endpoint to use for the call and sends the request with any necessary parameters (e.g., API key, country code, etc.).

[0684] Input: API endpoint URL, API key, request parameters

[0685] Output: News data in JSON format

[0686] Specifically, it sends an HTTP GET request to a URL such as https: / / newsapi.org / v2 / top-headlines?country=jp&apiKey=YOUR_API_KEY.

[0687] Step 2:

[0688] The server parses the received JSON-formatted news data to extract individual news articles. It then uses a JSON library to analyze the received data and obtain the title, body, and category of each news article.

[0689] Input: News data in JSON format

[0690] Output: List of news articles (title, body, category)

[0691] Specifically, it uses code such as json.loads(response.text) to convert the JSON data into a dictionary format and extract the required data fields.

[0692] Step 3:

[0693] The server categorizes the collected news articles by topic or category, and based on a predefined list of categories (e.g., business, technology, entertainment), it assigns the news articles to the corresponding category.

[0694] Input: List of news articles (title, body, category)

[0695] Output: A list of news articles sorted by category

[0696] Specifically, it classifies news articles into categories using conditional statements such as if article["category"] == "business":

[0697] Step 4:

[0698] The server uses a generative device to generate titles and summaries for news articles. It uses a generative AI model (e.g., GPT-3) to automatically generate optimal titles and summaries from the text of news articles.

[0699] Input: News article body

[0700] Output: Generated title and summary

[0701] Specifically, it calls functions such as gpt3.generate_summary(article["content"]) to obtain the generated results from the AI ​​model.

[0702] Step 5:

[0703] The server stores the generated title and abstract in a database. It executes an INSERT query to the database to store data such as title, abstract, body, and category.

[0704] Input: Generated title and summary, news article body, category

[0705] Output: News article information stored in a database

[0706] Specifically, it executes SQL statements such as INSERT INTO news_articles (title, summary, content, category) VALUES (...).

[0707] Step 6:

[0708] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It retrieves the user's browsing data from a database and uses analytical algorithms to extract personal preferences.

[0709] Input: User news browsing history data

[0710] Output: A personalized news distribution list

[0711] Specifically, the system aggregates data such as the number of category views per user and time periods, and calculates a priority score.

[0712] Step 7:

[0713] The server sends the created news distribution list to the device and delivers it to the user as a push notification or in-app message. The distribution list is packaged in JSON format and delivered using a messaging service (e.g., Firebase Cloud Messaging).

[0714] Input: Personalized news distribution list

[0715] Output: Notification to user terminal

[0716] Specifically, message sending code such as fcm.send_message(user_device_token, news_list_json) is used.

[0717] Step 8:

[0718] The device sends the user's reactions to the news to the emotion analysis engine, including facial expression recognition results and device touch data.

[0719] Input: User reaction data when browsing news

[0720] Output: Emotion analysis results

[0721] Specifically, the camera captures facial expressions and sends the images to an emotion analysis API.

[0722] Step 9:

[0723] The server processes the transmitted emotional data using an emotion analysis engine. The emotion analysis engine is used to identify the user's emotional state. The analysis results (e.g., stress level) are used to influence the next news delivery.

[0724] Input: User reaction data, results of sentiment analysis engine

[0725] Output: News distribution list based on emotional state

[0726] Specifically, it calls the sentiment analysis API and obtains the results.

[0727] Step 10:

[0728] Based on the analysis results, the server selects appropriate news articles and recommends them to the user. It also regenerates a news distribution list based on the emotional state and sends it to the device.

[0729] Input: Sentiment analysis results, news article list

[0730] Output: News distribution list based on emotional state

[0731] Specifically, the system generates and transmits a personalized news list again.

[0732] Step 11:

[0733] A user requests more information about a particular news item. Through the device interface, the user sends a request for more information about the particular article.

[0734] Input: News article ID

[0735] Output: Request status

[0736] Specifically, the request is sent when the "View details" button is clicked.

[0737] Step 12:

[0738] The server calls the news API again to get more information, which is then regenerated into a user-friendly format using generative artificial intelligence.

[0739] Input: News article ID

[0740] Output: Detailed information generated

[0741] Specifically, the API is called by adding the news article ID as a query parameter, and functions such as gpt3.generate_readable_format(article_detail_content) are used.

[0742] Step 13:

[0743] The server sends the regenerated details to the device for the user to view, and packages the generated details in JSON format and sends it.

[0744] Input: Generated details

[0745] Output: Notification to terminal

[0746] Specific operations include code such as send_to_device(user_device_token, detailed_article_json).

[0747] Step 14:

[0748] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a specific category. It combines a sentiment analysis engine to adjust the content and format of the alert according to the user's emotional state.

[0749] Input: User news browsing history, sentiment analysis results

[0750] Output: Bias alert

[0751] Specifically, it compares the viewing frequency of each category on a weekly basis and generates an alert if there is a bias.

[0752] Step 15:

[0753] The server sends the generated alert to the device to alert the user, and packages the alert content in JSON format and sends it to the device.

[0754] Input: Bias Alert

[0755] Output: Alert notification to terminal

[0756] Specific actions include using codes such as send_alert(user_device_token, alert_message_json).

[0757] (Application example 2)

[0758] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0759] In today's world, many users get their information from online news, but this presents several challenges. First, it is often impossible to efficiently gather the information that is truly important to users from the vast amount of news articles. Second, it is difficult to personalize news based on the user's interests and emotions. In particular, if a user is feeling stressed or has a bias toward certain categories, appropriate news delivery may not be provided. Furthermore, there is a lack of a mechanism for utilizing user feedback to improve the system.

[0760] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, and a means for generating titles and summaries using generative artificial intelligence. This enables efficient processing of collected news articles and personalized news delivery according to the user's needs. The server also includes an emotion engine that recognizes the user's emotional state and a means for dynamically adjusting the news delivery list based on the recognized emotion, a means for regenerating and providing news in a user-friendly format using generative artificial intelligence, and a means for accepting requests for detailed information on specific news. This configuration enables flexible and effective news delivery based on the user's emotions and interests, improving the user's information gathering experience.

[0761] A "news gathering means" is a means for gathering the latest news articles from news sources.

[0762] "Topic or category classification means" refers to a means for automatically classifying collected news articles into specific topics or categories, such as business, technology, entertainment, etc.

[0763] "Generative AI" is an AI technology used to automatically generate titles and summaries from collected news articles.

[0764] The "means for creating a personalized news distribution list" is a means for creating a news distribution list optimized for an individual user based on the user's past browsing history.

[0765] The "means for delivering to a user terminal" refers to a means for delivering the generated news title and summary to a device such as a user's smartphone or computer.

[0766] The "means for accepting a request for providing detailed information" is a means for accepting a request from a user for detailed information about a specific news article.

[0767] "Means for regenerating and providing in a user-friendly format" refers to means for obtaining detailed information about specified news and regenerating it in a user-friendly format using generative artificial intelligence.

[0768] "Means for generating and sending alerts" refers to means for analyzing a user's news browsing history, generating an alert if there is a bias in a specific category, and notifying the user.

[0769] The "emotion engine" is an engine that recognizes the user's emotional state (e.g., stress or frustration) and adjusts the content of news distribution lists and alerts based on that data.

[0770] A system embodying the invention includes the following major elements:

[0771] 1. News collection and classification

[0772] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The retrieved news articles are then classified by topic or category. For example, categories such as business, technology, and entertainment are set.

[0773] 2. Title and Abstract Generation Using Generative AI

[0774] The server uses generative artificial intelligence (AI) to automatically generate titles and summaries from collected news articles, which are then stored in a database.

[0775] 3. Create a personalized news distribution list

[0776] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on that information. The generated news distribution list is sent to the user's device as a push notification or in-app message.

[0777] 4. Emotional Engine Adjustment

[0778] The server is equipped with an emotion engine that analyzes users' reactions while browsing the news (e.g., facial expression recognition and device interaction data). This allows the server to recognize the user's emotional state and recommend the most appropriate news articles for each individual user. For example, if a user is feeling stressed, it will prioritize the delivery of relaxing entertainment articles.

[0779] 5. Requesting and Providing More Information

[0780] When a user wants to check the details of a specific article, they send a request for more information from their device. Upon receiving this request, the server calls the news API again to retrieve the details, and then uses generative artificial intelligence to regenerate them into a user-friendly format. The regenerated details are then sent to the device, where they can be viewed by the user.

[0781] 6. Generate and send alerts

[0782] The server periodically analyzes a user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function also incorporates an emotion engine, which adjusts the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotion into consideration.

[0783] Hardware and software used

[0784] Emotion recognition: Keras (TensorFlow backend), OpenCV

[0785] News gathering: requests library, news API

[0786] Data storage and management: Any database (e.g., PostgreSQL)

[0787] Emotion engine: Generative AI models (e.g. GPT-3)

[0788] Specific examples

[0789] A user launches a news app on their smartphone and browses recent news. If the emotion engine analyzes the user's facial expressions and recognizes that they are feeling stressed, the server automatically prioritizes relaxing entertainment news. If the user requests more information about a particular business news item, the server instantly generates and provides the detailed information.

[0790] Prompt Sentence Examples

[0791] When users are feeling stressed, introduce them to today's relaxing entertainment news.

[0792] This will realize a news distribution system that recommends appropriate news and products based on the user's emotions and history, and provides personalized feedback.

[0793] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0794] Step 1:

[0795] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The input is the response data from the news API, and the output is a list of extracted news articles.

[0796] Step 2:

[0797] The server classifies the collected news articles by topic and category. The input is the list of news articles obtained in the previous step, and the output is the news articles classified by category. Keyword analysis and natural language processing techniques are used for classification.

[0798] Step 3:

[0799] The server uses generative artificial intelligence to generate titles and summaries from news articles. The generated titles and summaries are stored in a database. The input is the news articles classified in step 2, and the output is the generated title and summary for each article. A generative artificial intelligence model (e.g., GPT-3) performs the text generation.

[0800] Step 4:

[0801] The server analyzes the user's past browsing history and creates a personalized news distribution list. The input is the user's browsing history data and the news articles obtained in step 3, and the output is a personalized news distribution list. A machine learning algorithm is used for the analysis.

[0802] Step 5:

[0803] The server uses an emotion engine to recognize the user's emotional state in real time. The input is the user's facial expression recognition data and interaction data, and the output is the user's emotional state information. Keras models and OpenCV are used for emotion recognition.

[0804] Step 6:

[0805] The server dynamically adjusts the news distribution list based on the user's emotional state. The input is the emotional state information obtained in step 5 and the news distribution list created in step 4, and the output is the adjusted news distribution list. The list priority is changed depending on the emotional state.

[0806] Step 7:

[0807] The device delivers the personalized news feed list to the user as push notifications or in-app messages. The input is the news feed list adjusted in step 6, and the output is the news articles displayed on the user's device.

[0808] Step 8:

[0809] When a user requests detailed information about a specific article, the terminal sends a detailed information request to the server. The input is the user's request for detailed information, and the output is the request being sent to the server.

[0810] Step 9:

[0811] When the server receives the request for more information, it calls the news API again to obtain the details and regenerates them in a user-friendly format using generative artificial intelligence. The input is the request for more information and the response data from the news API, and the output is the regenerated details.

[0812] Step 10:

[0813] The terminal displays the details received from the server to the user. The input is the details regenerated in step 9, and the output is the details displayed on the terminal.

[0814] Step 11:

[0815] The server periodically analyzes the user's news browsing history and generates and sends alerts if there is a bias towards a specific category. The input is the user's news browsing history data and current browsing data, and the output is the generated alert. An algorithm is used for the analysis, and the content and format of the alert reflects an emotional engine.

[0816] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0817] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0818] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0819] [Third embodiment]

[0820] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0821] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0822] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0823] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0824] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0825] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0826] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0827] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0828] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0829] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0830] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0831] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0832] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[0833] First, the server periodically calls the news provider's API to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses the received news data and classifies it by topic or category (for example, "business," "technology," "sports," etc.).

[0834] Next, the server uses generative artificial intelligence (AI) to generate titles and summaries for each news article based on the collected news data. The AI ​​uses the collected news headlines and text as inputs to automatically generate titles and summaries that are easy for users to provide. These generated titles and summaries are then stored in a database by the server.

[0835] The server analyzes the user's past news browsing history and creates a personalized news delivery list based on this. It collects each user's browsing history data and identifies their interests in specific categories and topics. This analysis makes it possible to deliver news optimized for each user. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[0836] When a user wants to know more information about a particular news article, the device sends a request for more information to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[0837] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[0838] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[0839] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will immediately respond to the request and provide the details. Furthermore, if the user only views business news, the server will send an alert to encourage the user to also view news in the "technology" and "sports" categories, encouraging the user to gather information from multiple angles. In this way, the system of the present invention helps users gather information more efficiently and gain a more diverse perspective.

[0840] The processing flow will be explained below.

[0841] Step 1:

[0842] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[0843] Step 2:

[0844] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[0845] Step 3:

[0846] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[0847] Step 4:

[0848] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[0849] Step 5:

[0850] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[0851] Step 6:

[0852] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[0853] Step 7:

[0854] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[0855] Step 8:

[0856] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[0857] Step 9:

[0858] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[0859] Step 10:

[0860] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[0861] Step 11:

[0862] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[0863] Step 12:

[0864] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[0865] Step 13:

[0866] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[0867] Step 14:

[0868] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[0869] Step 15:

[0870] The server generates alerts and sends them to the user's device, delivering them as push notifications or in-app messages.

[0871] Step 16:

[0872] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[0873] Step 17:

[0874] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[0875] Step 18:

[0876] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[0877] Example 1

[0878] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0879] Conventional news distribution systems often require manual processes for news collection, classification, and personalized distribution, resulting in a time-consuming and labor-intensive process. Furthermore, users may only view news from specific categories, and the systems lack appropriate alert functions to address this. Furthermore, the systems lack a mechanism for retraining generative AI based on feedback, which hinders user satisfaction.

[0880] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0881] In this invention, the server includes a news collection means, a means for classifying the collected news by topic and category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a user terminal, a means for accepting a request for detailed information on a specific news item, a means for acquiring detailed information on the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias in a specific category, and a means for collecting user feedback and retraining. This automates the process from news collection to personalized distribution, provision of detailed information, generation of alerts for category bias, and retraining of AI based on feedback, enabling efficient news distribution with high user satisfaction.

[0882] A "news gathering method" is a method for periodically calling the API of a news provider to gather the latest news articles.

[0883] "Means for classifying collected news by topic or category" refers to a means that has the function of parsing collected news data and classifying each news article by topic or category (e.g., "business," "technology," "sports," etc.).

[0884] "Means for generating titles and summaries using generative artificial intelligence" refers to means for automatically generating titles and summaries from the headlines and body text of news articles using a generative AI model.

[0885] "Means for creating a personalized news distribution list based on a user's past browsing history" refers to means for analyzing a user's past news browsing history and creating a news list optimized for each user based on the results.

[0886] "Means for delivering the generated news titles and summaries to the user's device" refers to means for sending the generated news titles and summaries to the user's device via push notifications or in-app messages.

[0887] The "means for accepting a request for providing detailed information on a specific news item" refers to a means for receiving a request when a user wants to know detailed information on a specific news item.

[0888] "Means for obtaining detailed information about specified news items and regenerating and providing it in a user-friendly format using generative AI" refers to means for obtaining detailed information about news items in response to a request and regenerating and providing it in a user-friendly format using a generative AI model.

[0889] "Means for analyzing a user's news browsing history and generating and sending an alert if there is a bias towards a specific category" refers to a means for periodically analyzing a user's news browsing history, generating an alert if there is a bias towards a specific category, and sending the alert to the user's device.

[0890] "Means for collecting user feedback and retraining" refers to a means for collecting feedback from users regarding the content of news distribution and the degree of personalization, and retraining the generative artificial intelligence based on that feedback.

[0891] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[0892] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses this data and classifies each news article by topic or category (e.g., "business," "technology," "sports," etc.). This operation uses techniques such as perlject and field analysis.

[0893] Next, the server uses a generative AI model (such as GPT-4) to generate a title and summary for each news article based on the collected news data. The generative AI model uses the collected news headlines and text as inputs to automatically generate a title and summary that is easy to provide to users. The generated title and summary are then stored in a database by the server.

[0894] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on this. Browsing history data is collected from databases such as MySQL to identify interests in specific categories and topics. As a result, news distribution optimized for each user becomes possible. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[0895] When a user wants to know more information about a particular news article, the device sends a detailed information request to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[0896] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[0897] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[0898] For example, if a businessperson receives the latest news at 7 a.m. and requests to know more about a particular business news item, the server will immediately respond to the request and provide more detailed information. Furthermore, if the person is only browsing business news, the server will send an alert to encourage them to also browse news in the "technology" and "sports" categories, encouraging them to gather information from a variety of angles.

[0899] The following are some examples of prompts that can be input to a generative AI model:

[0900] Write a summary of a news article that begins with "How new AI technologies will impact business." The news article should include the following:

[0901] [Enter news article text here]

[0902] Abstract length should be no more than 150 characters.

[0903] This system allows users to efficiently gather the latest news information and gain a multifaceted perspective.

[0904] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0905] Step 1:

[0906] The server periodically calls the news provider's API to collect the latest news articles. It uses the API endpoint and authentication token as input. The output is news data in JSON format. Specifically, it sends an HTTP request and receives a list of news articles as a response.

[0907] Step 2:

[0908] The server parses the collected news data and classifies it by topic and category. It uses the acquired JSON-formatted news data as input. The output is news data classified by category. Specifically, it parses the JSON data and sorts each news article into a category such as "business," "technology," or "sports."

[0909] Step 3:

[0910] The server uses generative artificial intelligence to generate a title and summary for each news article. It uses classified news data as input. The output is the generated title and summary. Specifically, it passes the headline and body of a news article to the AI ​​model and obtains the generated title and summary.

[0911] Step 4:

[0912] The server saves the title and summary in a database. It uses the generated title and summary as input. The output is news data saved in the database. The specific operation is to save the title and summary in the database together with the news data of the corresponding category.

[0913] Step 5:

[0914] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It uses the user's browsing history data as input, and the output is a personalized news list. Specifically, it extracts browsing history from a database and analyzes interest in specific categories and topics to create a news list.

[0915] Step 6:

[0916] The server delivers a personalized news list to the device. As input, it uses the personalized news list and the user's device token. The output is the news list sent to the user's device. Specific behavior is to send the news list as a push notification or in-app message.

[0917] Step 7:

[0918] When a device wants to know more information about a particular news article, it sends a request for more information to the server. The input is the ID of the requested news article. The output is receiving the requested information. In concrete terms, the user selects a news article and sends a request to the server.

[0919] Step 8:

[0920] The server receives the request and calls the news provider's API again to retrieve details about the specified news article. The input is the news article ID and API endpoint. The output is the retrieved details. The specific operation is to make an API call and receive the details.

[0921] Step 9:

[0922] The server uses generative artificial intelligence to regenerate the details into a user-friendly format and send it to the device. It uses the retrieved details as input. The output is the regenerated details in a user-friendly format. Specifically, it passes the details to the AI ​​model, retrieves the generated summary and clarification, and sends them.

[0923] Step 10:

[0924] The server periodically analyzes the user's news browsing history and generates and sends an alert if there is a bias in a particular category. The input is the user's browsing history data. The output is the generated alert message. Specifically, it extracts the browsing history from the database, analyzes the bias in categories, creates an alert message, and sends it to the device.

[0925] Step 11:

[0926] The server collects user feedback and retrains the generative AI based on that feedback. It uses user feedback as input and the output is a retrained AI model. Specifically, it collects feedback data and uses it as new training data to retrain the AI ​​model.

[0927] (Application example 1)

[0928] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0929] In modern society, users need to gather news from many sources, but in the process, they tend to be biased toward certain categories, making it difficult to gather information from multiple perspectives. Furthermore, systems that efficiently gather news and provide it to users in a personalized format are still insufficient. In particular, little consideration has been given to news distribution within virtual stores, leaving users with limited means to obtain important information in real time while shopping. To solve these issues, a personalized news distribution system is needed.

[0930] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0931] In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a virtual store user interface, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, and a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias toward a specific category, thereby enabling real-time news distribution and personalization in the virtual store.

[0932] "News gathering means" refers to a means of regularly gathering the latest news data using the API of a news provider.

[0933] "Means of classifying by topic or category" refers to a means of organizing collected news data into categories such as "business," "technology," and "sports."

[0934] "Generative AI" is an AI technology that automatically generates titles and summaries from collected news headlines and text.

[0935] The "means for creating a personalized news distribution list" is a means for analyzing a user's past news browsing history and generating an optimal news distribution list based on the user's interests.

[0936] A "virtual store user interface" is an interface device (such as smart glasses or a head-mounted display) that allows users to display and manipulate information within a virtual store.

[0937] The "means for delivering news titles and summaries" is a means for transmitting the generated news titles and summaries to a user terminal.

[0938] The "means for accepting a request for providing detailed information" is a means for accepting a request when a user wants to know detailed information about a particular news article.

[0939] "Means for regenerating using generative artificial intelligence" refers to artificial intelligence technology for regenerating the acquired detailed information in a format that is easy for the user to understand.

[0940] The "means for analyzing news browsing history" is a means for analyzing a user's past news browsing status and determining whether there is a bias toward a particular category.

[0941] "Means for generating and sending alerts" refers to the means for generating and sending alerts to the user interface when it is determined that there is a bias in a particular news category, encouraging the user to gather information from multiple angles.

[0942] "Means for collecting feedback from users" refers to means for collecting users' opinions and impressions regarding the content of news distribution and the degree of personalization.

[0943] The "means for retraining generative artificial intelligence" refers to a means for retraining generative artificial intelligence to improve its accuracy based on feedback data collected from users.

[0944] The "means for collecting the latest news articles" is a means for automatically calling the API of a news provider at a time interval specified by the user to collect the latest news articles.

[0945] This invention is a system that uses the API of news providers to collect the latest news articles, personalize them, and deliver them to a virtual store user interface. Below, we will explain the specific processing performed by the server and the hardware and software used.

[0946] Hardware and Software Overview

[0947] 1. Server: Communicates with news provider APIs to collect news data, parses and stores the news data, runs generative AI models, creates news distribution lists for each user, and collects user feedback.

[0948] 2. User interface device: Smart glasses or head-mounted display used in the virtual store. Users receive news feeds through this device.

[0949] 3. Generative AI model: Automatically generates titles and summaries for news data. The AI ​​model receives prompts to generate titles and summaries based on the latest news articles and generates responses.

[0950] 4. Database: Stores collected news data, generated titles and summaries, user browsing history, and feedback data.

[0951] Specific examples of processing

[0952] News gathering

[0953] The server periodically calls the news provider's API to collect the latest news data, allowing the server to always maintain the latest information.

[0954] Parsing and categorizing news

[0955] The acquired news data is generally provided in structured data format such as JSON. The server parses this data and organizes it into categories such as "business," "technology," and "sports."

[0956] Title and Abstract Generation

[0957] A generative AI model generates user-friendly titles and summaries from collected news data. Here are some example prompts:

[0958] "Create a summary of the latest business news."

[0959] This produces the following summary:

[0960] News: (Business Trends)

[0961] Summary: A Japanese company has announced plans to expand into a new market, which is expected to have a lot of potential.

[0962] Personalized news delivery

[0963] The server analyzes the user's news browsing history and creates an optimal news distribution list for each user based on that information. This list is then distributed to the user interface within the virtual store.

[0964] Providing more information

[0965] When a user wants to know more information about a particular news article, the user interface device sends a request to the server, which then calls the news provider's API again to obtain more information about the specified news article, regenerates it in a user-friendly format using the generative AI model, and provides it to the user.

[0966] Alerts for bias in news categories

[0967] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, sending it to the user interface, making it easier for users to gather information from multiple angles.

[0968] In this way, the system of the present invention provides news gathering, personalization, real-time distribution, detailed information provision, and alert functions for multifaceted information gathering in an integrated manner, allowing users to efficiently gather information even within a virtual store.

[0969] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0970] Step 1:

[0971] The server calls the news provider's API to collect the latest news data. The input is the API URL and any necessary authentication information, and the output is the news data in JSON format. Specifically, a regularly scheduled task sends requests to the news API and saves the retrieved JSON data.

[0972] Step 2:

[0973] The server parses the collected news data and classifies it by category. The input is news data in JSON format, and the output is a news list for each category. Specifically, it parses the JSON data and sorts it into lists based on the category information of each news article.

[0974] Step 3:

[0975] The server uses a generative AI model to generate a title and summary for each news article. The input is the organized news list, and the output is the generated title and summary. Specifically, for each news article, the server inputs the following prompt sentence into the generative AI model:

[0976] "Create a summary of the latest business news."

[0977] The generative AI model generates a summary as a response.

[0978] Step 4:

[0979] The server stores the generated title and summary in a database. The input is the generated title and summary, and the output is the data stored in the database. Specifically, the server associates the title and summary with the news article ID and category information and stores them in the database.

[0980] Step 5:

[0981] The server analyzes the user's past news browsing history and creates a personalized news distribution list. The input is the user's browsing history data, and the output is a personalized news distribution list. Specifically, it uses a machine learning algorithm to predict the user's interests and select the most suitable news articles.

[0982] Step 6:

[0983] The server delivers the generated news titles and summaries to the virtual store user interface. The input is a personalized news distribution list, and the output is the news displayed on the user interface. Specifically, the server sends the list contents to smart glasses or a head-mounted display, which notifies the user.

[0984] Step 7:

[0985] When a user wants to know more information about a particular news article, they send a request for more information from their device to the server. The input is the news article ID, and the output is detailed data from the news source. Specifically, the user operates the interface and clicks the "More Information" button for a particular article.

[0986] Step 8:

[0987] After receiving the request, the server calls the news provider's API again to obtain detailed information about the specified news article. The input is the news article ID, API URL, and authentication information, and the output is the detailed data of the news article. Specifically, the server makes the appropriate API call to obtain the detailed information.

[0988] Step 9:

[0989] The server uses the generative AI model to regenerate the acquired detailed information into a user-friendly format. The input is the acquired detailed information, and the output is the regenerated detailed information. Specifically, the acquired detailed information is input back into the generative AI model and reconstructed into an easy-to-understand format.

[0990] Step 10:

[0991] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, and sends it to the user interface. The input is news browsing history data, and the output is an alert message. Specifically, it uses a machine learning algorithm to analyze the history, and generates and sends an alert to users who only browse a particular category, such as "Please try browsing articles in other categories."

[0992] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0993] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by combining it with an emotion engine, it recognizes the user's emotions, and based on that information, it enhances the degree of personalization of the news delivery list and adjusts the content and format of alerts to address biases in news categories.

[0994] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles. This allows information such as the title, body, and category of each news article to be extracted.

[0995] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[0996] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[0997] One of the features of this invention is the incorporation of an emotion engine. The emotion engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[0998] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[0999] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[1000] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing the emotion engine makes it possible to deliver news that responds appropriately to the user's emotional state.

[1001] The system of the present invention makes users' information gathering more efficient, supports them in gaining a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[1002] The processing flow will be explained below.

[1003] Step 1:

[1004] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[1005] Step 2:

[1006] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[1007] Step 3:

[1008] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[1009] Step 4:

[1010] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[1011] Step 5:

[1012] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[1013] Step 6:

[1014] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[1015] Step 7:

[1016] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[1017] Step 8:

[1018] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[1019] Step 9:

[1020] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[1021] Step 10:

[1022] The emotion engine analyzes users' reactions to news articles, and the reaction data is provided by the device as facial expression recognition and interaction data.

[1023] Step 11:

[1024] The emotion engine recognizes the user's emotional state based on the reaction data it acquires, which can be categorized into categories such as stress, joy, and excitement.

[1025] Step 12:

[1026] The server updates the user profile with the recognized emotion data and uses that information to further personalize the news distribution list, selecting news articles according to the user's emotional state.

[1027] Step 13:

[1028] The server then transmits a personalized news distribution list based on the emotion engine back to the user's terminal, the distribution including news optimized based on the emotional state.

[1029] Step 14:

[1030] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[1031] Step 15:

[1032] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[1033] Step 16:

[1034] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[1035] Step 17:

[1036] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[1037] Step 18:

[1038] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[1039] Step 19:

[1040] The server generates alerts and sends them to the user's device. Alerts are delivered as push notifications or in-app messages. The content and format of the alerts are adjusted according to the user's emotional state, as recognized by the emotion engine.

[1041] Step 20:

[1042] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[1043] Step 21:

[1044] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[1045] Step 22:

[1046] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[1047] Example 2

[1048] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1049] Currently, many users gather news online, but personalized news delivery is not yet fully implemented. Furthermore, news recommendations do not take into account the emotional state of individual users, which can lead to information overload or bias. As a result, users feel stressed or dissatisfied with the news content. Furthermore, when users want to know more about a news item, they must manually search for additional information, which is time-consuming. There is a need to solve these issues and realize a more efficient and user-friendly news delivery system.

[1050] The identification process 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 a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to the user terminal, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for processing the user's reactions while browsing the news using an emotion analysis engine and adjusting the news distribution list based on the user's emotional state, and a means for analyzing the user's news browsing history and generating and sending an alert based on the emotion analysis results if there is a bias toward a specific category. This not only personalizes the user's news browsing experience, but also enables flexible news distribution that takes the user's emotional state into consideration.

[1051] A "news gathering means" is a mechanism for periodically calling the interface of a news provider to obtain the latest news articles.

[1052] "Classification method" refers to the algorithm or machine learning model used to sort collected news articles into specific topics or categories.

[1053] "Generative AI" refers to AI that uses natural language processing technology to automatically generate titles and summaries of news articles.

[1054] A "personalized news delivery list" refers to a list of news articles organized in an optimized manner based on a user's past browsing history and interests.

[1055] "User Terminal" means a device (e.g., a smartphone or tablet) used by a User to receive news articles and notifications.

[1056] A "means for accepting requests for more information" is a mechanism for receiving and processing user requests for additional information about a particular news article.

[1057] "Means of regenerating into a user-friendly format" refers to a mechanism that uses generative artificial intelligence to convert news articles containing academic or specialized information into a form that is easy for the general user to understand.

[1058] "Emotion analysis engine" refers to technology that analyzes reaction data (e.g., facial expressions and operational behavior) when a user browses the news and identifies their emotional state.

[1059] The "means for generating and sending alerts" refers to a mechanism for generating alerts that notify users of bias in specific categories based on their news viewing history and sentiment analysis results, and sending the alerts to the users' devices.

[1060] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by incorporating a sentiment analysis engine, it recognizes user emotions and uses that information to enhance the degree of personalization of news distribution lists and adjust the content and format of alerts to address biases in news categories. A specific embodiment for implementing this system is described below.

[1061] News gathering

[1062] First, the server periodically calls the interface of the news provider to collect the latest news articles. This collection process is performed using, for example, a news API. The collected news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles.

[1063] Category Classification and Generative AI

[1064] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[1065] Personalized News Generation

[1066] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[1067] Utilizing a sentiment analysis engine

[1068] One of the features of this invention is the incorporation of an emotion analysis engine. The emotion analysis engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[1069] Providing more information

[1070] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[1071] Alert function

[1072] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion analysis engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[1073] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion analysis engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing an emotion analysis engine makes it possible to deliver news that appropriately responds to the user's emotional state.

[1074] Example prompt sentence:

[1075] "Analyze the news browsing history of user A and identify areas of interest. Also, generate a news distribution list for when user A has recently been feeling stressed."

[1076] In this way, the system of the present invention makes the user's information gathering more efficient, supports the user in having a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[1077] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1078] Step 1:

[1079] The server periodically calls the interface of the news provider. For example, it can be configured to issue a request to the news API every hour. It specifies the URL endpoint to use for the call and sends the request with any necessary parameters (e.g., API key, country code, etc.).

[1080] Input: API endpoint URL, API key, request parameters

[1081] Output: News data in JSON format

[1082] Specifically, it sends an HTTP GET request to a URL such as https: / / newsapi.org / v2 / top-headlines?country=jp&apiKey=YOUR_API_KEY.

[1083] Step 2:

[1084] The server parses the received JSON-formatted news data to extract individual news articles. It then uses a JSON library to analyze the received data and obtain the title, body, and category of each news article.

[1085] Input: News data in JSON format

[1086] Output: List of news articles (title, body, category)

[1087] Specifically, it uses code such as json.loads(response.text) to convert the JSON data into a dictionary format and extract the required data fields.

[1088] Step 3:

[1089] The server categorizes the collected news articles by topic or category, and based on a predefined list of categories (e.g., business, technology, entertainment), it assigns the news articles to the corresponding category.

[1090] Input: List of news articles (title, body, category)

[1091] Output: A list of news articles sorted by category

[1092] Specifically, it classifies news articles into categories using conditional statements such as if article["category"] == "business":

[1093] Step 4:

[1094] The server uses a generative device to generate titles and summaries for news articles. It uses a generative AI model (e.g., GPT-3) to automatically generate optimal titles and summaries from the text of news articles.

[1095] Input: News article body

[1096] Output: Generated title and summary

[1097] Specifically, it calls functions such as gpt3.generate_summary(article["content"]) to obtain the generated results from the AI ​​model.

[1098] Step 5:

[1099] The server stores the generated title and abstract in a database. It executes an INSERT query to the database to store data such as title, abstract, body, and category.

[1100] Input: Generated title and summary, news article body, category

[1101] Output: News article information stored in a database

[1102] Specifically, it executes SQL statements such as INSERT INTO news_articles (title, summary, content, category) VALUES (...).

[1103] Step 6:

[1104] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It retrieves the user's browsing data from a database and uses analytical algorithms to extract personal preferences.

[1105] Input: User news browsing history data

[1106] Output: A personalized news distribution list

[1107] Specifically, the system aggregates data such as the number of category views per user and time periods, and calculates a priority score.

[1108] Step 7:

[1109] The server sends the created news distribution list to the device and delivers it to the user as a push notification or in-app message. The distribution list is packaged in JSON format and delivered using a messaging service (e.g., Firebase Cloud Messaging).

[1110] Input: Personalized news distribution list

[1111] Output: Notification to user terminal

[1112] Specifically, message sending code such as fcm.send_message(user_device_token, news_list_json) is used.

[1113] Step 8:

[1114] The device sends the user's reactions to the news to the emotion analysis engine, including facial expression recognition results and device touch data.

[1115] Input: User reaction data when browsing news

[1116] Output: Emotion analysis results

[1117] Specifically, the camera captures facial expressions and sends the images to an emotion analysis API.

[1118] Step 9:

[1119] The server processes the transmitted emotional data using an emotion analysis engine. The emotion analysis engine is used to identify the user's emotional state. The analysis results (e.g., stress level) are used to influence the next news delivery.

[1120] Input: User reaction data, results of sentiment analysis engine

[1121] Output: News distribution list based on emotional state

[1122] Specifically, it calls the sentiment analysis API and obtains the results.

[1123] Step 10:

[1124] Based on the analysis results, the server selects appropriate news articles and recommends them to the user. It also regenerates a news distribution list based on the emotional state and sends it to the device.

[1125] Input: Sentiment analysis results, news article list

[1126] Output: News distribution list based on emotional state

[1127] Specifically, the system generates and transmits a personalized news list again.

[1128] Step 11:

[1129] A user requests more information about a particular news item. Through the device interface, the user sends a request for more information about the particular article.

[1130] Input: News article ID

[1131] Output: Request status

[1132] Specifically, the request is sent when the "View details" button is clicked.

[1133] Step 12:

[1134] The server calls the news API again to get more information, which is then regenerated into a user-friendly format using generative artificial intelligence.

[1135] Input: News article ID

[1136] Output: Detailed information generated

[1137] Specifically, the API is called by adding the news article ID as a query parameter, and functions such as gpt3.generate_readable_format(article_detail_content) are used.

[1138] Step 13:

[1139] The server sends the regenerated details to the device for the user to view, and packages the generated details in JSON format and sends it.

[1140] Input: Generated details

[1141] Output: Notification to terminal

[1142] Specific operations include code such as send_to_device(user_device_token, detailed_article_json).

[1143] Step 14:

[1144] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a specific category. It combines a sentiment analysis engine to adjust the content and format of the alert according to the user's emotional state.

[1145] Input: User news browsing history, sentiment analysis results

[1146] Output: Bias alert

[1147] Specifically, it compares the viewing frequency of each category on a weekly basis and generates an alert if there is a bias.

[1148] Step 15:

[1149] The server sends the generated alert to the device to alert the user, and packages the alert content in JSON format and sends it to the device.

[1150] Input: Bias Alert

[1151] Output: Alert notification to terminal

[1152] Specific actions include using codes such as send_alert(user_device_token, alert_message_json).

[1153] (Application example 2)

[1154] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1155] In today's world, many users get their information from online news, but this presents several challenges. First, it is often impossible to efficiently gather the information that is truly important to users from the vast amount of news articles. Second, it is difficult to personalize news based on the user's interests and emotions. In particular, if a user is feeling stressed or has a bias toward certain categories, appropriate news delivery may not be provided. Furthermore, there is a lack of a mechanism for utilizing user feedback to improve the system.

[1156] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, and a means for generating titles and summaries using generative artificial intelligence. This enables efficient processing of collected news articles and personalized news delivery according to the user's needs. The server also includes an emotion engine that recognizes the user's emotional state and a means for dynamically adjusting the news delivery list based on the recognized emotion, a means for regenerating and providing news in a user-friendly format using generative artificial intelligence, and a means for accepting requests for detailed information on specific news. This configuration enables flexible and effective news delivery based on the user's emotions and interests, improving the user's information gathering experience.

[1157] A "news gathering means" is a means for gathering the latest news articles from news sources.

[1158] "Topic or category classification means" refers to a means for automatically classifying collected news articles into specific topics or categories, such as business, technology, entertainment, etc.

[1159] "Generative AI" is an AI technology used to automatically generate titles and summaries from collected news articles.

[1160] The "means for creating a personalized news distribution list" is a means for creating a news distribution list optimized for an individual user based on the user's past browsing history.

[1161] The "means for delivering to a user terminal" refers to a means for delivering the generated news title and summary to a device such as a user's smartphone or computer.

[1162] The "means for accepting a request for providing detailed information" is a means for accepting a request from a user for detailed information about a specific news article.

[1163] "Means for regenerating and providing in a user-friendly format" refers to means for obtaining detailed information about specified news and regenerating it in a user-friendly format using generative artificial intelligence.

[1164] "Means for generating and sending alerts" refers to means for analyzing a user's news browsing history, generating an alert if there is a bias in a specific category, and notifying the user.

[1165] The "emotion engine" is an engine that recognizes the user's emotional state (e.g., stress or frustration) and adjusts the content of news distribution lists and alerts based on that data.

[1166] A system embodying the invention includes the following major elements:

[1167] 1. News collection and classification

[1168] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The retrieved news articles are then classified by topic or category. For example, categories such as business, technology, and entertainment are set.

[1169] 2. Title and Abstract Generation Using Generative AI

[1170] The server uses generative artificial intelligence (AI) to automatically generate titles and summaries from collected news articles, which are then stored in a database.

[1171] 3. Create a personalized news distribution list

[1172] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on that information. The generated news distribution list is sent to the user's device as a push notification or in-app message.

[1173] 4. Emotional Engine Adjustment

[1174] The server is equipped with an emotion engine that analyzes users' reactions while browsing the news (e.g., facial expression recognition and device interaction data). This allows the server to recognize the user's emotional state and recommend the most appropriate news articles for each individual user. For example, if a user is feeling stressed, it will prioritize the delivery of relaxing entertainment articles.

[1175] 5. Requesting and Providing More Information

[1176] When a user wants to check the details of a specific article, they send a request for more information from their device. Upon receiving this request, the server calls the news API again to retrieve the details, and then uses generative artificial intelligence to regenerate them into a user-friendly format. The regenerated details are then sent to the device, where they can be viewed by the user.

[1177] 6. Generate and send alerts

[1178] The server periodically analyzes a user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function also incorporates an emotion engine, which adjusts the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotion into consideration.

[1179] Hardware and software used

[1180] Emotion recognition: Keras (TensorFlow backend), OpenCV

[1181] News gathering: requests library, news API

[1182] Data storage and management: Any database (e.g., PostgreSQL)

[1183] Emotion engine: Generative AI models (e.g. GPT-3)

[1184] Specific examples

[1185] A user launches a news app on their smartphone and browses recent news. If the emotion engine analyzes the user's facial expressions and recognizes that they are feeling stressed, the server automatically prioritizes relaxing entertainment news. If the user requests more information about a particular business news item, the server instantly generates and provides the detailed information.

[1186] Prompt Sentence Examples

[1187] When users are feeling stressed, introduce them to today's relaxing entertainment news.

[1188] This will realize a news distribution system that recommends appropriate news and products based on the user's emotions and history, and provides personalized feedback.

[1189] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1190] Step 1:

[1191] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The input is the response data from the news API, and the output is a list of extracted news articles.

[1192] Step 2:

[1193] The server classifies the collected news articles by topic and category. The input is the list of news articles obtained in the previous step, and the output is the news articles classified by category. Keyword analysis and natural language processing techniques are used for classification.

[1194] Step 3:

[1195] The server uses generative artificial intelligence to generate titles and summaries from news articles. The generated titles and summaries are stored in a database. The input is the news articles classified in step 2, and the output is the generated title and summary for each article. A generative artificial intelligence model (e.g., GPT-3) performs the text generation.

[1196] Step 4:

[1197] The server analyzes the user's past browsing history and creates a personalized news distribution list. The input is the user's browsing history data and the news articles obtained in step 3, and the output is a personalized news distribution list. A machine learning algorithm is used for the analysis.

[1198] Step 5:

[1199] The server uses an emotion engine to recognize the user's emotional state in real time. The input is the user's facial expression recognition data and interaction data, and the output is the user's emotional state information. Keras models and OpenCV are used for emotion recognition.

[1200] Step 6:

[1201] The server dynamically adjusts the news distribution list based on the user's emotional state. The input is the emotional state information obtained in step 5 and the news distribution list created in step 4, and the output is the adjusted news distribution list. The list priority is changed depending on the emotional state.

[1202] Step 7:

[1203] The device delivers the personalized news feed list to the user as push notifications or in-app messages. The input is the news feed list adjusted in step 6, and the output is the news articles displayed on the user's device.

[1204] Step 8:

[1205] When a user requests detailed information about a specific article, the terminal sends a detailed information request to the server. The input is the user's request for detailed information, and the output is the request being sent to the server.

[1206] Step 9:

[1207] When the server receives the request for more information, it calls the news API again to obtain the details and regenerates them in a user-friendly format using generative artificial intelligence. The input is the request for more information and the response data from the news API, and the output is the regenerated details.

[1208] Step 10:

[1209] The terminal displays the details received from the server to the user. The input is the details regenerated in step 9, and the output is the details displayed on the terminal.

[1210] Step 11:

[1211] The server periodically analyzes the user's news browsing history and generates and sends alerts if there is a bias towards a specific category. The input is the user's news browsing history data and current browsing data, and the output is the generated alert. An algorithm is used for the analysis, and the content and format of the alert reflects an emotional engine.

[1212] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1213] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1214] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1215] [Fourth embodiment]

[1216] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1217] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1218] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1219] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1220] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1221] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1222] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1223] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1224] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1225] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1226] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1227] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1228] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1229] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[1230] First, the server periodically calls the news provider's API to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses the received news data and classifies it by topic or category (for example, "business," "technology," "sports," etc.).

[1231] Next, the server uses generative artificial intelligence (AI) to generate titles and summaries for each news article based on the collected news data. The AI ​​uses the collected news headlines and text as inputs to automatically generate titles and summaries that are easy for users to provide. These generated titles and summaries are then stored in a database by the server.

[1232] The server analyzes the user's past news browsing history and creates a personalized news delivery list based on this. It collects each user's browsing history data and identifies their interests in specific categories and topics. This analysis makes it possible to deliver news optimized for each user. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[1233] When a user wants to know more information about a particular news article, the device sends a request for more information to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[1234] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[1235] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[1236] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will immediately respond to the request and provide the details. Furthermore, if the user only views business news, the server will send an alert to encourage the user to also view news in the "technology" and "sports" categories, encouraging the user to gather information from multiple angles. In this way, the system of the present invention helps users gather information more efficiently and gain a more diverse perspective.

[1237] The processing flow will be explained below.

[1238] Step 1:

[1239] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[1240] Step 2:

[1241] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[1242] Step 3:

[1243] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[1244] Step 4:

[1245] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[1246] Step 5:

[1247] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[1248] Step 6:

[1249] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[1250] Step 7:

[1251] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[1252] Step 8:

[1253] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[1254] Step 9:

[1255] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[1256] Step 10:

[1257] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[1258] Step 11:

[1259] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[1260] Step 12:

[1261] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[1262] Step 13:

[1263] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[1264] Step 14:

[1265] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[1266] Step 15:

[1267] The server generates alerts and sends them to the user's device, delivering them as push notifications or in-app messages.

[1268] Step 16:

[1269] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[1270] Step 17:

[1271] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[1272] Step 18:

[1273] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[1274] Example 1

[1275] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1276] Conventional news distribution systems often require manual processes for news collection, classification, and personalized distribution, resulting in a time-consuming and labor-intensive process. Furthermore, users may only view news from specific categories, and the systems lack appropriate alert functions to address this. Furthermore, the systems lack a mechanism for retraining generative AI based on feedback, which hinders user satisfaction.

[1277] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1278] In this invention, the server includes a news collection means, a means for classifying the collected news by topic and category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a user terminal, a means for accepting a request for detailed information on a specific news item, a means for acquiring detailed information on the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias in a specific category, and a means for collecting user feedback and retraining. This automates the process from news collection to personalized distribution, provision of detailed information, generation of alerts for category bias, and retraining of AI based on feedback, enabling efficient news distribution with high user satisfaction.

[1279] A "news gathering method" is a method for periodically calling the API of a news provider to gather the latest news articles.

[1280] "Means for classifying collected news by topic or category" refers to a means that has the function of parsing collected news data and classifying each news article by topic or category (e.g., "business," "technology," "sports," etc.).

[1281] "Means for generating titles and summaries using generative artificial intelligence" refers to means for automatically generating titles and summaries from the headlines and body text of news articles using a generative AI model.

[1282] "Means for creating a personalized news distribution list based on a user's past browsing history" refers to means for analyzing a user's past news browsing history and creating a news list optimized for each user based on the results.

[1283] "Means for delivering the generated news titles and summaries to the user's device" refers to means for sending the generated news titles and summaries to the user's device via push notifications or in-app messages.

[1284] The "means for accepting a request for providing detailed information on a specific news item" refers to a means for receiving a request when a user wants to know detailed information on a specific news item.

[1285] "Means for obtaining detailed information about specified news items and regenerating and providing it in a user-friendly format using generative AI" refers to means for obtaining detailed information about news items in response to a request and regenerating and providing it in a user-friendly format using a generative AI model.

[1286] "Means for analyzing a user's news browsing history and generating and sending an alert if there is a bias towards a specific category" refers to a means for periodically analyzing a user's news browsing history, generating an alert if there is a bias towards a specific category, and sending the alert to the user's device.

[1287] "Means for collecting user feedback and retraining" refers to a means for collecting feedback from users regarding the content of news distribution and the degree of personalization, and retraining the generative artificial intelligence based on that feedback.

[1288] This invention is a system that uses generative artificial intelligence to automatically deliver titles and summaries of the latest news articles, allowing users to dig deeper into the details of news they are interested in. The personalized news delivery also provides an alert function for bias in news categories.

[1289] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data obtained from the news API is generally provided in structured data format such as JSON. The server parses this data and classifies each news article by topic or category (e.g., "business," "technology," "sports," etc.). This operation uses techniques such as perlject and field analysis.

[1290] Next, the server uses a generative AI model (such as GPT-4) to generate a title and summary for each news article based on the collected news data. The generative AI model uses the collected news headlines and text as inputs to automatically generate a title and summary that is easy to provide to users. The generated title and summary are then stored in a database by the server.

[1291] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on this. Browsing history data is collected from databases such as MySQL to identify interests in specific categories and topics. As a result, news distribution optimized for each user becomes possible. The personalized news list is then sent from the server to the device and delivered to the user as a push notification or in-app message.

[1292] When a user wants to know more information about a particular news article, the device sends a detailed information request to the server. The server then calls the news provider's API to retrieve additional details about the specified news article. This information is then regenerated into a user-friendly format using generative artificial intelligence and sent to the device.

[1293] Furthermore, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. For example, if a user frequently browses only the "Business" category, an alert such as "Please try browsing articles in other categories" will be sent to the device. This alert function allows users to collect information from multiple angles.

[1294] The server also collects user feedback and uses it to retrain the generative AI, improving its accuracy and personalization capabilities and increasing user satisfaction.

[1295] For example, if a businessperson receives the latest news at 7 a.m. and requests to know more about a particular business news item, the server will immediately respond to the request and provide more detailed information. Furthermore, if the person is only browsing business news, the server will send an alert to encourage them to also browse news in the "technology" and "sports" categories, encouraging them to gather information from a variety of angles.

[1296] The following are some examples of prompts that can be input to a generative AI model:

[1297] Write a summary of a news article that begins with "How new AI technologies will impact business." The news article should include the following:

[1298] [Enter news article text here]

[1299] Abstract length should be no more than 150 characters.

[1300] This system allows users to efficiently gather the latest news information and gain a multifaceted perspective.

[1301] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1302] Step 1:

[1303] The server periodically calls the news provider's API to collect the latest news articles. It uses the API endpoint and authentication token as input. The output is news data in JSON format. Specifically, it sends an HTTP request and receives a list of news articles as a response.

[1304] Step 2:

[1305] The server parses the collected news data and classifies it by topic and category. It uses the acquired JSON-formatted news data as input. The output is news data classified by category. Specifically, it parses the JSON data and sorts each news article into a category such as "business," "technology," or "sports."

[1306] Step 3:

[1307] The server uses generative artificial intelligence to generate a title and summary for each news article. It uses classified news data as input. The output is the generated title and summary. Specifically, it passes the headline and body of a news article to the AI ​​model and obtains the generated title and summary.

[1308] Step 4:

[1309] The server saves the title and summary in a database. It uses the generated title and summary as input. The output is news data saved in the database. The specific operation is to save the title and summary in the database together with the news data of the corresponding category.

[1310] Step 5:

[1311] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It uses the user's browsing history data as input, and the output is a personalized news list. Specifically, it extracts browsing history from a database and analyzes interest in specific categories and topics to create a news list.

[1312] Step 6:

[1313] The server delivers a personalized news list to the device. As input, it uses the personalized news list and the user's device token. The output is the news list sent to the user's device. Specific behavior is to send the news list as a push notification or in-app message.

[1314] Step 7:

[1315] When a device wants to know more information about a particular news article, it sends a request for more information to the server. The input is the ID of the requested news article. The output is receiving the requested information. In concrete terms, the user selects a news article and sends a request to the server.

[1316] Step 8:

[1317] The server receives the request and calls the news provider's API again to retrieve details about the specified news article. The input is the news article ID and API endpoint. The output is the retrieved details. The specific operation is to make an API call and receive the details.

[1318] Step 9:

[1319] The server uses generative artificial intelligence to regenerate the details into a user-friendly format and send it to the device. It uses the retrieved details as input. The output is the regenerated details in a user-friendly format. Specifically, it passes the details to the AI ​​model, retrieves the generated summary and clarification, and sends them.

[1320] Step 10:

[1321] The server periodically analyzes the user's news browsing history and generates and sends an alert if there is a bias in a particular category. The input is the user's browsing history data. The output is the generated alert message. Specifically, it extracts the browsing history from the database, analyzes the bias in categories, creates an alert message, and sends it to the device.

[1322] Step 11:

[1323] The server collects user feedback and retrains the generative AI based on that feedback. It uses user feedback as input and the output is a retrained AI model. Specifically, it collects feedback data and uses it as new training data to retrain the AI ​​model.

[1324] (Application example 1)

[1325] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1326] In modern society, users need to gather news from many sources, but in the process, they tend to be biased toward certain categories, making it difficult to gather information from multiple perspectives. Furthermore, systems that efficiently gather news and provide it to users in a personalized format are still insufficient. In particular, little consideration has been given to news distribution within virtual stores, leaving users with limited means to obtain important information in real time while shopping. To solve these issues, a personalized news distribution system is needed.

[1327] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1328] In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to a virtual store user interface, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, and a means for analyzing the user's news browsing history and generating and sending an alert if there is a bias toward a specific category, thereby enabling real-time news distribution and personalization in the virtual store.

[1329] "News gathering means" refers to a means of regularly gathering the latest news data using the API of a news provider.

[1330] "Means of classifying by topic or category" refers to a means of organizing collected news data into categories such as "business," "technology," and "sports."

[1331] "Generative AI" is an AI technology that automatically generates titles and summaries from collected news headlines and text.

[1332] The "means for creating a personalized news distribution list" is a means for analyzing a user's past news browsing history and generating an optimal news distribution list based on the user's interests.

[1333] A "virtual store user interface" is an interface device (such as smart glasses or a head-mounted display) that allows users to display and manipulate information within a virtual store.

[1334] The "means for delivering news titles and summaries" is a means for transmitting the generated news titles and summaries to a user terminal.

[1335] The "means for accepting a request for providing detailed information" is a means for accepting a request when a user wants to know detailed information about a particular news article.

[1336] "Means for regenerating using generative artificial intelligence" refers to artificial intelligence technology for regenerating the acquired detailed information in a format that is easy for the user to understand.

[1337] The "means for analyzing news browsing history" is a means for analyzing a user's past news browsing status and determining whether there is a bias toward a particular category.

[1338] "Means for generating and sending alerts" refers to the means for generating and sending alerts to the user interface when it is determined that there is a bias in a particular news category, encouraging the user to gather information from multiple angles.

[1339] "Means for collecting feedback from users" refers to means for collecting users' opinions and impressions regarding the content of news distribution and the degree of personalization.

[1340] The "means for retraining generative artificial intelligence" refers to a means for retraining generative artificial intelligence to improve its accuracy based on feedback data collected from users.

[1341] The "means for collecting the latest news articles" is a means for automatically calling the API of a news provider at a time interval specified by the user to collect the latest news articles.

[1342] This invention is a system that uses the API of news providers to collect the latest news articles, personalize them, and deliver them to a virtual store user interface. Below, we will explain the specific processing performed by the server and the hardware and software used.

[1343] Hardware and Software Overview

[1344] 1. Server: Communicates with news provider APIs to collect news data, parses and stores the news data, runs generative AI models, creates news distribution lists for each user, and collects user feedback.

[1345] 2. User interface device: Smart glasses or head-mounted display used in the virtual store. Users receive news feeds through this device.

[1346] 3. Generative AI model: Automatically generates titles and summaries for news data. The AI ​​model receives prompts to generate titles and summaries based on the latest news articles and generates responses.

[1347] 4. Database: Stores collected news data, generated titles and summaries, user browsing history, and feedback data.

[1348] Specific examples of processing

[1349] News gathering

[1350] The server periodically calls the news provider's API to collect the latest news data, allowing the server to always maintain the latest information.

[1351] Parsing and categorizing news

[1352] The acquired news data is generally provided in structured data format such as JSON. The server parses this data and organizes it into categories such as "business," "technology," and "sports."

[1353] Title and Abstract Generation

[1354] A generative AI model generates user-friendly titles and summaries from collected news data. Here are some example prompts:

[1355] "Create a summary of the latest business news."

[1356] This produces the following summary:

[1357] News: (Business Trends)

[1358] Summary: A Japanese company has announced plans to expand into a new market, which is expected to have a lot of potential.

[1359] Personalized news delivery

[1360] The server analyzes the user's news browsing history and creates an optimal news distribution list for each user based on that information. This list is then distributed to the user interface within the virtual store.

[1361] Providing more information

[1362] When a user wants to know more information about a particular news article, the user interface device sends a request to the server, which then calls the news provider's API again to obtain more information about the specified news article, regenerates it in a user-friendly format using the generative AI model, and provides it to the user.

[1363] Alerts for bias in news categories

[1364] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, sending it to the user interface, making it easier for users to gather information from multiple angles.

[1365] In this way, the system of the present invention provides news gathering, personalization, real-time distribution, detailed information provision, and alert functions for multifaceted information gathering in an integrated manner, allowing users to efficiently gather information even within a virtual store.

[1366] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1367] Step 1:

[1368] The server calls the news provider's API to collect the latest news data. The input is the API URL and any necessary authentication information, and the output is the news data in JSON format. Specifically, a regularly scheduled task sends requests to the news API and saves the retrieved JSON data.

[1369] Step 2:

[1370] The server parses the collected news data and classifies it by category. The input is news data in JSON format, and the output is a news list for each category. Specifically, it parses the JSON data and sorts it into lists based on the category information of each news article.

[1371] Step 3:

[1372] The server uses a generative AI model to generate a title and summary for each news article. The input is the organized news list, and the output is the generated title and summary. Specifically, for each news article, the server inputs the following prompt sentence into the generative AI model:

[1373] "Create a summary of the latest business news."

[1374] The generative AI model generates a summary as a response.

[1375] Step 4:

[1376] The server stores the generated title and summary in a database. The input is the generated title and summary, and the output is the data stored in the database. Specifically, the server associates the title and summary with the news article ID and category information and stores them in the database.

[1377] Step 5:

[1378] The server analyzes the user's past news browsing history and creates a personalized news distribution list. The input is the user's browsing history data, and the output is a personalized news distribution list. Specifically, it uses a machine learning algorithm to predict the user's interests and select the most suitable news articles.

[1379] Step 6:

[1380] The server delivers the generated news titles and summaries to the virtual store user interface. The input is a personalized news distribution list, and the output is the news displayed on the user interface. Specifically, the server sends the list contents to smart glasses or a head-mounted display, which notifies the user.

[1381] Step 7:

[1382] When a user wants to know more information about a particular news article, they send a request for more information from their device to the server. The input is the news article ID, and the output is detailed data from the news source. Specifically, the user operates the interface and clicks the "More Information" button for a particular article.

[1383] Step 8:

[1384] After receiving the request, the server calls the news provider's API again to obtain detailed information about the specified news article. The input is the news article ID, API URL, and authentication information, and the output is the detailed data of the news article. Specifically, the server makes the appropriate API call to obtain the detailed information.

[1385] Step 9:

[1386] The server uses the generative AI model to regenerate the acquired detailed information into a user-friendly format. The input is the acquired detailed information, and the output is the regenerated detailed information. Specifically, the acquired detailed information is input back into the generative AI model and reconstructed into an easy-to-understand format.

[1387] Step 10:

[1388] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a particular category, and sends it to the user interface. The input is news browsing history data, and the output is an alert message. Specifically, it uses a machine learning algorithm to analyze the history, and generates and sends an alert to users who only browse a particular category, such as "Please try browsing articles in other categories."

[1389] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1390] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by combining it with an emotion engine, it recognizes the user's emotions, and based on that information, it enhances the degree of personalization of the news delivery list and adjusts the content and format of alerts to address biases in news categories.

[1391] First, the server periodically calls the API of the news provider to collect the latest news articles. The news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles. This allows information such as the title, body, and category of each news article to be extracted.

[1392] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[1393] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[1394] One of the features of this invention is the incorporation of an emotion engine. The emotion engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[1395] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[1396] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[1397] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing the emotion engine makes it possible to deliver news that responds appropriately to the user's emotional state.

[1398] The system of the present invention makes users' information gathering more efficient, supports them in gaining a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[1399] The processing flow will be explained below.

[1400] Step 1:

[1401] The server runs a scheduled task that calls the news provider's API at the specified interval to get the latest news data from the news API.

[1402] Step 2:

[1403] The server receives the raw data retrieved from the news API, which is typically provided in JSON format.

[1404] Step 3:

[1405] The server parses the received news data and extracts individual news articles, each of which contains information such as title, body text, publication date and category.

[1406] Step 4:

[1407] The server categorizes each news article by topic or category, such as business, technology, entertainment, etc.

[1408] Step 5:

[1409] The server uses generative artificial intelligence (AI) to generate a title and summary for each news article. The AI ​​inputs the article text and automatically generates a summary of just a few sentences.

[1410] Step 6:

[1411] The server stores the generated news titles and summaries in a database, including news ID, title, summary, category, publication date, etc.

[1412] Step 7:

[1413] The server retrieves the user's news browsing history from the database and analyzes it to identify the articles the user has recently viewed and the categories in which they have shown interest.

[1414] Step 8:

[1415] The server generates a personalized news distribution list based on the user's profile, listing articles optimized for each user.

[1416] Step 9:

[1417] The server sends the generated news distribution list to the user's device, delivering news titles and summaries as push notifications or in-app messages.

[1418] Step 10:

[1419] The emotion engine analyzes users' reactions to news articles, and the reaction data is provided by the device as facial expression recognition and interaction data.

[1420] Step 11:

[1421] The emotion engine recognizes the user's emotional state based on the reaction data it acquires, which can be categorized into categories such as stress, joy, and excitement.

[1422] Step 12:

[1423] The server updates the user profile with the recognized emotion data and uses that information to further personalize the news distribution list, selecting news articles according to the user's emotional state.

[1424] Step 13:

[1425] The server then transmits a personalized news distribution list based on the emotion engine back to the user's terminal, the distribution including news optimized based on the emotional state.

[1426] Step 14:

[1427] When a user browses the distributed news and wants to know more about a particular article, the user sends a request for more information from the terminal to the server.

[1428] Step 15:

[1429] The server receives the request for more information and calls the news provider's API again based on the ID of the relevant news article to obtain additional details.

[1430] Step 16:

[1431] The server uses generative artificial intelligence to regenerate the detailed information it has acquired into a user-friendly format, which is then visualized in an easy-to-understand format.

[1432] Step 17:

[1433] The server sends the generated detailed information to the user's device, allowing the user to view the detailed information in real time.

[1434] Step 18:

[1435] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias in a particular category. If a bias is detected, a warning message is generated.

[1436] Step 19:

[1437] The server generates alerts and sends them to the user's device. Alerts are delivered as push notifications or in-app messages. The content and format of the alerts are adjusted according to the user's emotional state, as recognized by the emotion engine.

[1438] Step 20:

[1439] The device will display a feedback form to the user regarding the content and personalization of the news delivered to them, and collect feedback in the form of active pop-ups and surveys.

[1440] Step 21:

[1441] Users fill out a feedback form and send it to the server from their device. The feedback includes opinions on news content, delivery frequency, and degree of personalization.

[1442] Step 22:

[1443] The server stores user feedback in a database and uses it to retrain the generative AI. The collected feedback is used to retrain the AI ​​model, improving its accuracy and functionality.

[1444] Example 2

[1445] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1446] Currently, many users gather news online, but personalized news delivery is not yet fully implemented. Furthermore, news recommendations do not take into account the emotional state of individual users, which can lead to information overload or bias. As a result, users feel stressed or dissatisfied with the news content. Furthermore, when users want to know more about a news item, they must manually search for additional information, which is time-consuming. There is a need to solve these issues and realize a more efficient and user-friendly news delivery system.

[1447] The identification process 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 a news collection means, a means for classifying the collected news by topic or category, a means for generating titles and summaries using generative artificial intelligence, a means for creating a personalized news distribution list based on the user's past browsing history, a means for distributing the generated news titles and summaries to the user terminal, a means for accepting a request for providing detailed information about a specific news item, a means for acquiring detailed information about the specified news item and regenerating it in a user-friendly format using generative artificial intelligence to provide it, a means for processing the user's reactions while browsing the news using an emotion analysis engine and adjusting the news distribution list based on the user's emotional state, and a means for analyzing the user's news browsing history and generating and sending an alert based on the emotion analysis results if there is a bias toward a specific category. This not only personalizes the user's news browsing experience, but also enables flexible news distribution that takes the user's emotional state into consideration.

[1448] A "news gathering means" is a mechanism for periodically calling the interface of a news provider to obtain the latest news articles.

[1449] "Classification method" refers to the algorithm or machine learning model used to sort collected news articles into specific topics or categories.

[1450] "Generative AI" refers to AI that uses natural language processing technology to automatically generate titles and summaries of news articles.

[1451] A "personalized news delivery list" refers to a list of news articles organized in an optimized manner based on a user's past browsing history and interests.

[1452] "User Terminal" means a device (e.g., a smartphone or tablet) used by a User to receive news articles and notifications.

[1453] A "means for accepting requests for more information" is a mechanism for receiving and processing user requests for additional information about a particular news article.

[1454] "Means of regenerating into a user-friendly format" refers to a mechanism that uses generative artificial intelligence to convert news articles containing academic or specialized information into a form that is easy for the general user to understand.

[1455] "Emotion analysis engine" refers to technology that analyzes reaction data (e.g., facial expressions and operational behavior) when a user browses the news and identifies their emotional state.

[1456] The "means for generating and sending alerts" refers to a mechanism for generating alerts that notify users of bias in specific categories based on their news viewing history and sentiment analysis results, and sending the alerts to the users' devices.

[1457] This invention is a system that efficiently collects news articles and delivers personalized news to users. Furthermore, by incorporating a sentiment analysis engine, it recognizes user emotions and uses that information to enhance the degree of personalization of news distribution lists and adjust the content and format of alerts to address biases in news categories. A specific embodiment for implementing this system is described below.

[1458] News gathering

[1459] First, the server periodically calls the interface of the news provider to collect the latest news articles. This collection process is performed using, for example, a news API. The collected news data is generally provided in JSON format, and the server receives and parses it to obtain individual news articles.

[1460] Category Classification and Generative AI

[1461] The server then classifies the collected news articles by topic and category. During this classification process, categories such as business, technology, and entertainment are established. Titles and summaries are automatically generated for the classified news articles using generative artificial intelligence (AI). The generated titles and summaries are then stored in a database by the server.

[1462] Personalized News Generation

[1463] The server analyzes the user's news browsing history and creates a personalized news distribution list based on that information. The optimized news distribution list is sent to the user's device and delivered to the user as a push notification or in-app message.

[1464] Utilizing a sentiment analysis engine

[1465] One of the features of this invention is the incorporation of an emotion analysis engine. The emotion analysis engine analyzes the user's reactions while browsing the news (e.g., facial expression recognition and device interaction data) to recognize the user's emotions. The recognized emotion data is processed by the server, which recommends news articles based on the individual user's emotional state. For example, if a user is feeling stressed, entertainment articles that will help them relax will be delivered preferentially.

[1466] Providing more information

[1467] If a user wants to know more information about a particular news item, they can send a request for more information from their device. The server that receives the request then calls the news API again to retrieve the details, and uses generative artificial intelligence to regenerate them into a user-friendly format. The details are then sent to the device, where they can be viewed by the user.

[1468] Alert function

[1469] In addition, the server periodically analyzes the user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function is combined with an emotion analysis engine to adjust the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotions into consideration.

[1470] As a concrete example, if a businessperson receives the latest news at 7 a.m. and requests details about a specific business news item, the server will respond immediately and provide the details. Furthermore, if the emotion analysis engine detects that the user is stressed from browsing only business news, the server will recommend entertainment news that will help them relax. In this way, utilizing an emotion analysis engine makes it possible to deliver news that appropriately responds to the user's emotional state.

[1471] Example prompt sentence:

[1472] "Analyze the news browsing history of user A and identify areas of interest. Also, generate a news distribution list for when user A has recently been feeling stressed."

[1473] In this way, the system of the present invention makes the user's information gathering more efficient, supports the user in having a multifaceted perspective, and realizes flexible delivery that takes into consideration the user's emotional state.

[1474] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1475] Step 1:

[1476] The server periodically calls the interface of the news provider. For example, it can be configured to issue a request to the news API every hour. It specifies the URL endpoint to use for the call and sends the request with any necessary parameters (e.g., API key, country code, etc.).

[1477] Input: API endpoint URL, API key, request parameters

[1478] Output: News data in JSON format

[1479] Specifically, it sends an HTTP GET request to a URL such as https: / / newsapi.org / v2 / top-headlines?country=jp&apiKey=YOUR_API_KEY.

[1480] Step 2:

[1481] The server parses the received JSON-formatted news data to extract individual news articles. It then uses a JSON library to analyze the received data and obtain the title, body, and category of each news article.

[1482] Input: News data in JSON format

[1483] Output: List of news articles (title, body, category)

[1484] Specifically, it uses code such as json.loads(response.text) to convert the JSON data into a dictionary format and extract the required data fields.

[1485] Step 3:

[1486] The server categorizes the collected news articles by topic or category, and based on a predefined list of categories (e.g., business, technology, entertainment), it assigns the news articles to the corresponding category.

[1487] Input: List of news articles (title, body, category)

[1488] Output: A list of news articles sorted by category

[1489] Specifically, it classifies news articles into categories using conditional statements such as if article["category"] == "business":

[1490] Step 4:

[1491] The server uses a generative device to generate titles and summaries for news articles. It uses a generative AI model (e.g., GPT-3) to automatically generate optimal titles and summaries from the text of news articles.

[1492] Input: News article body

[1493] Output: Generated title and summary

[1494] Specifically, it calls functions such as gpt3.generate_summary(article["content"]) to obtain the generated results from the AI ​​model.

[1495] Step 5:

[1496] The server stores the generated title and abstract in a database. It executes an INSERT query to the database to store data such as title, abstract, body, and category.

[1497] Input: Generated title and summary, news article body, category

[1498] Output: News article information stored in a database

[1499] Specifically, it executes SQL statements such as INSERT INTO news_articles (title, summary, content, category) VALUES (...).

[1500] Step 6:

[1501] The server analyzes the user's past news browsing history and creates a personalized news distribution list. It retrieves the user's browsing data from a database and uses analytical algorithms to extract personal preferences.

[1502] Input: User news browsing history data

[1503] Output: A personalized news distribution list

[1504] Specifically, the system aggregates data such as the number of category views per user and time periods, and calculates a priority score.

[1505] Step 7:

[1506] The server sends the created news distribution list to the device and delivers it to the user as a push notification or in-app message. The distribution list is packaged in JSON format and delivered using a messaging service (e.g., Firebase Cloud Messaging).

[1507] Input: Personalized news distribution list

[1508] Output: Notification to user terminal

[1509] Specifically, message sending code such as fcm.send_message(user_device_token, news_list_json) is used.

[1510] Step 8:

[1511] The device sends the user's reactions to the news to the emotion analysis engine, including facial expression recognition results and device touch data.

[1512] Input: User reaction data when browsing news

[1513] Output: Emotion analysis results

[1514] Specifically, the camera captures facial expressions and sends the images to an emotion analysis API.

[1515] Step 9:

[1516] The server processes the transmitted emotional data using an emotion analysis engine. The emotion analysis engine is used to identify the user's emotional state. The analysis results (e.g., stress level) are used to influence the next news delivery.

[1517] Input: User reaction data, results of sentiment analysis engine

[1518] Output: News distribution list based on emotional state

[1519] Specifically, it calls the sentiment analysis API and obtains the results.

[1520] Step 10:

[1521] Based on the analysis results, the server selects appropriate news articles and recommends them to the user. It also regenerates a news distribution list based on the emotional state and sends it to the device.

[1522] Input: Sentiment analysis results, news article list

[1523] Output: News distribution list based on emotional state

[1524] Specifically, the system generates and transmits a personalized news list again.

[1525] Step 11:

[1526] A user requests more information about a particular news item. Through the device interface, the user sends a request for more information about the particular article.

[1527] Input: News article ID

[1528] Output: Request status

[1529] Specifically, the request is sent when the "View details" button is clicked.

[1530] Step 12:

[1531] The server calls the news API again to get more information, which is then regenerated into a user-friendly format using generative artificial intelligence.

[1532] Input: News article ID

[1533] Output: Detailed information generated

[1534] Specifically, the API is called by adding the news article ID as a query parameter, and functions such as gpt3.generate_readable_format(article_detail_content) are used.

[1535] Step 13:

[1536] The server sends the regenerated details to the device for the user to view, and packages the generated details in JSON format and sends it.

[1537] Input: Generated details

[1538] Output: Notification to terminal

[1539] Specific operations include code such as send_to_device(user_device_token, detailed_article_json).

[1540] Step 14:

[1541] The server periodically analyzes the user's news browsing history and generates an alert if there is a bias towards a specific category. It combines a sentiment analysis engine to adjust the content and format of the alert according to the user's emotional state.

[1542] Input: User news browsing history, sentiment analysis results

[1543] Output: Bias alert

[1544] Specifically, it compares the viewing frequency of each category on a weekly basis and generates an alert if there is a bias.

[1545] Step 15:

[1546] The server sends the generated alert to the device to alert the user, and packages the alert content in JSON format and sends it to the device.

[1547] Input: Bias Alert

[1548] Output: Alert notification to terminal

[1549] Specific actions include using codes such as send_alert(user_device_token, alert_message_json).

[1550] (Application example 2)

[1551] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1552] In today's world, many users get their information from online news, but this presents several challenges. First, it is often impossible to efficiently gather the information that is truly important to users from the vast amount of news articles. Second, it is difficult to personalize news based on the user's interests and emotions. In particular, if a user is feeling stressed or has a bias toward certain categories, appropriate news delivery may not be provided. Furthermore, there is a lack of a mechanism for utilizing user feedback to improve the system.

[1553] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes a news collection means, a means for classifying the collected news by topic or category, and a means for generating titles and summaries using generative artificial intelligence. This enables efficient processing of collected news articles and personalized news delivery according to the user's needs. The server also includes an emotion engine that recognizes the user's emotional state and a means for dynamically adjusting the news delivery list based on the recognized emotion, a means for regenerating and providing news in a user-friendly format using generative artificial intelligence, and a means for accepting requests for detailed information on specific news. This configuration enables flexible and effective news delivery based on the user's emotions and interests, improving the user's information gathering experience.

[1554] A "news gathering means" is a means for gathering the latest news articles from news sources.

[1555] "Topic or category classification means" refers to a means for automatically classifying collected news articles into specific topics or categories, such as business, technology, entertainment, etc.

[1556] "Generative AI" is an AI technology used to automatically generate titles and summaries from collected news articles.

[1557] The "means for creating a personalized news distribution list" is a means for creating a news distribution list optimized for an individual user based on the user's past browsing history.

[1558] The "means for delivering to a user terminal" refers to a means for delivering the generated news title and summary to a device such as a user's smartphone or computer.

[1559] The "means for accepting a request for providing detailed information" is a means for accepting a request from a user for detailed information about a specific news article.

[1560] "Means for regenerating and providing in a user-friendly format" refers to means for obtaining detailed information about specified news and regenerating it in a user-friendly format using generative artificial intelligence.

[1561] "Means for generating and sending alerts" refers to means for analyzing a user's news browsing history, generating an alert if there is a bias in a specific category, and notifying the user.

[1562] The "emotion engine" is an engine that recognizes the user's emotional state (e.g., stress or frustration) and adjusts the content of news distribution lists and alerts based on that data.

[1563] A system embodying the invention includes the following major elements:

[1564] 1. News collection and classification

[1565] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The retrieved news articles are then classified by topic or category. For example, categories such as business, technology, and entertainment are set.

[1566] 2. Title and Abstract Generation Using Generative AI

[1567] The server uses generative artificial intelligence (AI) to automatically generate titles and summaries from collected news articles, which are then stored in a database.

[1568] 3. Create a personalized news distribution list

[1569] The server analyzes the user's past news browsing history and creates a personalized news distribution list based on that information. The generated news distribution list is sent to the user's device as a push notification or in-app message.

[1570] 4. Emotional Engine Adjustment

[1571] The server is equipped with an emotion engine that analyzes users' reactions while browsing the news (e.g., facial expression recognition and device interaction data). This allows the server to recognize the user's emotional state and recommend the most appropriate news articles for each individual user. For example, if a user is feeling stressed, it will prioritize the delivery of relaxing entertainment articles.

[1572] 5. Requesting and Providing More Information

[1573] When a user wants to check the details of a specific article, they send a request for more information from their device. Upon receiving this request, the server calls the news API again to retrieve the details, and then uses generative artificial intelligence to regenerate them into a user-friendly format. The regenerated details are then sent to the device, where they can be viewed by the user.

[1574] 6. Generate and send alerts

[1575] The server periodically analyzes a user's news browsing history and generates an alert if there is a bias toward a particular category. This alert function also incorporates an emotion engine, which adjusts the content and format of the alert according to the user's recognized emotional state. For example, if the user is feeling dissatisfied, the server will send an alert recommending news from other categories in a gentler way that takes their emotion into consideration.

[1576] Hardware and software used

[1577] Emotion recognition: Keras (TensorFlow backend), OpenCV

[1578] News gathering: requests library, news API

[1579] Data storage and management: Any database (e.g., PostgreSQL)

[1580] Emotion engine: Generative AI models (e.g. GPT-3)

[1581] Specific examples

[1582] A user launches a news app on their smartphone and browses recent news. If the emotion engine analyzes the user's facial expressions and recognizes that they are feeling stressed, the server automatically prioritizes relaxing entertainment news. If the user requests more information about a particular business news item, the server instantly generates and provides the detailed information.

[1583] Prompt Sentence Examples

[1584] When users are feeling stressed, introduce them to today's relaxing entertainment news.

[1585] This will realize a news distribution system that recommends appropriate news and products based on the user's emotions and history, and provides personalized feedback.

[1586] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1587] Step 1:

[1588] The server periodically calls the news provider's API to collect the latest news articles. The collected news data is provided in JSON format, which the server parses to retrieve individual news articles. The input is the response data from the news API, and the output is a list of extracted news articles.

[1589] Step 2:

[1590] The server classifies the collected news articles by topic and category. The input is the list of news articles obtained in the previous step, and the output is the news articles classified by category. Keyword analysis and natural language processing techniques are used for classification.

[1591] Step 3:

[1592] The server uses generative artificial intelligence to generate titles and summaries from news articles. The generated titles and summaries are stored in a database. The input is the news articles classified in step 2, and the output is the generated title and summary for each article. A generative artificial intelligence model (e.g., GPT-3) performs the text generation.

[1593] Step 4:

[1594] The server analyzes the user's past browsing history and creates a personalized news distribution list. The input is the user's browsing history data and the news articles obtained in step 3, and the output is a personalized news distribution list. A machine learning algorithm is used for the analysis.

[1595] Step 5:

[1596] The server uses an emotion engine to recognize the user's emotional state in real time. The input is the user's facial expression recognition data and interaction data, and the output is the user's emotional state information. Keras models and OpenCV are used for emotion recognition.

[1597] Step 6:

[1598] The server dynamically adjusts the news distribution list based on the user's emotional state. The input is the emotional state information obtained in step 5 and the news distribution list created in step 4, and the output is the adjusted news distribution list. The list priority is changed depending on the emotional state.

[1599] Step 7:

[1600] The device delivers the personalized news feed list to the user as push notifications or in-app messages. The input is the news feed list adjusted in step 6, and the output is the news articles displayed on the user's device.

[1601] Step 8:

[1602] When a user requests detailed information about a specific article, the terminal sends a detailed information request to the server. The input is the user's request for detailed information, and the output is the request being sent to the server.

[1603] Step 9:

[1604] When the server receives the request for more information, it calls the news API again to obtain the details and regenerates them in a user-friendly format using generative artificial intelligence. The input is the request for more information and the response data from the news API, and the output is the regenerated details.

[1605] Step 10:

[1606] The terminal displays the details received from the server to the user. The input is the details regenerated in step 9, and the output is the details displayed on the terminal.

[1607] Step 11:

[1608] The server periodically analyzes the user's news browsing history and generates and sends alerts if there is a bias towards a specific category. The input is the user's news browsing history data and current browsing data, and the output is the generated alert. An algorithm is used for the analysis, and the content and format of the alert reflects an emotional engine.

[1609] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1610] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1611] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1612] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1613] FIG. 9 illustrates an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and behaviors arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1614] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1615] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1616] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1617] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1618] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion sh...

Claims

1. news gathering means; A means to categorize collected news by topic or category, a means for generating a title and abstract using generative artificial intelligence; means for creating a personalized news distribution list based on a user's past browsing history; means for delivering the generated news titles and summaries to a user terminal; a means for accepting requests to provide more information about a particular news item; A means for obtaining detailed information about a specified news item, regenerating it into a user-friendly format using generative artificial intelligence, and providing it; A means of analyzing a user's news browsing history and generating and sending alerts if there is a bias towards a specific category; A system including:

2. a means for collecting user feedback regarding the content and personalization of news delivered; 10. The system of claim 1, further comprising means for retraining the generative artificial intelligence based on the collected feedback.

3. 2. The system according to claim 1, further comprising means for automatically calling an API of a news provider at a time interval designated by a user to collect the latest news articles.

Images